Textbook of Radiology and Imaging (Vol. 2)

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ CONTENTS ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ Previous Page CONTENTS Next Page SEVENTH EDITION TEXTBOOK OF RADIO...

Author: David Sutton

1847 downloads 7882 Views 76MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ CONTENTS ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ

Previous Page

CONTENTS

Next Page

SEVENTH EDITION

TEXTBOOK OF RADIOLOGY

AND IMAGING EDITED BY

DAVID SUTTON

MD, FRCP, FRCR, DMRD, FCan.AR (Hon)

Consulting Radiologist St Mary's Hospital and Medical School, London Director, Radiological Department (1963-1984)

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ

Consulting Radiologist, The National Hospital for Neurology and Neurosurgery, London, UK.

A SSOCIA tE EDITORS Nuclear Medicine

PHILIP J.A. ROBINSON

FRCP, FRCR

MRI

JEREMY P.R. JENKINS

FRCP DMRD, FRCR

CT RICHARD W. WHITEHOUSE

BSc, MB ChB, MD, FRCR

Ultrasound

PAUL L. ALLAN

MSc, MBBS, DMRD, FRCR, FRCP(Ed)

Cardiac Radiology PETER WILDE

BSc, MRCP, FRCR

Neuroradiology JOHN M. STEVENS

MBBS, DRACR, FRCR

I CHURCHILL LIVINGSTONE

Noordanesh Medical Publication

CONTENTS

Previous Page

CHURCHILL LIVINGSTONE An imprint of Elsevier Science Limited 2003, Elsevier Science Ltd. All rights reserved. The right of David Sutton to be identified as editor of this work has been asserted by him in accordance with the Copyright, Designs and Patent Act 1 988. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without either the prior permission of the publishers (Elsevier Science Limited, Robert Stevenson House, 1-3 Baxter's Place, Leith Walk, Edinburgh EH 13AF) or a licence permitting restricted copying in the United Kingdom issued by the Copyright Licensing Agency Ltd, 90 Tottenham Court Road, London W I T 4LP. First Edition 1969 Second Edition 1975 Third Edition 1980 Fourth Edition 1987 Fifth Edition 1993 Sixth Edition 1998 I SBN 0 443 071098 I nternational Student Edition ISBN 0 443 07108X

British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library


Library of Congress Cataloging in Publication Data A catalog record from this book is available from the Library of Congress

Note Medical knowledge is constantly changing. As new information becomes available, changes in treatment, procedures, equipment and the use of drugs become necessary. The editors, contributors and the publishers have, as far as it is possible, taken care to ensure that the information given in this text is

accurate and up to date. However, readers are strongly advised to confirm that the information, especially with regard to drug usage, complies with the latest l egislation and standards of practice.

The Publishers have made every effort to trace the copyright holders for borrowed material. If they have inadvertently overlooked any, they will be pleased to make the necessary arrangements at the first opportunity.

Commissioning Editor: Michael J. Houston Project Development Manager: Martin Mellor Project Manager: Nora Naughton (Aoibhe O'Shea) Designer: Sarah Russell

The publisher's policy is to use paper manufactured from sustainable forests

Printed in China by RDC Group Limited


Next Page

Previous Page

Next Page

CONTENTS Cover illustrations:

Vol 1: Front: Figs. 2.09; 1.60; 1.61A,B; 15.1 39B Back: Figs. 25.144; 2.8; 15.144; 26.43C Vol 2: Front: Fias. 59.26A; 59.39; 58.94A-D; 59.37 Back: Figs . 33.7113; 55.43C; 58.43C

SECTION 1 Respiratory system 1

SECTION 2 Cardiovascular system

The normal chest: methods of investigation and differential diagnosis 1 Janet Murfitt, Philip J. A. Robinson, Richard W. Whitehouse,

10 The normal heart: anatomy and techniques of examination 265 Peter Wilde, Mark Callaway

Andrew R. Wright, Jeremy P R. Jenkins

Roger H. S. Gregson, Richard W. Whitehouse,

11

Acquired heart disease I: the chest radiograph 283

Andrew R. Wright, Jeremy P. R. Jenkins

Mark Callaway, Peter Wilde


2 The mediastinum 57

12 Acquired heart disease II: non-invasive i maging 317

3 The pleura 87

Michael B. Rubens, Simon P G. Padley

Mark Callaway, Peter Wilde

4 Tumours of the lung 107

Michael B. Rubens, Simon P. G. Padley, Jeremy P. R. Jenkins

5 Pulmonary infections 131

13

I nvasive imaging and interventional techniques 347 Peter Wilde, Mark Callaway

Simon P. G. Padley, Michael B. Rubens

14 Congenital heart disease

6 Diseases of the airways: collapse and consolidation 1 61 Michael B. Rubens, Simon PG. Padley

7 Diffuse lung disease 187

363

Peter Wilde, Anne Boothroyd

15 Arteriography and interventional angiography 411 David Sutton, Roger H. S. Gregson, Paul L. Allan, Jeremy P. R. Jenkins

Simon P G. Padley, Michael B. Rubens

8 Miscellaneous chest conditions Simon P G. Padley, Michael B. Rubens

217

16 Phlebography

483

David Sutton, Roger H. S. Gregson, Paul L. Allan, Jeremy P. R. Jenkins

9 The paediatric chest

247

Catherine M. Owens, Karen E. Thomas

17 The lymphatic system

509

Graham R. Cherryman, Bruno Morgan


CONTENTS

Previous Page

6

Next Page

TEXTBOOK OF RADIOLOGY AND IMAGING

SECTION 3

32

Julian Kabala, Philip J. A. Robinson, Raj Persad, Robert Jones

Abdomen and gastrointestinal tract 18

33

The salivary glands, pharynx and oesophagus 533

Obstetric ultrasound

1 039

Roger Chisholm, Jeremy P R. Jenkins

34

A. H. A. Chapman, John A. Spencer, J. Ashley Guthrie, Philip J. A. Robinson

19

The urethra and male genital tract 1017

Gynaecological imaging

1 069

Mary Crofton, Jeremy P R. Jenkins

The stomach and the duodenum 575 A. H. A. Chapman, J. Ashley Guthrie, Philip J. A. Robinson

20

The small bowel and peritoneal cavity

615

Steve Halligan

21

The large bowel

SECTION 5 Skeletal system: soft tissue 35

635

Steve Halligan, Philip]. A. Robinson

Congenital skeletal anomalies: skeletal dysplasias, chromosomal disorders 1107 Peter Renton, Ruth Green

22 The acute abdomen 663

36

Stuart Field, lain Morrison

23

The abdomen and major trauma 691

Periosteal reaction; bone and joint i nfections; sarcoid 1153 Peter Renton

Otto Chan, loannis Vlahos

24 The biliary tract

37 Avascular necrosis; osteochondritis; miscellaneous bone lesions 1179

711

John Karani

25

Peter Renton, Ruth Green

The liver and spleen

737

Robert Dick, Anthony Watkinson, Julie F. C. Olliff,

38

Disease of joints

1 201

Peter Renton, Ruth Green

Philip J. A. Robinson, Richard W. Whitehouse

26 The pancreas

787


Janet Murfitt, Richard W. Whitehouse, Philip J. A. Robinson, Richard Mason, Paul A. Dubbins, Andrew R. Wright

27

39 Tumours and tumour-like conditions of bone (1) 1247

The adrenal glands

Mark Cobby, lain Watt

40

825

David Sutton, Philip J. A. Robinson

Tumours and tumour-like conditions of bone (2) 1287 Mark Cobby, lain Watt

28

The paediatric abdomen

849

Karen E. Thomas, Catherine M. Owens

41

Disorders of the lymphoreticular system 1321 and other haemopoietic disorders Mark Cobby, lain Watt

42

SECTION

Jeremy W. R. Young, Leonie Gordon

SECTION 4

43

Genito-urinary tract 29

Metabolic and endocrine disorders affecting bone 1351 Skeletal trauma: general considerations

1371

Jeremy W R. Young

The urogenital tract: anatomy and investigations

885

44

Julian Kabala, Tim Whittlestone, David Grier,

Skeletal trauma: regional

1389

Jeremy W R. Young

Philip]. A. Robinson

30

The kidneys and ureters

929

45

1 417

Jeremy P R. Jenkins, Janet Murfitt, Fritz Starer ,

Julian Kabala, Carl Roobottom

31

The soft tissues

Richard W Whitehouse, W Gedroyc

The bladder and prostate

989

Julian Kabala, Gary N. Sibley, Jeremy P R. Jenkins, Paul Hulse

46

The breast 1451 Michael J. Michell, Chris Lawinksi, Will Teh, Sarah Vinnicombe


CONTENTS

Previous Page

Next Page

CONTENTS

55 Angiography in neuroradiology 1673

SECTION 6

David Sutton, John M. Stevens, Katherine Mizkiel

56 I nterventional neuroradiology

1 707

Rolf Jager , Stefan Brew

57 Intracranial lesions (1) 47 The pharynx and larynx: the neck 1489 Peter D. Phelps, Philip J. A. Robinson, Richard W Whitehouse, Andrew R. Wright, Julie F. C. Olliff

48 The sinuses

1 723

David Sutton, John M. Stevens, Katherine Mizkiel

58 I ntracranial lesions (2)

1 767

David Sutton, John M. Stevens, Katherine Mizkiel,

1519

Philip J. A. Robinson, Keith Dewbury

Swarupsinh V Chavda, Julie F. C. Olliff

59 Recent technical advances

49 Teeth and jaws 1531

1 819

Richard W Whitehouse, Philip l. A. Robinson,

Peter Renton

Jeremy P. R. Jenkins, Paul L. Allan, Nicola H. Strickland,

50 Ultrasound of the eye and orbit 1551

Philip Gishen, Andrew R. Wright, Andrew P Jones

John A. Fielding

51 The orbit

1 573

Useful appendices

1 847

Michael 1. Rothman, Gregg H. Zoarski

A. Centres of ossification 52 The petrous temporal bone 1597 Peter D. Phelps

1 847

B. Glossary of CT terms 1850 C. Glossary of MR terms 1852

53 The skull

1 617

D. Radiopharmaceuticals for imaging

David Sutton

54 Neuroradiology of the spine John M. Stevens, Brian E. Kendall

1 643 I ndex to volumes 1 and 2



1 855

7

CONTENTS

Previous Page

Next Page

CONTRIBUTORS Paul L. Allan MSc, MBBS, DMRD, FRCR, FRCP (Ed) Honorary Consultant Radiologist Royal Infirmary Edinburgh, UK

Mark Cobby MBChB, MRCP, FRCR Consultant Radiologist Frenchay Hospital Bristol, UK

Stefan Brew MB, ChB, MHB (Hons), MSc, FRANZCR, FRCR Consultant Radiologist National Hospital for Neurology and Neurosurgery London, UK

Keith Dewbury BSc, DMRD, FRCR Consultant Radiologist Southampton General Hospital Southampton, UK

Anne Boothroyd MBChB, FRCR Consultant Radiologist Royal Liverpool Children's Hospital Liverpool, UK

Robert Dick MB, BS(Syd), FRCAR, FRCR Department of Radiology Royal Free Hospital London, UK

Mark Callaway BM, MRCP, FRCR Consultant Radiologist Bristol Royal Infirmary Bristol, UK

Paul Dubbins BSc, FRCR Consultant Radiologist I maging Directorate Derriford Hospital Plymouth, UK Ply

Otto Chan FRCS, FRCR Consultant Radiologist The Royal London Hospital London, UK

Stuart Field MA, MBBChir, DMRD, FRCR Consultant Radiologist Kent and Canterbury Hospital Canterbury, UK


Anthony H. A. Chapman Head of Clinical Radiology Leeds NHS Trust Consultant Radiologist St James's University Hospital Leeds, UK

FRCP, FRCR

Swarupsinh V. Chavda MBChB, DMRD, FRCR Consultant Radiologist St James's University Hospital Leeds, UK Graham R. Cherryman Professor of Radiology

M[3013, FRCR

Honorary Consultant Radiologist UHL NHS Trust Leicester, UK Roger Chisholm MA, MBBChir, MRCP, FRCR Consultant Radiologist Hope Hospital Salford, UK Mary Crofton FRCR, FRCP Consultant Radiologist Department of Radiology St Mary's Hospital London, UK

John A. Fielding MD, FRCP(Edin), FRCR Consultant Radiologist Royal Shrewsbury Hospital Shrewsbwy, UK W. Gedroyc MRCP, FRCR Consultant Radiologist St Mary's Hospital London, UK

Philip Gishen MB, BCh, DMRD, FRCR Consultant Radiologist and Director of Imaging Hammersmith Hospital UK Lond , f eic st r Roger H. S. Gregson MSc, MB, FRCR, DMRD Consultant Radiologist and Head of Training University of Nottingham Nottingham, UK

Ruth Green FRCR Consultant Radiologist Royal National Orthopaedic Hospital Middlesex. UK David Grier MBChB, MRCP, FRCR Consultant Radiologist Bristol Royal Hospital for Children Bristol, UK


CONTENTS

Previous Page

Next Page

CONTRIBUTORS

Leonie Gordon

Michael J. Michell

MD

FRCR

Professor of Radiology and Nuclear Medicine Medical University of South Carolina Charleston, South Carolina, USA

Consultant Radiologist King's College Hospital London, UK

J. Ashley Guthrie

Consultant Neuroradiologist National Hospital for Neurology and Neurosurgery London, UK

BA, MRCP, FRCR

Consultant Radiologist St James's University Hospital Leeds, UK

Steve Halligan

Bruno Morgan

MBBS, MD, MRCP, FRCR

Consultant Radiologist St Mark's Hospital London, UK

Paul Hulse

Consultant Radiologist National Hospital for Neurology and Neurosurgery London, UK

Honorary Senior Clinical Lecturer Manchester Royal Infirmary and University of Manchester Manchester, UK


BmedSci, BMBS, FRCS(Ed)

Simon PG. Padley

Brian E. Kendall

Peter D. Phelps

FRCR, FRCP, FRCS

Consulting Radiologist The National Hospital for Neurology and Neurosurgery and the Middlesex Hospital London, UK

Chris Lawinksi

BSc, MSc, MPhil

Consultant Physicist King's College Hospital London, UK

Richard Mason

ChM, FRCS, FRCS(Urol), FEBU

Consultant Urologist Bristol Royal Infirmary Bristol, UK

MRCP, FRCR

Consultant Radiologist Bristol Royal Infirmary Bristol, UK

MRCP, FRCR

Consultant Radiologist Chelsea and Westminster Hospital London, UK

Raj Persad

MSc, MBBS, FRCR

Consultant Radiologist King's College Hospital London, U K

BSc, MRCP, PFCR

Clinical Director Consultant Paediatric Radiologist Department of Radiology Great Ormond Street Hospital for Children London, UK

Urology Research Fellow Bristol Royal Infirmary Bristol, UK

Julian Kabala

B Med Sci, BM BS, MRCP, FRCR

Consultant Radiologist Honorary Senior Clinical Lecturer University of Birmingham Birmingham, UK

Catherine M. Owens MSc

Consultant Clinical Scientist Head of MR Physics Group Christie Hospital Manchester, U K

John Karani

MB BS, MRCP, FRCR

Consultant Radiologist and Director of Diagnostic Imaging St Bart's and The London NHS Trust London, UK

Julie F.C. Olliff FRCP, DMRD, FRCR

Consultant Radiologist

Robert Jones

MB BS, MRCP, FRCR

Janet Murfitt MD, FRCR

Andrew P. Jones

MA, MRCP, FRCR

Consultant Radiologist Kent and Canterbury Hospital Canterbury, UK

Consultant Radiologist Christie Hospital Manchester, UK

Jeremy P. R. Jenkins

BM(Hons), MRCP, FRCR

Senior Lecturer and Honorary Consultant Radiologist University Hospitals Leicester Leicester, UK

lain Morrison

MRCP FRCR

H. Rolf Jager

Katherine Mizkiel

MD, FRCS, FRCR

Former Consultant Radiologist Royal National Orthopaedic Hospital and University College Hospital Honorary Senior Lecturer I nstitute of Orthopaedics London, UK

Peter Renton

FRCR, DMRD

Consultant Radiologist Honorary Senior Lecturer Royal National Orthopaedic Hospital and University College London Hospitals London, UK

Philip J.A. Robinson

FRCS, MRCP, FRCR

Consulting Radiologist Middlesex Hospital University College of London Hospitals London, UK

FRCP, FRCR

Professor of Clinical Radiology University of Leeds Consultant Radiologist Leeds Teaching Hospitals Leeds, UK


9

CONTENTS

Previous Page

10

Next Page

TEXTBOOK OF RADIOLOGY AND IMAGING

M. I. Rothman MD Assistant Professor of Radiology, Neurosurgery and Otolaryngology /Head and Neck Surgery, Medical Director. Anna Gudelsky Magnetic Resonance Center Baltimore, Maryland, USA

Sarah Vinnicombe BSc, MRCP, FRCR Consultant Radiologist Department of Diagnostic Imaging St Bartholomew's Hospital London, UK

Carl Roobottom MSc, MBChB(Hon), MRCP, FRCR Consultant Radiologist Dcrriford Hospital Plymouth, UK

l oannis Viahos MSc, MBBS, MRCP, FRCR Research Fellow Department of Diagnostic Imaging St Bartholomew's Hospital London, UK

Michael B. Rubens MB, DMRD, FRCR Consultant Radiologist and Director of Imaging Royal Brompton Hospital London, UK

l ain Watt FRCP, FRCR Consultant Clinical Radiologist Bristol Royal Infirmary Bristol, UK

Gary N. Sibley FRCS Consultant Urologist Department of Urology Bristol Royal Infirmary Bristol, UK

Anthony Watkinson Bmet, MSc, MBBS, FRCS, FRCR Consultant and Senior Lecturer in Radiology Royal Free Hospital London, UK

John A. Spencer MA, MD, MRCP, FRCR Consultant Radiologist St James's University Hospital Leeds, UK John M. Stevens MBBS, DRACR, FRCR Consultant Radiologist Department of Radiology National Hospital for Neurology and Neurosurgery London. UK Nicola H. Strickland BM, BCh, MA(Hons)(Oxon), FRCP, FRCR Consultant Radiologist Hammersmith Hospital NHS Trust London, UK

Peter Wilde BSc, MRCP, FRCR Consultant Cardiac Radiologist Directorate of Clinical Radiology Bristol Royal Infirmary Bristol, UK Richard W. Whitehouse Consultant Radiologist Manchester Royal Infirmary Manchester, UK

BSc, MB ChB, MD, FRCR

Tim Whittlestone MA, FRCS (Ebg), MD, FRCS(Urol) Hunterian Professor of Surgery Specialist Registrar in Urology Bristol Royal Infirmary Bristol, UK


David Sutton MD, FRCP, FRCR, DMRD, FCan.AR (Hon) Consulting Radiologist St Mary's Hospital and Medical School, London Director, Radiological Department (1963-1984) Consulting Radiologist, The National Hospital for Neurology and Neurosurgery London, UK

Andrew R. Wright MA, MBBS, MRCP, FRCR Consultant Radiologist Honorary Senior Lecturer St Mary's Hospital I mperial College London, UK

Will Teh MBChB, MRCP, FRCR Consultant Radiologist Northwick Park Hospital Middlesex, UK

Jeremy W. R. Young MA, BM, BCh, FRCR Professor and Chairman of Radiology, Medical University of South Carolina Charleston South Carolina, USA

Karen E. Thomas MA, BM BCh, MRCP, FRCR Consultant Paediatric Radiologist Hospital for Sick Children Toronto Ontario, Canada

Gregg Zoarski MD Department of Diagnostic Radiology University of Maryland Medical Center Baltimore Maryland, USA


CONTENTS

Previous Page

Next Page

PREF ACE The First Edition of this Textbook was conceived in the 1960s and published in 1969. 1, like many of my contemporaries began my studies in Radiology at the end of the Second World War. My first post as an ex military service registrar was in the Radiology Department of the National Hospitals for Nervous Disease at Queen Square. It was pure serendipity that I should thus become associated with James Bull, the only British radiologist trained in Scandinavian neuroradiological techniques, i ncluding percutaneous angiography. and at that tine representing

syllabus for Specialist Radiologists, the change was highly successful from a purely practical point of view. Sales rose by 300°% (from 1000 to 4000 copies per annum). The first edition of this popular text-hook was published in 1 969 and this seventh edition in still growing strongly at the mature age of thirty three years. Review of the last five editions cover a period of exponential growth in radiological facilities and i maging. New fields were just beginning to open at the time of the first edition and these included ultrasound and nuclear medicine. Computer tomography began in the 1970s to he overtaken in the the most advanced aspects of radiology. 1 980s by magnetic resonance. It was generally felt that CT would My training in percutaneous cerebral angiography laid the foundation for other percutaneous techniques which I was able to soon be out-moded, but the last few years have seen a remarkable apply when appointed to St. Mary's hospital in 1952. Here again comeback from CT in the form of multi-slice spiral CT. As a pioneer work had already begun exploiting the potential for new result the versatility, speed and scope of CT examinations has been transformed. methods in vascular surgery. I n general, we hope this book reflects British Teaching Hospital As a result of this background we were able to publish in 1962 the first personal monograph based on an experience of more than Practice in the field of Imaging. The ISE edition remains very ten thousand cases. (See Ch. I5). popular with non British readers and the 6th Edition has also been X-Rays were discovered by Roentgen in 1895, and though the translated into two further languages, Greek and Portuguese. We i mportance of the discovery was immediately realised and widely believe that much of its success is due to the decision to discussed the impact on medical practice was surprisingly slow. concentrate on Clinical rather than Technical aspects of our The diagnosis and treatment of fractures and lesions of bones and rapidly expanding and evolving specialty. joints was the first area to be thoroughly studied and surveyed. At Whilst each new edition has emphasised clinical rather then the same tine, the dangers and potential hazards of the new rays t echnical progress, the student must also be aware of. and absorb. were becoming apparent for the first time, as was the therapeutic t he technical advances. The new edition therefore includes a chapter use of X-Rays. devoted to explaining this area. Other features of this new edition I n the Post war period, paining and experience of a specialist are the complete rewriting by mainly new authors of major sections radiologist was still a matter of considerable debate and concern. of the text. These include the Cardiac. GU. Paediatric, Small and Broadly speaking, there were those who favoured a technical Large bowel, Major Abdominal Trauma and Intcrventional approach and paining, usually pure scientists or physicists, and Neuroradiology chapters. Other chapters have been revised by others who preferred a largely clinical approach with a minimum deleting obsolete material or including new material. Recent clinical of technical training. Thus advanced paining in medicine or trends are also reflected in the revision. Thus imaging and Staging of malignant tumors has been revised and updated in many areas. surgery was regarded by many as essential for high quality radiology. The British Faculty of Radiologists was expanding and the opportunity has been taken to integrate the latest version of rapidly and soon became the Royal College of Radiologists. The the World Health Organisation (WHO) reclassification on a histopatholigical basis of primary cerebral tumoii s. The expansion FRCR thus became the essential higher radiological qualification on a par with the MRCP or FRCS, and the DMRD was of non invasive and minimally invasive angioaraphy is monitored, downgraded to a qualifying diploma. and discussed. However, this is to some extent balanced by the i ncreasing use of interventional techniques. At the time of this controversy, I took the opportunity to broaden my experience and expertise with the MD thesis, Membership of Radiology is a graphic subject. and images and illustrations are the Royal College of Physicians, London and Fellowship of the i ts vital tool. This edition contains no less then 5600 illustrations. Faculty of Radiologists. This was undoubtedly the clinical, rather some 2000 of which are new. than technical approach to radiological expertise. As in previous editions, we would remind the student that large I n 1955 1 was appointed Editor to the Faculty .Journal, and took t extbooks. like large animals, have a longer period of gestation. It i s therefore important to keep up with the current literature and the opportunity to persuade the Editorial Board to change its attend up to date seminars. name to Clinical Radiology. Apart from showing where my own i nterest lay in the continual medico-political controversy between pure scientists (mainly physicists) and clinicians which many felt David Sutton could adversely affect the future paining and examination '_ 002



CONTENTS

Previous Page

Next Page

29 THE UROGENITAL TRACT: ANATOMY AND INVESTIGATIONS J ulian E. Kabala with contributions from Philip J. A. Robinson, Tim Whittlestone and from David Grier

ANATOMY The kidneys and retroperitoneum, ureters, bladder and male genitalia are discussed.

The kidneys and retroperitoneum The kidneys develop from three structures which succeed each other: the pronephros and the mesonephrie and metanephrie duets. The pronephros develops during the 3rd week of gestation and regresses during the 4th to the 8th week, leaving no adult correlate. It is replaced by the mesonephrie (wolffian) duet, which develops in males into the efferent ductules of the testes, the epididymis and the vas deferens and in females into the epioophoron and paraoophoron. In the 5th week the kidney begins to develop from two separate cell lines arising from the metanephrie duet: the ureterie bud and the metanephrie blastema. The ureterie bud develops into the ureter, renal pelvis and (by multiple divisions of the upper end) the calyces and collecting duets. The metanephrie blastema develops into the Bowman's capsules, proximal and distal convoluted tubules, Henle's loops and the remainder of the renal parenchyma. The division of the ureterie bud into calyces gives rise to the underlying lobar structure of the kidney, each lobe consisting of a calyx and its associated collecting ducts and renal cortex. During development about 14 lobes develop which fuse initially to form an obviously lobulated kidney. Renal cortical cellular multiplication continues in early childhood, gradually smoothing out the lobar outline by around 5 years of age. In approximately 5% of individuals the lobar outline persists into adulthood. The kidneys develop in the upper half of the pelvis and migrate eranially during the 4th to the 8th week of gestation. At the same ti me the kidneys rotate 90' medially so the renal pelves lie on the anteromedial aspect of the kidney. Initially the kidneys are supplied by lateral sacral branches of the aorta but during ascent they acquire successively higher lateral branches of the aorta up to the definitive renal arteries at the level of first lumbar disc (L1-L2 ). Failure of

regression of the inferior arteries is common and gives rise to accessory renal arteries. The kidneys come to lie in the retropcritoneal space high on the posterior abdominal wall. They are bean-shaped with their concave aspect pointing medially. The right kidney lies a variable distance l ower than the left in most subjects. owing to the presence of the liver. In quiet respiration they move up and down approximately 2-3 em but this may more than double with deep inspiration. The hilum of the kidney is a vertical opening on the medial aspect which contains the renal pelvis. The renal vein and one or two branches of the renal artery pass through the hilum to enter the kidney anterior to the renal pelvis, a further branch of' the renal


artery passing through the hilum posterior to the renal pelvis. The hilurn also contains fat, sympathetic nerve fibres and lymphatic channels that drain to the lateral aortic lymph nodes around the origins of the renal arteries from the aorta. tissue is divided into a peripherally placed cortex and the Renal central medulla. The medulla consists of 8-16 pyramids, which contain the descending and ascending tubules and the collecting duets. The apex of each pyramid projects into a calyx as a renal papilla. The cortex contains the glomeruli and the proximal and distal convoluted tubules. The superior end of the ureter expands to form the renal pelvis which divides into 2-4 major calyces, each of which divide into 2-l minor calyces. Each calyx is indented by a papilla, which is the apex of a medullary pyramid. The calyces therefore have a characteristic shape with a well-defined extension around the convexity of the papilla (the forniees). The minor calyces drain into a major calyx via a neck (infundibulum). Often at the renal poles (especially the upper pole) several papillae drain into one large calyx. These are referred to as compound calyces and are particularly vulnerable to damage from reflux nephropathy (Fig. 29.1). The ureter runs retroperitoneally down the anterior aspect of the psoas muscle, separated from it by the transversalis fascia. As it enters the pelvis it crosses over the anterior aspect of the common iliac artery bifurcation immediately in front of the sacroiliac joint. It descends along the lateral pelvic wall, just medial to the obturator

885


Previous Page

886

CONTENTS

Next Page

A TEXTBOOK OF RADIOLOGY AND IMAGING

space). The perinephrie space is one of a number of spaces within the abdomen and pelvis that are important determinants of the direction of disease spread. The transversal's fascia lines the inside of the abdominopelvie cavity, including the inferior aspect of the diaphragm, the posterior and medial surfaces of the anterior and lateral abdominal wall muscles respectively, the anterior aspect of the spinal column, psoas and paraspinal muscles and the superior aspect of the pelvic diaphragm. The major organs and their associated fascia develop within the space defined by the transversal's fascia. During embryol ogical development, as the kidneys ascend from the pelvis the surrounding (perinephrie) fascia forms a cone, with its apex superiorly. The perinephrie fascia anterior to the kidney is often referred to as Gerota's fascia, the posterior fascia as Zuekerkandl's fascia and the enclosed space as the perinephrie space (Fig. 29.2). It contains the kidney, the adrenal gland (anteromedial to the kidney on the left, superomedial to the kidney on the right), the upper ureter and the Fig. 29.1 Structure of the kidney. 1, Cortex; 2, compound calyx; 3, minor calyx; 4, medullary pyramid; 5, papilla; 6, renal sinus; 7, renal pelvis; 8, infundibulum of major calyx; 9, ureter.

i nternis, to the level of the isehial spine, from where it runs anteromedially until it enters the superolateral angle of the bladder base. The vas deferens crosses over the ureter, separating it from the bladder just before the ureters enter the bladder wall. The ureters run obliquely through the bladder wall for around 2 em. A fibrous capsule is closely applied to the renal cortex over the entire kidney apart from the hilum. The kidney is surrounded by perinephrie fat and lies within a space partly enclosed by layers of fascia, traditionally referred to as the perinephrie space (perirenal

nephrie fascia and the adjacent transversal's fascia which contains only fat (the posterior pararenal space). This space is of interest to the interventional radiologist as a potential area in which to kink a guide-wire and mistakenly position a drain intended for the renal collecting system. There is a more substantial space anterior to the perinephrie spaces between the anterior perinephrie fascia and the posterior layer of the peritoneum, the anterior pararenal space. This contains the pancreas and duodenum centrally, with the ascending colon on the right and the descending on the left. These organs are therefore in direct contact with the anterior perinephrie fascia. Laterally the anterior and posterior perirenal faseiae fuse with the l ateroeonal fascia at the fascia) trifureation. The lateroeonal fascia continues laterally and anteriorly to fuse with the parietal peri-



Previous Page

CONTENTS

Next Page

THE UROGENITAL TRACT: ANATOMY AND INVESTIGATIONS

toneum. Consequently the posterior pararenal spaces are continuous laterally with the fat-containing lateral extraperitoneal spaces (the properitoneal fat line) which lie between the parietal peritoneum and the transversalis fascia. Medially the posterior perinephric fascia fuses with the transversalis fascia over the paraspinal muscles, closing the posterior pararenal spaces medially. The line of fusion between the posterior perinephric and transversalis fasciae varies from cranial to caudal, being over the psoas muscle at the level of the superior renal pole and, more laterally, over the quadratus lumborum at the level of the inferior poles. Parts of both the perinephric and posterior perirenal spaces therefore lie directly over the psoas muscles, separated from them only by the transversalis fascia. This is an important route of spread of renal disease, particularly inflammatory processes, into the psoas muscle compartment and hence inferiorly into the pelvis and the iliacus muscle. The anterior perinephric fascia and the combined posterior perinephric and transversalis fasciae run medially across the midline to meet with the corresponding fasciae on the contralateral side. This would imply the perinephric spaces can communicate with each other across the midline. Observation of the extension of disease within the perinephric space, however, shows that this does not occur as readily as would be expected if there was free communication. This problem is overcome by considering the layers of fasciae as laminated structures that arise embryologically from more than one layer and in certain areas contain potential spaces and communications. The anterior and posterior perirenal fasciae thicken medially and to some extent fuse, acting as a relative barrier to medial extension of disease. There is a potential space within the l aminated anterior perinephric fascia (the anterior interfascial or retromesenteric space) that can eventually transmit disease processes across the midline. The posterior perinephric fascia is also laminated and contains a potential space (the posterior interfascial or retrorenal space). Fluid in either of these potential spaces may track to the other and also via the fascial trifurcation into the l ateroconal fascia. Fibrous septae are found within the perinephric space running from the renal capsule to the perinephric fascia and act as potential conduits between the perinephric and interfascial potential spaces.

887

Superiorly (the apex of the cone of the perinephric fascia) the perinephric space is closed. Since the perinephric fascia originally develops as a cone around the kidneys, as they ascend it would be expected that the perinephric spaces remain open inferiorly. This, however, remains a matter of controversy and it has been suggested that the anterior and posterior layers of perinephric fascia fuse i nferiorly in the iliac fossae, closing the perinephric spaces and forming a combined laminated fascia that continues inferiorly along the anterolateral aspect of the psoas muscles. In routine radiological practice the distinction does not appear to be important, as both descriptions offer mechanisms for disease to spread from the region of the kidneys inferiorly into the pelvic extraperitoneal spaces. The relationships of the kidneys are as follows. Posteriorly the kidneys lie over the psoas and paraspinal muscles and the 12th (and on the left the 11 th) ribs. Anterosuperiorly the kidneys are in contact with the adrenal glands and anteroinferiorly with the jejunum. Anterolaterally the kidneys abut the liver, and below it the hepatic flexure on the right (Fig. 29.3A); the spleen and below it the splenic flexure on the left (Fig. 29.3B). The other important anterior relationship is with the duodenum anteromedially on the right and with the pancreas and stomach anteromedially on the left. The arterial supply and venous drainage of the kidneys are both extremely variable. The renal arterial supply is most commonly provided by a single lateral branch of the aorta, the renal artery, originating at the level of the first lumbar disc. The right renal artery has the longer course and runs posterior to the inferior vena cava. The renal arteries run posterior to their respective renal veins. The renal artery gives off the inferior adrenal and the renal capsular arteries and then classically divides into three branches as it enters the renal sinus, two running anterior to the renal pelvis, one posterior. The main renal branches subsequently divide into the interl obar arteries, which run centrifugally within the cortical tissue between the medullary pyramids. At the level of the base of the pyramids the interlobar arteries give off the arcuate arteries, which run along the line of the corticomedullary junction. The venous drainage mirrors the arterial tree except that the kidneys are drained by five or six major venous branches, which most commonly combine to form a solitary renal vein that drains


Fig. 29.3 Anterior relations of (A) the right kidney, and (B) the left kidney. 1, Hepatic flexure; 2, small intestine; 3, right adrenal gland; 4, liver; 5, duodenum; 6, left adrenal gland; 7, spleen; 8, stomach; 9, pancreas; 10, splenic flexure.


CONTENTS

Previous Page

888

Next Page


i nto the inferior vena cava, also around the level of the first lumbar disc. The lymphatic drainage of the kidneys is to the lateral aortic l ymph nodes around the origin of the renal arteries.

The bladder The transversalis fascia continues inferiorly into the pelvis, where it i s often referred to as the pelvic fascia. It lines the walls of the pelvis and overlying muscles, notably including the obturator i nternis laterally and the pelvic diaphragm inferiorly. The parictal peritoneum of the abdomen also continues inferiorly and is reflected over the pelvic viscera and fat-containing spaces of the pelvis, leaving most of the pelvic organs outside the peritoneal cavity. The umbilicovesical fascia is a large triangular fascia with its apex at the umbilicus. It extends inferiorly, dividing the prevesical space anteriorly from the perivesical space posteriorly and centrally. The cave of Retzius is the part of the prevesical space i mmediately posterior to the pubis. The prevesical space continues posterolaterally into the paravesical spaces (Fig. 29.4). These spaces are continuous superiorly with the properitoneal fat stripes (lateral extraperitoneal spaces of the abdomen) and the posterior pararenal spaces. The perivesical space contains the bladder, the urachus and the seminal vesicles. It is limited posteriorly by the reetovesical septum (reetovaginal in the female). The pelvic organs are supported by the pelvic floor, which is formed by the pelvic diaphragm. This separates the pelvic cavity superiorly from the perineum inferiorly. The pelvic diaphragm consists of the large levator ani and the small coccygeus muscles and their covering fascia. The puboreetalis sling is a prominent part of the levator ani which runs around the lateral and posterior aspects of the anorectal junction and sweeps forward along the lateral margins of the prostate and inserts into the posterior aspect of the body of the pubis (Fig. 29.5).

Fig. 29.5

The relationships of the prostate.

The urinary bladder lies immediately posterior to the pubic bones. When full it is roughly spherical hut when empty it approximates to a pyramidal shape, having an apex and four roughly triangular surfaces: posterior (base), superior and two inferolateral. The apex lies immediately behind the upper margin of the symphysis pubis and gives rise to the urachus, which is the fibrous remnant of the allantois. The urachus runs superiorly in the extraperitoneal fat to the umbilicus as the median umbilical ligament. The superior surface of the bladder is covered in peritoneum and loops of ileum and/or sigmoid colon. The inferolateral surfaces relate anteriorly to the retropubic fat and pubic bones and posteriorly to the obturator i nternis (superiorly) and the levator ani (inferiorly). The base of the bladder receives the ureters at its superolateral angles and gives rise to the urethra at its inferomedial angle. This rests on the prostate inferiorly, and the muscle fibres of the bladder wall run continuously into the prostate. The muscle fibres are thickened around the origin of the urethra as the (internal) sphincter vesicae. This area of the bladder is referred to as the bladder neck. The bladder mucosa is thrown into folds when the bladder empties.



Previous Page

CONTENTS

Next Page

THE UROGENITAL TRACT: ANATOMY AND INVESTIGATIONS except over the base where it is firmly adherent to the underlying muscle layers. This region is referred to as the trigone. The superior border of the trigone is defined by a muscular ridge running between the ureteric orifices (the interureteric ridge). The upper part of the base is covered by peritoneum and lies anterior to the rectovesical pouch (uterovesical pouch in the female), the lower part is subperitoneal and separated from the rectum by the vasa deferentia and seminal vesicles (vagina in the female). The arterial supply is from the superior and inferior vesical arteries which arise from the anterior division of the internal iliac arteries, as described below. The common iliac arteries terminate at the pelvic inlet anterior to the sacroiliac joints by dividing into the external and internal iliac arteries. The external iliac artery runs along the medial border of the psoas muscle, following the pelvic brim to exit the pelvis under the inguinal ligament to become the common femoral artery as it enters the leg. The internal iliac artery enters the pelvis anterior to the sacroiliac joint with the ureter running inferiorly and crossing it anteriorly. At the level of the upper border of the sciatic notch the artery bifurcates into anterior and posterior divisions. Further branches are variable but classically i nclude the umbilical artery (giving rise to the artery to the vas deferens and the superior vesical artery), the inferior and middle vesicle arteries, the internal pudendal, inferior gluteal and uterine arteries. Branches of the posterior division of the internal iliac artery are essentially musculoskeletal, being the iliolumbar, lateral sacral and superior gluteal arteries. The venous drainage of the bladder is to the vesical venous plexus, which communicates inferiorly with the prostatic venous plexus and drains superiorly into the internal iliac veins. Lymphatic draina g e is into the internal and external iliac lymph nodes lying along the course of the respective arteries. These in turn drain into the common iliac nodes. The medial chains of the external iliac lymph nodes (obturator nodes) lie next to the lateral pelvic walls at the level of the acetabulun. They are frequently involved early in metastatic spread of bladder or prostatic cancers and are therefore often referred to as sentinel nodes. Small pelvic lymph nodes may be seen on CT and MR1 and are regarded as normal when less than 10 mm in diameter.

and lateral aspects of the proximal urethra. The central zone lies against the posterior aspect of the proximal urethra and encases the transition zone posteriorly and superiorly. The peripheral zone surrounds the central zone and the distal urethra. Radiologically the central and transitional zones are indistinguishable and are commonly referred to collectively as the central gland. In young males the largest component of the prostate is the peripheral zone (70%), the transitional and central zones constituting 5% and 25% of the gland, respectivcly . The zonal architecture in young males cannot be clearly seen radiologically but it becomes more evident with advancing years. The transitional zone is the site of development of benign prostatic hyperplasia and with age increases in size and eventually dominates the gland. The prostate i s supplied by the inferior vesical and middle rectal arteries. Venous drainage is into the periprostatic venous plexus which communicates inferiorly with the dorsal vein of the penis and superiorly with the perivesical venous plexus, which in turn drains into the internal iliac veins. Lymphatic drainage is into the i nternal iliac nodes. The seminal vesicles are paired lobulated sacs, approximately 5 cm long in the young adult male (becoming atrophic with age), lying on the posterior surface of the bladder. They run inferomedially into a narrow duct which joins with the duct of the vas deferens to form the ejaculatory duct. The ejaculatory duct pierces the prostate to enter the prostatic urethra on the verumontanum. The male urethra is approximately 20 cm long and runs from the bladder neck to the external meatus, where it dilates to form the fossa terminalis (navicular fossa). It can be considered in three sections: prostatic, membranous and penile. The prostatic urethra is 3 cm long and runs from the bladder neck through the prostate. It is normally the widest part of the urethra. Its most notable feature is a l ongitudinal mound (verumontanum) along the posterior wall. At the centre of the verumontanum is a small diverticulum (prostatic utricle) which receives the ejaculatory ducts at its lateral margins. The urethra exits from the prostate just above the apex to become the membranous urethra. The membranous urethra is the shortest ( I-1.5 cm) and least distensible section of the urethra. It is enclosed by the sphincter urethrae (external sphincter), which, along with the deep transverse perineal muscles and their covering fasciae, form the urogenital diaphragm. The urogenital diaphragm closes the arch formed by the pubic hones and separates the cavity of the pelvis from the perineum below. The penile urethra can be subdivided into bulbar (proximal) and pendulous (distal) sections which relate to the root and body of the penis, respectively. The root is made up of three masses of erectile tissue, the centrally placed bulb and paired laterally placed crura. The bulb is attached to the underside of the urogenital diaphragm: it encases the proximal penile urethra and is surrounded by the bulbospongiosus muscle. The crura run along the posteromedial aspects of the pubic arch and are covered by the ischiocavernosus muscles. The bulb and the crura continue anteriorly into the body of the penis as the corpus spongiosum and the dorsolaterally placed paired corpora cavernosa, respectively. The urethra continues into the body of the penis encased by the corpus spongiosum and its distal expansion (the glans). The scrotum is a pouch of skin that contains the testes, epididymis and spermatic cord. It has three layers from superficial to deep: skin, dartos muscle and Collcs' fascia (continuous anterosuperiorly with the anterior wall membranous fascia-Scarpa's fascia). It is divided into two compartments by a midline fibrous


The prostate has roughly the shape of an inverted rounded cone/ pyramid. It is a glandular organ enclosed by a fibrous capsule. It contains a substantial fibromuscular component which surrounds the prostatic urethra. It has a base, apex, anterior, posterior and two lateral surfaces. Superiorly its base relates to the bladder. Inferiorly i ts apex lies on the upper surface of the urogenital diaphragm. Anteriorly it relates to extraperitoneal fat in the retropubic space (cave of Retzius). Posteriorly the prostate is separated from the rectum by the rectovesical septum (fascia of Denonvillier), which is the fused walls of the inferior end of the rectovesical pouch. Laterally the prostate relates to the puborectalis, which is an important radiological landmark seen running backwards close to the lateral margin of the prostate and encircling the anorectal junction. The neurovascular bundle of the prostate lies in the angle between these two structures (Fig. 29.5). The classical lobar anatomy of the prostate is unhelpful. The zonal anatomy is more useful clinically and pathologically. Three zones are described histologically: central, transitional and peripheral. The transition zone has a C shape lying against the anterior

889


Previous Page

890

CONTENTS

Next Page


septum (median raphe). Each compartment contains a testicle, epididymis and proximal spermatic cord. The testicles originate on the posterior abdominal wall around the level of the first lumbar vertebra. They descend retroperitoneally through the abdomen and with their associated neurovascular structures, lymphatics and ducts exit the pelvis via the inguinal canal around the 7th to 8th month of uterine life. They are preceded by a peritoneal diverticulum (processes vaginalis), remnants of which may persist into adult life. As they leave the pelvis they acquire a covering from each layer of the anterior abdominal wall which surrounds the spermatic cord and the testes. The transversalis fascia gives rise to the internal spermatic fascia, the internal oblique muscles to the cremasteric muscle and fascia and the external oblique aponeurosis to the external spermatic fascia. Deep to these three layers is the tunica vaginalis. This is a serous membrane (derived from peritoneum) with two layers: parietal, against the i nternal spermatic fascia, and visceral, which closely invests the testis and the epididymis. The tunica vaginalis envelops the testicle and applies it to the posterior scrotal wall. There is a potential space between these two layers which extends a little way along the spermatic cord and normally may contain up to 1-2 ml of serous fluid. In disease states a more substantial collection may be seen (hydrocele). The outermost layer of the testicle is the tunica albuginea, a dense fibrous capsule. Posteriorly there is a vertical linear extension of the tunica albuginea into the testicular parenchyma (mediastinum of the testis). The testicle is divided into 200-400 lobules by numerous trabeculae that extend from the mediastinum. The l obules consist of interstitial tissue (which includes the androgenproducing Leydig cells) and 2-4 seminiferous tubules (which i nclude the supporting Sertoli cells) that drain via a common tubule. These converge into a network of tubules at the mediastinum (the rete testis) that drain via 10-15 efferent tubules out of the testicle i nto the epididymis. The epididymis consists of a convoluted common tubule with supporting tissue. Its upper, mid and lower portions may be referred to as the head, body and tail. Drainage is from the tail i nto the vas deferens. An embryological remnant a couple of millimetres long (so-called appendix or hydatid) may be identified related to the upper pole of the epididymis (derived from the mesonephric duct) or the testis (derived from the milllerian duct). The testicle is supplied by an end artery, the testicular artery, which originates from the aorta just below the renal artery. This runs inferiorly along the posterior abdominal wall on the surface of the psoas muscle and enters the scrotum via the inguinal canal within the spermatic cord. At the mediastinum of the testis the dominant testicular artery (sometimes referred to as the inferior testicular artery) divides into several capsular arteries which run around the testicular parenchyma just deep to the tunica albuginea. This layer has been referred to as the tunica vascularis. Branch arteries run from this layer towards the mediastinum, close to which further branching occurs, running hack in the opposite direction. There may be an additional branch (transtesticular artery) which enters the mediastinum directly from the testicular artery and is often accompanied by a large vein. Before reaching the mediastinum the testicular artery also gives rise to the epididymal artery and a minor testicular branch (internal testicular artery), which enters the upper pole of the testicle and contributes to the capsular branches.

Venous blood drains from the posterior border of the testicle into an extensive venous plexus (the pampiniform plexus). The pampiniform plexus ascends within the spermatic cord with the testicular artery and the vas deferens. Classically the branches of the pampiniform plexus unite to form a single testicular vein (the internal spermatic vein) as it enters the abdomen via the inguinal canal. The testicular vein runs superiorly along the posterior abdominal wall in the retroperitoneum and terminates in the renal vein on the l eft and in the inferior vena cava just below the renal vein on the right. This, however, is extremely variable and often there are multiple collaterals to the renal or retroperitoneal veins, multiple terminal ostia and multiple testicular veins. The spermatic cord runs from the testicle to the abdominal cavity via the inguinal canal. It contains the vital supply and drainage structures of the testicle surrounded by continuations of the three deep fascial layers of the scrotum ((he internal spermatic, cremasteric and external spermatic fasciae). The contents include the testicular artery, pampiniform plexus, lymphatic vessels and the vas deferens. The veins are routinely visualised and are normally no more than 1-2 mm maximum diameter. The lymphatic vessels follow the testicular artery and vein retroperitoneally to terminate in the first-echelon nodes (sometimes referred to as the sentinel lymph nodes) in the upper abdomen. On the left these are immediately lateral to the para-aortic lymph nodes and medial to the renal lymph nodes around the level of the first to the second lumbar vertebrae, while on the right they are in the paracaval region just below the renal vessels around the level of the first to the third lumbar vertebrae. Drainage from these nodes is into the nodes between the renal vessels and the aortic bifurcation (paracaval from the right. paraaortic from the left and both potentially into the aortocaval centrally). Drainage from these nodes is superiorly to the para-aortic and paracaval nodes above the renal hila and extending retrocrurally into the mediastinum. In a minority the lymphatics drain directly into nodes at the level of the aortic bifurcation and occasionally directly into the common iliac nodes. Testicular lymph node drainage is usually to the ipsilateral nodes but bilateral drainage may occur, more often from the right than the left. These considerations are important when staging testicular malignancy. The vas deferens is a muscular duct responsible for the transport of the sperm from the testicle to the urethra. It originates from the tail of the epididymis and leaves the scrotum in the spermatic cord. After passing through the inguinal canal it enters the pelvis at the deep inguinal ring, hooks round the lateral aspect of the inferior epigastric artery and runs posteroinferiorly on the lateral wall of the pelvis (on the obturator internis) to the level of the ischial spine, where it turns infcromedially to run along the posterior aspect of the bladder immediately anterosuperior to the distal ureters. The distal vas dilates to form the ampulla. This curves inferiorly where it joins the duct of the seminal vesicle to form the ejaculatory duct. which joins the urethra at the verumontanum.

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ I NVESTIGATIONS

The standard plain radiographic imaging of the urinary tract is the KUB (kidneys, ureters and bladder), which consists of a full l ength abdominal film and an upper abdominal (cross-kidney) film ( Fig. 29.6). The films are taken with the patient supine using a low


Previous Page

CONTENTS

Next Page


Box 29.1

891

Calcification on the KUB

Urinary tract

Renal: calculi, renal cell carcinoma, tuberculosis, arterial (atheroma or aneurysm) Ureter: calculi, tuberculosis, schistosomiasis Bladder: calculi, schistosomiasis, transitional cell carcinoma 5

Outside the urinary tract

Musculoskeletal: costal cartilage calcification Hepatobiliary: gallstones, hepatic granuloma Pancreas: chronic pancreatitis Adrenal: tuberculosis, Addison's disease Spleen: granuloma Aorta: atheroma, aneurysm Venous: phlebolith Uterine: fibroid Lymphatic: calcified lymph nodes (presumed postinfective)

There are numerous causes of calcification visible on the plain radiograph, with considerable overlap in their appearances, and a specific diagnosis is often not possible. The more common causes arc listed in Box 29.1. The KUB is most usefully employed as part of an intravenous urogram (IVU) or to follow up a previously proven calculus.

Intravenous urography The intravenous urogram (IVU) is the classic routine investigation of' uroradiology. With the advent of ultrasound its role is now much diminished and its future is the subject of considerable debate as other modalities, particularly spiral computed tomography (CT), become more widely available. At this date, however, it is still widely used and requires consideration. Currently the main indications are the investigation of persistent or frank haematuria, renal and ureteric calculi (particularly prior to endourological procedures), ureteric fistulas and strictures and complex urinary tract i nfection (including tuberculosis). The IVU consists of a series of plain films taken after adminis-


Fig. 29.6 The KUB. (A) Full length and (B) cross kidney films from a KUB examination. Several phleboliths are seen in the pelvis. They are characteristically smooth small rounded calcific areas with a tiny central radiolucency.

voltage technique (60-65 kV) to maximise soft-tissue contrast. The full length filth is taken in inspiration using a 35 x 43 cm cassette positioned with the lower border at the symphysis pubis to ensure the urethra (particularly the prostatic urethra) is included on the film. The cross-kidney filth is taken in expiration using a 24 x 30 cm cassette with the lower border 2.5 cm below the iliac crests. The outline of several anatomical structures can be seen on the KUB, including renal, psoas and bladder outlines, much of the axial skeleton, the bowel gas pattern and the lung bases, all of which should he routinely inspected. A range of pathology may he observed in these and other structures but the KUB is a relatively unreliable diagnostic tool and, despite the occasional incidental finding in other systems, its principal use is in the assessment of urinary tract calculi. It is, however, extremely unreliable in the diagnosis of ureteric calculi, with an accuracy of only around 50%.

tration of an intravenous injection of a water-soluble iodinecontaining contrast medium. There is considerable variation in the exact details of how an IVU is performed in different departments, although there should be general agreement on the underlying principles. Traditionally the patient was prepared with a period of 4 h starvation and fluid deprivation and the bowel purged with a strong l axative. Bowel preparation is now generally regarded as unhelpful and it was unpleasant for the patient: it has now largely been dispensed with. Occasionally the patient will feel nauseated after the I VU injection and rarely there will be a severe reaction with the need for cardiovascular and occasionally cardiopulmonary support.

With this in mind, it seems reasonable to persist with avoidance of food for 2-4 h prior to the procedure. There has been a considerable body of work on the potential adverse effects of intravenous contrast on renal function and its relationship to the patient's fluid status. Traditionally fluid was restricted prior to the IVU in order to i mprove opacification oh the collecting system. However, it has long been accepted that dehydration is associated with an increased risk of nephrotoxicity, which may be permanent in patients with diabetes mellitus, myeloma, hyperuricaemia, sickle-cell disease and pre-existing renal disease. The risk of irreversible damage to renal function in a previously healthy kidney due to the contrast injection i s very low. This appears to he further reduced with avoidance of


Previous Page

892

CONTENTS

Next Page


̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ Fig. 29.7 Postcontrast IVU films. (A) Immediate, (B) 5 min, (C) 15 min film with compression producing calyceal distension; (D) full length release; (E) full length postmicturition. The phleboliths noted on the plain films (Fig. 29.6A) are nicely shown to lie outside the urinary tract.


CONTENTS

Previous Page

Next Page

THE UROGENITAL TRACT: ANATOMY AND INVESTIGATIONS dehydration. Modern non-ionic contrast agents do not provoke an

893

of these a routine inquiry about previous contrast exposure and allergy is recommended before administering the contrast medium.

osmotic diuresis and the degree of opacification is unlikely to be significantly altered by dehydration. Fluid restriction should there-

The injection should be through some form of indwelling cannula

fore he avoided and if there is a risk that the patient is dehydrated

or needle that can he taped into place for the duration of the investi-

before the IVU this should be corrected first.

gation. This allows emergency treatment to he administered if

The classical series of plain films (immediate, 5 and 15 min, full

required, or a further injection of contrast if opacification is seen to

l ength release and postmicturition) is described, with mention of some of many potential modifications. A preliminary postmicturi-

be inadequate. There should also be a doctor (usually the radiol ogist) available in the X-ray department throughout the investiga-

tion plain film (KUB) is performed. This should he examined to

tion. Most adverse events are likely to take place within the first

check exposure factors, centring and obvious pathology, particu-

few minutes after the injection. Emergency drugs, oxygen and resuscitation equipment should also be readily available. Treatment

l arly urinary tract calcification. Intravenous contrast is given relatively rapidly by hand. The standard dose is 50 ml of 350-370

of contrast reactions is discussed on page 928, as is the specific

strength water-soluble contrast. Some understanding of the under-

problem of metformin interaction. A cross-kidney film is taken immediately after contrast injection

l ying structure of water-soluble contrast agents is desirable and a knowledge of potential adverse reactions and their treatment is

and at 5 min after the injection (Fig. 29.7). Abdominal compres-

i mperative. These issues are discussed under water-soluble contrast ( p. 926). At this point it is worth emphasising some safety features.

sion is applied as soon as the 5 min film has been taken (a variety of

Although modern contrast medium is exceptionally safe, there is a

promote distension of the pelvicalyceal systems, optimising their

small risk of serious reactions. The most dangerous of these are the

visualisation. A further cross-kidney film is taken at 12-15 min to

anaphylaeloid-type hypersensitivity reactions. To minimise the risks

demonstrate this.

devices are available for this) to inhibit ureteric drainage and

Plain films

Additional obliques or tomograms

To assist the location of potentially intrarenal opacities. Rarely required; ultrasound and other imaging manoeuvres usually preferable

Nephrogram

Thick slice tomogram

To improve definition of the renal outlines

Omit along with the 5 min film and take a solitary 3 min film

To reduce radiation dose

5 min film

Second injection of contrast

• i mprove opacification of the pelvicalyceal systems if inadequate

1 5 min compression film

Series of 1 cm thick tomograms

• differentiate between overlying shadows and filling defects within the collecting systems

• delineate the renal outlines when inadequately seen (better done with

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ ultrasound)

1 5 min release film

Additional bladder views

When the bladder is poorly filled on the release film (as is often the case) delayed films of the fuller bladder may be performed. When equivocal filling defects are seen in the bladder area oblique films may be performed The above additional bladder views are rarely indicated as they increase the radiation burden to the patient and the relevant clinical problems are better answered by ultrasound or cystoscopy. However, occasionally a small suspected calculus in the distal ureter may be confirmed with the appropriate oblique

Bladder area only

I f the upper tracts have already been adequately imaged then imaging the bladder area alone will reduce the patient's radiation burden

Omit

The preceding films may have already provided all the information required from the investigation

Additional view

Where the renal pelvis is dilated, contrast may be slow to pass into the ureter; this can be accelerated by positioning the patient prone when the heavier contrast will run anteroinferiorly into the ureter, often to the level of the obstruction. Simply asking the patient to sit or stand for a few minutes first may improve the result

Erect images

Additional radiograph or fluoroscopy

If it is difficult to determine whether or not there is a small ureteric calculus, an erect oblique radiograph of the ureter or screening the ureter in this position may be useful on rare occasions

Frusemide IVU

Administration of 20mg of frusemide i ntravenously after the 15 minute film with a further film 15 minutes later

If suspected pelviureteric junction obstruction is being investigated and there i s no evidence of this on the standard IVU, this manoeuvre can be performed. It may provoke hydronephrosis and pain. It is rarely necessary if the patient is to be investigated with radionuclide renography, as is often the case in this situation

Full length postmicturition film

Prone full length film


Previous Page

894

CONTENTS

Next Page


Compression is routinely omitted in children and should be avoided in a number of situations. It is contraindicated if there is an aortic aneurysm. If the patient's abdomen is tender it should also be avoided. This contraindication includes recent abdominal surgery and the acute painful abdomen, the latter including renal colic. In the case of renal colic, even if the patient can tolerate compression it is best avoided because it is unlikely to contribute to the investigation and potentially may exacerbate the tendency of the acute obstruction to provoke extravasation of urine at the fornices. If this does occur, however, it is not a cause for undue alarm as it is not uncommonly encountered spontaneously in renal colic and does not appear to be associated with adverse sequelae. If the 15 min film is satisfactory the compression is removed and a full length film is taken immediately to offer the best opportunity of demonstrating the ureters. The patient is then asked to empty the bladder and a further full length film is taken to demonstrate drainage of the upper tract and the postmicturition bladder volume. This classical series can he modified to deal with particular circumstances, to attempt to increase the sensitivity of the procedure or to reduce the radiation dose to the patient. Some of the most frequently employed modifications are detailed in Table 29.1. Three circumstances in particular are worthy of further comment. When there is significant acute obstruction, usually due to calculi, there is delay in opacification of the collecting system. The delay may be considerable, up to 24 h or more. It is then necessary to perform the minimum number of additional films. The time interval between films traditionally is approximately doubled, with films taken at 0.5, 1, 2, 4, 16 and 24 h, as necessary; however, in order to minimise the radiation exposure, if there is no opacification of an acutely obstructed kidney at 30 min it is usually unhelpful to perform the next film before around 4 h after contrast injection. A further manoeuvre to minimise radiation dose in patients with a strong clinical suspicion of ureteric colic is to omit all films after contrast until a full length 15 min film is performed. Patients with proven or suspected ureteric calculus may require one or more follow-up IVUs. These should have the minimum number of films required to answer the specific question, for example a full length plain and 15 min postcontrast film may be sufficient. It is very rarely necessary to perform an IVU on a pregnant patient. If it is required, the radiation exposure should be minimised. The collecting systems in pregnancy are capacious and the ureters exhibit poor peristalsis. Consequently a single full l ength preliminary film and a delayed solitary full length film around 30-45 min may well be enough. The stereotypical appearances of the normal IVU are as follows. It takes approximately 12-20 s for contrast to reach the renal arteri es following its intravenous injection. At this stage its concentration is maximal in the vascular compartment; however, this falls rapidly as the contrast medium begins to escape into the extracelluJ ar compartment and also undergoes rapid glomerular filtration and enters the renal tubules. In the first minute of the IVU healthy kidneys (assuming a normal cardiovascular system) show diffuse enhancement. This is referred to as the nephrogram. During the nephrogram phase the renal size (normally at least three vertebrae i n length but no more than four) and outlines arc best seen. In roughly the first half minute, contrast in the vascular compartment dominates and therefore the cortex is more enhanced than the medulla: this differentiation is sometimes visible on the immediate film of the IVU series (but regularly visible on CT performed at this stage). In the second half minute, contrast in the tubules

increases and enhancement of the kidneys is more diffuse. Contrast begins to appear in the calyces from around 1 min. From this time, in good quality films, contrast may also be visible in the collecting ducts as fine linear opacities running along the medullary pyramids towards the calyces. This may be referred to as pyelotubular stasis or, when less defined, as the medullary or pyramidal blush, and is a normal phenomenon. Contrast in the normal calyces will begin to drain immediately i nto the pelvis and ureter and this phase may be referred to as the pyelogram. The successful application of compression impedes ureteric drainage and distends the pelvicalyceal system, producing optimal visualisation of the pelvicalyceal system around 12-15 min. After compression is released there is a transient increase in flow down the ureters and the release film offers the best chance of demonstrating the ureters. The normal ureters exhibit continual peristalsis and on a single film it is uncommon to demonstrate the entire l ength of both (or even either) ureters. They will often demonstrate smoothly narrowed areas (especially at the pelviureteric junctions and as they cross the iliac vessels in the pelvis) and more relaxed capacious areas. This is normal and how vigorously further efforts are made to demonstrate the entire length of the ureter depends on the clinical situation. In most situations partial visualisation of a non-obstructed but otherwise normal ureter is acceptable. If there is persistent hacmaturia or abnormal cytology further efforts to demonstrate the ureter are required and may include prone films, repeat IVU after a short interval and/or retrograde pyclography.

Ultrasound of the urinary tract can be considered under three headi ngs: ultrasound of kidneys and bladder, ultrasound of the male genitalia and ultrasound of the prostate. Routine ultrasound of the urinary tract consists of examination of the kidneys and the full urinary bladder and is probably now the most frequently performed radiological investigation of the urinary tract. The commonest indications are urinary tract infection and prostatism. Other indications include haematuria, obstruction, masses, calculi, congenital abnormalities. renal failure and assessment of transplants. It is frequently used to guide diagnostic or therapeutic procedures. Assessment of the postmicturition bladder volume is often also performed as part of the same examination, especially for prostatism. Most modern probes emit a broad band of ultrasound that can be biased towards high (for thin patients) or low frequency, usually within the 2-6 MHz range for abdominal and pelvic work. The position of the normal kidneys is extremely variable and the optimal window for visualisation has to be determined for each patient. Conventionally each kidney is examined subcostally from the loin with the patient in the lateral position, the side of interest uppermost. Often, however, complementary views can be obtained with the patient supine with a more lateral or intercostal approach. On ultrasound the normal renal cortex appears slightly hypoechoic compared to the liver and spleen. It lies as a peripheral rim with normal invaginations (columns of Bertin) projecting inwards between the medullary pyramids (Fig. 29.8). The pyramids are markedly hypoechoic compared to the cortex in the first 6 months of life, becoming less hypoechoic with ageing and in some disease states. Corticomedullary differentiation is therefore most obvious i n young adults and children, and the hypoechoic pyramids are Kidneys and bladder



Previous Page

CONTENTS

Next Page


Fig. 29.8 Normal renal ultrasound; note the relatively hypoechoic pyramids and markedly hyperechoic fat containing sinus (centrally) compared to the cortex. occasionally mistaken for the dilated calyces of a hydronephrosis. Arcuate arteries can be identified in 25% of the adult population as focal hyperechoic areas at the corticomedullary junction. The centrally placed renal sinus contains the calyces, infundihula, part or all of the renal pelvis, blood vessels, lymphatics, fibrous tissue and fat. It is the most hyperechoic part of the kidney, owing to the presence of fat. The normal renal pelvis may be seen within the sinus fat as an echo-free structure, especially with a full bladder in the female. Often it will decrease in size when the patient empties her bladder. This, and the absence of dilated calyces separating the sinus fat, allows differentiation from hydronephrosis. Most normal adult kidneys have a maximum length of 10-12.5 cm, although substantial numbers of normal kidneys may be seen within the 9-13.5 cm range. Kidneys are roughly related to the patient's size and show some decrease in length with age (especially above

895

decrease substantially with age and he associated with an increase i n the central sinus fat, a phenomenon also visible on IVU. Doppler assessment of the renal artery flow in order to identify significant renal artery stenosis may be requested as part of the investigation of hypertension (discussed further under renal arterial disease) and early in the assessment of renal transplant viability. The pelviureteric and vesicoureteric junctions may be seen on ultrasound, especially when diseased, but otherwise the normal ureter is not routinely demonstrated. Routine examination of the bladder requires it to be moderately full. The normal bladder has a triangular shape in the sagittal plane, and that of a square with the corners rounded off in the transverse plane. The normal wall thickness is 2-3 mm when the bladder is moderately full (Fig. 29.9). In the investigation of bladder outflow obstruction an assessment of postmicturition residue is often requested. This is fraught with inaccuracies, related as it is to the bladder volume initially (both over- and underdistension may preclude normal emptying) and variability in the shape of the bladder. It is, however, a broadly useful tool and simple to employ. The bladder is assumed to be roughly elliptical and the maximum transverse, sagittal and craniocaudal dimensions are measured. There is usually a volume function on the ultrasound machine that will generate a volume from these figures, or they can he multiplied together and by the constant for an ellipse, which is 0.533 (or roughly a half). The prostate is routinely identified at the bladder base, especially when enlarged. The seminal vesicles are also often seen posterior to the bladder. Male genitalia This investigation is most commonly requested for assessment of the scrotum, occasionally of the penis. The major indication for scrotal ultrasound is the investigation of mass l esions, including a hydrocele, when there may be concern about an underlying tumour. Ultrasound of the scrotum and groins may also be performed to localise and assess an ectopic testicle. A high-frequency probe, sometimes disparagingly referred to as a small-parts probe, is used. Again, modern probes will emit a broad range of frequencies within approximately the 7-12 MHz range. The normal testicle is ovoid, approximately 3.5 cm in length and 3 cm in diameter. It has a speckled intermediate echogenicity (Fig. 29.10). The mediastinum (the invagination of the tunica albuginea) along the posterior aspect is seen as an echogenic line parallel to the epididymis. The lobular structure of


80 years). A difference of 2.0 cm or more between the two kidneys raises the possibility of unilateral disease. Measurements of renal length on ultrasound are of the order of 0.5 cm less than on IVU, which suffers from radiographic magnification. Although areas of scarring and diffuse cortical loss are appreciated subjectively, it is difficult to measure cortical thickness objectively with any degree of accuracy. In young healthy adults cortical thickness is of the order of 2.5-3 cm at the poles and 1.5-2.0 cm elsewhere. This may

Fig. 29.9 Normal bladder ultrasound; note the thin smooth wall and the shape approximating to a rounded-off square in the transverse view (A) and a rounded-off triangle in the sagittal view (B).


Previous Page

896

CONTENTS

Next Page


Fig. 29.10 Scrotal ultrasound. (A) Longitudinal image showing the intermediate slightly speckled appearance of the normal testicle. The thin ecbogenic line represents the mediastinum of the testicle. (B) Longitudinal view showing a prominent vascular channel (linear echopoor structure)-normal variant. (C) Longitudinal view showing the body of the epididymis as a thin echo-poor line posterior to the testicle. Transverse view of the testicle through the mid (D) and upper (E) parts, the latter showing the bead of the epididymis as a small area of soft tissue adjacent to the upper pole of the testicle.


the testicle is very occasionally identifiable, with the interlobular septations visible as thin echogenic lines. The head (upper pole) of the epididymis lies lateral to the superior pole of the testicle, appears slightly more echogenic than the remainder of the epididymis and has a maximum diameter of 5-I5 mm diameter. The epididymis tapers down through the body to the tail, which has a maximum diameter of 1-3 mm. Occasionally, small protuberances are seen at the head of the epididymis or the superior pole of the testicle and are the appendices of these structures. The scrotal skin and subcutaneous tissue is seen as a hyperechoic structure 5-7 mm thick. Ultrasound of the penis is occasionally performed to assess mass lesions of the shaft, most commonly Peyronie's disease, although the clinical usefulness is debatable. Doppler ultrasound of the penis i s most commonly performed in the assessment of impotence and is discussed under the relevant section.

Prostate The prostate can be seen at the bladder base on routine transabdominal scanning of the bladder. Apart from a rough idea of its size, little useful information is obtained using this route. Prostate ultrasound is performed using a dedicated medium- to high-frequency cavity probe (predominant frequencies around 4-7 MHz) positioned within the rectum. It is used in the investigation of suspected prostatic cancer, usually with multiple ultrasound-guided biopsies. It can also be used to assess the prostate

and seminal vesicles in infertility and (although its value is questionable) in prostatitis. The normal prostate is roughly beanshaped, the concavity of the bean facing posteriorly. The transitional and central zones have the same intermediate echogenicity and are inseparable on ultrasound (Fig. 29.11). They can be referred to as the central gland, and increase in volume and heterogeneity with age as benign prostatic hyperplasia develops in the transitional zone. The peripheral zone is seen as an echogenic l ayer lying posteriorly. It is the site of most prostatic tumours, which usually appear as relatively echo-poor areas. Normal prostatic dimensions in the adult male are approximately 4 cm in the craniocaudal and transverse planes and 3 cm in the anteroposterior plane, with a maximum volume of 20-25 ml. The seminal vesicles appear as paired lobulated echo-poor structures which often contain multiple small echo-free areas. The vasa deferentia appear medial to the seminal vesicles and superior to the prostate. They dilate normally near their termination (the ampull ae). The periprostatic venous plexus is seen anterior to the prostate and is of variable prominence.

Direct contrast investigations In this group of investigations water-soluble contrast is injected via some form of catheter directly into part of the urinary tract. The


Previous Page

CONTENTS

Next Page


897

Fig. 29.11 Transrectal ultrasound of the prostate. (A-C) Young adult prostate. (A) Transverse section. There is a little calcification appearing as speckled echogenicity in the periurethral area, presumed to be related to previous prostatitis and urethritis. The prostate at this age shows poor zonal anatomy. (B) Transverse view of the normal paired echo-poor seminal vesicles. (C) Longitudinal view showing the subjacent puborectalis sling as a laminated, relatively echo-poor structure. (D, E) Transverse and longitudinal views of the prostate around middle to old age. The zonal anatomy is clearly seen with a relatively hyperechoic peripheral gland compared to the hypoechoic central gland. following are considered under this heading: retrograde pyelography, cystography, loopography, stentography and urethrography. A further group of procedures, including antegrade pyelography, also i nvolve direct injection of contrast into the collecting system but because they involve a percutaneous approach they are considered under the heading of non-vascular interventional procedures. All of these procedures may potentially introduce infection into the urinary tract and care should he take to use a sterile technique, and broad-spectrum antibiotics should be administered prior to the procedure.

Spot films should be taken prior to injection of contrast and then of the opacified pelvicalyceal system and ureter (Fig. 29.12). The ureteric catheter can be withdrawn to allow contrast injection at a site of concern after adequate images have been obtained higher up.


Retrograde pyelography This investigation aims to optimally

opacify the pelvicalyceal system and ureter. It usually follows an I VU and is indicated when there is persistent uncertainty about the diagnosis, particularly if there is haernaturia and/or suspicious cytology. It is indicated to confirm or refute the presence of one or more filling defects within the collecting system, or to improve demonstration of the collecting system, either when there has been inadequate demonstration of part or all of the system or when the IVU is normal but the abnormal laboratory findings persist. It is occasionally used to demonstrate the lower end of an obstructed ureter. The urologist positions catheters within one or both ureters cystoscopically and the patient is transferred to the X-ray department. Under screening control 5-20 ml of a 150 strength watersoluble iodine-containing contrast agent is injected via each catheter in turn. It is important to avoid injecting air bubbles, which can be mistaken for filling defects. The pelvicalyceal system and ureter should be adequately opacified but not overdistended.

Fig. 29.12 Retrograde pyelogram demonstrati ng the pelvicalyceal system and ureter down to the vesicoureteric junction.


CONTENTS

Previous Page

898

Next Page

A TEXTBOOK OF RADIOLOGY AND IMAGING drain urine from the ureters to a cutaneous stoma. The standard investigation for ileal conduits is the loopogram. It can be performed in the immediate postoperative period to demonstrate the i ntegrity of the surgical anastamoses, or later (after months or years) to differentiate between reflux (common) or obstruction (usually at the ureteroconduit junction, usually benign rather than recurrent tumour) when dilatation of the upper tracts has developed. A Foley catheter (12-16 gauge) is positioned so that its balloon lies a couple of centimetres into the conduit and is then cautiously inflated so as to produce a reasonable seal without overdistending the conduit. Between 20 and 40 ml of 150 strength contrast is injected under direct screening to outline the conduit, which is normally of the order of 12-15 cm long. Usually there is free reflux into the ureters and pelvicalyceal systems. A series of spot films are taken to record this. Obliques may he useful and the most important areas to study are the ureteroconduit anastamoses (Fig. 29.14). A rare complication of this procedure in patients with spinal injuries is the development of severe hypertension (autonomic dysreflexia) due to overdistension of the ileal loop. Urologists routinely leave narrow gauge hollow stents running from the ureters into reconstructed bladders or ileal conduits in the immediate postoperative period. The stents run to the exterior either via the urethra or a cutaneous stoma. Stentography is frequently requested to demonstrate the integrity of the distal ureteric anastamoses within a few days of surgery. Between 1 0 and 20 ml 150 strength contrast is injected under direct screeni ng via each stent in turn, which opacities the upper tracts first, followed by drainage around the stents down into the bladder or diversion. Spot films are taken, again paying particular attention to the distal ureteric anastamoses (Fig. 29.15). Stentography

Retrograde pyelogram with pyelosinus extravasation due to overdistension of the collecting system. Fig. 29.13

Overvigorous injection of contrast may lead to reflux of contrast i nto the collecting ducts (pyelotubular reflux) and forniceal rupture with contrast extravasation into the renal sinus (pyelosinus extravasation) or more extensively into the regional lymphatics or veins (pyelolymphatic and pyelovenous extravasation) (Fig. 29.13).

Cystography can be classified into three groups: mieturiting cystourethrography (MCUG), dynamic cystography and simple cystography. The MCUG is primarily performed for the assessment of vesicoureteric reflux and is therefore essentially

The classical urological diversion procedure is the fashioning of an ileal conduit, i.e. the use of a loop of ileum to

Cystography

Normal loopogram (conduitogram). Contrast outlines the il eal loop (which still demonstrates the typical small-bowel mucosal pattern) and freely refluxes into both ureters.

Fig. 29.15

Loopography


Fig. 29.14

Normal stentogram demonstrating intact anastamoses between the ureters and the afferent loop of small bowel leading to the reconstructed neobladder.


Previous Page

CONTENTS

Next Page

THE UROGENITAL TRACT: ANATOMY AND INVESTIGATIONS an investigation of childhood. It is discussed in the paediatric investigations section below. Dynamic cystography is part of the urodynamics investigation of the lower urinary tract, and again is discussed below in the appropriate section. Simple cystography is a relatively frequently performed and straightforward investigation in the adult. It is used to assess the i ntegrity of the bladder following trauma or surgery or to investigate suspected fistulas involving the bladder (usually into the gastrointestinal tract, occasionally elsewhere such as the vagina). In the context of trauma the patient should be referred with a catheter already in place (either suprapubic or urethral if the referring surgeon can safely position one). A spot film should be obtained before contrast is administered. Approximately 250 ml of 150 strength contrast is infused into the bladder via a giving set. This should be done under frequent intermittent screening control so that extravasation can he identified as soon as it occurs. When the bladder has been filled or when extravasation is identified a spot f il m is obtained in the supine position. Ideally 45° oblique and l ateral spot films should be obtained but the patient will often have significant pelvic trauma and this may not he possible. If a C-arm is available this may be useful but external splinting of the pelvis may degrade the images. In modern trauma management many patients will undergo CT scanning of the abdomen and pelvis after intravenous contrast. Under these circumstances repeati ng the images of the bladder 20 min after contrast is administered i s at least as accurate as direct contrast cystography and obviates the need to transfer the patient to a different table for a further i nvestigation. When a patient has undergone radical prostatectomy or cystcctomy, with preservation of the sphincter and reconstruction of the bladder using small bowel, a cystogram is often performed around 10 days postoperatively to demonstrate the integrity of the surgical anastamoses prior to removal of the urethral catheter. Only 1 00-150 nil of contrast is required and the patient will often cornplain of fullness at a significantly smaller volume, at which stage the infusion should stop. The patient is likely to be ambulant and adopt easily the positions required for the full series of images ( Fig. 29.16). When a patient has a suspected vesical fistula, 250 ml of contrast can he used to maximise the chance of demonstrating the fistula. Again the patient usually has no trouble adopting the required positions.

899

Urethrography This can be performed via an ascending or descending approach. With advances in urethroscopy these procedures are required much less than formerly. Descending urethrography is usually part of the micturating cystogram and is rarely indicated in adults. When it is performed in adults the bladder should be adequately filled (with at least 200 ml of 1 50 strength contrast). The screening table should be positioned erect. Imaging is performed directly anteroposterior in females (Fig. 29.17) and in a 45° oblique projection in males (Fig. 29.18). Males are generally used to micturating while standing, often in unusual situations, and can manage with a bottle while screening is performed and spot films taken of the urethra and bladder base. Females are provided with a special drainage receptacle that is held between the thighs.


Fig. 29.17 Normal female micturating cystourethrogram showing the typical short open-necked female urethra.

Fig. 29.16 Postprostatectomy cystogram. Supine (A), oblique (B) and lateral (C) views demonstrate a substantial extravasation arising from the right posterolateral aspect of the vesicourethral anastamosis.


Previous Page

900

CONTENTS

Next Page


Fig. 29.18 Descending urethrogram in a male. The entire length of the urethra is demonstrated as the bladder empties. A post-operative urethrorectal fistula is present with contrast tracking posteriorly from the prostatic beds.

Ascending urethrography is essentially confined to the male. It is used in the investigation of trauma, stricture and fistulas. The patient is positioned in a 45° oblique position with the dependent hip partly flexed to provide stability and ensure the urethra is not projected over hone. A I2-16 gauge Foley catheter is positioned with its balloon a Couple of centimetres into the distal urethra. The balloon is gently partially inflated to provide a seal without undue trauma. Between 5 and 10 nil 150 strength contrast is injected gently into the urethra under direct screening and spot filets arc taken (Fig. 29.19). The urethra is usually easily opacifed hack to t he urogenital diaphragm. In a minority of patients contrast will reflux i nto the posterior urethra and bladder. Usually, however, with ascending urethrography the prostatic urethra is not demonstrated. Overenlhusiastic instillation of contrast into the urethra can he painful and produce extravasation of contrast into the corpora cavernosa ( Fig. 29.20). Female urethrography is rarely required, virtually all urethral pathology being better demonstrated on urethroscopy or trailsvaginal sonography.


Percutaneous interventional procedures

The common feature of all these procedures is the positioning of a cannula or drain percutaneously into the upper urinary tract under some form of image guidance. Included in this category are antegrade pyclography, percutaneous nephrostomy and nephrostograms, antegrade stent positioning and ureteric dilatation and percutaneous nephrolithotomy. The commonest complication of these procedures i s infection, which is occasionally fife-threatening. Where there is no evidence of infection at the start of the procedure a single intravenous injection of a broad-spectrum antibiotic (say cefuroxime 750 mg i.v.) 1-6 Ii prior to the procedure should he sufficient. if the urinary tract is known to he infected in advance there is a real risk of the procedure provoking Gram-negative septicaemia and hypotensise collapse. It is imperative the patient is then on an adeCiliate course of broad-spectrum antibiotics and ideally. if circum-

Fig. 29.19 The normal male ascending uretbrogram usually shows contrast flowing retrogradely as far as the junction of the bulbar and membranous urethra (A). Sometimes contrast will flow into the bladder and demonstrate the prostatic urethra (B). Note the smooth filling defect of the verumontanum within the prostatic urethra.

stances permit. should have been on antibiotics for at least 24 h prior to the procedure. should he taken to avoid overdistending Care t he collecting system or any unnecessary manipulations, either of which may provoke hacteracmia. Less common complications i nclude hacmonrhage and bowel perforation. 11r the patient has known hepatic disease or other predisposition to coagulation defects. the platelet count and prothrombin time should he assessed prior to the procedure and ahnormalitics corrected. The risk of all complications increases with the complexity of the procedure and t he size of the cannulas and drains ultimately employed. Percutaneous nephrolithotomy is performed under general anaesthesia but otherwise these procedures arc performed using local anaesthetic and are potentially painful. Adequate intravenous analgesia and sedation with appropriate pulse oximetry according to l ocal protocols are therefore required. Following the procedure the patient requires an adequate period of observation, which usually i nvolves at least overnight hospital admission. although simple


Previous Page

CONTENTS

Next Page


901

Antegrade pyelography This is a relatively simple procedure. It

i s used to evaluate the cause and level of ureteric obstruction in the minority of patients in whom non-invasive imaging has not provided the information. It may also be the first step in performing a nephrostomy or the Whitaker procedure (discussed in the section on urodynarnics). The patient is positioned approximately 45° semiprone and the pelvicalyceal system cannulated with a fine (22 gauge) needle. The general rule in gaining access to the pelvicalyceal system in all these procedures is that a puncture directly into the renal pelvis risks lacerating it. Ideally the puncture should be directed through the renal parenchyma into a suitable calyx and then into the pelvis. Confirmation of cannulation of the collecting system is obtained by aspirating urine. If this appears infected, the system should be drained and infection treated before proceeding with an antegrade pyclogram: 150

Ascending urethrogram. A vigorous infusion of 150 strength contrast has been performed. The external sphincter has not relaxed (as is usually the case) and the anterior urethra is therefore overdistended with contrast extravasating and entering the corpora cavernosa. Fig. 29.20

diagnostic antegrade pyelography and some selected nephrostomies may he performed on an outpatient basis. The initial percutaneous cannulation of the collecting system is usually performed under ultrasound guidance, fluoroscopy or a comhination of both. Complex manoeuvres such as stent placement require fluoroscopy. The other general point worth making is the value of an adequate i ncision in the skin. This is unnecessary if using only a line needle ( 21-22 gauge) but for most of these procedures a suhstantial catheter will he introduced and a skin incision of at least 5-I0 mm i s required; it should he at least 10-15 mm deep, traversing the superlieial fascia. For pereutaneous nephrolithotomy the length of the incision should be doubled.

strength contrast is infused into the system to opacity it (Fig. 29.21). If the procedure is being performed for diagnostic purposes, a series of spot films of the ureter down to the level of the obstruction is taken. The flow of contrast down the ureter can he assisted by elevating the head of the screening table. Percutaneous nephrostomy I n this procedure a pigtail drain (usually 6-8F) with multiple holes is positioned in the pelvical yccal system. It is indicated to drain urine from an obstructed system. Occasionally it is used to drain unobstructed systems to divert the urine flow and allow ureteric or vesical fistulas to heal. There are various methods and devices available to position a drain within the pelvicalyccal system of the kidney and the interventional radiologist is likely to be acquainted with a number of t hem but have a preferred method for routine use. A single-stab technique with a nephrostomy tube mounted on a cannula is available and, when successful, is the quickest method. This is most commonly performed under ultrasound guidance, although CT or even MRI have been used on occasion. Alternatively a sheathed needle (18 gauge) may be positioned (under ultrasound guidance or fluoroscopy after initial antegrade pyelogram) within the col-


Antegrade pyelography. The dilated pelvicalyceal system has been cannulated with a 22 gauge needle and opacified with 1 50 strength contrast (A). The ureter is dilated, shows marked medial displacement distally and tapers to a complete occlusion (B). Fig. 29.21


Previous Page

90 2

CONTENTS

Next Page


Fig. 29.22 Nephrostomy procedure using fluoroscopy and dual puncture technique. Having opacified the system through a fine-gauge needle a second puncture is made with a sheathed needle (A). A guide-wire is advanced through this into the pelvicalyceal system (and ideally for stability into the ureter) (B). The fine needle is removed and the sheathed needle exchanged for a dilator over the wire (C). The calibre of the dilator depends on the nephrostomy drain to be positioned and often comes as part of a nephrostomy set. For an 8F drain a dilator of 8 or 9F is usually employed. The dilator is exchanged for the drain, which is positioned, coiled up, in the renal pelvis (D).

l ecting system and a conventional 35 gauge wire introduced through the sheath into the calyces and renal pelvis. This is used to guide one or more dilators and then the nephrostomy tube (a pigtail drain with multiple drainage holes) into position (Fig. 29.22). A further alternative is to position a fine wire through an initial fine needle puncture and dilate until a cannula sufficient to take a 35 gauge wire can be positioned. Drainage of a perinephric or renal abcess is essentially a variation of this procedure. If the abscess is small, simple aspiration of pus via a sheathed needle may be sufficient. A multiple-hole pigtail drain can be positioned in a larger abscess using any of the above techniques, most commonly as a single-stab technique under ultrasound or CT guidance.


nephrostomy tube is positioned for therapeutic purposes but while it remains in position a nephrostogram can be performed. It is indicated to assess the obstructing lesion, either to evaluate its status (continued presence of a calculus, response of a malignancy to treatment) or obtain further diagnostic information if the cause of the obstruction has not yet been determined. It is also used to monitor the progress of fistulas. It is a very similar procedure to the antegrade pyelogram: 150 strength contrast is infused i nto the collecting system and spot films taken of the ureter down to the level of the obstruction. Where infection is present the procedure should be deferred until this has been adequately treated. Nephrostogram A

If the cause of the obstruction is unlikely to resolve within a few days of the pcrcutaneous nephrostomy a ureteric stent can he positioned antegradely to relieve the patient of an external drain. This is most often performed for malignant obstruction. It can also be performed for non-obstructive probl ems such as a postoperative ureteric fistula or anastamotic breakdown. Under fluoroscopic guidance a 35 gauge wire is manipulated through the nephrostomy tube into the collecting system. It is convenient to replace the nephrostomy tube with an Antegrade stenting

angled catheter, such as a cobra, and use this and a suitable angled guide-wire to advance into the ureter and then across the obstructed area into the bladder. It is usual to dilate the constricted area, most simply with a rigid plastic dilator, and then position a double J stent from the pelvicalyceal system to the bladder (Fig. 29.23). Percutaneous nephrolithotomy (PCNL) was developed as a minimally invasive alternative to open surgery for the treatment of renal calculi. Its role has been much reduced by the development of the even less invasive procedure of lithotripsy but it maintains an important role in patients with calculi resistant to lithotripsy (such as pure calcium phosphate or cystine calculi), where the size of the calculus would preclude sucessful lithotripsy (above 2.5 cm) and/or where there i s likely to he a drainage problem (for example ureteric stricture). PCNL is performed under general anaesthesia; usually the patient i s placed supine on the operating table and the urologist performs an initial cystoscopy with a view to positioning a thin catheter (5F) retrogradely through the ureter and into the pelvis of the kidney containing the target calculus. The patient is then repositioned semiprone with the side of the kidney of interest rotated approximately 45° upwards. Ideally the operating table will be articulated to allow some degree of flexion of the patient to separate maximally the ipsilateral subcostal margin and iliac crest. Contrast (150 strength) is infused via the ureteric catheter to opacify and distend the collecting system. A small amount of methylene blue is traditionally added to assist differentiation of blood-stained urine from a vascular puncture. Fluoroscopic screening of the kidney should identify the optimal site for access. This relates to the site of the stone. Calculi within the renal pelvis and/or mid or lower pole calyces can be accessed via a mid or l ower pole calyx, which usually permits a subcostal approach. The desired calyx can usually be brought below the costal margin Percutaneous nephrolithotomy


Previous Page

CONTENTS

Next Page


903

Fig. 29.23 Antegrade stent positioning. A guide-wire has been manipulated down the right ureter to the level of the obstruction (A). Note the distal end of a previously postioned left-sided stent is visible within the bladder. The guide-wire is advanced across the obstruction into the bladder. Following a dilatation step with a long 9F dilator the stent has been advanced along the wire so its distal end li es within the bladder (B). Once the wire is removed the distal end of the stent will adopt a pigtail configuration within the bladder. The upper end will behave simil arly within the renal pelvis, provided care has been taken to ensure the upper end has been advanced sufficiently and does not lie within renal parenchyma.

by the anaesthetist holding the patient in an adequate inspiration while the initial percutaneous puncture is performed. This is made with a standard 18 gauge sheathed needle. A guide-wire (usually a 3 mm angled wire) is positioned through the needle into the target calyx and manipulated into the renal pelvis and ideally into the ureter. Initial dilatation to 9F is performed using plastic dilators, and then more significant dilatation is performed up to 28-30F. At this stage a hard plastic (Amplatz) sheath of this calibre is positioned over the dilatation system into the collecting system. The dilatation system is removed, allowing the urologist to gain access to the collecting system with a nephroscope. If an upper pole calyceal approach is required, the puncture is likely to be above the 12th rib. This is generally safe since the pleural reflection is usually around the 11 th rib but punctures above this level should he avoided because of the risk of pleural trauma. Different dilatation systems are available. One of the quickest and safest systems is a high-pressure balloon catheter that is i nflated to 10 mm diameter with a screw-type syringe (Fig. 29.24). Occasionally there is difficulty in completely dilating the mid part of the balloon (usually at the level of the renal capsule) and an ancillary method must he employed. Graded plastic dilators are useful for this. On their own, however, multiple plastic dilators may prove troublesome and traumatic. Multiple metal dilators that lit sequentially over each other (telescopic dilators) are an effective and rapid method hut should he used with care as they can cause considerable trauma if they drift forwards during the procedure. Once adequate access to the system has been obtained, the urologist will inspect the inside of the collecting system, identify the calculus and remove it. Often the calculus must be broken up in situ, which is performed using some form of direct contact lithotripsy with an instrument introduced through a channel within the nephroscope and placed in direct contact with the calculus. After the cal-

and collateral organ damage. Haemorrhage is frequent but not normally severe. It usually originates from renal parenchymal trauma or rupture of vessels around the calyceal necks and stops relatively readily, either spontaneously or in response to a brief period of tamponade if the pelvicalyceal system fills with thrombus. More dramatic haemorrhage may result from damage to the vessels at the renal hilum if the dilatation system is allowed to drift too deep into the patient; this may be life-threatening and on rare occasions require emergency nephrectomy. Traumatic arteriovenous fistulas have also been reported and may produce florid haematuria and require selective arterial embolisation. Any of the surrounding organs may sustain damage and sequelae may include bowel perforation and pncumothorax. Although rare, these conditions should be considered during the postoperative period. A nephrostogram is usually performed 3-4 days after the procedure to demonstrate normal drainage (no obstructing stone fragments or ureteric damage) and good clearance of calculus.


culus has been removed a percutaneous drain is positioned within the renal pelvis and the sheath is removed. Major complications include infection (broad-spectrum antihiotics should be administered before the procedure). haemorrhage

Other percutaneous procedures

There are several other percutaneous uroradiological procedures that involve techniques similar to those described above. Having established a large percutaneous track, other forms of endoscopic intervention can be performed. Inspection, biopsy and ablation of are the/jul masses may he performed but is rarely indicated and is associated with the risk of seeding tumour along the access track. More commonly endoscopic pvelolvsis (endopyelotomy) is performed for the treatment of pelviureteric junction (PUJ) obstruction. A vertical incision is made through the full thickness of the wall of the PUJ. The danger of this procedure is the potential association of the condition with aberrant renal arteries that cross the PUJ and are at risk when the incision is made. They usually run anteriorly and therefore the incision should be made posterolaterally. A preoperative search for aberrant renal arteries should he performed, usually with contrast CT ( CT angiography). Attempts have also been made to treat PUJ


CONTENTS

Previous Page

904

Next Page


̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ Fig. 29.24 Percutaneous nephrolithotomy. A faintly opacified staghorn calculus is present (A). A retrograde catheter is in position. After the system is opacified and dilated via the retrograde catheter a sheathed needle is used to cannulate one of the calyces (B). A guide-wire is advanced through the sheath into the calyx and manipulated into the renal pelvis adjacent to the calculus (C). An initial dilatation step is performed with an 8F dilator over the wire (D). A high-pressure balloon is positioned with its distal end in the collecting system. As it is inflated a waist at the level of the renal capsule is often the l ast part to dilate (E). Following inflation an Amplatz sheath is positioned over the balloon (F-example from a second patient). The balloon is deflated and removed, allowing the urologist access to the calculus. After the procedure fluoroscopy demonstrates adequate clearance of the stone (G).

obstinelion with balloon dilatatlion

, either via antegrade

(through

Perculaneous renal cyst asp iration for

-

diagnostic purposes. Once

a percutaneous track) or retroerade approaches, but the success

common. is now relatively rare. as Faith modern cross-sectional

rate is controversial.

i maging the vast majority can be confidently


categorised

as benign .

Previous Page

CONTENTS

Next Page


905

Fig. 29.25 Percutaneous cyst aspiration. A 22 gauge needle has been postioned within a heavily calcified cyst under CT guidance for diagnostic aspiration.

Fig. 29.26 CT-guided biopsy. An 18 gauge cutting needle has been positioned in a solid mass in the right renal bed under CT guidance. The core of tissue obtained confirmed recurrent renal cell carcinoma.

Fluid can be aspirated under ultrasound, CT or MRI guidance from complex cysts where there is a concern about potential malignancy. This is a simple procedure performed on an outpatient basis using a thin (21-22 gauge) needle and followed by a brief period of posti ntervention observation (Fig. 29.25). Simple cysts are virtually always asymptomatic but on rare occasions the patient and the referring urologist put forward a reasonably convincing argument that a particular cyst, often large, is causing distressing loin pain. It is straightforward to position a needle within the cyst, usually under ultrasound guidance, and aspirate to dryness. This can be very timeconsmning with a narrow-gauge needle and a plastic sheathed 1 8 gauge needle is recommended. When the cyst has been aspirated to dryness, a sclerosant can be infused into it to prevent recurrence. Absolute alcohol has been used but it is recommended that the needle is replaced by a small pigtail catheter to secure the position and drain the sclerosant after It) ruin. Contrast should also be i njected into the cyst first to demonstrate the absence of extravasation (into the retroperitoneum, collecting system or vascular compartment). Arguably, a safer alternative is tetracycline (100 mg). Percutaneous renal biopsy i s easily performed under ultrasound, CT or MRI guidance. A core of tissue is obtained using a cutting needle. It is indicated in the assessment of diffuse renal disease. It is not routinely indicated for renal masses, which are usually charactensed by cross-sectional imaging, but is useful in selected cases,

sion of the channel to the balloon, which may occur in long-term catheters. An initial attempt to clear the channel with the stiff end of a guide-wire should be made, as this may solve the problem. If this fails, a line needle can be introduced under ultrasound guidance through the anterior pelvic wall into the bladder and into the catheter balloon, puncturing it (often quite dramatically) and deflating it, allowing the catheter to be removed.

Lithotri s Extracorporeal shock wave lithotripsy (ESWL) is the technique of shattering urinary tract calculi into fragments of 3 mm diameter or l ess with high-power ultrasound (shock waves). The resultant fragments are then passed in the urine. Early ESWL machines caused considerable discomfort to the patient and were relatively inefficient, requiring multiple sessions to obtain an adequate response. Conventional machines are much better tolerated (although this varies between patients) and achieve satisfactory results more rapidly. Renal calculi above 25 mm diameter respond to ESWL poorly because of the numerous fragments that are produced. These may remain predominately in the upper tract and provoke subsequent stone reformation or overwhelm ureteric drainage and lead to obstruction. They are often better treated with nephrolithotomy. Shock waves are directed at the calculus under fluoroscopic or ultrasound guidance. Although ureteric and renal calculi can be treated with ESWL, a distal ureteric calculus may become difficult to image and inacessible to the shock waves if it comes to rest anterior to the sacrum. Prior to ESWL, patients undergo IVU examination, partly to assess the stone burden but more importantly to demonstrate adequate urinary drainage. If this is impaired (say due to ureteric strictures or bladder wall hypertrophy), the passage of the calculus fragments will be difficult and painful obstruction is likely to follow treatment. Under these circumstances passage of a double J stent retrogradely prior to ESWL may be necessary, or treatment of the calculus with nephrolithotomy may be preferred. Usually the stone fragments pass without trouble but in up to 10% of cases there may be some degree of ureteric obstruction, as multiple stone fragments build up in the distal ureter (referred to as steinstrasse or stone street). This situation usually resolves with analgesia and an adequate fluid intake but in approximately a quarter of cases rctrograde stenting or nephrostomy insertion is required.


for example when there is metastatic malignancy with a solid renal mass. These situations may warrant treatment with interferon for renal cell carcinoma but alternative chemotherapy for other malignancies and histological characterisation is required. Similarly, renal bed masses following nephrcctorny or malignancies of unknown histology invading the kidney may require guided biopsy using similar techniques (Fig. 29.26). Percutaneous procedures targeted to the bladder are less common because urologists have good access to the bladder lumen cystoscopically and are also confident at inserting suprapubic catheters clinically. Occasionally radiologists are involved in ultrasound-guided catheter placement and should he aware that the bladder wall is remarkably tough. For this reason, puncture of the full bladder with a sheathed needle, positioning of a stiff guide-wire i nto the bladder and multiple dilatation steps may he preferred to a single-stab technique. A simpler and quicker procedure that radiologists are required to perform from time to time is deflation of the balloon of an indwelling catheter. The problem arises from occlu-


Previous Page

90 6

CONTENTS

Next Page


The only other common complications are failure of treatment (inadequate fragmentation or poor clearance of fragments and subsequent reaggregation) or rapid recurrence of calculus, presumably related to residual fragments acting as a nidus. Less common complications include renal contusion, perinephric haemorrhage, fractare due to misdirected shock waves (usually of a transverse process) and possibly an increased incidence of long-term hypertension.

Renal angiography This procedure originally developed a role in the assessment of renal masses (particularly suspected renal cell carlesions, such as arteriovenous malformations and fistulas and aber rant or accessory renal arteries in patients undergoing live kidney donation. The vast majority of this diagnostic role (particularly the tumour work) is now better performed by cross-sectional imaging. Renal angiography is still useful in selected cases, usually combined with subsequent intervention, notably transluminal angioplasty for renal arterial stenosis or selective embolisation. The latter is performed for arteriovenous abnormalities and some cases of traumatic renal damage with continued bleeding. It may also be used in some i noperable renal tumours as a palliative treatment for haematuria. Standard angiographic technique is employed and there is a choice of embolisation materials, including feathered coils (Fig. 29.27), particulate material and absolute alcohol. Angiography of the bladder and prostate is also largely performed as a prelude to intervention for troublesome or life-threatening haematuria, either due to inoperable tumour or postsurgical (usually transurethral prostatectomy) bleeding (Fig. 29.28).

CT has a wide variety of indications, including the characterisation of renal masses. staging of urinary tract tumours and the assessmeat of inflammatory and traumatic processes, calculi and the causes of obstruction. It is also used to direct biopsies and the positioning of percutaneous drains. It may also be used to assess renal artery stenosis. Protocols vary between institutions but should conform to the same principles. They are tailored towards the clinical problem and common protocols are described below. The gastrointestinal tract should be opacified before imaging for renal masses: 40 ml of 150 strength (or 20 ml of 300 strength) water-soluble iodine-base contrast diluted to I litre with fruit squash taken orally half an hour prior to the scan is a suitable protocol. An initial plain scan from the dome of the diaphragm to the iliac crests is performed, a suitable protocol being collimation of 8 mm, pitch of 1.5 and slice thickness of 7-8 mm. This is repeated after an intravenous injection of contrast. The tinting is important, aiming towards imaging the kidneys while contrast is in the renal veins and inferior vena cava. This is usually optimal around 20-30 s from the start of the injection. With fast modern scanners 50 ml of 300 strength contrast is usually adequate. For lower tract tumours both the plain and postcontrast scans should be from the dome of the diaphragm to the symphysis pubis. A further I litre of dilute oral contrast taken 4 h before the scan is useful to image the large bowel and rectum. Alternatively, the same strength contrast can be administered immediately before the scan as an enema. All patients being investigated for malignancies should have a chest radiograph and some departments will also perform routine CT of the chest. CT is commonly performed as a relative emergency for suspected acute urinary tract problems. When CT is performed for urinary tract calculi, both intravenous and oral contrast are avoided. Relatively narrow collimation (5-7 mm) with a pitch of 1.5-2.0 is recommended, bearing in mind that excessively narrow collimation increases the radiation close to the patient. Using the wider collimation and higher pitch does not appear to significantly reduce the accuracy of the scan. The mAs should also he kept to a minimum. In the context of trauma, a protocol similar to the postnephrogram phase described above for the investigation of renal masses is suitable. Repeat scanning after 6-8 min may demonstrate urinary extravasation, e specially if there is suspected ureteropelvic disruption. In the assessment of renal or perinephric abscess, 8 mm slices are obtained prior to intravenous contrast and then again during the nephrogram phase. Opacification of the gastrointestinal tract with dilute oral contrast is useful in these cases. Spiral CT with a contrast infusion and relatively narrow collimation and image reconstruction (for example 3 mm and 1.5 mm respectively; pitch of 1-1.6) is suitable for demonstration of the renal arteries and major branches in the assessment of renal artery stenosis or in the preoperative work-up of patients for partial nephrectomy. The normal renal parenchyma is of intermediate density, measuring between 30 and 60 HU (Fig. 29.30). The renal sinus and perinephric fat are low density, around -10 to -50. Following intravenous contrast the cortex enhances more rapidly than the medulla, appearing more dense from around 20 s to 60 s when there is maximum corticomedullary differentiation. Around 60-180 s the renal parenchyma becomes homogeneously high density, around 80-120 HU. After this, contrast appears in the collecting system. The ureter can he identified as it runs retroperitoneally along the


Inferior vena cavography This procedure is obsolete for the assessment of vascular spread of renal cell carcinoma. Occasionally renal phlebography is performed to obtain blood for p hormone (Usually renin assay. Testicular phlebograp hyy is per) y' ° p p formed as a prelude to testicular vein embolisation for the treatment of varicoceles (Fig. 29.29). Although either the femoral or jugular vein can be used to gain access to the inferior vena cava, it i s pleasanter for the patient to work from the groin; with modern l ow-friction catheters and wires, gaining access to the renal vein and then the testicular vein from the direction of the femoral vein s rarely a problem. It has to be emphasised, however, that the testicular vein is very prone to spasm if excess catheter and wire manipulation is performed, and this may preclude a successful outcome. The success rate of embolisation of the testicular vein (usually with metal coils) is of the order of 80%, which is comparable with surgery, the radiological procedure having a substantially lower morbidity. Abnormalities of penile vascular supply are an important cause of erectile dysfunction and have been assessed with pudenda) angiography, cavernosography and duplex ultrasound. These are discussed in Chapter 32. Vasography I n this procedure the vas deferens is exposed surgically in the scrotum and cannulated. Two to three millilitres of water-soluble contrast is injected into the vas and spot films are taken to demonstrate areas of narrowing and occlusion. Its use in the investigation of male infertility is described in the appropriate section in Chapter 32.


Previous Page

CONTENTS

Next Page


90 7

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ Fig. 29.27 Right renal angiogram showing a small upper pole accessory artery (A). Selective cannulation of the dominant renal artery demonstrates a small peripherally placed renal cell carcinoma (B). A larger renal cell carcinoma is shown on this left renal angiogram (C) with a characteristic malignant circulation (irregular disorganised vessels). Selective embolisation of part of the renal cell carcinoma has been performed (D).


Previous Page

908

CONTENTS

Next Page


̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ Fig 29.28 Pelvic angiography and tumour embolisation. The left internal iliac artery has been cannulated, by manipulating a catheter from the right femoral artery across the bifurcation, and its two major divisions (anterior and posterior) demonstrated (A). The patient is suffering severe immediate postprostatectomy haemorrhage. A coil has been positioned in the anterior divison (B) and the subsequent angiogram demonstrates successful obstruction of flow in this vessel (C). It is rare actually to demonstrate the bleeding point and, given the life threatening situation, it is often worth embolising both sides. Consequently in this case the right internal iliac artery has also been cannulated from the same femoral artery (D) and the anterior division of the internal iliac selectively cannulated (E). This injection shows the inferior vesical artery from which most of the bleeding is probably occurring and therefore this has also been occluded with a coil (F) (see over)


CONTENTS

Previous Page

Next Page


Fig. 29.29 Testicular vein embolisation. The left renal vein has been cannulated via the inferior vena cava (A). A testicular phlebogram performed once the catheter has been advanced down the testicular vein reveals a double testicular vein with multiple origins unprotected by valves, one of a number of variants associated with varicocele formation (B). The upper end of the varicocele is just visible once the catheter has been advanced to the inguinal ring (the varicocele itself should not be screened, to avoid irradiating the testicles) (C). The main testicular vein is occluded with multiple (2-4) coils along its length. Two are illustrated here (D). The most important is the distal one that also occludes the entry of the accessory testicular vein.

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ Fig. 29.28

(F) with satisfactory clinical response.


909

CONTENTS

Previous Page

91 0

Next Page


Fig. 29.30 Transverse CT through the midpart of the kidneys. Prior to contrast the kidneys are of intermediate density (A), similar to liver, spleen and blood vessels. During the nephrogram phase the cortex is seen to enhance earlier than the medullary pyramids (B).

Fig. 29.31 Contrast-filled ureters entering the pelvis at the level of the iliac crest, running along the anterior aspect of the psoas muscles (A) and then around midsacral level crossing the iliac vessels and beginning to turn anteriorly (B).


medial aspect of the psoas muscle, especially when dilated or contract-filled (Fig. 29.31). After entering the pelvis the ureter runs posterolaterally to the l evel of the ischial spine, where it turns anteromedially and can be i dentified running in front of the seminal vesicles or vaginal fornices to reach the bladder base. The bladder is seen as a thin-walled structure between the urine and the fat. The seminal vesicles appear as tubular structures related to the superior aspect of the prostate, posterior to the lower bladder and anterior to the rectum. There is a fat plane between the seminal vesicle and the bladder (Fig. 29.32). The prostate is usually of homogeneous soft-tissue density but often shows multiple foci of calcification, increasingly from early adulthood onwards (Fig. 29.33). On CT, normal lymph nodes are seen as soft-tissue density similar to unenhanced renal parenchyma, homogeneous apart from the possibility of a small area of fat at the hilum. Normal nodes are up to 10 mm in maximum transverse diameter in the para-aortic and iliac chains and 6 mm in the rctrocrural region.

Magnetic resonance imaging On T,-weighted images the renal cortex is of similar signal intencity to liver. The medulla (containing more water) is slightly lower signal, similar to the spleen. On T,-weighted and STIR sequences the kidneys are diffusely high signal, similar to spleen, although

Fig. 29.32 CT of the pelvis showing the distal ureters just above the vesicoureteric junctions lying anterior to the seminal vesicles (A). The right ureter happens to contain a bolus of contrast at the time of the scan. (B) shows the ureters at the vesicoureteric junctions with a jet of opacified urine projecting from the right ureter into the bladder. Further interiorly the prostate is seen just projecting into the bladder base (C).

Fig. 29.33 Transverse CT of the prostate with focal calcification. The puborectalis sling is well demonstrated as a thin intermediate-density line encircling the anorectal junction and the prostate.

sometimes the medulla is seen to be of higher signal than cortex (Fig. 29.34). Gadolinium produces intense enhancement, the cortex first and then the medulla. This emphasises the corticomedullary differentiation, which is maximum around 20-60 s following contrast injection (Fig. 29.35). At around 60-180 s both the cortex and


Previous Page

CONTENTS

Next Page


91 1

MRI scan of the normal kidneys. On the T,-weighted sequence (A) the renal cortex is seen to be of higher signal than the medulla. The reverse is seen on the T2 -weighted (B) and STIR (C) sequences, with the medulla being of higher signal than the cortex (although both are relatively high signal). Fig. 29.34

urine rises the signal intensity may fall dramatically due to the hyperparamagnetic effect of concentrated gadolinium. This may also be observed in the region of the apices of the medullary pyraminds (renal papillae), where the same phenomenon can be seen due to the high concentration of gadolinium in the collecting ducts. The bladder wall is seen as a thin low-signal intensity line on wTeightdm2asjc-noheigalftusdenri inside the bladder (Fig. 29.36). It is seen as an intermediate signal line on T,-weighted images, similar to urine. The bladder wall enhances intensely with intravenous gadolinium. Differential enhancement of urine is seen in horizontal layers as contrast is excreted, again highly concentrated gadolinium causing a paradoxical fall in signal intensity. Lymph nodes within the abdomen and pelvis appear as well-


defined ovoid structures of intermediate signal on the T 1

MRI scan of the kidneys 30 s (A) and 90 s (B) following intravenous gadolinium showing the intense enhancement of the cortex and then both the cortex and the medulla. Fig. 29.35

medulla are intensely enhanced and there is little corticomedullary differentiation. Towards the end of this time contrast appears in the calyces, initially enhancing the urine but as the concentration in

-weightdsqunca igloT,-wehtdanSIR sequences. A small amount of fat is occasionally demonstrable at the lymph node hilum. Normal figures for size are the same as on CT. The prostate zonal anatomy is seen from about the age of 5 years, before which there is insufficient central gland for contrast. The peripheral gland shows high signal on the T,-weighted sequence, compared to the intermediate signal of the central gland (similar to skeletal muscle). The capsule is seen as a thin highsignal line, best seen with an endorectal coil. On the T,-weighted sequence the prostate appears as an intermediate signal mass without visible zonal anatomy. Benign prostate hyperplasia produces nodularity and heterogeneity on all sequences, especially the T,-weighted sequence. The seminal vesicles appear multicystic and high signal on the T,-weighted sequence and intermediate signal on the T i - weighted sequence. The testes are homogeneous high signal on the T 2 - weighted sequences, with low signal at the mediastinum and somewhat heterogeneous high signal within the epididymis. On T,-weighted sequences the testes and epididymis are homogeneous and intermediate, similar to muscle.


CONTENTS

Previous Page

91 2

Next Page



Fig. 29.36 MRI of the pelvis on the T,-weighted sequence at the level of the bladder (A) and the prostate (B). The same levels (on a different patient) are seen on the T 2 - weighted sequence (C,D) demonsrating the increased conspicuity of the bladder wall and the zonal anatomy of the prostate. Both sequences show the puborectalis sling well. The same levels are also shown on the STIR sequences (E, F).

99

RADIONUCLIDE IMAGING Philip J.A. Robinson

The investigation of renal function is a key element in the management of most disorders of the urinary tract. Radioactive tracers that

are excreted by the kidney are used to assess renal perfusion and excretory functions, while tracers that are retained in the kidney are used to demonstrate renal anatomy.

The 'intact nephron' concept I n most renal disorders nephrons behave as independent functioning units, so that when any part of the nephron is diseased, the function of the whole nephron is affected. Whatever the primary site of pathology, whether at arteri al, arteriolar, glomerular, tubular or ureteric level, the functional deficit can be assessed by the same techniques. Exceptions to this general rule occur in acute tubular necrosis and in some enzyme deficiencies when filtration and excretory functions may be dissociated.

Renal tracers Currently the most satisfactory tracer for imaging renal anatomy is dimercaptosuccinic acid (DMSA) labelled with technetium-99m. For investigating renal perfusion and excretion,

"'Tc-labelled mercaptoacetylglycylglycylglycine (MAG3) is now usually preferred to earlier tracers excreted solely by glomerular filtration (e.g. diethylenetriaminepenta-acetic acid. DTPA) or by tubular secretion (orthoiodohippurate labelled with i odine-131 or iodine- 123). Stressing the system As with physiological investigation of other body systems, early disease is detected with greater sensitivity and accuracy by stressing the functions which are being investigated. Pharmacological stress may be conveniently regarded as the nuclear medicine equivalent of contrast enhancement in radiographic imaging. Diuretics are used to increase urine flow rates so as to reveal minor degrees of obstruction, and angiotensinconverting enzyme (ACE) inhibitors are used to expose and magnify the physiological disturbances associated with renal arterial disease.

Measurement of glomerular filtration rate (GFR) GFR is used as a measure of global renal function in patients with known or suspected renal disease. It is also used to help plan treat-


CONTENTS

Previous Page

Next Page


913

ment in patients receiving ncphrotoxic drugs, e.g. some chemotherapeutic agents for cancer. GFR is the volume of blood from which a solute which is wholly excreted by glomcrular filtration is cleared in unit time, conventionally expressed as millilitres per minute. Tracers that are excreted almost entirely by glomerular filtration include the chelating agents EDTA and DPTA. EDTA can be labelled with chromium-51, DTPA with technetium-99m, and either of these radiopharmaceuticals can be used to estimate GFR. I n order to avoid the inconvenience and potential inaccuracy of urine collection, GFR can be estimated from serial blood samples obtained at carefully timed intervals after injection of the tracer. The method assumes that uniform distribution of the injected tracer within the blood and extracellular fluid compartments occurs within 2 h after injection, and the subsequent rate of fall of activity in the blood, corrected for radioactive decay, is proportional to the rate of renal excretion. Assuming that the blood curve becomes linear after the initial phase of redistribution and equilibration, extrapolation of the linear part of the curve back to zero time will indicate the effective volume of distribution. GFR can then be calculated as the product of the distribution volume and the slope of the linear part of the curve.

anterior view and express split function as the geometric mean of percentage counts in the anterior and posterior views. Geometric mean is the square root of the product of the background-subtracted i ndividual kidney counts in the two views.

DMSA scintigraphy

Applications of DMSA scintigraphy

Radiopharmaceutical `'`'"'Tc-DMSA is extracted by and bound within the parenchymal cells of the proximal convoluted tubule. The extraction efficiency is rather low, so in the normal subject it takes about 3 h for approximately 50% of the injected tracer to become localised in the kidneys. Most of the other 50% is excreted unchanged in the urine. With renal failure, more of the i njected activity accumulates in the liver and there is then appreciable biliary excretion. The state of hydration of the patient has little influence over the kinetics of DMSA, so no patient preparation is needed. The typical adult activity of 80 MBq should be scaled down for children according to body surface area. Adverse reactions are very rare.

DMSA scintigraphy is used to locate the functioning renal tissue, differentiate renal cortex from soft-tissue masses in or adjacent to the kidney, to find scars or non-functioning areas of renal parenchyma and to establish the functional contribution of an

99

Fig. 29.37 Normal '°Tc-DMSA study. LPO = l eft posterior oblique; POST = posterior; RPO = right posterior oblique. expressed as a percentage of the total counts from both kidneys, after background has been subtracted. With symmetrical kidneys, this estimate can be made satisfactorily from a single posterior view, hut with asymmetric or ectopic kidneys it is important to obtain an

abnormal kidney. Renal ectopia and anomalies With horseshoe kidney (Fig. 29.38) the lower poles of both kidneys are joined by a bridge of renal tissue which lies anterior to the aorta and vena cava. This causes the ureters to take a more anterior course than usual, and there may be a degree of obstruction where the ureters cross the bridging renal tissue. DMSA is used to assess individual renal function, as well as to demonstrate the anatomical abnormality. The most common location for an ectopic kidney is in the midline of the pelvis (Fig. 29.39). Pelvic kidneys often show parenchymal distortion or dysplasia. so again the estimation of divided renal function offers a useful supplement to the demonstration of the abnormal location and anatomy of the kidney. When multicystic


Acquisition Because the concentration of DMSA in the renal parenchyma continues to increase while the levels in blood and in the extracellular fluid compartment gradually decline, the optimum time for imaging, with maximum kidney : background ratio, occurs 2-4 h after intravenous injection of the tracer. Images

should be obtained at 2-3 h in children and 3-4 h in adults. For demonstration of renal anatomy, posterior and posterior oblique views are obtained. With marked renal asymmetry, and particul arly with ectopic or transplant kidneys, an anterior view should also be obtained.

Interpretation The resolution of the images should be sufficient to demonstrate the demarcation of renal cortex and medulla (Fig. 29.37). The size, shape and location of both kidneys should be shown. Mass lesions of all types appear as non-functioning areas. Defects, deformities or scarring of the cortex should be recognised. Divided renal function Because of the relatively high kidney: background ratio. and the stability of the tracer kinetics during

the few minutes needed to acquire the images, "I'Tc-DMSA studies offer the most accurate non-invasive method for the estimation of the distribution of function between the two kidneys. The contribution of each kidney is measured as its individual count rate

99

Fig. 29.38 Horseshoe kidney. "'Tc-DMSA showing bridging renal tissue between the lower poles of both kidneys.


CONTENTS

Previous Page

91 4

Next Page


Fig. 29.39 Ectopic kidney. 99mTc-DMSA study showing right kidney in normal position; left kidney lying in the midline of the pelvis superior to the bladder.

dysplastic kidney is suspected, DMSA scintigraphy is used to confirm the total absence of function that is typical of this condition, and to differentiate it from other forms of unilateral cystic disease in infancy that are associated with preservation of some degree of renal function. I n a minor developmental anomaly, a portion of cortex may be infolded so as to lie adjacent t o the renal pelvis, producing an apparent mass or `pseudomass' on ultrasonography. Also known as `prominent columns of, Renal masses and pseudomasses

by their concentration of DMSA at scintigraphy. Tumours. cysts or abscesses show no uptake of DMSA so they appear as photondeficient areas in the renal images. The majority of true renal masses will be investigated s ated satisfactorily by anatomical imafing but occasionally, particularly with bilateral disease, functional i maging with DMSA will be needed to establish the distribution of functional renal tissue. This may indicate which of two kidneys with bilateral tumours should be operated on first (Fig. 29.40), or whether the residual function on one side is sufficient to justify a conservative surgical procedure (e.g. partial nephrectomy) rather than the simpler but more drastic procedure of full nephrectomy.

99

Fig. 29.41 Post infective scarring. mTc-DMSA study showing normal left kidney; scarred right upper pole (arrows).

more severe cases, focal scarring may ensue and the local functional deficit becomes permanent. Scarring is recognised as an area of indentation or discontinuity in the cortical image (Fig. 29.41 ). The severity of scarring may be expressed in terms of the extent of the lesion, the presence of multiple scars and by the distribution of renal function between affected and unaffected kidneys. Acute infection may produce total non-function in the affected area, indistinguishable from scarring, but in less severe cases may appear as an area of reduced function, with preservation these areas can be confirmed as normal renal parenchyma tion of the normal renal outline. In kidneys with a duplex collecting system and two ureters, the upper pole cortex is often damaged by obstruction, or the lower moiety may be injured by reflux and i nfection. In such cases DMSA imaging can be used to establish the functional contribution of the upper and lower moieties in order to assist decisions on surgery to the kidney. When to investigate? Although focal areas of diminished or absent function are well shown by DMSA scintigraphy, it must be remembered that acute infection and permanent scarring cannot be distinguished by this method. If the presence or extent of an acute l obar nephronia is under question, it is appropriate to carry out DMSA imaging at the time of the infection, or shortly afterwards. However, if the aim of DMSA imaging is to detect and monitor renal scarring, it is important to delay the examination for long enough to allow functional recovery to reach a maximum. The ti me interval required for this is variable, but it is recommended that DMSA for scarring should he delayed for at feast 3 months (and probably 6 months) after the infective episode.


Infection and scarring Infection in the renal parenchyma (acute pyefonephritis or acute nephronia) produces diminished function i n the affected areas, often a small subsegmental wedge of the renal cortex. Under favourable conditions the infection may heal by resolution, and in such cases the anatomical integrity of the affected area is maintained and normal function is restored. In

Dynamic renal studies Because nephrons under most circumstances act as individual functional units, it is not necessary to i nvestigate perfusion, filtration, reabsorption and transit functions i ndependently. In most circumstances, tracers which are handled differently by the kidney will produce similar results in the estimation of individual renal uptake and clearance. 9mTc-DPA (excreted almost entirely by glomerular filtration) and 1 2 h3ipuran(eIxctd-vylgbuarsection)hvb superseded by 99mmTc-MAG3 . 99 mTc-MAG3 i s loosely protein-hound in the circulation. A small proportion is filtered but a much higher proportion is secreted by the tubules, and although the extraction efficiency is not as good as Radiopharmaceuticals

99 Fig. 29.40 Bilateral Wilms' tumours. mTc-DMSA study showing extensive replacement of the left kidney; smaller tumour at the hilum of the right kidney.


Previous Page

CONTENTS

Next Page


915

that of hippuran, the rate of excretion is much faster than with DTPA. Labelling with technetium-99m gives a lower radiation dose to the patient, with higher count rates giving better imaging and measurement statistics. The typical amount of activity administered for an adult is 100 MBq for excretion studies, or up to 200 MBq if perfusion imaging is also being performed. i s less costly and also quicker and easier to prepare DTPA than99mTc-MAG3 , and so it may be used in emergent situations where a rapid result is needed, particularly in respect of perfusion, e.g. with acute oliguria in the renal transplant recipient. Excretion is slower than with MAG3, count statistics are interior, and a larger amount of activity needs to be administered (300 MBq for excretion studies and up to 800 MBq for perfusion imaging). Diphosphonate agents used for bone scintigraphy (e.g. 99 mTcmedronate) are also excreted almost entirely by glomerular filtration. In patients having scintigraphic surveys for hone metastasis, where there is no indication for imaging in the early vascular and soft-tissue phases, a satisfactory renal study may be obtained in the first 30-40 min after injection. Conventional bone imaging can be carried out 2-3 h later as normal, allowing both investigations to he completed at single visit.

urine flow rate in order to stress the system. Furosemide (frusemide) (normal adult dose 20-40 mg, scaled down by body surface area for children) produces a maximal diuretic response in the normal kidney within 5-10 min after injection. Currently there is no consensus on the optimum timing of diuretic injection. The most widely used technique requires intravenous furosemide to he given 15-20 min after starting the study (F+20 study). Image data are then collected for a further 15-20 min to determine the response of the kidneys to an increased flow rate. Alternatively, furosemide may be given 15 minutes before the injection of the radiotracer, in order to ensure that the patient is already undergoing a maximum diuresis at the start of data acquisition (F-15 study). Finally, some users prefer to give the furosemide and radiotracer simultaneously, so that a diuresis develops rapidly during the early part of this study. Each of these methods can produce satisfactory results, but each requires appropriately specific criteria for interpreting the results. The F+20 method has the advantage that it demonstrates a diuretic response, i.e. the change from resting to high urine flow rates, while the F-15 approach is thought to he more decisive in detecting minor degrees of obstruction.

Patient preparation Minor degrees of ureteric or outflow obstruction will not be detected if the rate of urine flow is slow. Also, giving a diuretic may be hazardous if the patient is already

Data analysis The image data is analysed by deriving time-

dehydrated. For these reasons it is essential that the patient is well hydrated at the start of the procedure. Preparatory instructions should encourage the patient to take plenty of fluid before attendi ng for the procedure, and a further 300-500 ml of clear fluid should he given 30 min before injection.

from the heart (if included). After correcting the renal curves for background activity, the early part of the curves may be used to calculate split renal function. The kidney contains a mixture of short and long nephrons but it is known that the shortest parenchymal transit time within the kidney, i.e. the time taken for filtrate in Bowman's capsule and the proximal tubule to reach the renal pelvis, is 2.5-3 min, so it is safe to assume that none of the tracer entering the kidney during this period after injection will have reached the outflow tract. The relative renal function can then be expressed as the ratio of the area under the renogram curves of the two kidneys obtained during the period 40 s - 2 min 40 s after bolus injection of the tracer. The first 40 s are usually excluded in order to avoid errors that may be introduced by vascular structures overlapping the renal and background areas during the first pass of the tracer. If furosemide has been given during the procedure (F+20 technique), a numerical estimate of the diuretic response is calculated, e.g. by measuring the residual counts in the kidney 20 min after diuretic injection as a proportion of the maximum renal counts. It may be helpful to confirm that an adequate diuresis has been achieved by measuring the average urine flow rate during the exami nation. This is achieved by having the patient empty the bladder

Acquisition With the patient in a reclining (dentist's chair) or

supine position, posterior view images are obtained with the gamma camera positioned so as to include both kidneys and the bladder. If deconvolution analysis is to be performed (see below), the heart should also be included in the field of view. Transplanted kidneys are normally located in the left or right iliac fossa so an anterior view is appropriate in these cases. If the patient is catheterised, the bag should also he within the field of view. The radiopharmaceutical is given as a bolus injection into a peripheral vein. If an assessment of perfusion is required, it is i mportant to deliver the bolus as tightly as possible. The volume of the injection is typically small (about I ml or less) and several techniques have been described for ensuring a speedy transit of the bolus to the central circulation. Rapid image acquisition is required to demonstrate the perfusion phase, for example one frame per second for the first 30-40 s after i njection. For the excretion phase, sequential images are acquired over periods of 10-30 s each for 20-30 min. The supine position is preferred to sitting because it is more easily reproducible and there is less likelihood of the patient moving during the procedure. The disadvantage of the supine position is that a dilated but unobstructed (baggy) renal pelvis may not drain freely, so it is important to obtain a further spot view of the urinary tract with the patient sitting, after the dynamic acquisition has been completed. Additionally, a final spot view obtained after micturition is helpful to demonstrate the completeness of bladder emptying and to confirm prolonged hold-up in the obstructed upper tract.

activity curves from regions of interest placed over each kidney, from a background area around each kidney, from the bladder and


Use of diuretics The sensitivity of dynamic renal imaging for detecting urinary tract obstruction is maximised by increasing the

i mmediately before starting the study, and also at the end of the study, and measuring the volume of urine generated during that ti me. Residual urine volume after micturition can also be assessed by comparing the counts in the bladder before and after micturition, and measuring the volume of urine passed. Deconvolution analysis is a mathematical method of calculating what the output curve of an organ (in the case of the kidney, the renogram) would have looked like if the injection had been made as a bolus in the renal artery rather than into a peripheral vein. This is done by considering the observed renogram as a complex sum of a series of individual renograms that would have been obtained by sequentially injecting different amounts of tracer into the renal artery. The instantaneous concentrations entering the renal artery


Previous Page

916

CONTENTS

Next Page


are derived from the time-activity curve of the heart, so, knowing i nput of the tracer at each moment during the study, it is possible to differentiate the observed output function (the actual renogram) to derive a theoretical curve which represents the range of transit ti mes between the renal artery and the renal pelvis. Although the calculation of mean parenchymal transit times offers a novel aspect of function in renal disease, the successful use of deconvolution methods requires meticulous attention to technique, and the added diagnostic value of their use in routine clinical practice remains uncertain. Interpretation With MAG3, the images obtained during the first pass of the tracer illustrate vascular anomalies in the abdomen as well as the perfusion of the kidneys (Fig. 29.42). Images obtained 3-5 min after injection illustrate the size, shape and location of the kidneys. Serial images obtained during the excretion phase give a visual impression of relative renal function, rate of transit of activity through the renal pelvis, drainage from the pelvis in the sitting position, and the residual bladder volume after micturition. The renogram gives several different indications of renal function. The initial, steeply rising part of the curve represents the delivery of circulating tracer to the kidney, and so the slope of this part of the curve is proportional to renal perfusion-the steeper the slope, the better the perfusion. As mentioned above, the area under the curve between 40 s and 2 min 40 s may be taken as a measure of the contribution of each kidney to overall renal function. The normal kidney produces a curve which peaks within a few minutes after i njection and then falls (Fig. 29.42). The height of this peak repre-

sents the maximum amount of tracer which accumulates in the kidney. If the anatomical region from which the renogram is derived includes the renal pelvis, then the height of the peak is influenced by the volume of the pelvis-the reservoir effect of a baggy renal pelvis will produce a much higher peak on the renogram curve than a kidney of equal function with a small volume pelvis. However, the height of the peak is also affected by the uptake function of the kidney-with diminished function the magnitude of the peak activity is less. The time interval between the injection and the peak of the renogram (Tmax ) is prolonged in renovascular disease because the filtration pressure is reduced and the intrarenal transit times are slowed. However, a delayed TmaX can also result from obstruction or from the reservoir effect of a dilated renal pelvis. Response to diuresis The distinction between obstruction and the

reservoir effect of an unobstructed but dilated renal pelvis can usually be made from the response to diuresis. With the normal kidney, administration of furosemide 20 min after MAG3 produces a rapid increase in urine flow so that activity retained within a baggy renal pelvis is displaced by urine containing much lower concentration of activity, so the observed count rate from the kidney falls. Various indices have been described to express the rate of elimination of tracer from the renal pelvis in response to diuretic stimulus, but there is no single index which is universally accepted. One simple approach is to say that if renal activity falls by more than 50% within 20 min of diuretic injection then obstruction is very unlikely. If the activity remains constant or increases, then obstruction is extremely likely. Activity falling by

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ Fig. 29.42 Normal 99mTc-MAG3 dynamic renal study: (A) Part of the first-pass acquisition showing perfusion of aorta and both kidneys; (B) selected i mages from the excretion series with symmetrical uptake and clearance from both kidneys; (C) renogram curves from the perfusion sequence in the first 30 s (left) and the excretion curves up to 30 min (right). LK = left kidney; RK = right kidney; Ao = aorta.


CONTENTS

Previous Page

Next Page

THE UROGENITAL TRACT: ANATOMY AND INVESTIGATIONS l ess than 50% in 20 min may suggest low-grade obstruction but it is important also to consider the level of function within the affected kidney. Usually the first aspect of renal function to be affected in disease is the ability of the kidney to concentrate the urine, so the failing kidney is typically unresponsive to the doses of diuretic agents that would be effective in the normal kidney. Distinguishing between a kidney which is sufficiently damaged to prevent a normal diuretic response, and one with lowgrade obstruction, remains difficult (see below). Low-grade obstruction is also suggested by a renogram which shows an i ncomplete fall-off in response to a diuretic stimulus, and then starts rising again-the delayed double-peak sign described by Homsy.

Captopril scintigraphy With renal arterial disease, the perfusion pressure of the afferent glomerular arterioles is reduced, and filtration pressure is sustained by efferent arteriolar vasoconstriction mediated by the reninangiotensin system. Administration of angiotensin-converting enzyme (ACE) inhibitors such as captopril blocks the efferent vasoconstriction, causing a fall in filtration pressure, which leads to a reduction in the GFR of the affected nephrons. Because of the reduced volume of glomerular filtrate, the mean transit time is prol onged and this in turn allows a greater reabsorption of water. These mechanisms produce in the captopril renogram the typical changes of diminished uptake, flattened peak and delayed Tmax . In severe cases the clearance may be so slow that the curve continues rising throughout the period of observation. Preparation For a captopril test, the patient should be off captopril and diuretics for 2 days, and off other ACE inhibitors for 7 days. Obtaining a baseline study without captopril intervention i mproves the specificity of the test. Pre- and post-captopril studies can be obtained on consecutive days, or even on the same day if the patient has good renal function. The patient should be normally hydrated. The effect of captopril on renal function is not mediated through a fall in systemic blood pressure, and such a fall is undesirable because it may cause changes in the renogram of the unaffected kidney. A small dose (25 mg) of captopril is given orally I h before injection of the radiopharmaceutical. The patient should be kept supine throughout and blood pressure should he monitored before, during and after the procedure. Because some patients are very sensitive to even small doses of captopril, a recommended precaution is to maintain an intravenous saline infusion with a period of the study to allow rapid treatment of any sudden hypotension.

917

Grade 2a

More prolonged delay in T ax (greater than 1 1 min) but still with an excretion phase Grade 2b Continually rising or flat curve Grade 3 As grade 2b, with marked reduction in function of the affected kidney. With MAG3, Tma x occurs earlier than the DTPA time quoted above.

Renal transplantation The clinical setting of the transplant kidney provides differing and specific problems which require modifications in the technique for renal dynamic studies. Typically, the kidney is placed in the right iliac fossa, or sometimes on the left. An anterior view of the pelvis will encompass the kidney, ureter and bladder. Since there is only a single kidney, the question of split function estimation does not arise. Dynamic studies are obtained as normal, with perfusion and excretion imaging. In the context of post-transplant oliguria, calcul ation of a perfusion index is often helpful. This is derived by obtaining a time-activity curve from the parenchyma of the kidney and also a curve from the adjacent iliac or femoral artery. A fast bolus injection of the tracer results in a steeply rising arterial curve with a sharp peak (T ,,,, x ). After background subtraction, the areas under the arterial and renal curves are measured up to T111,11. The ratio of these two areas, after normalisation for differences in the size of the regions of interest from which the curves were derived, defines the perfusion of the renal parenchyma relative to the blood flow through the adjacent large artery. Transplant perfusion index = (flow in renal parenchyma/flow in iliac artery) x 100%. Using this method of calculation, a perfusion index greater than 50% will usually be associated with good renal function.


99m Acquisition A 30 min dynamic study using Tc-MAG3 is carried out. Images are obtained at 1, 2, 5, 10, 20 and 30 min, with renogram curves for the whole study obtained in the usual way.99mTc-DTPA may be used as an alternative in patients who have good renal function.

Interpretation As with unenhanced renal studies, interpretation of the captopril study is based on inspection of the images and analysis of the curves. The critical features indicative of renal vascular disease are reduced function and delayed transit. Affected kidneys are often small, although this is not a specific feature. The following grading system was used in a European multicentre study: Grade I

Mild delay in Tmax (6-11 min using a falling excretion phase

9 99m

Tc-DTPA) with

Radionuclide cystography

This is the radionuclide analogue of X-ray micturating cystography with iodinated contrast media. In outline, the bladder is filled with fluid containing a non-absorbable radiotracer, and dynamic imaging of the urinary tract is carried out during micturition in order to detect vesicoureteric reflux. This may be achieved with direct instillation of the tracer into the bladder, but this approach does not capitalise on the major advantage of the radionuclide technique, which i s the possibility of avoiding bladder catheterisation. Much more clinically useful is the technique of indirect cystography, which is basically similar but which uses as a starting point the full bladder 99 at the end of a conventional dynamic renal study. Using mTcMAG3, renal activity normally falls off to a considerable extent by the time the bladder is full, so the presence of baseline counts in the background and in the kidneys at the start of micturition is not a significant disadvantage. Preparation and technique This is basically as for diuretic renography, described earlier. After completing the initial part of the study, including the response to diuresis, the patient is placed in a sitting or standing position and the camera repositioned to include both kidneys and the bladder within the field of view. Image frames of 5-10 s duration are collected during micturition. Time-activity curves are derived from the region of the bladder, each kidney individually, and also from regions along the line of each ureter.


Previous Page

918

CONTENTS

Next Page


Interpretation Vesicoureteric reflux is confirmed if there is an i ncrease in kidney counts during micturition. Lesser degrees of reflux may appear as increased counts in the ureters, which may be transitory.

Applications of dynamic radionuclide studies Urinary tract obstruction Obstruction implies increased resistance to flow, and resistance can only be assessed accurately by si multaneous measurements of pressure and flow. Such measurements require invasive techniques and so are unsuitable for routine clinical use. Anatomical imaging methods rely on demonstrating the structural abnormalities which cause obstruction and those which result from its effects, while radionuclide techniques demonstrate the alteration in flow caused by obstruction, as well as the consequent functional deficit that may follow. With acute ureteric obstruction (e.( , . by a stone), renal perfusion and glomerular filtration is maintained at first, and increased tubular reabsorption of water compensates, at least in part, for the cessation of urine flow, so that radiotracers continue to accumulate in the kidney. This mechanism gives rise to the characteristic obstructive renogram which shows a continuing rise (Fig. 29.43). A rather si milar rising curve may also he produced by an unobstructed kidney if the extrarenal pelvis is large, causing a reservoir effect. However, in such cases an effective diuresis will wash out the contents of the renal pelvis, causing a rapid fall in the renogram (Fig. 29.44). It may be argued that there is no such thing as partial obstruction-resistance to flow is either normal or increased. However, obstructions may be low-grade or high-grade, with minor or major changes in intrarenal pressure. High-grade obstruction is usually clearly shown on the renogram and its associated images. Recognising lesser degrees of obstruction is more difficult. The renal

Dilated but unobstructed renal pelvis. 99mTc images at 5, 10 and 15 min show rapid uptake and clearance from the left kidney; slower clearance from the right kidney. Renogram curves (bottom right) show normal left side and delayed peak on the right with rapid washout followi ng furosemide (frusemide) i njection. LK = left kidney; RK = right kidney; B = bladder. Fig. 29.44

with low-grade obstruction the kidney may show a transient and i ncomplete diuretic response. In some cases this results in the renogram beginning to rise again after a diuretic-induced fall ( Homsy's double peak sign; Fig. 29.45), while in other cases the rate of fall of the curve may he less than expected. A similar problem arises in patients with diminished renal function-the


response to a diuretic is not an àll or nothing' phenomenon, and

Unilateral PUJ obstruction. 99 mTc-MAG3 i mages at 1 min (A), 5 min (B) and 15 min (C), showing typical left hydronephrosis with normal clearance from the right kidney. Renogram curves (bottom right) show normal clearance on the right and an obstructed left side. LK = left kidney; RK = right kidney; B = bladder. Fig. 29.43

99 nTc-MAG3 diuretic study images at Fig. 29.45 Low-grade obstruction. 1, 5 and 20 min show normal uptake and clearance on the right; normal

uptake on the left but incomplete clearance. Renogram curves (bottom right) show normal right side and normal uptake on the left but after an i nitial fall the excretion curve rises again (Homsy's sign). LK = left kidney; RK = right kidney; B = bladder.


Previous Page

CONTENTS

Next Page


919

Fig. 29.46 Right renal artery stenosis. 99m -Tc-MAG3 dynamic study shows reduced blood flow to the smaller right kidney on the perfusion series (A), delayed excretion on the 15 min image (B), and the renogram curve (C) shows reduced uptake, delayed T max and slower clearance from the right kidney. LK = left kidney; RK = right kidney. affected kidneys may be unresponsive to diuretics, so the usual rapid fall in the renogram does not occur. When trying to distinguish between impaired but unobstructed kidneys that are unable to respond normally to diuretics and those with chronic obstruction and functional impairment, the concept of òutput efficiency' is helpful. This requires the diuretic response to be considered in relaLion to the uptake function of the kidney. If the input to the kidney i s diminished, a `normal' diuretic response cannot be expected. The rate of fall of the renogram during the diuretic phase should he compared with the rate of rise during the uptake phase-if the relationship between the two is the same as in a normal kidney, then obstruction is very unlikely. Renovascular hypertension The main reason for investigating the renal tract in patients with hypertension is to detect the minority of patients with renal vascular disease that is treatable by anatomical means angioplasty or surgery. However, it must be emphasised that renovascular hypertension may also result from s mall-vessel disease, which shows no specific features at angiography. Radionuclide studies will select out those hypertensive patients who have renovascular disorders, but cannot distinguish between large-vessel and small-vessel disease, as, whatever the l evel of arterial obstruction, the functional disturbance is similar. Characteristic changes in the dynamic renal study are those of prolonged transit due to increased water reabsorption in the distal tubules, and diminished renal function associated with decreased blood flow caused by arterial narrowing. These changes are manifest by reduced gradient in the uptake part of the renogram, reduced and delayed Tmax and prolonged transit times (Fig. 29.46). The affected kidney may be small or normal in size. Enhancement of the dynamic study with captopril magnifies these functional changes and unmasks some additional cases in which the baseline study is not diagnostic. Captopril-enhanced renography has good results as a screening test for renovascular hypertension (sensitivity and specificity 80-90%) when the selection of patients for investigation is appropriate a renal cause is more likely in patients who arc young and those with severe or resistant hypertension. Discrepancies between scintigraphy and angiography occur because some patients with renovascular hypertension have only small-vessel disease, while angiographic renal artery stenosis occurs incidentally in a substantial proportion of elderly patients, and so may be associated with essential hypertension in occasional cases.

versible (including acute or chronic renal failure). The rare but i mportant occurrence of acute renal failure as a consequence of urinary tract obstruction must also be recognised. A99m Tc-MAG3 or DTPA study in ATN typically shows that renal perfusion is well maintained, even in the presence of anuria or severe oliguria. The GFR is relatively well preserved, but the filtered or secreted solutes either remain in the kidney or are reabsorbed (abnormally) from damaged tubules. Dynamic images show fairly good perfusion, with subsequent fading of the renal images and little or no excretion (Fig. 29.47). The renogram shows an initial vascular peak, which usually then falls in parallel with blood clearance curve. With acute renal failure caused by intrinsic parenchymal disease or major vessel obstruction, perfusion is severely impaired or absent. With acute or chronic renal failure, a low level of excretion may persist, giving renograms with markedly impaired uptake phases and a fiat excretion phase. In acute renal failure caused by obstruction, renal perfusion is well maintained. Renal transplantation Although ultrasonography is the first-line technique for monitoring renal transplants, there are several indications for scintigraphy as a supplementary test when ultrasound is technically difficult or when its results are inconclusive or


Acute renal failure The aim of scintigraphy is to discriminate between the potentially reversible condition of acute tubular necrosis (ATN) and those conditions that are progressive or irre-

unexpected.

Fig. 29.47 Acute tubular necrosis. 99mTc-MAG3 study shows perfusion of both kidneys is reduced (A) and excretion images at 1 min (B) and 20 min (C) show persistent retention of the tracer in the kidney with no excretion. Renogram curves (D) show immediate uptake but no clearance. LK = left kidney; RK = right kidney.


Previous Page

920

CONTENTS

Next Page


Fig. 29.48 Acute oliguria after renal transplantation. Anterior perfusion phase images from 99 mTc-DTPA study show that blood flow to the transplanted kidney in the right iliac fossa is well maintained. Diagnosis: acute tubular necrosis.

Fig. 29.49 Acute oliguria after renal transplantation. Anterior view dynamic images 2, 5 and 10 minutes after injection of 99 mTc-DTPA show a photon deficient area which represents the totally ischaemic graft in the left iliac fossa; diagnosis: renal vein thrombosis.

Fig. 29.51 Suspected obstruction in renal transplant. 99mTc-DTPA study shows good uptake at 2 min (left) with activity reaching the bladder via a distended ureter and collecting system at 20 min (right). Diagnosis: obstruction confirmed.

Fig. 29.52 Sudden deterioration in renal transplant function.9mTc-DPA study shows normal uptake at 2 min (left); at 20 min (right) some activity is draining via a bladder catheter (C) but much of the activity is leaking into the peritoneal cavity (L). Diagnosis: ureteric leak. period after cadaveric transplantation, perfusion is relatively well maintained (Fig. 29.48). If oliguria results from thrombosis of the transplant renal artery or vein, the kidney becomes avascular and appears on scintigraphy as a photon-deficient area (Fig. 29.49). Acute rejection is associated with vascular obstruction at arteriolar l evel, and scintigraphy shows severely impaired perfusion, hut enough to demonstrate the kidney is viable (Fig. 29.50). Although renography is not the method of choice for detecting outflow probl ems, the demonstration of delayed clearance of tracer from the renal pelvis supports the diagnosis of obstruction at the vesicoureteric anastomosis (Fig. 29.51). Another cause of sudden deterioration in function, particularly in the postoperative period, is leakage of urine from the bladder or ureter, and again scintigraphy may be occasionally helpful in this situation (Fig. 29.52). Transplant renal artery disease is best demonstrated by magnetic resonance angiography.


Reflux and infection The major advantage of indirect radio-

Fig. 29.50 Acute oliguria after renal transplantation. Anterior view perfusion images (A) and excretion images at 1 min (B) and 20 min (C) show severely impaired perfusion of the transplant kidney in the right iliac fossa with slowly increasing uptake but hardly any excretion to the bladder. Renogram curve (D) shows the typical flat curve of a badly-functioning kidney but without the initial vascular spike usually seen with acute tubular necrosis. Diagnosis: severe rejection. In acute oliguria, particularly in the postoperative period, a perfu99n sion study using Tc-MAG3 or DTPA gives an image of the distribution of blood flow within the kidney, and an estimation of its perfusion. With ATN, which is very common in the postoperative

nuclide cystography is that it avoids the need for bladder catheterisation. The disadvantage is its lack of anatomical detail. For infants and children with suspected reflux disease, an appropriate strategy is to carry out X-ray micturating cystography at the initial assessment in order to detect anatomical abnormalities of the outflow tract as well as looking for reflux. Indirect radionuclide cystography can then be used as the follow-up examination to monitor response of reflux to treatment (Fig. 29.53). The i ndirect radionuclide method also allows monitoring of individual renal function. Using 99 Tc-MAG3 as the tracer, renal i mages about 2 min after injection should he of sufficient quality


CONTENTS

Previous Page

Next Page


92 1

detrusor smooth muscle (P,,.,) and any contribution from abdominal wall straining (Pabd ). Hence.

Common urodynamic disorders Lower urinary tract dysfunction may be caused by: 1. A disturbance of nervous control 2. Disorders of muscle function 3. Structural abnormalities.

Fig. 29.53 Radionuclide cystography showing reflux. Selected images from the dynamic sequence obtained during micturition (left) show tracer appearing in the right kidney and ureter as the bladder empties (arrows). Time-activity curves over left and right ureters (right) show no reflux on the left and mild but prolonged episodic reflux on the right.

to show major scars, but DMSA scintigraphy produces better anatomical detail and a more accurate estimate of individual renal function.

Tim Whittlestone

Disturbances of nervous control The clinical picture is dependent upon the extent of denervation of the lower urinary tract. Complete lesions of the lower spinal cord decentralise the bladder and urethra; the bladder becomes acontractile and the urethral sphincter inactive. Bladder emptying then relics on abdominal straining. Lesions above the fifth lumbar segment preserve spinal reflexes but the coordination between the bladder and urethra is l ost (detrusor sphincter dyssynergia). Cerebral damage may result i n one of many Urodynamic disorders. Damage to the frontal l obes results in incontinence. Damage to the basal ganglia leads to i nvoluntary bladder contractions (detrusor overactivity). Disorders of muscle function Overactivity of the detrusor smooth muscle in the absence of any discernable cause (idiopathic detrusor instability) is a common Urodynamic finding and l eads to urgency, frequency and urge incontinence. Many disease processes that reduce bladder compliance, radiotherapy being one common example, may compromise muscle function and reduce bladder capacity. Bladder outflow obstruction caused by prostatic enlargement may lead to either under- or overactivity of the bladder muscle.

Urodynamic studies provide accurate and objective information on the pathophysiology of the lower urinary tract in patients with symptoms suggesting dysfunction of the bladder and/or urethra. The basis of urodynamics is the recording of pressure within the bladder (the cystometrogram) or urethra (the urethral profile) and the flow of urine (during voiding (the flow rate). The relationship between pressure and flow has been studied for 50 years but refinements in techniques and advances in technology, such as the addition of ultrasound and real-time imaging, have made urodynamics an accurate science. I n d973 the multidisciplinary International Continence Society (ICS) realised the need for uniform standards and terms in dynamic studies. The ICS has since published a series of recommendations that should he used when dealing with lower urinary tract function.

Structural abnormalities Pelvic floor dysfunction. bladder neck descent and urethral sphincter incompetence arc common urodynamic findings in postpartum women with stress incontinence. Bladder divcrticula, often the result of chronically raised bladder pressure, are clearly seen with videourodynamics and may affect both the pressure and flow characteristics. Finally, there is a trend for urologists to reconstruct the Tower urinary tract using bowel segments after total cystectomy. Both storage and voiding abnormalities may arise following reconstruction and synchronous cystometry and cystography can he invaluable in addressing these problems.

Lower urinary tract physiology

Urodynamics encompasses a number of complementary techniques of varying degrees of complexity, the application of which needs to he tailored to meet the clinical requirements of each patient. Those with particular relevance to the radiologist arc:

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ 1_11- 0-

There are two distinct phases in lower urinary tract function: storage and micturition. During storage, the normal detrusor function is characterised by an increase in bladder volume as it fills with urine but no significant rise in bladder pressure or involuntary cont ractions. During this phase the urethral closure mechanism maintains a positive pressure and continence is achieved. When the bladder is full to its maximum functional capacity, sensory mechanisms signal to the cortex and the subsequent conscious desire to void may be inhibited or acted upon. During micturition, normal detrusor contraction effectively empties the bladder as the urethral sphincter mechanism relaxes. The pressure within the bladder during micturition (P,,.,) is the sum of the pressure generated by the

Urodynamic techniques

• Ultrasound cystodynamogram • I ntravenous urodynamogram • Videocystometrography. Ultrasound cystodynamogram (USCD) The simplest investigation in the assessment of voiding dysfunction is the measurement of urinary flow rate using a volumetric flow meter. The device measures the volume of urine passed per unit time and the flow rate is expressed in millilitres per second. When the flow rate is combined with ultrasound the resulting USCD reveals more detailed information.


CONTENTS

Previous Page

922


The subject usually attends for the investigation with a full bladder. The bladder is scanned with any ultrasound probe that allows adequate visualisation and the volume of urine is calculated; t his is the functional bladder capacity. After the subject has voided i nto a suitably prepared flow meter, in private, the bladder is scanned once again to provide an assessment of the residual volume. This simple technique provides a non-invasive measurement of bladder emptying and urinary flow. It is particularly useful i n the diagnosis of bladder outflow obstruction, the monitoring of patients with acontractilc bladders and following surgical procedures on the lower urinary tract. Intravenous urodynamogram (IVUD)

The IVUD is a combination of the conventional intravenous urogram and the assessment of urinary flow rates. It has largely been replaced by USCI). In addition to the upper tract films of an intravenous urogram the patient performs a flow rate when the bladder feels full. The subsequent postmicturition film after voiding allows an assessment of the residual volume. Videocystometrography (VCMG) Generally, the majority of L[1 0dynamic units have neither fluoroscopic screening nor video recording facilities. However, the combination of synchronous cystography and cystometry with video recording (VCMG) is recommended in the assessment of complex cases, younger patients, neuropathic patients, those for whom corrective surgery has failed and patients with reconstructed bladders. Radiological screening provides real-time anatomical information during bladder filling and during micturition. It also allows the assessment of ureteric reflex, bladder neck support, the level of outflow obstruction and the competence of the urethral sphincter mechanism. With the patient supine on the X-ray screening table the external urethral meatus is cleaned with antiseptic solution and I % or 2% lignocaine gel is instilled into the urethra. A ]OF filling catheter is i nserted into the bladder and any residual urine drained. Fine-bore pressure catheters are inserted into the bladder and rectum (or vagina) and connected to fluid-filled pressure transducers -

(Fig. 29.54). Subtraction of the rectal pressure from the bladder pressure allows the true detrusor pressure to be constantly displayed. The pressure transducers are zeroed to atmospheric pressure and raised or lowered to the level of the symphysis pubis. Care i s taken to exclude air bubbles from the pressure lines and transducers. Contrast medium at room temperature is instilled into the bladder by a peristaltic pump at a constant rate (typically 50mi/min) . When the patient senses that the bladder is full (first desire to void) the X-ray table is tilted to the standing position. Filling may continue until the maximum bladder capacity is reached (strong desire to void). The filling catheter is removed prior to voiding and the table turned to the oblique position. The patient voids into a standard flow meter. X-ray screening may be continuous throughout the investigation or at intervals, depending on the i nformation sought by the investigator. The residual volume can be calculated at the end of voiding. It is common to ask the patient to i nhibit micturition during the test-the 'stop test'. In men. contrast medium will he milked hack from the distal sphincter proximally through the bladder neck and into the bladder. If there is bladder neck obstruction, contrast will be trapped within the prostatic urethra. Videourodynamic equipment is becoming increasingly sophisticated but all systems allow the pressure and flow data to be displayed alongside the real-time cystogram. Video recording allows analysis later and demonstration of the results to other specialists.

Upper urinary tract urodynamics Radiologists are often called upon to confirm upper urinary tract obstruction. Modern imaging techniques have largely dispensed with the need for upper tract urodynamics (the Whitaker test): however, in the absence of any definite obstructing lesion or calcul us, where there is pelvicalyceal dilatation and persisting symptoms upper tract studies may he of use. Under normal circumstances urine accumulates in the renal pelvis at a resting pressure of less than 5 cmH2O. The pelvic pressure rises to d0 cmH2O once dis-


Fig. 29.54

Next Page

Set-up for lower urinary tract urodynamics.


Previous Page

CONTENTS

Next Page

THE UROGENITAL TRACT: ANATOMY AND INVESTIGATIONS t ended and urine enters the ureter to be transported to the bladder by ureteric peristalsis. In obstruction, peristalsis becomes disordered and the pressure in the renal pelvis will rise above physiol ogical levels. The Whitaker test is performed following insertion of a nephrostomy tube and can be supplemented by fluoroscopy to define the anatomy of the upper tracts simultaneously. Bladder pressure is constantly measured by an indwelling urethral pressure catheter. Dilute contrast is infused into the nephrostomy tube via one arm of a Y connector at d0 ml/min. The other arm of the Y is connected to a pressure transducer. The bladder pressure is continuously subtracted from the pelvic pressure. A pressure difference between the renal pelvis and bladder of less the d5 cmH,O excludes obstruction. If the renal pelvic pressure exceeds bladder pressure by more than 22 cmH,O obstruction is confirmed. Between 15 and 22 cmH,O the result is equivocal. If both the bladder and renal pelvic pressure rise equally together. vesicoureteric reflex has occurred.

923

Fig. 29.55

Renal ultrasound. Corona) (A) and transverse (B) images of a neonatal kidney demonstrating prominent hypoechoic medullary pyramids and no renal sinus fat.

This section will review the common imaging investigations of the urinary tract in children, together with relevant anatomy. Many of these techniques are similar to those used in adults but need some modification when applied to children, and a few are peculiar to children. Examinations in children may take longer to perform than equivalent studies in adults, but it is important to gain the trust and confidence of children and their parents if the examinations are to be successful. An awareness of radiation protection is important. Children are relatively more vulnerable to the deleterious effects of ionising radiation as they are still growing and have a longer life expectancy than adults. Examinations involving ionising radiation need to he justified and optimised if they are to be performed. Every attempt should be made to answer the clinical question using techniques that do not employ ionising radiation, especially ultrasound.


Ultrasound The two most frequent indications for renal sonography are the evaluation of an antenatal abnormality (hydronephrosis or cystic change) and the investigation of a child who has had a urinary tract i nfection (UTI). Less common reasons include children with conditions associated with renal anomalies. Renal anatomy The kidneys of neonates and young children differ from those of older children and adults in many respects, not just in size. There is a gradual transition in the sonographic appearance with age. The kidneys of a term neonate measure approximately A Z cm in length. Because there is no significant fat i n the renal sinus around the hilar vessels and pelvis they lack the typical increased central hilar echogenicity seen in older children and adults. The renal contour may be quite lobular, reflecting persistent fetal lobulation. The medullary pyramids are usually very prominent and are of much lower echogenicity than the overlying cortex (Fig. 29.55). This may be so marked as to simulate dilated calyces unless this variation is appreciated. Some neonates in the first few days of life may demonstrate paradoxically echogenic 1

Renal ultrasound. Corona) (A) and transverse (B) images of the kidney of a 10-year-old child. The medullary pyramids are less conspicuous than in the neonatal period and there is now some echogenicity around the renal hilum. Fig. 29.56

medullary pyramids, which resolve once normal urine flow is established. This is due to deposition of proteinaceous deposits in the renal tubules (Tamm-Horsfall proteinuria). which are cleared as urine flow increases. The above appearances may be quite


CONTENTS

Previous Page

924

Next Page


striking in neonates, but become less obvious with increasing age as the kidney takes on a more typical adult configuration. Renal l ength also inca- eases throughout childhood to the average adult l ength of I I cm. In neonates the renal pelvis is often visible, although it should he less than 5 nun in transverse (AP) diameter. There is often a considerable ext arenal component to the pelvis. I n older children perisinus fat is laid down, producing the central hilar echogenicity within which is the renal pelvis. The medulla and cortex become more similar in echogenicity and are usually of slightly lower echogenicity than the adjacent fiver and spleen. The contour of the kidney becomes smoother as fetal lobufation resolves ( Fig. 29.56). Technique The ultrasound examination should atlempt to evaluate the whole urinary tract (bladder, ureters and kidneys). In all children, and especially those who are not toilet-trained, the examination should hegira with the bladder in case it spontaneously empties before it has been imaged. It should he thinwalled (2 cm but < 5 cm in greatest dimension, or multiple lymph nodes, none >5 cm in greatest dimension Distant metastasis

ment (i.e. inercrising the bandwidth reduces the chemical shift artefact but the downside is a reduced signal to noise) or fat suppression techniques.

patients restlessness during the examination. An antiperistaltic went such as glucagon can be administered to reduce bowel motion artefact. If available, a pelvic phased-array coil should be used for evaluation of the whole pelvis and abdomen. Transverse and coronal large field of view (typically 38-49 cm) T,-weighted i mages of the abdomen and pelvis arc obtained in order to evaluate the upper abdominal viscera, particularly the kidneys, for the presence of hydronephrosis, to identify any lymph node enlargement and the presence of bone metastases. Transverse and coronal small field of view (typically 20-24 cm) T,-weighted images of the pelvis are obtained in order to demonstrate the local and regional extent of tumour and to reconfirm the presence of lymph node metastases. Additional sagittal images can be obtained for tumours that are present in the bladder base or vault (Fig. 31.49). The use of intravenous gadolinium-chelate is not mandatory. It can be helpful to demonstrate small lesions and the extent of tumour spread through and beyond the bladder wall and invasion of adjacent organs, as the tumour enhances to a greater extent than the bladder wall (Fig. 31.50). A dynamic contrastenhanced technique produces the best results, as tumour enhances during the early phase following injection, contrasted with the later

There are three main reasons for imaging patients with established bladder cancer: (i) to stage the tumour; (ii) to assess response to treatment; and (iii) to identify recurrent disease. Purpose of imaging

I n the United Kingdom and Europe, the TNM system is used to stage bladder cancer. The extent of spread within the bladder wall, the perivesical fat and adjacent organs (T stage), spread to pelvic and abdominal lymph nodes (N stage), and distant metastases (M stage) are assessed (Box 31.2). Staging


Technique Prior to imaging it is essential to have a comprehensive and accurate clinical history. Recent endoscopic resection or biopsy gives rise to haemorrhage and oedema, which can produce misleading appearances on MRI. The imaging technique is outlined in the Royal College of Radiologists' booklet A Guide to the Practical Use of MRI in Oncology. The patient should have a moderately full bladder in order to distend the bladder wall. An overfull bladder will result in

Enhancing bladder carcinoma (t) in the posterolateral part of the bladder on a transverse T 1 -weighted spin-echo (SE 740/40) postgadolinium-DTPA image. Note the three layers-u no pacified urine (u), signal void posteriorly in the bladder due to the superparamagnetic effect of high concentrations of gadolinium-DTPA, with a high-signal layer in between due to the paramagnetic effect of normal concentrations of gadolinium-DTPA. Also note the heterogeneity of signal within the middle l ayer due to urine flow. r = rectum; arrows = fluid levels. Fig. 31.50

Extensive T3b bladder tumour shown on (A) a transverse T 2 -weighted spin-echo (SE 3000/100) image; (B) transverse (at the same anatomical level as (A)) and (C) parasagittal T 2- weighted spin-echo (SE 3000/100) images with fat suppression. The tumour (curved arrow) is infiltrating through the bladder wall (straight arrows), which is better appreciated on the fat suppression scans. p = prostate. Fig. 31.49

Previous Page

CONTENTS

Next Page THE BLADDER AND PROSTATE

1009

Multiple confined bladder tumours (t) on transverse T 1 - weighted spin-echo (SE 740/40) (A) pre- and (B) early postgadolinium-DTPA images. Note the enhancing tumour anterolateral to the left is well visualised but the posterior-based lesions are obscured due to the high-signal effect of normal concentrations of gadolinium-DTPA. r = rectum; s = seminal vesicles; u = urine. Fig. 31.51

enhancement of normal structures (Fig. 31.51). Gadolinium-chelate i s useful in the differentiation of enhancing tumour from nonenhancing intraluminal blood clots. The use of intravenous contrast has drawbacks. Enhancement within a tumour can make it more difficult to identify its outer margin separate from the high-signal perivesical fat. Enhancing pelvic lymph nodes and lytic bone metastases can become less conspicuous on postgadolinium-chelate T 1 weighted images. The use of fat suppression sequences helps to rectify these drawbacks and increase the sensitivity for the detection of hone metastases. MR findings in bladder carcinoma Lesions less than 1.0 cm in diameter cannot be reliably detected on MRI. On T,-weighted i mages the tumour is usually of medium signal intensity, higher than that of surrounding urine but lower than perivesical fat ( Fig. 31.52). The tumour size, its intravesical component and extension into perivesical fat are best demonstrated on T,weighted or fat-suppressed T2-weighted scans (Figs 31.49, 31.52). Tumours on T2-weighted images, particularly with fat suppression technique, show intermediate signal intensity lower than adjacent urine. T2-weighted images show extension through the bladder

wall and infiltration of adjacent organs, the former being recograised by disruption of the normal low-signal band (Fig. 31.53). Bladder carcinomas involving only the mucosa and submucosa (TNM stage TI) cannot be differentiated from those with superficial muscle infiltration (stage T2). It is sometimes possible to distinguish lesions involving only the superficial muscle layer (stage T2a) from those invading the deep muscle (stage T2b), based on the preservation of the integrity of the black line of the bladder wall remaining intact on T,-weighted images when the deep muscle layer is not involved. However, the important clinical question is whether sufficient bladder muscle is present deep to the tumour for endoscopic resection to be performed. The ability to obtain this information preoperatively can make some stage T3 l esions, at present treated by radiotherapy, amenable to endoscopic resection. Stage T3b tumour represents infiltration through the bladder wall into the perivesical fat, with stage T4a disease showing invasion of local organs and stage T4b indicated by pelvic or abdominal wall infiltration (Fig. 31.54). As with CT, MRI relies on enlargement of lymph nodes to demonstrate the presence of metastatic disease within them. This has the drawback that metastatic tumour will not be identified in


Fig. 31.52 Histological stage 13b bladder carcinoma (curved open arrow) on transverse (A) T,-weighted spin-echo (SE 700/40) and (B) moderately T 2 weighted spin-echo (SE 1800/80) images. The intraluminal component of the tumour is well shown in (A) but the destruction of the low-signal bladder wall (straight arrow) is better demonstrated in (B) Left-sided pelvic lymphadenopathy (n) shown in (A) is obscured in (B). b = bladder; r = rectum.

Previous Page 1010

CONTENTS

Next Page


Fig. 31.53 Extensive stage T3b bladder tumour (straight arrows) with bilateral obstructive uropathy on (A) transverse and (B) parasagittal T 2 -weighted spinecho (SE 3500/100) images with fat suppression. The dilated ureters (u) are clearly shown to contain highsignal urine with no spread of the transitional cell carcinoma into the ureters. This was suggested on CT (not shown). There is also ascites (curved arrows) present.

for identification of metastatic disease. Ten millimetres is usually taken as the upper limit of normal for lymph nodes within the pelvis (with more refined measurements noted for certain nodal sites common iliac 9 mm, internal iliac 7 mm, and obturator 8 mm). It should be noted that the maximum short axis diameter of the lymph node is used for measurement purposes, as it has been found that this measurement remains relatively constant irrespective of the orientation of the node in the vertical plane. Central and posterior pelvic nodes (paracervical, perirectal and presacral) are not usually visualised in normal individuals, therefore when they are i dentified they are highly suspicious of nodal metastases (Fig. 31.55). Thorough scrutiny of the visualised portions of the lung bases, liver and bones should be performed to identify distant metastatic disease. Bone metastases are most conspicuous on T,-weighted i mages and can be characterised by correlation with the accom-

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ panying T2- weighted images.

Fig. 31.54 Metastatic transitional cell bladder carcinoma on a transverse TI - weighted spin-echo (SE 520/15) image showing a soft-tissue mass in the perineum (arrow) involving the left corpus cavernosum and left inferior pubic ramus.

normal-sized nodes, and tumour may be incorrectly diagnosed within enlarged hyperplastic nodes. In addition, in bladder cancer l ymph nodes can frequently be completely replaced by tumour without becoming enlarged, resulting in high false-negative rates

Post-treatment appearances and recurrent disease Following

treatment with radiotherapy or surgery, there is frequently oedema, inflammation and fibrosis, with consequent thickeni ng of the bladder wall and distortion of the perivesical fat (Fig. 31.56). Radiotherapy change can persist for up to 2 years following treatment. Because of this, using MR] or CT to differentiate between post-treatment change and recurrent tumour is frequently unreliable. Comparison with pretreatment imaging is

Fig. 31.55 Pelvic lymphadenopathy (n) demonstrated on T,-weighted spin-echo (SE 740/40) images in (A) coronal and (B,C) transverse planes in different patients. The soft-tissue signal from the lymphadenopathy contrasts well with the signal void from flowing blood within adjacent vessels and high signal from surrounding fat. b = bladder; p = psoas muscles; t = bladder tumour; arrow = dilated left ureter filled with low-signal urine.

Previous Page

CONTENTS


1011

mandatory in order to identify a new soft-tissue mass at the site of suspected tumour recurrence, or the presence of new metastatic disease within lymph nodes or distant organs to indicate progressive disease. Dynamic imaging following gadolinium-chelate has been shown to help differentiate between radiotherapy change and recurrent tumour but it requires the construction and analysis of ti me-signal intensity curves and is not routinely employed in clinical practice.

Resected stage T1 /T2 bladder tumour with a focal area of wall thickening (straight arrows) and ulcer crater (curved arrow) with no residual tumour on a coronal T 2 - weighted spin-echo (FSE 3000/100) image with fat suppression. (1.5 T CE Horizon using a phased-array pelvic coil.) Fig. 31.56

MRI versus CT MRI has been consistently shown to be superior to CT in demonstration of tumour invasion of the deep muscle layer of the bladder (stage T2b) and early invasion of the perivesical tissues (stage T3a). In more advanced disease, CT and MRI are of similar accuracy for staging bladder cancer. However, while both CT and MRI rely on alterations in size and contour of adjacent organs in order to identify pathological changes, MRI has the advantage that changes in signal intensity also occur within those organs. MRI and CT both rely on lymph node enlargement to identify metastatic disease but MRI has the advantage over CT i n that it can differentiate between ectatic pelvic vessels and enlarged nodes without the need for intravenous contrast enhancement. A recent advance of MRI is in the use of ultrasmall iron oxide particles (USIOP) (particle size of 20 nm), which are i njected intravenously and taken up by normal and hyperplastic lymph nodes, producing a signal void (a superparamagnetic effect). Tumour-replaced nodes fail to take up the contrast and retain their signal characteristics (see p. 1007).

Prostate Normal appearances The normal prostate has a homogeneous

l ow-intermediate signal on T i - weighted images (Fig. 31.57). Zonal anatomy is best demonstrated on T,-weighted images. It comprises a low-signal central zone and a higher signal peripheral zone (Fig. 31.58). The transitional and central zones appear of si milar signal intensity and are loosely termed the central gland. The prostate does not have a well-defined capsule but is enveloped by a variable band of condensed fibromuscular tissue. The periprostatic venous plexus can be visualised as a thin rim of high signal intensity anterolateral to the peripheral zone (Figs 31.59, 31.60). Denonvilliers' fascia can be resolved on sagittal images separating the pr static from the rectum. The neurovascular


Cyst of Cowper's gland (c) on spin-echo (A) sagittal intermediate-weighted (SE 1100/26) and (B) transverse T2 -weighted (SE 2000/80) i mages. Note that the urethra is well visualised between the anterior and posterior parts of the cyst in (A). b = bladder; p = prostate; r = rectum; s =symphysis pubis; v = seminal vesicles; straight arrow = urethra; curved arrow = bulb of penis. Fig. 31.57

Fig. 31.58 Stage T3a prostate carcinoma (arrow) posterolateral to the right in the peripheral zone infiltrating through the capsule into the adj acent neurovascular bundle. p = peripheral zone; c = central gland.

Previous Page 1012

CONTENTS

Next Page


Box 31.3 Tl T2 T2a T2b T3a T3b T4

N1 M1

TNM staging of prostate carcinoma

Clinically inapparent tumour not palpable nor visible by imaging Confined to the prostate Involves one lobe Involves both lobes Extracapsular extension Invades seminal vesicle Fixed or invades adjacent structure other than seminal vesicles: bladder neck, external sphincter, rectum, levator muscles, and/or pelvic wall Regional lymph node metastases Distant metastatic disease

more extensive disease are usually treated with external beam 31.59 Transvese T -weighted spin-echo (SE 2000/80) image 2 through the prostate of an 80-year-old patient. Note that there is no differentiation between the central and peripheral zones due to benign prostatic hypertrophy. p = prostate; r = rectum; s = symphysis pubis arrows, periprostatic venous plexus. Fig.

Staging The International Union Against Cancer (IUCC) system i s used to stage prostate cancer (Box 313). The usual pattern of spread is direct extension through the prostatic capsule into the bladder base and seminal vesicles. The presence of Denonvilliers' fascia prevents direct spread to the urethra and rectum unless there is extensive disease. The deficiency of the capsular condensation of fibromuscular tissue at the apex of the gland increases the likelihood of extracapsular extension occurring here. Lymph node metastases initially occur in the local pelvic nodes. particularly the obturator nodes, and in the retroperitoneal nodes in more advanced disease. Distant metastases occur most frequently to bone, particularly that in the lumbar and thoracic spine because of the communication between the periprostatic venous plexus and the presacral veins. Bone metastases also occur in the appendicular skeleton, ribs and skull. The majority (9590) of prostate cancers are adenocarcinomas. although other malignant tumour types do occur. notably transitional and squamous cell carcinoma. Within one gland, foci of tumour with different degrees of differentiation can occur. Poorly differentiated tumours tend to have a worse prognosis. The Gleason system of grading the degree of malignancy calculates the sum of the two predominant histological types within the gland and attrihutes each one a score out of 5, giving a sum of between 2 and 10, Glands with a score of 10 contain the most undifferentiated honour and therefore these patients have the worst prognosis.


Benign prostatic hypertrophy (p) on a fat-suppressed coronal spin-echo (FSE 3500/100) image. Arrows = periprostatic venous plexus. (1.5 T IGE Horizon using a phased-array pelvic coil.) Fig. 31.60

radiotherapy and hormone treatment. Following treatment, monitoring of the .serum prostate-specific antigen (PSA) is performed. Patients with rising PSA levels or others who are suspected of having recurrent disease have imaging evaluation of the skeleton using isotope bone scintigraphy, and of the lower spine and pelvis using MR1.

T 2 -weighted

bundles are sited posterolaterally at 5 and 7 o'clock positions on transverse sections of the prostate. Within this neurovaseular bundle are the nerves supplying the corpora cavernosa of the penis and the external urethral sphincter, the preservation of which is critical for nnaintaining erectile function and urinary continence.

Prostate cancer

Purpose of imaging MRI is used to stage the extent of prostate cancer once the diagnosis has been established and to identify the presence of recurrent disease following treatment. It is important to realise that MRI is not used in the primary diagnosis of prostate cancer. This is usually established following biopsy at transrectal ultrasound. The main role of MRI in staging prostate cancer is to determine whether disease is limited to the prostate gland. Debate still exists about the optimum treatment modality for limited disease, the choice being between nerve-sparing radical prostatectomy , radical external beam radiotherapy, or prostate brachytherapy. Patients with

Technique I maging technique is outlined in the Royal College of Radiologist booklet A Guide to the Practical Use o/ MR in empty Oncology. The patient should his bladder prior to the examination. If available, a pelvic phased-array coil should he used for evaluation of the whole pelvis and abdomen. An endorectal coil can be employed. This improves the spatial resolution within the area adjacent to the coil and can improve the accuracy i n determining extracapsular extension of tumour: therefore, endorectal coils are most valuable in the evaluation of early stage disease. They have the drawback of a limited field of view and near-field artefacts, and being invasive. Some MRI systems combine pelvic phased-array and endorectal coils and. although not widely employed, these should help to resolve some of the drawbacks of endorectal coil imaging. Transverse and coronal

Previous Page

CONTENTS


1013

large field of view T,-weighted images of the abdomen and pelvis are obtained in order to evaluate the upper abdominal viscera, particularly the kidneys, for the presence of hydronephrosis, to identify any lymph node enlargement and the presence of hone metastases. 'Transverse, coronal and sagittal small field of view T2-weighted images of the pelvis are obtained in order to demonstrate the local and regional extent of tumour and to reconfirm the presence of lymph node metastases. Sagittal and coronal images are particularly useful for evaluating the bladder base and extension of spread into the bladder neck and seminal vesicles. As in bladder cancer, the use of gadolinium-chelate is not mandatory. Tumours in the central zone can appear more conspicuens following contrast enhancement. A dynamic examination may help to identify localised extracapsular spread as prostate cancer enhances earlier than normal tissue. In addition, work has shown that there is a correlation between enhancement characteristics and the grade of the tumour.

Fig. 31.62 Prostatic carcinoma (stage T2) in the right peripheral zone (arrow) abutting onto the capsule on a fat-suppressed transverse T z -weightdspnco(TSE350/1)san.Notehcidalbng prostatic hypertrophy. (1 .5 T Philips Gyroscan using the body coif.)


Fig. 31.61 Muftifocaf prostatic carcinoma (curved arrows) in the left peripheral zone with associated benign prostatic hypertrophy on a transverse T 2 - weighted spin-echo (FSE 4000/104) image. Note the area of high signal (straight arrow) in the more posteriorfy placed tumour due to haemorrhage from recent biopsy.

Fig. 31.63 Stage T4a prostatic carcinoma (arrow) infiltrating the bladder base on a fat-suppressed coronaf T 2 -weighted spin-echo (TSE 3500/100) i mage. This tumour was not detected on transrectaf ultrasound examination (not shown). (1 .5 T Philips Gyroscan using the body coif.)

Fig. 31.64 (A,B) Mucin-secreting adenocarcinoma of the prostate (straight arrows) extending through the capsule with a large right pelvic nodal metastasis (n) on transverse T 2 -weighted spin-echo (SE 2000/100) images. Note the similar high signal from the primary prostatic tumour and nodal disease.

Previous Page 1014

CONTENTS

Next Page


MR findings in prostate cancer

It is re-emphasised that MRI is not used for the diagnosis of prostate carcinoma. A normalappearing prostate gland on MRI does not exclude the presence of a tumour. Conversely, a number of conditions can produce findings that mimic the presence of tumour. Benign prostatic hypertrophy which occurs in the transitional zone can be marked and give rise to a very heterogeneous appearance of the gland. Prostatitis and artefacts from haemorrhage and oedema, secondary to recent biopsies, can also cause confusion (Fig. 31.61). Approximately 70% of tumours arise in the peripheral zone, 20% i n the transitional zone, and 10% in the central zone. The usual pattern is of a single focus of low signal intensity on the T2-weighted images seen within the peripheral zone but tumours can or diffusely infiltrating (Figs 31.61-31.63). be multifocal Confusingly, up to a third of low-signal areas within the peripheral zone on T,-weighted images may not be due to tumour and represent fibrosis or other benign pathology. In addition, the carcinoma may be isointcnse or hyperintense compared with the peripheral zone, the latter found in rare mucin-producing adenocareinomas (Fig. 31.64). Tumour occurring in the transitional zone cannot be reliably distinguished from the heterogeneous signal due to benign prostatic hypertrophy. Extension of the tumour through the capsule is usually adjacent to the neurovascular bundles, at the sites of nerve penetration of the capsule (perineural infiltration). Macrocapsular penetration (arhitrarily defined as tumour extending more than 2 mm beyond the capsule) can be assessed on MRI, and is shown as focal thickening and/or bulging of the capsule (Fig. 31.65). Periprostatic infiltration can be demonstrated on T,-weighted images as a low signal within the periprostatic fat or as an intermediate signal using T 2 - weighted fat-suppressed scans. Tumour extension into the seminal vesicles, recognised as an alteration of signal as well as surface contour, is Extension of tumour into the bladder and rectum is best appreciated on transverse and sagittal T 2 - weighted scans. Direct infiltration of

Recurrent rectal carcinoma (t) infiltrating an enlarged prostate (p) from benign prostatic hypertrophy. b = bladder, curved open arrow = perforation into the perirectal fat space; straight arrows - thickened left perirectal fascia. Fig. 31.66

the prostate from primary tumours of the bladder or rectum can also be demonstrated (Fig. 31.66). T 1 - and T,-weighted images can be used for the detection of lymph node metastases. As with CT, MRI relies on enlargement of lymph nodes to demonstrate the presence of metastatic disease within them. Thorough scrutiny of the visualised portions of the lung bases, liver and bones should be performed to identify metastatic disease and to evaluate the presence of hydronephrosis. Sclerotic bone metastases from prostate cancer are frequently of low signal on T1-best demonstrated on transverse and coronal T,-weighted images. and T -2weightdmages. post-treatment appearances and recurrent disease After irra-


diation or hormonal therapy, the prostate often becomes low signal on T-,-weighted images, an appearance that should not be confused with tumour recurrence. The seminal vesicles can also become deformed and show low-signal change but may retain small focal areas of high-signal fluid. After a time, the peripheral zone of the prostate may revert to its normal high signal on T_ weighted images. Following treatment, a patient with prostatic carcinoma is monitored for recurrent tumour by serial serum PSA levels. Patients with rising levels normally have evaluation with isotope bone scintigraphy, to identify disease in the skeleton, and MRI, to identify disease i n the residual prostate or prostatic bed, and lymph node enlargement within the pelvis and retroperitoneum. The presence of an irregular new low-to-intermediate signal intensity mass on the T2-weighted images is highly suspicious of recurrent disease. Reference to previous imaging is, of course, very useful and recommended.

MRI versus transrectal ultrasound

Fig. 31.65

Stage T3a prostatic carcinoma extending through the capsule (arrow) posterolaterally to the left with involvement of the adjacent neurovascular bundle on a fat-suppressed transverse T.-weighted spin-echo (TSE 5000/117) i mage. (1.5 T IGE Horizon using a phased-array pelvic coil.)

Transrectal ultrasound has established an essential role in the initial evaluation and biopsy of suspected prostate cancer; however, the results for its accuracy in staging, particularly predicting capsular penetration and seminal vesicle invasion, have been variable. Transrectal ultrasound has the limitation of being operator-dependent and dependent on the quality of the ultrasound machinery available. It cannot evaluate l ymph node enlargement on the pelvic side-wall or retroperi-

Previous Page

CONTENTS


toneum or identify the presence of distant metastatic disease. Because of this, MRI has been adopted as the imaging modality of choice in staging prostatic carcinoma, although with reported accuracies ranging from 70 to 90% improvements are required, particularly in the staging of early disease. REFERENCES AND SUGGESTIONS FOR FURTHER READING Balaji, K. C., McGuire, M., Grotas. .1. (1999) Upper tract recurrences following radical cystectorny: an anaysis of prognostic factors, recurrence pattern and stage at presentation. Journal of Urology, 162, 1603-1606. Brossner, C., Bayer, G.. Madersbacher, S., Kuber, W., Klingler, C., Pycha, A. (2000) Twelve prostate biopsies detect significant cancer volumes (>0.5 ml). British Journal of Urology International, 85, 705-707. Brossner, C., Madershacher, S., Klingler, H. C., Pycha, A., Kuber, W., Marherger, M. (1998) A comparative study of a double-line versus a fanshaped technique for obtaining transrectal ultrasound-guided biopsies of t he prostate. European Urology, 33, 556-561. Charnsangavej, C. (1990) Lymphoma of the genitourinary tract. Radiologic Clinics of North A merica, 28, 865-877. Clements, R. (1996) The changing role of transrectal ultrasound in the diagnosis of prostate cancer. Clinical Radiology, 51, 671-676. Clements, R. (2001) Contemporary ultrasound guided biopsy in the diagnosis of prostate cancer. Imaging, 13, 18-26. Corrigan, N. T., Crooks, J., Shand, J. (2000) Are dedicated bladder films necessary as part of intravenous urography for haematuria? British Journal of Urology International, 85, 806-8 10. Crundwell, M. C., Cooke, P. W.. Wallace, D. M. A. (1999) Patient's tolerance of transrectal ultrasound-guided prostatic biopsy: an audit of 104 cases. British Journal of Urology, 83, 792-795. Fontaine, E., Leaver, R., Woodhouse, C. R. J. (2000) The effect of intestinal urinary reservoirs on renal function: a 10 year follow-up. British Journal of Urology, 86, 195-198. Ganaway, W. M., Alexander. F. E. (1997) Prostate disease: epidemiology, natural history and demographic shifts. British Journal of Urology, 79(suppl. 2), 3-8. Grossfeld, G. D., Coakley, M. B. (2000) Benign prostatic hyperplasia: clinical overview and value of diagnostic imaging. Radiologic Clinics of North Anmerica, 38. 31-47. Hakenburg, 0. W., Linne, C., Manseck, A., Wirth, M. P. (2000) Bladder wall t hickness in normal adults and men with mild lower urinary tract symptoms and benign prostatic enlargement. Neurology and Uroclvnamics, 19, 585-593. Hall, T. B., MacVicar, A. D. (2001) Imaging of bladder cancer. Imaging, 13, I -I0. Halpern, E. J.. Rosenberg, M.. Gomella, L. G. (2001) Prostate cancer: contrast-enhanced US for detection. Radiology, 219, 219-225. Herman.T E., McAlister, W. H. (1991) Radiographic manifestations of congenital anomalies of the lower urinary tract. Radiologic Clinics of North A merica, 29, 365-382. Hession, P., Flynn, P. N., Goodfellow, J. Murthy, L. N. S. (1999) Intravenous urography in urinary tract surveillance in carcinoma of the bladder. Clinical Radiology, 54, 465-467. Khadra, M. H., Pickard, R. S. Charlton, M.. Powell, P. H., Neal, D. F. (2000) A prospective analysis of 1930 patients with haematuria to evaluate current diagnostic practice. Journal of Urology, 163, 524-527. Leventis, A. K.. Shariat, S. F. (2001) Local recurrence after radical prostatectomy: correlation of US features with prostatic fossa biopsy findings. Radiology, 219, 432-439. Salah, M. A. (2000) Ultrasonography of urinary tract lesions caused by hilharziasis in Yemeni patients. British Journal of Urology International, 86.790-793. Samuel, M., Misra, D., Larcher, V., Price, E. (2000) Schi.stosoma haernatohium i nfection in children in Britain. British Journal of Urology International, 85, 316-318. Sandler, C. M., Hall, J. T., Rodriguez, M. B., Corriere, J. N. (1986) Bladder injury in blunt pelvic trauma. Radiology, 158, 633-638. Smeulders, N., Woodhouse. C. R. J. (2001) Neoplasia in adult exstrophy patients. British Journal of Urology International, 87, 623-628. Vaccaro, J. P., Brody, J. M. (2000) CT cystography in the evaluation of major bladder trauma. Radiographics, 20, 1373-1381.

1015

Wilkinson, B. A., Hamdy, F. C. (2001) State-of-the-art staging in prostate cancer. British Journal of Urology International, 87, 423-430. Yousem, D., Gatewood, O. M. B., Goldman, S. M., Marshall, F. F. (1988) Synchronous and metachronous transitional cell carcinoma of the urinary tract: Prevalence, incidence and radiographic detection. Radiology. 167. 613-618. Yu, J. S., Kim, K. W., Lee, H. J., Lee, Y. J., Yoon, C. S.. Kim, M. J. (2001) Urachal remnant diseases: spectrum of CT and US findings. Radiographics, 21, 451-461. MRI Aerts, P., Van Hoe, L., Bosmans, H., Oyen, R., Marchal, G., Baert, A. L. (1996) Breath-hold MR urography using the HASTE technique. A merican Journal of Roentgenology. 166, 543-545. Barentz, J. 0., Berger-Hartog, 0.. Witjes, J. A.. et al (1998) Evaluation of chemotherapy in advanced urinary bladder cancer with fast dynamic contrast-enhanced MR imaging. Radiology, 207, 791-797. Barentz, J. O., .lager, G. J., van Vierzen, P. B. J., et al (1996) Staging urinary bladder cancer after transurethral biopsy: value of fast dynamic contrastenhanced MR imaging. Radiology, 201, 185-193. Bluemke, D. A. (ed.) (2000) Oncologic MR imaging. Magnetic Resonance Imaging Clinics of North A merica, 8, 691-963. Brawer, M. K., Deering, R. E., Brown, M. S., et al (1994) Predictors of the pathological stage in prostate carcinoma. The role of neovascularity. Cancer, 73, 678-687. Brown, G., MacVicar, D., Ayton, V.. Husband, J. E. (1995) The role of i ntravenous contrast enhancement in MR staging of prostatic carcinoma. Clinical Radiology, 50. 601-606. Chan, T. W., Kressel, H. Y. (1991) Prostate and seminal vesicles after irradiation: MR appearance. Journal of Magnetic Resonance Imaging. 1, 503-506. Chen. M., Hricak, H., Kalbhen, C. L., et al (1996) Hormonal ablation of prostatic cancer: effects on prostate morphology, tumor detection, and staging by endorectal coil MR imaging. A merican Journal of Roentgenology, 166. 1 1 57-1163. Clare, M. C., Tempany, M. D. (eds) (1996) The male pelvis. Magnetic Resonance Imaging Clinics of North A merica, 4, 439-587. Epstein, J. I., Carmichael, M. J., Pizov, G., Walsh, P. (1993) Influence of capsular penetration on the progession following radical prostatectomy: a study of 196 cases with long term follow-up. Journal of Urology, 150, 1 35-141. Fernback, S. K., Feinstein, K. A. (1994) Abnormalities of the bladder in children: imaging findings. A merican Journal of Roentgenology. 162, 1 1 43-1150. Hamm, B., Laniado, M., Saini, S. (1990) Contrast-enhanced magnetic resonance imaging of the abdomen and pelvis. Magnetic Resonance Quarterly, 6, 1 08-135. Harris, R. D., Schned, A. R, Heaney, J. A. (1995) Staging of prostate cancer with cndorectal MR imaging: lessons from a learning curve. Radiographics, 15, 813-829. Hawnaur, J. M., Johnson, R. J., Isherwood, L. Jenkins, J. P. R. (1990) GdDTPA in MRI of bladder carcinoma. In Bydder, G., et al (eds) Contrast Media in MRI, International W orkshop, Berlin, pp. 357-363. Netherlands: Medicorn Europe. Hricak, It.. Carrington, B. M. (1991) MRI of the Pelvis--A Text A tlas. London: Martin Dunitz. Hricak, H., White, S., Vigernon, D., et al (1994) Carcinoma of the prostate gland: MR imaging with pelvic phased array coils versus integrated endorectal-pelvic phased array coils. Radiology, 193, 703-709. Huch Boni, R. A., Boner, J. A., Debatin. J. F., et al (1995) Optimisation of prostate cancer staging: comparison of imaging and clinical methods. Clinical Radiology, 50, 593-600. Husband, J. E. (1992) Review: staging bladder cancer. Clinical Radiology, 46,153-159. Husband, J. E. S.. Reznek, R. H. (eds) (1998) Imaging in Oncology. Oxford: I SIS Medical Media. Husband, J. E., Padhani. A. R., MacVicar, A. D., Revell. P. (1998) Magnetic resonance imaging of prostate cancer: comparison of image quality using endorectal and pelvic phased array coils. Clinical Radiology. 53. 673-681. Jenkins, J. P. R. (1990) Magnetic resonance imaging in oncology. In Johnson, R. J., Eddleston. B., Hunter, R. D. (eds) Radiology in the Management of Cancer, Ch. 2, pp. 25-58. Edinburgh: Churchill Livingstone. Jenkins, J. P. R., Zhu. X. P.. Isherwood, 1. (1990) Magnetic resonance i maging H: A quantitative approach. In Peeling, W. B. (ed.) The Role of Imaging in Prosstatic Cancer, pp. 1 9-26. Oxford: Medicine Group (UK).


Previous Page 1016

CONTENTS

Next Page


Johnson, R. J., Carrington, B. M., Jenkins, ,1. P. R., Barnard, R. J., Read, G., Isherwood. I. (1989) Accuracy in staging carcinoma bladder by magnetic resonance imaging. Clinical Radiologv, 41, 258-263. MacVicar, A. D. (2000) Bladder cancer staging. British Journal of Urology hiternational, 86 (suppl. 1), 111-122. Mirowitz, S. A., Brown, D. G., H eiken, J. P. (1993) Evaluation of the prostate and prostatic carcinoma with gadolinium enhanced cndorectal coil MR i maging. Radiologv, 186, 1 53-157. Outwater, E. K., Petersen, R. 0., Siegelman, E. S., et al (1994) Prostate carcinoma: assessment of diagnostic criteria for capsular penetration on endorectal coil MR images. Radiologv, 193, 333-339. Padhani, A. R., Gapinski. C. J., MacVicar. 1). A., et al (2000) Dynamic contrast enhanced MR[ of prostate cancer: correlation with morphology and tumour stage, histological grade and PSA. Clinical Racliologr, 55, 99-109. Passe, T. J., Blucmke, D. A., Siegelman, S. S. (1997) Tumor angio g enesis: t utorial on implications for imaging. Radiology, 203, 593-600. Quinn, S., Franzini, I)., Denlow, T., et al (1994) MR imaging of prostate cancer with endorectal surface coil technique: correlation with whole mount specimens. Radiology, 190, 323-327. Rifkin. M. D., Zerhouni, E. A., Gatsonis, C. A., et al (1990) Comparison of magnetic resonance innaging and ultrasonography in staging early prostate cancer-results of a mufti-institu(ional cooperative trial. New England Journal o/ Medicine, 323, 62 1-626. Rohinson, P.. Collins, C. D.. Ryder. W. 1). J., et al (2000) Relationship of MRI and clinical stage to outcome in invasive bladder cancer treated by radiotherapy. Clinical Radiology, 55, 301-306.

Rgrvik. J., Halvorsen, O. J., Albrektsen, G., Ersland. L., Deehlin, L., Haukaas, S. (1999) Use of pelvic surface coil MR imaging for assessment of clinically localised prostate cancer with histopathological correlation. Clinical Radiology, 54, 1 64-169. Saini, S.. Modic. M. 7'.. Hamm, B.. Hahn. P. F. (1991) Advances in contrast-enhanced MR imaging. Anteriran./ourna/ of Roenigenologv, 156, 235-254. Scheidler, J., Hricak, H., Vigneron, D. B., et al (1999) Prostate cancer: l ocalisation with three-dimensional proton MR spectroscopic imaging clinicopathologic study. Radiology, 213, 473-480. Schnall. M. D., Bezzi, M., Pollack, H. M.. Kressel, H. Y. (1990) Magnetic resonance imaging of the prostate. Magnetic Resonance Quarterly, 6, 1 -16. Stark, D. D., Bradley, W. G. (eds) (1999) Magnetic Resonance Imaging, 3rd edn, Ch. 27 (Urinary bladder). Ch. 30 (Prostate and seminal vesicles). St Louis: Moshy. Tentpany, C.. Zhou , X.. Zerhouni. E.. et al ( 1994) Staging of prostate cancer with MRI: results of Radiology Diagnostic Oncology Group projcctcomparison of different techniques including endorectal coil. Radiology, 192,47-54. Vinniconihe, S. J.. Normal, A. R.. Nicholson, V., Husband. J. E. ( 1995) Normal pelvic lymph nodes: evaluation with CT after hi pedal l ymphangiography. Radiology, 194, 349-355.


Previous Page

CONTENTS

Next Page

32 Cowper's glands These paired glands are embedded in the uroPosterior

urethral valves I n this condition there is anomalous

development of mucosal folds that obstruct the distal prostatic urthra i n males. They manifest as bladder outflow obstruction of varying severity, the most extreme producing bilateral hydronephrosis , renal failure and Potter's syndrome. The milder forms present with lesser degrees of outflow obstruction and poor urinary stream. Micturating cystourethrography demonstrates dilatation oh the prostatic urethra and a well-defined linear filling defect obstructing the urine flow. Mullerian duct remnants The caudal part of the mullerian duet

i n the male becomes obliterated (apart from the caudal tip, which persists as file verumontanum). Incomplete obliteration may give rise to cysts between the bladder and the rectum. These may be ac mptomatic or present with frequency, infertility, urinary obstructive symptoms or deep pcrineal pain on micturition or defecation. A cavity communicating with the posterior urethra derived from the caudal mullerian duct may he seen (prostatic utriele) and is considered the male homologue of the uterus. Again this may be asymptonnatic but is often associated with , maldevelopment of the genitalia (hypospadias. eryptorehidism intersex . etc.) and nay lead to urinary tract infection and ealeuli due to the presence of stagnant urine. These lesions may he demonstrated as cystic structures on CT, mRl and transreetal ultrasound.urethrography may demonstrate an extrinsic mass with a mullerian cyst or a communicating narrownecked, often irregular. cavity within the prostate utriele.

genital diaphragm and drain via duets of variable length (from a few millimetres to a few centimetres), which run anteroinleriorly on to the ventral surface of the perineal part of the bulhar urethra. They may he opacilied during urethrography in the presence of strictures or inflammation. The duets to the glands may he congenitally dilated (Cowper's duct syringoeele ). usually asymptomatieally but on rare occasions associated with urcthral obstruction. Other congenital urethral lesions The urethra is rarely duplicated congenitally, usually in the sagittal plane. The accessory urethra terminates along the shaft oh the penis and is usually hlind-ending, although if it conunutnicates with the bladder it is associated with continuous incontinence. Variable narrowing of the urethra may occur, ranging from mild stenosis to complete atresia. The more severe forms are lethal, associated with obstructive renal failure and potter's syndrome unless there is a persistent patent urachus or rectovesical fistula to decompress the system.


Hypospadias I n this congenital malformation the urethra let - -

urinates on the ventral surface oh the penis, abnormally proximath. The meatus is found anywhere from just proximal to its normal site back as far as the perineum in severe eases. There i s an association with other urinary tract anomalies and undescended testes. The anatomy is best demonstrated with tireI hroaraphy but this may be difficult as the meatus is abnormal and usually narrowed, resulting in upstream urethral dilatation, often uithsecondaryinfeetion.

I n developed countries these are most often traumatic in origin,

sometimes inflammatory and rarely neoplastic .

Inflammatory I nflammatory urethral strictures may develop fol-

l owing urethritis. Classically it follows gonoeoeeal urethritis but i s occasionally seen following non-specific urethritis with organi sms such as Chlamydia trachomrtis . Urethrography should he avoided while there is active urethritis but on rare occasions the features of urethritis are observed on urethrography and include contrast filling of the periurethral glands and multiple small sessile filling defects due to polypoidal mueosal hype - plasia (urethritis eystica) especially with gonorrhoea. The stricture is typically situated in the anterior urethra (Fig. 32.1 ), most commonly in the proximal bulbar urethra. Strictures may be multiple and of any length. Tuberculosis is a rare cause, being seen in the urethra in 2 1/ or less of urinary tract tubereulosis. It almost always arises from a focus elsewhere in the

1017

Previous Page 1018

CONTENTS

Next Page

A TEXTBOOK OF RADIOLOGY AND IMAGING Accidental injuries to the posterior urethra are frequently associ ated with subsequent stricture development, usually associated with complete transection of the urethra. They are generally short (less t han 2 cm long) and solitary. Strictures may he associated with pseudodiverticula and peri urethral abscesses developing upstream. Fistulas, sinuses, urinals extravasation and false passages (especially following iatrogenie trauma) may also occur.

This is almost entirely restricted to the male, being exceptionally rare in the female urethra unless there is very major pelvic trauma. It should be suspected if pelvic trauma is associated with urinary retention or haematuria, especially blood at the urethral meatus. I nvestigation is with careful ascending urethrography, which will demonstrate the presence and extent of the injury and atlas classification into trauma to the posterior or anterior urethra.

Urethrogram demonstrating short postinflammatory stricture at the junction of the bulbar and penile urethra. Fig. 32.1

Descending urethrogram in a male. The entire length of the urethra is demonstrated as the bladder empties. The prostatic urethra is a little distended in this example due to a short stricture at the junction of the membranous urethra and the bulbar urethra following a traumatic urethroscopy. Fig. 32.2

urinary tract, usually renal, sometimes as a result of direct extension of infection from the prostate or perineum. If severe, this may be associated with multiple urethropcrincal fistulas. Rarely strictures may be seen following severe, usually chronic, urethritis from almost any cause, including seronegative athritides (Reiter's syndrome) and syphilitic and chemical urethritis. This latter condition arises from therapeutic instillation of chemicals into the urethra (for example, podophyllin or 5-fluorouraeil for the treatment of condyl oma accuminata). Formerly common, it is now rare. Traumatic This may follow iatrogenic or accidental trauma. l atrogenic strictures most commonly follow catheterisation (traumatic, failed or prolonged) and are usually at fixed narrowed areas, particularly the level of the membranous urethra and penoserotal j unction. They may also he seen following instrumentation and urethral surgery (Fig. 32.2). They may he single or multiple (Fig. 32.3) and of variable length but are usually short.

Posterior urethra This is traumatised in around I0%%: of anterior pelvic fractures. A formal classification for posterior urethral i njuries has been described on the basis of urethrographic findings ( Table 32.1). In the mildest type of injury (I) there is only contun sion or partial tear of the membranous urethra and no extravasation. The bulhar urethra appears normal or stretched if the bladder i s elevated by hacniatoma. The commonest injury is the type II. with rupture at the prostatic apex (prostatomembranous junction) imm ediately above the urogenital diaphragm (UGD) with sparing of the UGD itself. Extravasation therefore occurs into the retro. pubic space but cannot extend into the perineum and the bulbar urethra is intact. In the type III injury the membranous urethra ruptures at the bulbomembranous junction below the UGD. which itself is disrupted. The bulbar urethra is often also ruptured. Contrast extravasation is in the direction of least resistance and i nto the perineum (Fig. 32.4). There may he some bladder tilling on the retrograde urethrogram, allowing a diagnosis of partial or incomplete posterior uret hral rupture to he made. Complete rupture is associated with absence of bladder tilling. Around two-thirds of type II injuries i nvolve complete urethral rupture: one-third are partial. Type III are more severe and virtually always complete. While partial urethral tears usually heal uneventfully, complete tears are much more likely t o he associated with adverse sequelae such as urethral strictures ( usually less than I cm or so in length). impotence (up to around I0%) or incontinence (2°/r).

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ Table 32.1 Type

Male posterior urethral injuries

Membranous

Bulbar

urethra

urethra

Contrast extravasation Perineum

Retropubic space

I

II

Multiple short tight anterior urethral strictures following attempted self-catheterisation with a knitting needle and resultant gross urethritis. Fig. 32.3

Micturating cystogram following extensive posterior urethral trauma showing gross urethral disruption at the level of the urogenital diaphragm and florid extravasation into the perineum. Fig. 32.4

I II

Contusion or partial tear Rupture above urogenital diaphragm Rupture below urogenital diaphragm

Normal or stretched Normal or stretched

No

No

No

Yes

Ruptured

Yes

No

Previous Page

CONTENTS

Fig. 32.5 Ascending urethrogram following a classical straddle injury showing marked extravasation of contrast from a partial anterior urethral tear. Anterior urethra ' this is most commonly damaged during attempted catheterisation or instrumentation. usually iatrogenically occasionally self-inflicted). In the communhity it is usually associated with blunt perineal trauma when the bulbar urethra and corpus spongiosum are compressed against the inferior aspect of the anterior pelvic ring (straddle injury). The injury is not usually associ, aled with a fracture . then may he urethral contusion only or somc degree of rupture. either partial or complete (Fib. 32.5). In partial rupture the proximal urethra is demonstrated, while in complete rupture there is an abrupt discontinuity at site of rupture. Extravasation i s associated with both partial and complete rupture and contrast may enter the corpus spongiosum and veins. Even cont usion alone may lead to subsequent .stricture but this is more eommon with any degree of rupture, especially complete.

Next Page

In the female it has a peak incidence around 40-60 years and is usually associated with chronic inflammation and may be associated with fibrous polyps and caruncles. It presents with blood appearing at the urethral meatus and symptoms of irritation and obstruction. Seventy-five per cent are squamous cell carcinomas and most of the rest are transitional cell carcinoma. Adenocarcinoma is occasionally encountered and arises from urethral divcrticula, presumably reflecting the origin of the diverticulum from an infected pararurethral (Bartholin's) gland. Carcinoma appears as one or more serial irregular filling defects within the urethra on urethrography or as an irregular stricture. Lymph node metastases occur before blood-borne spread. t he distal third of the urethra metastasising to superficial and deep i nguinal lymph nodes, the proximal two-thirds to the external and i nternal iliac lymph nodes. I n the shale the peak incidence is above the age of 50. It can he regarded as two different conditions. In the UK the commoner is carcinoma of the prostatic urethra, usually transitional cell carcinoma. It has a strong association with bladder transitional cell carcinoma, especially with previous resection. Carcinoma in the remainder oh the urethra (membranous,bulbar and penile) is strongly associated with chronic urethral inflammation (including sexually transmitted disease) and strictures (present in up to 75% of patients). As with carcinoma of the female urethra, approximately 75% are squamous cell carcinoma, most of the rest are transitional cell carcinoma and adenocarcinoma is occasionally seen. Presentation is usually with insidious onset of poor urinary stream, haematuria and serosanguinous discharge. On urethrography carcinoma appears as a well-defined filling defect: it may he eccentric or circumferential and may he associated with irregular stricturing . Metastatic spread is virtually always via the lymphatics. the penile urethra metastasising to the deep inguinal and external iliac lymph nodes. the bulbar and posterior urethra to the internal iliac and obturator lymph nodes. MRI is the favoured modality for staging uretIral carcinoma in both sexes because of its better demonstration of the soft-tissue anatomy of' the perineum. A formal TNM staging system is available, the salient features of which are: TI t umour invading subepithelial connective tissue: 'I'2-tumour invading any of periurethral muscle, corpus spongiosum or prostate: T3 tumour invadi ng any of bladder neck, corpus cavernosum, anterior vagina: and T4 tumour invading any of the other adjacent organs. including the bladder (Fig. 32.6).


Urethral metastases Metastases from distant primmaries to the

Benign tumours Fibroepithelial polyps are occasionally encountered, usually in children, sometimes in young adults. They appear as finger-like processes 1-2 cm long originating near t he verumonlanum and present with intermittent outflow ohstruction. Transitional and squamous cell papillomas are benign mesenehymal ncoplasms lined with urothelium which may he metaplastic. f n the male they may he seen in the prostate urethra (and may he associated with bladder papillomas) but are most often situated in the fuse nasicularis and parameatal area. In the female they are usually found in the distal third of the urethra (urethral cal-uncles).

urothelium of the urethra is excessixely rare. The urethra is more

They usually present between 20 and 40 years of age with haema-turiaand/or discharge. They are seen as smooth solitary or multiple tilling defects on urethrography. Urethral carcinoma This rare tumour is also unusual in being encountered more frequently in females than males.

Fig. 32.6 Transverse post-gadolinium T I - weighted image showing invasion of the penile bulb and muscles of the right side of the pelvic floor by urethral carcinoma.

Previous Page

1020

CONTENTS

Next Page


Fig. 32.7 (A) Postmicturition film from an iVU series showing a urethral calculus and gross detrusor hypertrophy due to secondary bladder outflow obstruction. (B) Urethrogram (different patient) showing a filling defect due to a urethral calculus within the prostatic urethra.

commonly involved by direct spread of turnout- from adjacent organs, most often carcinoma of the bladder or rectum, occasionally prostate, vagina or cervix.

The commonest urethral tilling defects are calculi and tumours (usually benign, occasionally carcinoma and rarely metastases). Other less common causes include the occasional urethral equivalent of inflammatory conditions seen elsewhere in the urinary tract (malacoplakia, urethritis cystica and the swollen mucosa of urethritis) and condyloma acuminata. Condyloma acuminata (venereal wart) is due to a DNA papillomavirus and appears as a dermal wart on the genitals and around the anus. Condylomas may extend into the anterior urethra, where they appear as multiple sessile filling defects on urethrography. This procedure, however. may be associated with retrograde seeding of the infecting agent into the bladder and is best avoided if the diagnosis is suspected in advance (as is any form of urethral catheterisation). There is an association with squamous cell carcinoma of the vulva.

Urethral diverticula These are rarely congenital, when they are classically saccular and situated in the midpenilc urethra. They are associated with high-grade obstruction in children and are difficult to distinguish from urethral valves. Acquired diverticula are strictly speaking pseudodiverticula, as t hey are lined with fibrous tissue or inflammatory cells rather than urothelium. They are inflammatory in origin and may develop from


Urethral calculi Urethral calculi are uncommon and over 90% originate in the kidney or bladder. They may present with sudden obstruction of the urine flow or more chronic symptoms of urethritis, dysuria and poor stream. A minority originate in the urethra in association with a structural abnormality (usually strictures or diverticula). They enlarge slowly and present with symptoms of prostatism. Although most urethral calculi are radiopaque, they are often overlooked as they are frequently not suspected and are situated over hone on radiographs, or even off the film entirely. In the female urethra they are usually seen in a midurethral diverticulum; i n the male they are generally found around the junction of the prostatic and membranous urethra (Fig. 32.7).

Fig. 32.8 Cystourethrogram showing a substantial sinus from the posterior aspect of the urethra following rectal surgery.

Previous Page

CONTENTS

Next Page

THE MALE GENITALIA AND URETHRA

1021

an abscess in the paraurethral glands or within the prostate.

Urethral fistulas and sinuses

Meehanical causes play a part and they may develop immediately

ment to closure after urethral rupture or laceration. They may be

upstream of a urethral stricture or following pressure necrosis with a

secondary to rectal or gynaecological surgery, obstetric injury,

l ong-term urethral catheter, usually on the inferior wall of the urethra at t he penoscrotal junction. Stasis of urine may lead to secondary

radiotherapy and some inflammatory lesions (Crohn's disease.

i nfection and calculus development. If they become large enough

They may end in the soft tissues of the perineum or continue on to

they may elevate the bladder base. They may be demonstrated on

the skin of the perineum or penis. Alternatively they may commu-

urethrography or ultrasound (transvaginally in the female), the latter

nicate with the rectum, colon, vagina or uterus. Conventionally

having the advantage of avoiding radiation.

diagnosis is with sinography or micturating cystourethrogram

Fig. 32.9

These form if there is an impedi-

periurethral abscess, tuberculosis) or with strictures or carcinoma.

Coronal T2 - weighted MR image showing a sinus from the base of the penile urethra down into the scrotum.


Fig. 32.10

Sagittal T,-weighted (A) and T.-weighted (B) MR images in a patient with Nunan's syndrome showing male external genitalia and a uterus .

Previous Page

1022

CONTENTS

Next Page


(Fig. 32.8), which will show an irregular track that may lead into an irregular cavity or into another viscera. MRI is also useful to show the fistulous track (Fig. 32.9) and (together with CT) may demonstrate the underlying cause.

Ambiguous genitalia and intersex This is a complex situation with a wide variety of phenotypes. Specialist chromosomal, endocrine and clinical assessment are required and the radiologist's role is generally to assist with the determination of which internal pelvic structures are present and which associated congenital abnormalities exist, in particular within the urinary tract. Ultrasound and MRI (Fig. 32.10) are the preferred modalities, both to avoid ionising radiation and to demonstrate optimally the soft-tissue anatomy. Undescended testicle (cryptorchidism) I n the majority of prepubertal males the testes may normally be retractile into the groin because of the cremasteric muscle reflex. If, however, the testicle can never be located within the scrotum, it can be considered undescended. An undescended testicle most commonly lies in the inguinal canal (canalicular). Otherwise it may lie higher up along the normal line of descent (abdominal testicle) or in a site away from the normal line of descent (ectopic). Most abdominal testicles lie just proximal to the inguinal ring, although they may lie further cranially within the pelvis or retroperitoneum. Ectopic testicles are uncom uncommon and are most often found in the superficial inguinal pouch, occasionally coming to lie in the femoral canal, suprapubic fat pad, the perineum and the opposite scrotum. Failure of descent by the age of 2-3 years is associated with abnormal development of the testicle and this is particularly severe i f it continues beyond puberty. Consquently undescended testes may be atrophic with poor spermatogenesis. A small proportion of

Fig. 32.11 Transverse ultrasound showing an atrophic undescended testicle (arrows) lying in the inguinal canal.

Ultrasound can be regarded as the first-line investigation to locate an undescended testicle, being quick and able to locate the testicle at its commonest sites (within the inguinal canal orjust proximal to it ). The testicle may look relatively normal, although the longer it bas been undescended the more likely it is to be small, atrophic and echo poor (Fig. 32.11 ). If the testicle cannot be identified on ultra sound, a more extensive search may be performed with MRI. This is a better modality than CT as it avoids radiation and the testicle shows a conspicuous high signal on T weighted and STIR sequences (Fig. 32.12). Testicular pblebograpby or artcriography bas been employed in the search for undescended testes. If both ultrasound and MRI are negative it is unlikely that these angio graphic procedures will detect the missing organ, as it is probably absent or extremely atrophic. This is particularly the case if the patient is above 30 years old.


undescended testes (5%) are not found even at Surgical exploration. While some of these may be true agenesis, it is distinctly possible that a number are so severely atrophic that they cannot be located. Testes that remain undescended (especially abdominal testes) in boys above the age of 5 years also suffer from an increased incidence of malignant neoplasia, up to 40 times normal, usually with the development of seminoma. There is also an increased risk in the contralateral, normally descended testicle. Other associations of undescended testes include abnormalities of Wolfian duct-derived structures, for example seminal vesicle cysts and agenesis.

Patent processus vaginalis Failure of closure of the processes vaginalis after testicular descent results in a persistent communication between the peritoneal cavity and the scrotum. This may

Fig. 32.12 Transverse STiR i mages from MRI examinations of patients with undescended testicles (arrow) in the proximal end of the inguinal canal (A), suprapubic pouch (B) and pelvis (C).

Previous Page

CONTENTS

transmit disease processes (for example ascites) or become the site of a hernia. Incomplete closure of the processus may lead to a developmental cyst, usually in the upper scrotum or inguinal region.

hydrocele is the formation of fluid between the two layers (visceral and parietal) of the tunica vaginalis. Ultrasonographically it is seen as an echo-free area (Fig. 32.13) partly surrounding the testicle (in comparison to cysts, which do not Hydrocele A

Next Page

surround the testicle). A hydrocele may develop as a result of infection (epididymo-orchitis), trauma, malignant testicular tumour or infarction (including torsion). In most cases, however, an underlying cause for the hydroccle cannot be found. When infection is a cause there are usually obvious clinical and radiological features of epididymitis. The hydrocele may demonstrate internal echoes (Fig. 32.14), especially on the rare occasions when infection is sufficiently severe for frank pus (pyocele) to develop. The commonest infective organisms are usually bowelrelated Gram-negative bacilli but a variety of other agents may be responsible, although now tuberculosis is rare. Chronic infective hydrocele (especially tuberculous) may be associated with considerable calcification of the tunica (Fig. 32.15). Following trauma a haematocele (haernorrhagic hydrocele) will also exhibit considerable echogenicity. Cysts Simple cysts are extremely common in the scrotum. They may be seen at any age from adolescence onwards but are most common in the elderly. The majority are seen in the epididymis. particularly in the upper pole, and they may be single or multiple.

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ Fig. 32.14 hydrocele.

Ultrasound showing an infected, partly septated echogenic

Fig. 32.13 Ultrasound demonstration of a hydrocele seen as an echofree area partly surrounding a normal testicle.

Fig. 32.15 fil m (B).

Ultrasound (A) showing dense peripheral calcification around the exterior of a chronic inflammatory hydrocele. This is also visible on the plain

Previous Page

102 4

CONTENTS

Next Page


Fig. 32.16 Ultrasound showing (A) a classical echo-free well-defined thin-walled solitary epididymal cyst and (B) a cluster of simple cysts.

Fig. 32.19

Ultrasound of varicocele seen as echo-free serpiginous structures.

The vast majority of varicoceles are described as primary and are

Fig. 32.17

Ultrasound of infected epididymal cyst showing debris and

fluid level. Ultrasound of well-defined spermatocele with slightly echopoor contents. Fig. 32.18

On ultrasound they show the classical features of cysts elsewhere i n the body, being anechoic, showing distal acoustic enhancement and having no appreciable wall thickness (Fig. 32.16). Haemorrhage or infection may alter the appearance of a preexisting cyst to show some degree of echogenicity, occasionally with a visible fluid level (Fig. 32.17). Alternatively infection or trauma (including vasectomy) may provoke the creation of a cystic l esion with internal echoes, partly due to the presence of spermatazoa (spermatocele, Fig. 32.18). These again are more common in the upper pole of the epididymis.

presumed to be due to developmental abnormalities of the valves and/or the veins themselves. This is far more likely on the left, where at least 95% are encountered. A minority occur secondary to a lesion compressing or occluding the testicular vein. The classical cause is a left-sided renal cell carcinoma extending along the renal vein as far as the termination of the testicular vein, hut benign (for example, hydronephrosis) and other malignant conditions (includi ng abdominal lymphadenopathy) can provoke a varicocele on either side. There is a higher risk of an underlying cause, particul arly a tumour if the varicocele is of recent onset with an acute presentation, the patient is older than 40 years, it is right-sided and is unchanged with provocative manoeuvres. Treatment of varicoceles is usually for persistent discomfort or as part of the treatment of suhfertility and is by percutaneous embolisation or surgery. Direct surgical ligation of the testicular vein (by open operation or laparoscopy) should have a low primary failure rate (of the order of I % ) with a recurrence rate around I6% . Surgery, however, is more traumatic than percutaneous embolisation and is associated with specific risks not encountered with the alternative procedure, in particular damage to the testicular artery


Varicocele Dilatation of the network of veins draining the testicle

i s described as a varicocele. They are extremely common (up to 8-16% of the male population), and usually asymptomatic, often being discovered incidentally on clinical or ultrasonographic examination for other conditions. There is a possible association with subfertility, reported in 21-39% of males being investigated for subfertility. They are most frequent between 15 and 25 years of age, almost always left-sided and, when symptomatic, present with scrotal aching and/or soft scrotal mass. Classically these symptoms worsen during the day while the patient is upright. Occasionally they present relatively acutely as a manifestation of a renal carcinoma. Treatment may be offered for discomfort or as part of the management of subfertility. On ultrasound varicoceles are seen as a leash of predominantly echo-free serpiginous structures measuring more than 2 mm maximum diameter (Fig. 32.19). Visible flow may he seen within l arger varicoceles on conventional ultrasound. Confusion with cysts i s unlikely, even when the varicocele is small, but confirmation of their nature can he obtained with slow-flow Doppler Ultrasound. Their prominence is increased in the upright position and with the Valsalva manoeuvre, although this is not routinely required to make the diagnosis.

and resultant infarction. The total failure rate for percutaneous embolisation (Fig. 32.20) i s comparable with that for surgery, although the technical failure rate may be higher (6-I2%% ) and the recurrence rate similar or lower (4-16%). Morbidity should be minor and infrequent, up to 10% experiencing transient discomfort in the loin or scrotum on the side of the varicocele, with few other symptoms. Postvaseetomy epididymis Numerous individuals have had a

vasectomy and many will undergo subsequent scrotal ultrasound for one reason or other, often unrelated to the vasectomy. In these patients the epididymis often appears thickened and slightly echopoor (Fig. 32.21). These findings can almost be regarded as variations of normal and are not usually related to any specific symptoms. Epididymitis I nflammation of the epididymis, usually due to

ascending infection with Gram-negative bacilli (E. coli, etc.) or Chlamydiu , i s extremely common. It is particularly seen in young adult males, with a second peak in late middle age and the elderly,

Previous Page

CONTENTS

Next Page


1025

Fig. 32.20 Testicular phlebography and percutaneous embolisation of varicocele. The testicular vein in this patient has a wide termination unprotected by a valve and at least one small accessory connection to the renal vein (A). The catheter is manipulated into the distal testicular vein (B) to commence deploying the coils (C).


when it may be associated with cystitis and/or prostatitis in association with benign prostatic disease and bladder outflow obstruc- fig 32.21ged Ultrasound showing typical postvasectomy echo-poor epi epididymis. Lion. The clinical presentation varies in severity and acuteness On ultrasound the epididymis shows swelling, often nodular and from mild pain, tenderness and scrotal swelling to a severe pyrexparticularly affecting the lower or upper pole. There is often diffuse ial illness with marked scrotal pain and swelling. Although or patchy reduction in echogenicity, although sometimes the epiusually asymmetrical, epididymitis is not uncommonly bilateral.

Previous Page

1026

CONTENTS

Next Page


Fig. 32.22 Ultrasound of aggressive epididymitis showing a heterogeneous mass with areas of reduced and increased echogenicity adjacent to the lower pole of the testicle.

Fig. 32.23 Ultrasound of severe orchitis. The bulk of the testicle shows diffuse reduction in echogenicity. The heterogeneous area in the lower pole represents a developing abscess.

didymis shows considerable heterogeneity (Fig. 32.22) and even rarely a diffuse increase in echogenicity. A heterogeneous pattern of predominant increase in echogenicity is more frequently associated with chronic epididymitis. Doppler ultrasound often demonstrates hypervascularity. Rarely the condition progresses so that there is demonstrable abscess formation within the epididymis. There is often an associated hydrocclc of variable size, which may be septated and contain echogenic fluid. Severe cases may be associated with demonstrable oedematous thickening of the overlying skin and (in up to 20% of cases) there may be coexisting orchitis.

These are rare and usually benign; the most common is the adenomatoid tumour of the epididymis, which is a benign hamartoma, most frequently seen in the tail. It varies in size from about 5 to 50 mm and is of similar or slightly i ncreased echogenicity compared with normal testicle. Others include benign neoplasms. such as mesenchymal tumours (leiomyoma, fibroma, lipoma), adrenal rests and benign mesothelioma. This latter condition arises from the tunica vaginalis. It produces small paratesticular hypcrechoic masses and a disproportionately large hydrocele. Extratesticular malignancies are extremely rare. Although carcinoma and metastases have been described, the least rare malignant neoplasia in this site is sarcoma of the spermatic cord, most commonly embryonal rhabdomyosarcoma (in younger patients), but sarcomas of other mesenchymal tissue are seen occasionally, for example liposarcoma. Extratesticular tumours


Orchitis Inflammation of the testicle itself may be seen in systemic viral illness (classically mumps) or in association with bacterial epididymitis. Up to 25% of postpubertal males with mumps will suffer some degree of orchitis, usually 7-10 days after the parotitis. Approximately two-thirds are unilateral. Other viruses i mplicated include echoviruses, group B arboviruses and the lymphocytic choriomeningitis virus. In the acute phase ultrasound may show testicular swelling with patchy or diffuse reduction in echogenicity. Following resolution the testicle may return to Testicular torsion I n this condition the spermatic cord becomes normal but in severe orchitis there may be atrophy, with reduction t wisted as a result of testicular rotation, producing testicular in size and echogenicity, usually apparent within 6 months of the i schaemia. It may occur at any age but is most fi equent i n the first acute insult. year of life or in adolescence, when the testicle is rapidly enlarge Bacterial orchids shows similar acute features (Fin. 32.23) but inn . Different patterns of torsion have been described. The loose usually with marked changes of epididymids. Doppler ultrasound attachment of the testicle and spermatic cord to the scrotum in may show increased vascularity but severe orchids may be associi nfants and neonates predisposes to torsion of the entire cord ated with ischaemia and infarction with reduced or absent vascularabove the level of the scrotum (extravaninal torsion). Rotation of ity. In particularly severe orchids heterogeneous areas may develop the cord within the tunica vaginalis (intravaninal torsion) is the with a potential for intratesdcular abscess formation (irregular areas commonest situation in the older age group. This is most often of liquefaction containing debris). Again there is a substantial risk due to poor testicular attachment to the posterior scrotal wall by of subsequent atrophy leading to a small echo-poor testicle with an abnormally narrow mesentery, which predisposes to rotation little or no spermatogenesis. The underlying organisms are usually of the cord. This anatomical situation has been referred to as the the same as for epididymids, although rarely chronic epididymitis, bell-and-clapper malformation and may be bilateral in up to 80°Ic especially with abscess formation, may be due to tuberculosis. of cases, with simultaneous bilateral torsion in up to 5%. Surgical Calcific scars may be seen following any form of orchids. treatment of testicular torsion should therefore include detorsion

Previous Page

CONTENTS

of the affected testicle and fixation of the contralateral testicle to prevent subsequent torsion. Torsion may be complete (i.e. rotation of at least 360°) or incomplete and spontaneous torsion and detorsion may occur. The degree of torsion determines the severity of the ischaemia and the rapidity with which irreversible changes occur within the testicle. Surgical treatment of complete torsion within 5 b is associated with a testic1 ular salvage rate of 80%, which falls to 20 7- or less after 12 h. If torsion is incomplete, the testicle remains viable longer: 80% up to 1 2 h and 40% up to 24 h. Even in this situation, however, viability falls to 10% after 24 h. I n the acute stage ultrasound may be normal or demonstrate a swollen testicle with patchy or diffuse hypoechogenicity ( Fig. 32.24). The epididymis may also become swollen and echopoor. There may be a reactive hydrocele and the overlying scrotal skin may be thickened and oedematous. Doppler ultrasound bas a claimed sensitivity of upwards of 85 17c i n the diagnosis of torsion by virtue of reduced vascularity (absence or poor colour flow, redueed peak systolic velocities) compared with the unaffected side. False positives (for example ischaemia associated with severe epididymo-orchids) and false negatives (due to difficulty in obtaining an adequate colour flow, intermittent nature of torsion, etc.) are encountered and there is still considerable controversy as to the role of this investigation in the management of these patients. Given the i mportance of operating within a few hours oh the onset of symptoms, it remains axiomatic that neither the performance nor interpretation of an ultrasound examination should delay surgical treatment. If there is doubt the urologist should operate on clinical !rounds. The testicular appendix (often referred to as the hydatid of Morgagni) and rarely other vestigial scrotal appendices (such as the epididymal appendix) may undergo torsion. This is associated with acute scrotal pain and localised swelling and tenderness. On ultrasound there is a focal soft-tissue mass adjacent to the upper pole of the epididymis which is often heterogeneous with a central echo-poor area and an associated

Next Page

hydrocele. This may slough off and become calcified, giving rise to mobile highly echogenic shadowing foci between the layers of the tunica vaginalis (scrotoliths or scrotal pearls, Fig. 32.25).

Blunt scrotal trauma most commonly results in haemorrage around the testicle (haematocele) and intratesticular hacmatoma, which may be associated with a tear of the tunica albuginea. More significant trauma may be associated with demonstrable fragmentation of the testicle. The role of imaging depends on the surgical approach. Where there is substantial trauma, surgical exploration, drainage of hacmatoma and repair of testicular tears may enhance the subsequent viability of the testicle. Ultrasound is the most appropriate imaging modality. The commonest finding is a complex haemorrhagic hydrocele (haematocele) with areas of intermediate and low echogenicity representing thrombus and liquid blood or serum. The underlying testicle may show areas of contusion or hacmatoma, visible as echo-poor, someti mes echogenic. areas, often with a relatively linear configuration. The testicle may be deformed by subcapsular haemorrhage. There may be rupture of the capsule (tunica albuginca ) with disruption of the underlying testicle and associated hacmatoma (Fig. 32.26). The testicular tear may be linear or complex, and in severe trauma there may be fragmentation of the testicle. Scrotal hacmatoma may he complicated by abscess formation. and severe testicular damage may be followed by subsequent atrophy.

Appendix torsion

Testicular cysts As with cysts elsewhere in the body, testicular cysts are well-defined, thin-walled and anechoic on ultrasound, with distal acoustic enhancement. They vary in size from a millimetre or two to I-2 cm and lie within testicular tissue (Fig. 32.27). They are occasionally multiple, are often situated at or close to the testicular hilum and may be associated with epididymal cysts. Cysts directly arising from the tunica albuginea have been described with similar ultrasound features (Fig. 32.28) and benign nature. Multiple small cystic areas are not infrequently seen at the testicular hilum. These are often referred to as dilatation of the retc testes (Fig. 32.29) and again are often seen with epididymal cysts. All

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ these lesions are entirely benign.

Ultrasound of an infarcting testicle. There are extensive areas of reduced echogenicity within the substance of the testicle. The adjacent epididymis is also markedly diseased and swollen. Fig. 32.24

Ultrasound showing a classical highly echogenic scrotolith with marked distal acoustic shadowing and small hydrocele. Fig. 32.25

Fig. 32.26 Ultrasound showing traumatic disruption of upper pole of testicle and small adjacent haematocele.

CONTENTS

Previous Page 1028

Next Page


Fig. 32.27 Ultrasound of a small (A) and a l arge (B) testicular cyst, both showing an

echo-free area without any significant solid elements.

Ultrasound of epidermoid cyst seen as a well-defined echo-poor nodule. Fig. 32.30

Ultrasound of a small echo-free tunica albuginea cyst. Fig.

32.28

Approximately 95% of testicular neoplasms are of germ cell origin, 4% lynrphomas and V/( rarities such as metastases, Levdig cell and Sertoli cell tumours. Approximately 50%- of germ cell tumours are pure seminomas, which arc less aggressive and present i n an older age group (peak 30-45 years). The remaining 50% are ( NSGCTs), classified as non-.seminomatous germ cell tumours although up to a fifth of these may contain some seminomatous element. They present in the younger age group (peak 15-30 years) and a majority produce one or more biochemical tumour markers ( beta-human chorionic gonadotrophin or alpha-fetoprotein). There are two main histological classifications of NSGCTs, the World


Health Organization (WHO) and the British Testicular Tumour Panel systems, the former being more widely used in the North

Ultrasound showing multiple tiny echo-free areas at the testicular hilum (dilated rete testis). Fig. 32.29

Epidermoid cysts These uncommon benign tumours arc thought

American literature. The two classifications are broadly similar. Both classify the yolk-sac tumours as a separate subgroup. In the British system, the rest are teratomas of varying pattern (differentiated, intermediate, undifferentiated, trophoblastic). In the WHO system, mixed germ cell tumours, embryonal carcinoma and choriocarcinoma are separated from the teratomas of varying histology. Risk factors for the development of testicular malignancy include undescended testes, testicular atrophy, microlithiasis, previous tes ticular malignancy, infantile hernia, low birth weight and maternal exposure to diethylstilboestrol. There is a weak genetic tendency. with a family history of the disease seen in 2% of patients with tes titular cancer. Testicular cancer most commonly presents as a pain

to derive from epithelial rests or inclusions. They consist of layers of keratin within a fibrous capsule lined with squamous epithelium. On ultrasound they are well-defined auricular rounded lesions and may be solitary, multiple or bilateral. Calcification may be present and adopt a variety of patterns (multiple foci less scrotal mass inseparable from the testicle, although the patient may sometimes also complain of a dull ache or sensation of heaviwithin the cyst, curvilinear along the wall or a target pattern of ness. Occasionally the condition presents with widespread pal nes.cirl)Othweyausldif chomonary or abdominal metastases, rarely when the primary is no Ppoor (Fig,. b • 32.30). Occasionally Y they Y may y appear highly echol onger evident. The rare Leydig and Sertoli cell tumours may genie, presumably if calcification becomes widespread throughout produce sex hormones, which can provoke premature virilisation the lesion. If the typical ultrasound features are present, the urolo (especially Leydig cell tumor s) or gynaecomastia (especially gist may be able to perform partial rather than total orchidectomy. Sertoli cell tumours). Testicular cancer Although uncommon (representing approxScrotal ultrasound is the investigation of choice, with a sensitivi mately I % of cancers in males), testicular cancer is the commonity approaching 100%. Tumours are seen as intratesticular lesions est malignancy in males between 20 and 40 years of age, with the with replacement of the normal architectural pattern by material of i ncidence apparently rising, predominantly low echogenicity. Small tumours may show no mass

Previous Page

CONTENTS

Next Page


1029

dicted from the ultrasound appearances, which are usually nonspecific. However, classically seminomas are well defined and uniformly modestly echo-poor (Fig. 32.32). NSGCTs more often show

Fig. 32.31

Ultrasound of small testicular malignancy with hydrocele.

effect. When larger than 2-3 cm they are associated with deformity and expansion of the outline of the testicle. Occasionally there is an associated hydrocele (Fig. 32.31). The histology cannot be pre-

significant echogenic areas of fibrosis or calcification and echo-free cystic areas (Fig. 32.33). Larger tumours (above approximately 1.5 cm) may show increased vascularity on colour flow Doppler; smaller ones are more often hypovascular. The differential diagnosis includes focal or diffuse acute orchitis, abscess, infarct and haemorrhage. The clinical presentation and course usually permits differentiation. It is, however, important to exercise extreme caution, and an echo-poor lesion within the testicle must always be regarded as potentially representing a malignancy. Therefore, if an i ntratesticular lesion is not showing progressive resolution within a short time (a fortnight or less), consideration of orchidectomy would he recommended and the dilemma discussed with the patient. Occasionally a lesion of a centimetre or less is encountered within the testicle. Those that have been observed over a prolonged period of time appear to show a benign course. Again it would be prudent to ensure the patient is aware of the dilemma and seeks help at the slightest change. The management of testicular cancer depends heavily on the detection and treatment of lymph node metastases. The lymphatic vessels from the testes follow the spermatic cord through the inguinal canal, cross the pelvis and ascend to drain into lymph nodes along the inferior vena cava and aorta in the lumbar region around the L2 level. Lymph node involvement virtually always starts on the side of the affected testicle. The first group of nodes affected are sometimes referred to as the sentinel nodes, being the nodes around the renal vessels (renal perihilar), the para-aortic nodes immediately below this level on the left and the paracaval nodes immediately below the renal vein on the right. Bilateral i nvolvement is uncommon when the primary is left-sided, but is more often encountered when the disease starts in the right testicle. It does not usually occur until there is considerable ipsilateral disease. Contralateral lymphadenopathy with normal ipsilateral nodes is extremely rare. Following initial lymphadenopathy, disease may spread cranially and caudally along the para-aortic, paracaval and aortocaval nodes. Iliac and inguinal lymph nodes may he involved by retrograde spread, especially on the right, or directly in circumstances where the tumour gains access to the ilioinguinal l ymphatic system. This may occur with disease originating from an intra-abdominal testicle, with disruption of the lymphatics followi ng surgery or, if tumour is grossly invasive, within the scrotum (epididymal involvement may lead to external iliac nodal disease, skin invasion leading to inguinal lymphadenopathy). Tumour may also extend retrogradely into the mesenteric lymph nodes and antegradely into the retrocrural, mediastinal and supraclavicular lymph nodes. Drainage into the thoracic duct may occur, with subsequent haematogenous metastases to the lungs and, less often, the liver, brain and skeleton. Haematogenous spread is rare before lymphatic spread, except with choriocarcinoma. It is generally more common with non- semi nomatous tumours. Although a TNM staging system has been formulated for testicul ar malignancy, the Royal Marsden staging system is more widely used in routine practice (Box 32.1). Its importance is that it not only describes the site of metastases but also the volume of tumour, which is likely to he at least as important. In stage I there is no evidence of metastases. Abdominal lymph node metastases are seen in stage II, supradiaphragmatic in stage 111. Distant metastases (extral ymphatic) are present in stage IV.

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ Fig. 32.32 Ultrasound of seminoma showing homogeneous reduction in echogenicity.

Fig. 32.33 Ultrasound of NSGCT which is echo-poor but relatively illdefined and containing at least one area of prominent calcification.

CONTENTS

Previous Page

1030 Box 32.1

Next Page


Royal Marsden staging system for testicular cancer

Stage I IM

No evidence of metastases Rising serum tumour markers

Stage II A B C

Abdominal lymph node metastases 2 cm diameter or less Between 2 and 5 cm diameter 5 cm diameter or above

Stage III M N 0 ABC

Supradiaphragmatic nodal metastasis Mediastinal Cervicat or axillary Without abdominal lymph node metastases Relates to size as described in stage II

Stage IV Lung Ll L2 L3 H Br Bo ,:,,...

Extralymphatic metastases 3 metastases or less More than 3 metastases, all 2 cm diameter or less More than 3 metastases, at least one above 2 cm diameter Liver metastases Cerebral metastases Bone metastases r, ,

Fig. 32.35 CT demonstrating a huge left para-aortic mass of metastatic NSGCT showing areas of reduced density. The mass encases the aorta, which is displaced forward and invades posteriorly into the psoas muscle and anteriorly into the root of the mesentery. visible node less than 1.0 cm in diameter is therefore best rekarded as equivocal, especially if above 8 min, and the scan should be a

repeated in 3 months (or guided biopsy performed). Metastatic

CT is the conventional staging method (Fig. 32.34). CT detection of lymph node metastases depends on size. Lymph nodes above 1.0 cm are usually involved (specificity 879)-), and above 1.5 cm are almost always involved (specificity 98%). Most nodes less than 1 cm in diameter are uninvolved but the fact that 25-30 9/ of untreated stage 1 disease develop subsequent metastases indicates that there is a significant number of lymph node micrometastases. A

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ Fig. 32.34 CT demonstrating para-aortic metastases from a left testicular malignancy. (A) At this level a small lymph node deposit is seen immediately lateral to the aorta and a larger one is visible anterior to the left psoas muscle. Higher up (B), just below the level of the renal hila, there is a large l eft-sided metastatic mass partly encasing the aorta.

Fig. 32.36 Transverse T 1 - weighted MR images on two different patients. (A) In this case there is a solitary left para-aortic lymph node tumour deposit from a left-sided testicular primary. (B) In this case there is a more substantial right-sided deposit from a NSGCT showing considerable heterogeneity. It has encased and displaced the inferior vena cava.

Previous Page

CONTENTS

Next Page


1031

Fig. 32.37 Complex deposits of NSGCT i n two different patients seen on MRI. (A) I n the first case there are tumour deposits in the left retrocrural area, close to the left renal hilm, and a large mass invading into the mesentery. (B) In this case an undescended testicle is the site of grossly metastatic NSGCT which is seen throughout the pelvis on this T 2 - weighted image. It is interesting to note that the pattern of spread is similar to ovarian cancer because of the atypical site of the primary.


Fig. 32.38 Post-treatment lymph node masses. (A) CT showing substantial mass with substantial cystic (low-density) areas. (B) Transverse T,-weighted MR scan showing a large predominantly cystic lymph node mass with extensive low-signal areas. These are seen on the transverse STiR squence (C, higher level) as intensely high-signal areas. Note the tumour has displaced the aorta anteriorly and is extending into the left renal hilum associated with some renal obstruction, as demonstrated by the perinephric high signal. (D) CT showing an ill-defined lymph node mass between the aorta and inferior vena cava (and inseparable from both) and a second mass to the left of the aorta showing dense areas of calcification.

Previous Page

1032

CONTENTS

Next Page


masses may become enormous but in the case of seminomas they usually remain of homogeneous soft-tissue density on CT, although they do occasionally demonstrate central necrosis. NSGCT deposits, however, often become heterogeneous as they enlarge, with lower density and even cystic elements (Fig. 32.35). The role of MRI scanning is still a matter of controversy. The same size criteria for the detection of lymph node metastases described for CT apply to MRI scanning, and from that point of view the modalities are similar (Fig. 32.36). MRI has the advantage that it avoids ionising radiation but the disadvantage that it is probably inferior for the detection of pulmonary metastases and, especially with the advent of spiral and multislice CT, is significantly slower and hence prone to movement degradation. It does have better soft-tissue contrast and might have a role in difficult cases to assist differentiation of potential tumour deposits from native structures (large gonadal or ascending lumbar vessels, anomalous inferior vena cava or renal vessels, unopacified bowel loops), or areas of solid tumour from fibrosis or cystic degeneration (Fig. 32.37). Approximately 75% of patients with seminomas have stage I disease. About 80% of these are cured with orchidectomy. The remainder are at risk of developing metastases. A surveillance policy is more difficult in this situation because of the absence of reliable biochemical tumour markers, so initial adjuvant radiotherapy to the para-aortic area can be offered, which raises the cure rate to 97%. Patients who present with metastatic disease are also treated with radiotherapy or chemotherapy (the latter particularly for bulky stage II disease). Seventy per cent of stage I NSGCT patients are cured by orchidectomy alone; 30%, however, will demonstrate subsequent metastatic disease. This can be almost entirely prevented by retroperitoneal lymph node dissection at the time of staging (a policy pursued in North America). Alternatively, close radiological surveillance can he instituted to allow early detection of metastatic disease, with chemotherapy. Around 25% of NSGCT patients have metastases (stage II-IV) at presentation and these are also treated with chemotherapy, which is usually associated with a complete response. Around 25%, however, will demonstrate persistent lymph node masses following chemotherapy (Fig. 32.38). These often contain areas of fibrosis, haemorrhage, cystic degeneration and sometimes calcification, but some will contain mature teratoma and up to 10-30% of these will contain active malignancy. No imaging modality is currently completely reliable in determining which masses contain significant disease and generally these are treated with surgical excision (retroperitoneal lymph node dissection). CT or MRI scanning should cover the pelvis, abdomen and chest. Follow-up scans should be performed relatively frequently for the first 2 years; for example, 3-monthly for the first year and 6monthly in the second. Yearly scanning should be continued for at l east 5 years for seminomas of all stages and stage I NSGCTs. There is an annual relapse rate of I-2% for treated metastatic NSGCTs for more than 10 years and these tumours therefore require longer follow-up. Sometimes a male patient presents with widespread metastases ( which may include abdominal, pelvic and mediastinal lymphadenopathy and pulmonary metastases) that are confirmed histol ogically to originate from a malignant teratoma but without a palpable testicular tumour. In some of these patients the teratoma may have originated outside the testicle. In a proportion, however, a small primary lesion, with the classical echo-poor ultrasound appearances described above, is identified in a testicle on ultra-

Fig. 32.39 Testicular ultrasound showing a small echogenic area with distal acoustic enhancement. The patient presented with widespread abdominal and mediastinal lymphadenopathy and pulmonary metastases, histologically shown to be teratoma. The testicular lesion is presumed to be the site of a burnt-out primary. sound. Rarely ultrasound demonstrates a small echogenic area within a testicle (Fig. 32.39), which is believed to represent the scar of a burnt-out primary, a situation that has been described with seminoma and NSGCT.


Metastases to the testicle are rare. The ultrasound appearance is of mixed, often predominantly reduced, echogenicity (Fig. 32.40); they are indistinguishable from primary malignancies except that they generally affect a significantly older age group. They often originate from cancer elsewhere in the urinary tract, especially the prostate. Occasionally carcinoma of the bronchus or malignant melanoma metastasise to the testicle. Others include lymphoma (the commonest testicular malignancy over the age of 60 years) and leukaemia (in childhood, the testicle being a classical sanctuary site for relapse). Testicular lymphoma This condition is rare, accounting for less than 5% of malignancies of the testicle, although it is the cornmonest testicular tumour in men over the age of 60. It is bilateral i n up to 25% and is associated with extranodal disease, especially i n the pharynx, skin and central nervous system. It is virtually always non-Hodgkin's disease. The ultrasound appearances are of a well-defined mass of low echogenicity expanding the testicle. It

Fig. 32.40 Testicular ultrasound showing virtually complete replacement of normal testicular tissue by metastatic prostate cancer. An associated hydrocele is also demonstrated.

Previous Page

CONTENTS

Next Page


103 3

often extends to involve the epididymis and spermatic cord. Tile dominant feature of the mass is its homogeneity. Multiple nodules or foci of haemorrhage or necrosis are uncommon.

Coarse areas of calcification (above 3 mm diameter) that are relatively few in number are seen within some malignant testicular tumours, particularly non-seminommatous germ cell tumours. A solitary substantial focus may be associated with the so-called burntout primary tumour (Azzopardi tumour). Infection, trauma or i nfarct may also produce dystrophic calcific areas. They are usually solitary or few in number and vary from 1-2 min in diameter to several millimetres. Numerous small areas of calcification (1 -2 mm diameter) seen throughout the testicle are referred to as testicular microlithiasis. These are thought to represent intratubular microcalcifications assoeiated with degenerate cells within the seminiferous tubules and appear to he a primary process.

This is usually bilateral but may be asymmetrical. It is associated , testicular with cryptorchidism, male pscudobcrmaphroditism atrophy and subfertility, as well as Down's syndrome and Klinefelter's syndrome. It appears to have an association with testicular eancer. although the Strength of this association is still a subject of' controversy. The mechanism is also still under investigation but t here is a potential link with intratubular germ cell neoplasia(IGCN). IGCN (formerly refer ed to as carcinoma in situ) is likely to he the precursor of germ cell tumours. It is found adjacent to most tumours, and is seen in the contalateral testicle in up to 5 ,-1( of cases. It is associated with testicular atrophy and infertility and progresses to carcinoma in approximately 50`/c of' cases in 5 years. IGCN itself' is not identifiable on ultrasound but it is thought that it nay he associated with testicular microlithiasis.

Fig. 32.42 Testicular ultrasound showing microtithiasis and the development of a seminoma. Classical testicular microlithiasis (CTM) has been defined as live or more microliths visible on at least one ultrasound image ( Fig. 32.41 ). More than one but less than live has been referred to as limited testicular microlithiasis (LTM). The frequency of tumour occurrence with microlithiasis (Fig. 32.42) and the risk of tumour development remains uncertain. initial reports suggested a very high risk, with up to 40c/c of CTM associated with testicular cancer at the time of the initial ultrasound scan. More recently large c studies have suggested a significantly lower range of 15-27 /(.


Subsequent tumours have been reported as developing in CTM and I ,TM but the magnitude of the risk is still unclear. The most recent studies have not demonstrated any malignancies on limited followup of 3-4 years. It has become routine practice to suggest annual ultrasound surveillance of these patients, although regular testicular self-examination is probably more important and may even be sufficient.

Carcinoma of the penis This uncommon malignancy is essen-

Fig. 32.41 calcific foci.

tially a dermatological disease, being a kcratinised squamous cell presentation. It may be carcinoma (SCC), but with a urological classified into three anatomical subsites (prepuce. glans penis and shaft). bill in practice it almost always starts in the sulcus and proceeds with local invasion into the clans and along the shaft. Advanced tumours invade into the urethra, penile aura. prostate and perineum. There is a .strong tendency, like all SCCs, to early l ymph node metastases and late Iiaematogenous spread. and deep inguinal The lymphatic drainage is to thesuperficial nodes in the first instance. MRI is useful for loosing both t he extent of the primary tumour and the regional lymph nodes ( Fig. 32.43). Unfortunately these nodes are often hyperplastic due t o inflammatory stimulus and may normally be up to 15 mm in diameter, so unenlarged involved lymph nodes are easily Ultrasound of testicular microtitbiasis showing numerous tiny maxnmun overlooked. Nodes above 15 mm in diameter arc highly likely to be

Previous Page

CONTENTS

Next Page

Longitudinal ultrasound of a penis in Peyronie's disease showing a small calcified echogenic plaque with distal acoustic shadowing. Fig. 32.44

a snap at the time of injury, with immediate detumescence and marked ecchymosis. Ultrasound and cavernosography have been employed to image this condition but they contribute little, as the diagnosis is essentially clinical and the treatment (relatively acute surgical repair) is not affected by the imaging findings. Peyronie's disease I n this condition there is focal induration and fibrosis of the corpus cavernosum. It is of unknown aetiology and may present at any age, but particularly in middle age. It is of variable severity and often resolves spontaneously over a period of months or y ears. It presents with a palpable nodule and may be associated with painful erections and/or difficulty with sexual i ntercourse. It is seen on ultrasound as an echogenic plaque. Calcification may develop, leading to distal acoustic shadowing (Fig. 32.44). The disease may also be demonstrated on CT or MRI but imaging findings rarely alter the diagnosis or management, which again rests on the clinical picture.


Scintigraphic demonstration of testicular perfusion offers a non-

invasive approach for distinguishing between the ischaemia associated with testicular torsion and the hyperaemia associated with acute epididymo-orchitis (Figs 32.45, 32.46).

Fig. 32.43

MR scan of extensive penile carcinoma. Transverse postgadolinium T,-weighted image (A) shows destruction of the normal anatomy of the glans and shaft by the irregular enhancing mass of tumour. This is seen on the transverse STIR sequence (B), which also demonstrates upstream dilatation of the urethra, a finding generally only seen with advanced tumours. The coronal post-gadolinium T 1 -weighted image (C) demonstrates the presence of inguinal lymph nodes. These are not particularly enlarged but the node on the right (arrow) shows central necrosis characteristic of squamous cell carcinoma metastasis.

i nvolved by tumour. As with other SCCs, involved lymph nodes may develop irregular central necrosis, a finding highly suggestive of metastatic disease. This is seen as unenhancing areas on contrast MR1 (or CT) or very high signal areas on T,-weighted and STIR sequences.

Trauma Rupture of the tough fascia over the corpora cavernosa (the penile tunica albuginea) is a rare injury, referred to as fracture of the penis. It is usually due to lateral trauma during athletic or unusual sexual activity, generally presenting with a history of a sensation of

Acute epididymo-orchitis (two cases). Anterior 99mTc i mages obtained 1 min after injection showing diffusely increased uptake in the right testis of case A and the left testis of case B. Case A also shows diffusely i ncreased activity along the spermatic cord on the affected side. A l ead marker has been used to indicate the midline of the scrotum. Fig. 32.45

Previous Page

CONTENTS

Next Page

THE MALE GENITALIA AND URETHRA Table 32.2

Technique The patient is positioned supine, with the scrotum ele-

vated by a towel placed between his legs, or better, a sling of adhesive tape placed across the thighs. It is also helpful to place a l ead-rubber shield beneath the sling and the scrotum in order to block out activity from the patient's thighs. A rapid bolus injection of about 200 MBq of 99mTc-pertechnetate i s given, and a rapid sequence of images of the scrotal region is obtained (e.g. one frame per second for 40 seconds). After the initial dynamic series, one or two further spot views with high count density should he obtained, but typically the most informative data are those derived from summating the first-pass frames to a single image. Interpretation With torsion, the testis appears avascular or

severely ischaemic. This is manifest on the images as a photondeficient area, which is often surrounded by a rim of increased activity representing surrounding hyperaemia. The line of the spermatic cord also often shows increased activity. With acute epididymo-orchitis, activity within the testis itself is usually normal, but the inflammatory response in the epididymis produces a C-shaped margin of markedly increased activity, and in some cases this may appear to extend across the testis as well. The contralateral normal testis is used as a benchmark to indicate whether the affected side shows increased or reduced activity.

Radiological investigations of erectile dysfunction

Investigation

Indication

Comments

Duplex ultrasound

Screening test for vascular function

PSV >35 cm/s normal PSV 35 cm /s predicts normal arterial inflow and that PSV3 mm between 10 and 14 weeks gestation) there is a well-established association with chromosome abnormality (Nicolaides et al 1992) as well as with other structural anomalies (in particular cardiac abnormalities, diaphragmatic hernia, omphalocele and body stalk abnormalities). In centres where first trimester nuchal transl ucency screening takes place, an abnormal result necessitates formal chromosomal analysis by either early amniocentesis or CVS. Where the chromosome complement proves normal there should be detailed transvaginal scanning in the first trimester, with further second trimester transabdominal scanning at around 21 weeks in the search for structural abnormalities. The advantages of such screening are, of course, the earlier detection of abnormality and the possibility that suction termingtion may be possible in appropriate cases. One major disadvantage is the anxiety and grief caused to parents making decisions about whether or not to progress a pregnancy when there is already a high natural loss rate in fetuses with chromosomal abnormalities and major structural anomalies. Others include difficulties in obtaining a reliable nuchal measurement in a significant proportion of fetuses, determining the significance of minor abnormalities, and the absence of any pathological confirmation of the diagnosis.

1049

Fetal gender This may be accurately determined at or after 20 weeks gestation in 95-100% of fetuses. More recent work has used both TAS and TVS for earlier assessment. Overall success at 1 2-14 weeks was 80% in one study (Whitlow et al 1999), identification increasing with gestational age and being highest at 14 weeks (90%). In relation to genetic screening, the clinical value of sex determination is in deciding whether to carry out prenatal invasive testing in pregnancies at risk of sex-linked genetic abnormalities, where testing would be necessary only with male fetuses.

The brain The fetal brain was one of the first areas examined for abnormality, not only because the head was routinely imaged for t he biparietal diameter, but also because CNS anomalies are among the commonest birth defects. Furthermore, such anomalies frequently carry severe clinical implications and their early recognition thus allows for the possibility of early therapeutic terminalion. With TVS it is possible to distinguish the fetal head and torso from 7 weeks gestation. At this time the fetal brain is seen as a single ventricle, while by the ninth week two lateral ventricles largely filled with choroid are apparent, surrounded by a thin cerebral mantle. The thalamus, third ventricle, midbrain, brainstem and cerebellar hemispheres have formed and change little throughout the rest of pregnancy, other than to enlarge progressively. Significant facial development takes place during the middle of the first trimester, with the bony elements of the mandible, maxilla and orbits being seen by 10 weeks (Fig. 33.23). By 18 weeks there has been marked thickening of the hypoechoic cerebral cortex and the ventricles appear less prominent. The spine The earliest spinal structures may be seen at 7-8 weeks with TVS (Fig. 33.24). Although the spine is seen clearly with TAS as two parallel echogenic lines from 12 weeks on, developmental changes that take place during the first half of the second tri mester make 18-20 weeks a more satisfactory time for the assessment of normality. The fetal vertebra is composed of three ossification centres: the anterior, which forms the vertebral body (centrum), and two posterior centres, which form the posterior neural arch. Full assessment requires imaging in three planes: the oblique coronal (Fig. 33.25 through the anterior centrum and one


'Soft markers' of chromosome abnormality There is some association of minor but relatively common sonographic abnormalities (including choroid plexus cysts, mild renal pyelectasis, nuchal (hickening, shortened femur or humerus, echogenic bowel and mild ventriculomegaly) with chromosome abnormalities, although this has not been fully assessed in unselected populations (Chitty 1998). The majority of reported studies are based on selected populations and do not allow for known risk factors such as maternal age or serum markers. Uncertainty surrounds the value of these markers, with some units offering amniocentesis to affected individuals whereas others do not even mention or document the findings unless more than one marker is present. In general, prior risk based on maternal age and serum markers may be modified by the presence of soft sonographic markers. Based on a detailed search for specific sonographic markers in patients at high risk of Down's syndrome, sensitivities and specificities in excess of 90% have been reported. Equally, where no markers were found, prior risk could be reduced by up to 50%.

Fig. 33.23 A normal 10 week fetus. This coronal image of the face shows the orbits (o), maxilla and mandible (m).

CONTENTS

Previous Page

Next Page

Fig. 33.24 A normal fetus with a menstrual age of 8 weeks 3 days. The parallel echogenic lines of the developing spine (S) can be seen.

Fig. 33.26 Transverse images. (A) Lumbar spine with three ossification centres forming a complete 'ring' (large arrows) in a third-trimester fetus. The two kidneys are seen on either side. (B) Lumbosacral junction. LW = iliac wings; A = lower abdomen. the fatal condition anencephaly and various forms of spina bifida, ranging from the severe lumbar myelomeningocele to the potentially surgically correctable meningocele. Neural tube defects are disorders of polygenic multifactorial inheritance, with an overall i ncidence of 1-2 per thousand live births in the USA and as many as 4.5 per thousand live births in the UK. Not only are they relatively common compared with other fetal anomalies but they are also associated with a high rate of perinatal mortality, morbidity and long-term disability. Current technology, both immunological and sonographic, has meant that most neural tube defects can be recognised sufficiently early for therapeutic termination to be a realistic possibility.


Oblique coronal images: cervical (A) and lumbar (B) spines. 0 = occiput; S = intact skin surface; st = nuchal soft tissues; sc = spinal cord. Fig. 33.25

posterior centre), the true coronal (through both posterior centres) and the transverse (Fig. 33.26). The centra are ossified first in the l ower thoracic and upper lumbar regions, followed by progressive ossification in both the cephalic and caudal directions. In contrast, ossification in the posterior centres follows sequentially in a crani ocaudal direction from the cervical region downward, and may not be seen in the lumbar spine until 22-24 weeks and in the sacral spine until 25 weeks or more. Awareness of these features is necessary to prevent false-positive diagnoses of spinal dysraphism.

Diagnosis A positive family history in either parent results in an i ncreased risk of occurrence in the offspring (5%); in a couple with a prior affected child, the risk of recurrence is approximately 2%. Over 90% of neural tube defects occur in families without a positive family history. Reference has already been made to the use of maternal serum a-fetoprotein screening and amniocentesis in the diagnosis of neural tube defects. If screening is applied to all pregnant women, few cases will be missed. Such a test works well for a very large group of fetuses with a low prevalence of disease, but there will be both false-positive and false-negative cases. Sonography subsequently provides an effective second-line test for discriminating the normal from the abnormal (and neural tube defects from other causes of a high MSAFP) in a small group of fetuses with a high prevalence of disease. Anencephaly This accounts for 50-65% of all neural tube defects

Neural tube defects result from failure of the neural tube to close in early embryogenesis. These malformations include defects such as

and is the commonest anomaly affecting the CNS. There is a failure of closure of the neural tube at its cephalic end between

Previous Page

CONTENTS

Next Page OBSTETRIC ULTRASOUND

105 1

Fig.

33.28 Encephalocele (small arrows) with visible sulci. D = calvarial defect.

but are commoner in the lumbar or lumbosacral regions (90%) than in the thoracic (6%,('%) or cervical (3%) spines. Isolated meningoceles are unusual compared with the commoner myelomeningocele. Prognosis tends to be worse for higher and larger lesions. or when there are other associated anomalies. Despite this, prediction of neurological deficit is frequently unreliable and demonstration of lower limb movement and bladder emptying in utero is no guide to function after birth. Sonographic diagnosis may be difficult. The spine should be i maged in all three planes (sagittal, coronal and transverse), as any one of these may show abnormalities not demonstrated on the other two. The spectrum of abnormality ranges from complete spinal disorganisation (Fig. 33.29), through isolated abnormalities at different l evels (Fig. 33.30), to subtle splaying of the posterior neural arches demonstrated only on the coronal image. Classically, a divergence of the posterior neural arches on longitudinal imaging (Fig. 33.31 A) is Fig. 33.27 (A,B) Anencephaly in two fetuses showing prominent orbits (arrows) and an absence of cerebral tissue more cranially. the second and third weeks of development, which results in an absence of the cerebral hemispheres but relative preservation of the brainstem and portions of the midbrain. The main sonographic features are symmetric absence of the cranial vault and normal brain structures above large and prominent orbits (Fig. 33.27).


This may be seen in 40-50% of cases. Although the diagnosis may be suspected by 12-13 weeks, it is more reliable by around 15-16 weeks, at which time the ossification in normal calvarial bones is more obvious. Associated anomalies include spinal defects (in up to 50%) and midline facial defects. Polyhydramnios

The least common neural tube defect, encephalocele results in a bony calvarial defect which allows herniation of meninges alone (cranial meningocele) or both brain and meninges (encephalocele). Approximately 75% occur in the occipital midline, t he remainder being divided between the frontal midline and parietal regions. Sonographically detected as a fluid- or solid-filled structure protruding from the calvarium, an encephalocele is usually fairly easy to demonstrate (Fig. 33.28). It may be more difficult conclusively to show a bony calvarial defect (necessary to distinguish an encephalocele from the commoner cystic hygroma) and conversely to guard against the possibility of producing bony defects by artefact ual acoustic shadowing, a mistake which could lead to incorrect management decisions regarding the fetus. Encephalocele

Spina bifida This results from a failure of the neural tube to close at 3-4 weeks. Lesions may occur anywhere along the spine

Fig. 33.29 Cervicothoracic myelomeningocele: longitudinal (A) and transverse (B) images with grossly abnormal distal spine (C, arrows). (D) Equivalent radiograph.

Previous Page

1052

CONTENTS

Next Page


Fig. 33.32 Meningocele: transverse image. S = sac containing only fluid; poc = splaying of the posterior ossification centres.

Fig. 33.30 Spina bifida with defects in the cervical (C), thoracic (T) and l umbar (L) spines. Longitudinal (A) and magnified (B) images. accompanied by a 'V'-shaped profile on transverse imaging (Fig. 33.31 B), due to outward flaring of the two posterior ossification centres. The presence of an intact sae seen as an extension from the posterior aspect of the spine makes the diagnosis easier; conversely, when a sae is not seen, either because it has ruptured or is compressed against adjacent structures, the diagnosis must be made

Fig. 33.33 Meningomyelocele: transverse image. Short arrows = sac containing neural elements; poc = splaying of the posterior ossification centres.


by demonstrating the 'bony' defect of spina bifida. With a simple

Fig. 33.34 The 'lemon sign'. There is flattening of the frontal bones (arrows) on this transverse image through the Bead. meningocele the sae is fluid-filled (Fig. 33.32) but in the ease of a myelomeningoccle it also contains solid tissue (Figs 33.29, 33.33). Other pointers to the diagnosis which have attracted interest

recently include abnormalities of fetal cranial contour (the 'lemon sign'; Fig. 33.34) and of the cerebellum (the 'Banana

sign'

Fig. 33.35). Initially described by Nicolaides et

al (1986), these signs have been shown to Be useful predictors of spina bifida in a Fig. 33.31 Cervical spina bifida: longitudinal (A) and transverse (B) images. Note the 'V' shape rather than the normal 'ring' of ossification centres on the transverse image.

high-risk population, although in a routine screening population their positive predictive values are much less (Filly 1988). Effacement of the eisterna magna is another important pointer to the

•

Previous Page

CONTENTS


105 3

Fig. 33.35 The 'banana sign'. Note the cerebellum (arrows) stretched around the brainstem (B) with effacement of the cisterna magna. Arnold-Chiari malformation, as is ventriculomegaly, seen in 80-90% of cascs of spina hifida. Full assessment of the fetal cranium after the first trimester thus requires imaging in the transcerebellar, transthalamic and transventricular planes.

Fetal hydrocephalus has many causes and a variable, although generally poor. prognosis. Neonatal survival is less than 30% and many of t hese infants arc mentally or physically handicapped. Although it may prove impossible to distinguish non-obstructive ventriculomegaly from hydrocephalus, in all cases a careful search should he made for other anomalies, both of the CNS and other organ systems. Up 30% of all cases of hydrocephalus may be associated with p a meningomyelocclc either or, less commonly, an encephalocele . Nyberg et al (1987c) reviewed 61 fetuses with hydrocephalus over a e S; year period and found that 84/c had one or more major CNS malformations or extra-CNS malformations. Fetal mortality was directly related to the presence of extra-CNS anomalies, which in t his series was associated with a uniformly fatal outcome. conversely, the prognosis for mild and occasionally transient lateral ventricular dilatation in the absence of other sonographically detected abnormalities (idiopathic ventricular dilatation) may be good. Accurate assessment is therefore essential. The lateral cerebral ventricles become easily visible at around 1 3-14 menstrual weeks. The diagnosis of hydrocephalus has traditionally been based on the lateral ventricular ratio (Fig. 33.36) and a number of charts have been developed for normal fetuses which relate the ratio of the lateral ventricular width to the hemispheric width as a function of gestational age. Unfortunately the wide range of normality, particularly early in the second trimester, limits their usefulness.

Fig. 33.37 One of bilateral choroid plexus cysts (cursors), measuring 13 x 8 mm. There were additional forearm abnormalities and a small cystic hygroma in this fetus, which karyotyping showed to be a trisomy 18. Other criteria for the diagnosis of vcntriculomegaly have been suggested. An absolute measurement of lateral ventricular width ( measured from the midlinc echo to the leading edge of the lateral wall of the lateral ventricle) of I. I cm or less and choroid plexus filling the body and atria of the lateral ventricle both suggest normali ty. Most commonly used today, the ventricular atria/ diameter has been shown to be largely age independent through the second and third trimesters (Cardoza et al 1988), a value greater than Ii) mm indi eating significant ventriculomegaly. The attraction of this measurerent is its simplicity, requiring neither a knowledge of the gestational age nor reference to charts relating ventricular size to gestational age to judge normalcy. Although the medial atrial wall is a better specular reflector than the lateral wall, the gravity dependent orientation of the choroid plexus serves as a useful pointer to the position of the latter interface. It has been suggested that assessment of the ventricular atrium, in combination with assessment of the posterior fossa and cisterna magna, constitutes the most appropriate sonographic screen mg procedure for major CNS malformations (Filly et al 1991). Care must he taken not to confuse vcntriculomegaly with choroid plexus cysts. These relatively common structures are almost always transient and of no clinical significance. Associated chromosomal abnormalities, in particular trisomy 18. have been reported but are uncommon (Fig. 33.37).

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ Fig. 33.36 Ventriculomegaly: transverse image through fetal head. a/b = lateral ventricular ratio.

Fig. 33.38 Dandy-Walker syndrome. The posterior fossa cyst measures 12 x 9 mm in this fetus with a gestational age of 19 weeks. A hypoplastic cerebellum was identified and there was also ventriculomegaly with a ventricular atrium measuring 14 mm across.

Previous Page

1054

CONTENTS

Next Page


Dandy-Walker syndrome ( DWS) This results from abnormal development of the cerebellum and fourth ventricle. Atresia of the foramina of Luschka and Magendie, with hypoplasia or aplasia of the cerebellar vermis, leads to dilatation of the fourth ventricle and expansion of the posterior cranial fossa (Fig. 33.38). Hydrocephalus is usually seen but is variable in extent. Associated anomalies include agenesis of the corpus callosum, encephaloceles, polycystic kidneys and cardiovascular defects. The sonographic appearances are characteristic, with separation of the cerebellar hemispheres by the enlarged fourth ventricle. Distinction must be

made from a posterior fossa arachnoid cyst, in which case there is displacement but no separation of the cerebellar hemispheres by the cystic lesion. Aqueduct stenosis This may be inherited as an X-linked receslive trait, but infectious, teratogenic and neoplastic causes have all been implicated. Although the diagnosis is suggested by the finding of dilated third and lateral ventricles with a normal fourth ventricle, these features are non-specific. Many other conditions of both non-obstructive ventriculomegaly and true hydrocephalus can give a similar sonographic appearance; even in the most experienced hands the diagnosis remains one of exclusion. Holoprosencephaly This is a complex developmental abnormality of the forebrain occurring in early embryonic life. Chromosomal anomalies are frequent and there is both a familial ten tendency and also an association with teratogenic agents. The condition results from a failure of cleavage of the prosencephalon which gives rise to the cerebral hemispheres and diencephalon. Associated facial anomalies include cyclopia, cebocephaly and a median cleft lip. Three types are recognised alobar, semilobar and lobar depending on the degree of forebrain cleavage. Alobar holoprosencephaly, the most severe form, is associated with a

Semilobar holoprosencephaly: transverse i mage. MV = monoventricle; small arrows = anterior cortex with no interhemispheric fissure; large arrow = posterior falx. On this image there is no distinction from the alobar form. Fig. 33.40

monoventricular cavity, fusion of the thalami, and absence of the corpus callosum, falx cerebri, optic tracts and olfactory bulbs (Fig. 33.39). In semilobar holoprosencephaly the two cerebral hemispheres are partially separated posteriorly but there is still a single ventricular cavity and partial fusion of the thalami (Fig. 33.40), whereas in the lobar variety the interhemispheric fissure is well developed and there is some separation of the thalami. In all types the posterior fossa is normal. Reliable prenatal sonographic diagnosis is possible in cases of alobar and semilobar, but not usually lobar, holoprosencephaly. The prognosis for the first two of these is very poor. Hydranencephaly This is the result of a destructive intrauterine i nsult rather than a developmental anomaly. Vascular occlusion of the destrucon internal carotid arteries has been postulated, leading to destruoti cerebral hemispheres, occa,ionally with some sparing

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ Hydranencephaly: sagittal (A), anterior coronal (B), transverse (C) and posterior coronal (D) i mages. Note the brainstem surrounded by fluid (arrows, A,B) and an absence of cerebral cortex. fn (D) the tentorium (small arrows) can be seen above a normal posterior fossa. c = cervical spine. Fig. 33.41

33.39 Alobar holoprosencephaly: coronal (A) and transverse (B) images. T = fused thalami; MV = monoventricle; arrows = cortical mantle. Fig.

Previous Page

CONTENTS


of the temporal and occipital cortex and basal ganglia. There is preservation of the brainstem, thalami and cerebellum, although these may be smaller than usual. The falx cerebri may be absent or incomplete and the choroid plexus is preserved, thus allowing secondary hydrocephalus and macrocrania to ensue. The head is filled with CSF, contained in a cavity lined with leptomeninges. The sonographic appearances of the brainstem protruding into this fluid-filled cavity are fairly characteristic (Fig. 33.41 A-C), but occasional confusion may occur with severe hydrocephalus (a thin ri m of compressed cortical tissue would differentiate the two) or alobar holoprosencephaly (but the thalami are not fused in hydranencephaly). The posterior fossa appears normal (Fig. 33.41 D). Prognosis and management are as for holoprosencephaly. of the corpus callosum This may be associated with other CNS and non-CNS abnormalities, including holoprosencephaly, the Dandy-Walker malformation, microcephaly, macrocephaly, median cleft lip syndrome, and cardiovascular, gastrointestinal and genitourinary anomalies. Sonographic features include elevation of the third ventricle, an abnormal lateral yentricular axis, and localised dilatation of the atria and occipital horns. In addition, an abnormal pattern of gyral development occurs along the interhemispheric fissure, where sulci radiating Agenesis

toward the third ventricle may be seen as a wavy appearance of the midline in axial sections. The prognosis is variable and is frequently determined by associated anomalies. In cases where the agcnesis is an isolated finding normal obstetric management applies.

Congenital heart disease is the most common severe congenital anomaly found in neonates and may be associated with serious morbidity and mortality. Examination of the fetal heart is an integral part of any prenatal obstetric ultrasound examination. Specific indications include the previous occurrence of congenital heart disease in mother or sibling (both with a recurrence risk of between 2 and 5%), a maternal disease known to affect the fetus, such as diabetes mellitus, and the use by the mother of drugs such as lithium or alcohol which are also known to affect the fetus. Fetal heart activity has been reported at 40 days and should always he present at 46 menstrual days with TVS. Cardiac motion is generally seen as soon as the embryo is visualised and always

1055

when the crown-rump length is greater than 5 mm. Normal cardiac rates vary with menstrual age, with values anywhere between 90 and 128 beats/min at 6 weeks, reaching a peak of up to 174 beats/ min at 9 menstrual weeks, before declining to around 150 beats/min at 14 weeks. The basic view of the fetal heart, and in many centres the only view obtained during the routine obstetric sonographic examination, is the four-chamber view (Fig. 33.42). This can be obtained in most fetuses older than 16 weeks but a more reliable assessment is possible between 18 and 22 weeks of gestation because the cardiac valves are then well developed and the size of the fetal heart and position of the fetus generally allow better visualisation at this stage. Evaluation of the anatomical detail provided by this view may require several different approaches (McGahan 1991). The apical four-chamber view allows good visualisation of the atrioventricular valves, while the subcostal view will better demonstrate the i ntegrity of the atrial and ventricular septa. The four-chamber view i s also excellent for defining the comparative size of the cardiac chamhers but does not demonstrate the aorta and pulmonary arteries. Other views are therefore necessary for establishing diagnoses such as transposition of the great vessels, tetralogy of Fallot, truncus arteriosus, aortic and pulmonary valve stenoses, and coarctation of the aortic arch ( Allen 1986). Long-axis views may be obtained through both the left and right ventricles, allowing assessment of their respective outflow tracts, while the aortic arch is best imaged in the longitudinal axis of the fetus. A short-axis view allows for detailed assessment of the great vessels and can rule out a transposition. Both M-mode and Doppler echocardiography have been used in the assessment of fetal cardiac anomalies. Not only can cardiac chambers and great vessels be measured but also dvsrhvthmias and stenotic and regurgitant valvular lesions may be diagnosed and quantified. With increasing experience of the technique, it is likely that fetal echocardiography for congenital heart defects and other cardiac anomalies will become an important part of fetal assessment.

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ Fig. 33.42 Fetal heart: subcostal four-chamber image. S = spine; ra = right atrium; la = left atrium; rv = right ventricle with moderator band; fv = left ventricle.

In addition to the assessment outlined above, the first trimester nuchal translucency measurement should also be mentioned, as documented above (p. 1041). While important in the assessment of chromosomal abnormalities, an increased value may also be a strong pointer to major congenital cardiac defects. In one study (Hyett et al 1999), 55% of major abnormalities of the heart and great vessels were associated with an increased nuchal translucency thickness at 10-14 weeks gestation. They concluded that an i ncreased thickness constituted an indication for specialist fetal echocardiography, even within the first trimester, where the relevant expertise exists.

Fig. 33.43 Cardiac rhabdomyoma (R) in a 33 week fetus: oblique (A) and transverse (B) images. The tumour measured approximately 4 cm i n diameter. Note small pericardial effusion (arrows).

CONTENTS

Previous Page 1056

Next Page


Fig. 33.44 Hydrops fetalis, transverse image. P = pleural effusions; H = heart; SP = spine.

Diaphragmatic hernia: transverse (A) and longitudinal (B) i mages. S = stomach, herniated into chest; H = heart; arrows - hypoechoic diaphragm. Note the polyhydramnios. Fig. 33.46

Hypoplastic chest in a fetus with multiple abnormalities including hydrocephalus and short limbs. C = chest; A = abdomen (with ascites). Fig. 33.45

Apart from the anomalies mentioned above, ultrasound may also diagnose congenital cardiac tmnou s. These are rare and generally benign. The commonest arc rhahdonivontas (Fig. 33.43); less common lesions include fihrumas, ntv.vomas and haetnangiomas. Although such tumours are generally isolated anomalies, there is a recognised association with tuberous sclerosis i n up to 86% of cardiac rhabdomyomas.

Fig. 33.47 Pulmonary sequestration: longitudinal (A) and transverse (B) images. M = hyperechoic mass lying at left lung base.

tality ranges from 50 %r to 80%, being due mainly to pulmonary hypoplasia arising secondary to the compression of the lungs by the adjacent hernia. Many anomalies may be associated with CDH, particularly in the central nervous and cardiovascular systems, and these are thought to account for many antenatal deaths. A detailed sonographic examination, together with chromosomal analysis, is essential in these patients.


Fetal lungs are non-aerated and demonstrate a homogeneous and solid appearance on ultrasound examination. Despite attempts to correlate the echogenicity and compressibility of lung with pulmonary maturity, the lecithin:sphingunwelin ratio within amniotic

fluid remains the most reliable means of assessment at present. The fetal thorax can be scanned transversely from the clavicles at the l ung apices to the echo-poor diaphragmatic crura, noting the shape and size of the chest, the size and position of the heart, and the appearance of the lungs. While small pericardial e/jusions have been reported in the normal fetus, pleural fluid is abnormal at any gestational age and may occur either in isolation or as part of a more general condition, such as fetal hydrops (Fig. 33.44) or urinary ascites secondary to posterior urethral valves. Pulmonary hvpoplasia may be evident (Fig. 33.45) and is a common sequel to oligohydramnios of any cause. The commonest thoracic abnormalities diagnosed with ultrasound are the followin7.5 ml More than seven 2-5 mm follicles; mainly peripheral distribution I ncreased stroma

Normal size

Small

Several follicles of 5-10 mm

No evidence of follicular activity

No increased stroma Normal/small uterus with thin endometrium

No increased stroma Small uterus with thin endometrium

Large/normal uterus with thick endometrium

CONTENTS

Previous Page

Next Page GYNAECOLOGICAL IMAGING

Fig. 34.12

EVS.polycyst uir ovar

1073

y with peripheral cysts.

Fig. 34.15 TAS. Thickened irregular endometrium (arrows) due to tamoxifen. Note the small cysts (arrowheads) at the myometrial endometrial interface.

Fig. 34.13

TAS. Polycystic

ovary. Central and peripheral cysts.

Sequential HRT i.e. oestrogens followed by progesterones for at l east 12 days/month, are usually prescribed for perimenopausal and early postmenopausal patients who have persistent but unpredictable ovarian function. Most patients have regular monthly bleeds on such a regimen so should have an endometrium that varies in thickness during the cycle. The maximum thickness should still be less than 8 min. Any patient found to have a mildly thickened endometrium on sequential HRT should have a repeat scan following a withdrawal bleed. Continuous combined hormone replacement involves taking oestrogen and progesterone together continuously and should result i n endometrial atrophy within 6 months, i.e. endometrial thickness l ess than 4 mm.

Tamoxifen Tamoxifen, used in the treatment of breast cancer, has both antioestrogenic and oestrogenic effects and is associated with endometrial hyperplasia, an increased incidence of polyps and endometrial carcinoma. The uterus has a characteristic appearance with increased thickness of the cavity echo with multiple tiny cysts (Fig. 34.15). These cysts were originally thought to be part of the endometrium but recently it has been suggested that they are actually in the most superficial layer of the myometrium rather than the endometrium. Doppler studies in patients with endometrial thickening due to tamoxifen show low impedance flow, i.e. low values for the RI and PI.


TAS. Multifollicular ovary (arrows) in a patient with amenorrhoea due to anorexia nervosa. Fig. 34.14

levels of circulating oestrogens and arc seen in athletes and in association with weight loss and anorexia nervosa. Primorr ovarian,%ciilure or premature menopause is the onset of dre menopause before the age of 35 years and is a rare cause of amenorrhoea.

Value of ultrasound screening for patients on HRT Routine screening is not indicated for patients on HRT but should be performed in those with abnormal vaginal bleeding.

Combined oral contraceptive pill The uterus and ovaries are

suppressed. The endometrium should be thin and the ovaries small with no evidence of follicular activity.

The progesterone-only pill does not necessarily inhibit follicular activity but the uterus should look small, with a thin endometrium. Progesterone-only pill

replacement therapy ( HRT) The effect of HRT on the uterus and ovaries depends on the type of HRT taken. Unopposed oesnr,gen admini.stration causes thickening of the endometrium due to endometrial hyperplasia. This is a precursor to endometrial carcinoma, therefore unopposed oestrogen regimens are only recommended if the patient has had a hysterectomy. Hormone

Conventional IUCDs An IUCD is a common incidental finding i n patients undergoing pelvic ultrasound. They are easily visualised as highly echogenic structures with acoustic shadowing and should be entirely within the uterine cavity, not protruding into the nryometrium or endocervical canal (Fig. 34.16). Large fibroids, particularly if calcified, can occasionally cause difficulty. Expulsion during menstruation does occur but most patients referred with 'missing coil' will have the IUCD present normally i n the cavity, the threads having retracted into the cervical canal. Complications such as migration of the coil into or through the

Previous Page 1074

Fig. 34.16

CONTENTS

Next Page


TAS. IUCD (arrow) in the cervical canal.

Fig. 34.17 TAS. IUCD embedded in the myometrium (arrow), well outside the cavity (arrowheads).

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ Fig. 34.18 1 6-year-old girl with cyclical abdominal pain found to have a didelphys uterus with a complete vaginal septum and imperforate hymen on the left. (A) Longitudinal scan demonstrates a large heterogeneous mass (arrows) due to blood-filled distended vagina. A small uterus and cervix (arrowheads) can be seen at the upper margin of the mass. (B) Longitudinal scan to the right of (A) shows a further uterus (arrowheads mark the position of the endometrium). The normal right vagina is compressed and hence not visible. (C) Transverse scan confirming two uterine bodies. (D) Intravenous urogram. Solitary but duplex kidney on the left.

Previous Page

CONTENTS

Next Page

myometrium usually occur at the time of insertion and are associated with pain (Fig. 34.17). It is rarely possible to see an IUCD outside the uterus on ultrasound, so if there is doubt about the presence of an IUCD on the ultrasound a plain abdominal X-ray should he performed to exclude uterine perforation. Ultrasound is also indicated to look for causes of difficulty inserting a coil, e.g. large fibroid distorting the cavity or a severely retroflexed uterus. The IUCD is associated with an increased incidence of pelvic nfection. In particular colonisation with Actinomycosis may occur i with long-term use and may well be asymptomatic. The Mirena IUCD contains progestogens and is used for both contraception and to treat dysfunctional uterine bleeding. Although similar in shape to most other coils, the Mirena coil is less echogenic than standard IUCDs and hence is harder to visualiseon ultrasound, even when this is performed endovaginally. It is often seen by virtue of its acoustic shadow rather than direct visualisation of the coil itself. Mirena coil

Minor duplication abnormalities of the uterus are only of relevance i n the investigation of subfertility and recurrent miscarriage and are discussed in more detail in the section on hysterosalpingography. However, severe duplication anomalies with obstructed menstrualion present in adolescence and early adulthood. Typical presenting symptoms are primary amenorrhoea, cyclical abdominal pain, pelvic mass and severe dysmenorrhoea. Ultrasound reveals a thickwalled cystic mass, owing either to an obstructed vagina (imperforate hymen) or an obstructed uterus (hacmatomctrium), with usually be reached if the possibility is considered and care taken to look for accessory pelvic organs. The examination must also include visualisation of the urinary tract as there is a high incidence of associated single kidneys (Fig. 34.18). A non-obstructed double uterus (uterus didelphys) may simulate a solid adnexal mass on pelvic examination but can be differentiated on ultrasound due to the presence of a central endometrial echo

Fig. 34.20 EVS. Two endometrial echoes (arrows) within the uterus, suggestive of a uterine septum.

Fibroids are present in up to 20-50% of women. They are particul arly common and present at a younger age in black women. Although frequently asymptomatic, they may present with abnormal bleeding, pain, abdominal distension, subfertility or recurrent miscarriage, the symptoms depending to a certain extent on the l ocation and size of the fibroids. Their location in the uterus is described as: Submucous-arising within the cavity. Rarely they can form a fibroid polyp and protrude through the os (Fig. 34.21). ' Mural within the myometrium. They may or may not abut or distort the shape of the cavity, depending on their precise l (Fig. 34.22). ocatinerlhsduob.Tecrtdiagnos ' Subserosal arising deep to the serosa and causing a bulge on the surface of the uterus (Fig. 34.23). a pedicle usually from the serosal surface. . Pedunculated- on • Cervical rare, i.. less than 5% fibroids. Fibroids are hormone-dependent, increasing in size and becoming l ess echogenic during pregnancy, and decreasing in size following the menopause. However, regression following the menopause is prevented and may be reversed in patients taking HRT. On ultrasound, most fibroids are round, well-defined hypoechoic masses with a characteristic internal architecture showing recurrent shadowing (Fig. 34.24). They can be hyperechoic and may be calcified, particularly in postmenopausal patients. Degeneration within fibroids appears as either areas of increased echogenicity or i rregular cystic areas (Fig. 34.25). Fibroids can he very vascular so Doppler studies may show very low impedance flow. Malignancy (leiomyosarcoma) is rare but should be suspected if a fibroid suddenly increases in size. There are no specific ultrasound features of malignancy. The role of ultrasound is to confirm the diagnosis and determine the number, size and location of fibroids, as this will help determine their likely significance and appropriate treatment. See Table 34.2 for differential diagnosis of fibroids.


(Fig. 34.19). The shape of the endometrial echo, particularly on a transverse scan, can also help diagnose the less severe duplication anomalies (Fig. 34.20). Differentiation of a septate from a bicornuate uterus depends on identification of a fundal notch and requires visualisation of the uterus en face. This can be difficult with conventional ultrasound but is more easily achieved with 3D ultrasound. Recognition of a unicornuate uterus is also difficult on ultrasound but can be inferred by its small size and abnormal lateral position.

Fig. 34.19 TAS. Two separate uterine horns indicative of bicornuate uterus. Arrowheads indicate position of endometrium.

Adenomyosis or endometriosis interna is the presence of endometrial glands within the myometrium associated with adjacent myometrial hyperplasia. It is usually a diffuse process but may form a localised mass or adenomyoma. Clinical findings are dysmenor-

Previous Page 1076

CONTENTS

Next Page


Fig. 34.23 TAS. Two subserosal fibroids (arrows). Arrowheads indicate position of uterine cavity.

Fig. 34.21 (A) TAS. Fibroid polyp (arrows) within the cervical canal. The stalk of the polyp (arrowheads) can be seen in the uterine cavity. (B) TAS. Same patient 6 months earlier. The fibroid polyp (arrows) is now seen within the uterine cavity.

Fig. 34.24

TAS. Fibroids. Typical recurrent shadowing (arrowheads).


Fig. 34.25 TAS. Irregular area of increased reflectivity in the centre of a fi broid due to degeneration.

Fig. 34.22 TAS. Typical mural fibroids (arrows) abutting but not displacing the cavity (arrowheads). The larger fibroid shows typical recurrent shadowing.

The ultrasound features of diffuse adcnomyosis are poorly defined

areas

of decreased echogenicity and heterogeneity in the

myometrium, associated in approximately 509c with small (2-5 mm) cystic spaces in the myometrium (Fig. 34.26). Using these criteria.

rhoea and menorrhagia with a tender bulky uterus. Most cases are

sensitivities and specificities of 80-90°/r; have been achieved.

diagnosed following pathological examination of hysterectomy

Focal adenomyomas cause focal bulges in the myometrium and

specimens: however, cndovaginal ultrasound and MRI have been

may he hyper- or hypoechoic. They are generally less well defined

shown to he of value.

than fibroids but it is difficult to differentiate the two.

CONTENTS

Previous Page


Table 34.2

Differential diagnosis of fibroids

Location

Differential diagnosis

Comments

Submucosal

Endometrial polyps Retained products of conception (RPC) Endometrial carcinoma

Usually hyperechoic, i.e. similar to endometrium History helpful. RPC usually of mixed echogenicity Postmenopausal patients. Less well defined, possibly with myometrial invasion Heterogeneous endometrium with little mass effect. Cavity may contain fluid (or gas if fistula present)

Endometritis / fluid collection

Mural/subserosal

Adenomyoma / area of adenomyosis Leiomyosarcoma Myometrial contraction Metastatic deposits

Pedunculated

1077

Any cause of solid adnexal mass

Less well defined area of heterogeneity with no recurrent shadowing effect Rapid growth but otherwise impossible to differentiate. Very rare Poorly defined virtually isoechoic mass; changes with ti me-rarely a problem Very rare, no distinguishing features Usually possible to see connection with uterus. Doppler of no value

• retained products of conception • trophoblastic turnouts-usually following a pregnancy. Fluid, including pus or blood, can also distend the cavity but should not be included in the endometrial measurement. A tiny amount of fluid (1-3 nom depth) is occasionally seen in a postmenopausal uterus and is of no significance; however, a cavity distended by fl uid usually indicates an obstructed uterus. An attempt should he made to identify the underlying cause. e.g. cervical or uterine carcinoma, previous radiotherapy to the cervix, uterine synechiae due to Asherman's syndrome, previous cervical surgery, etc. A pyometrium due to a uterocolic or utcrovesical fistula can cause a si milar appearance. Fig. 34.26 EVS. Adenomyosis. Coarse myometrial texture with small cysts due to blood lakes.

Enr/antetria/ polyps arc common and typically measure 5-15 min (Fig. 34.27). There is an increased incidence in patients on tamoxifen or HRT. Most polyps are benign and cause intermenstrual

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ bleeding, with or without pain.

i s a precursor to endometrial carcinoma The endometrium is considered abnormally thick if it measures Endomtetrial hyperplasia so must he recognised and endometrial sampling performed. Causes more than 14 mm in a premenopausal patient and more than 4 mm of endometrial hyperplasia include polycystic ovaries, obesity, nopau salpatient Causes of thickening of the cavity in a i exogenous hormones, endogenous excess oestrogen production, e.g. echo include: e: due to functioning ovarian tumours. • endometrial polyps Differentiation of endometrial polyps from hyperplasia can be • suhnurcous fibroids

• endomctrial hyperplasia • endometrial carcinoma

Fig. 34.27 EVS. Multiple endometrial polyps. Note the midline echoes due to the endometrial interface (arrows) are displaced by the polyps. This is a useful feature when trying to differentiate hyperplasia from polyps.

difficult but can be helped by performing an endovaginal ultrasound examination during intrauterine injection of saline, so-called sonohysterography (Figs 34 28, 34.29).

Fig. 34.28 EVS. Endometrial thickening. This looks like hyperplasia but subsequent saline hysterography demonstrated it to be a polyp.

Previous Page 1078

CONTENTS

Next Page

A TEXTBOOK OF RADIOLOGY AND IMAGING than 50% myometrial invasion) and stage 2 disease (more than 50% myometrial invasion); however, MRI is far better at looking for extrauterine spread, so this is the preferred technique. Doppler ultrasound of endometrial or intrauterine vessels is of little value in premenopausal patients because of cyclical changes and the effect of other uterine pathology, such as fibroids. However, measurement of Doppler indices may help in postmenopausal patients. It has been reported that malignancy is never found if the RI is greater than 0.83.

Fig. 34.29

Saline hysterography. Endometrial polyp outlined by saline.

Endometrial carcinoma occurs mainly in postmenopausal women. The risk factors are the same as those described for endometrial hyperplasia, i.e. increased oestrogen levels, in addition to hypertension, diabetes and nulliparity. The most common presenting symptom is abnormal uterine bleeding. Ultrasound appearances vary from moderate endometrial thickening, to an irregular hypoechoic intracavitary mass, to an enlarged diffusely infiltrated uterus (Figs 34.30, 34.31). Endovaginal ultrasound is said to be good at differentiating between stage I (less

The cervix is most effectively examined by direct inspection, Papanicolau smears and colposcopy, so the role of ultrasound is very li mited. It has no role in screening or routine staging of cervical carcinoma; however, occasionally patients present to the ultrasound department prior to vaginal examination so the radiologist needs to be aware of the appearances of cervical disease. Early tumours are undetectable but advanced tumours show an irregularly enlarged cervix (Fig. 34.32), with or without an area of highly reflective echoes due to necrosis. There may be an obstructed, distended uterine cavity. Advanced disease may also show invasion of the bladder wall and hydronephrosis due to ureteric involvement-either due to nodal i nvolvement or direct ureteric invasion (Fig. 34.33). Cervical polyps and cervical fibroids are also occasionally seen.

Simple adnexal cysts Simple adnexal cysts are most commonly functional in origin. They vary in size, reaching up to 7 cm in diameter, and yet still resolve spontaneously. Haemorrhage into

Fig. 34.30

carcinoma.


EVS. Poorly defined intrauterine mass due to endometrial

Fig. 34.32

TAS. Carcinoma of cervix. Large irregular cervix (arrowheads) with a small tongue of tumour (arrow) extending towards the bladder.

Fig. 34.31 TAS. Obstructed uterus. The cavity (arrows) is distended by blood with a polypoid mass just above the internal os due to endometrial carcinoma.

Fig. 34.33

TAS. Cervical carcinoma invading bladder base (arrows) and causing an obstructed uterus (arrowheads).

CONTENTS

Previous Page

Next Page GYNAECOLOGICAL

Fig. 34.34 eve. fimmoral

cyst adjacent

to ovary

I MAGING

1079

Fig. 34.36 EVS. Typical endometrioma with diffuse moderately high-level echoes (arrows). The differential diagnosis of a complex adnexal mass should i nclude:

• Haemorrhagic cyst-contains diffuse internal echoes or an

Fig. 34.35

EVS. Hydrosalpinx (arrows) adjacent to the ovary (arrowheads).

the cyst may cause pain and give rise to diffuse internal echoes or a clump of solid echoes within the cyst due to clot. Follow-up ultrasound will show a reduction in the size of the cyst and a change in the appearance of the internal echoes confirming its benign nature. Small simple cysts (1-3 cm) are also relatively common (3-5°h;) n i post menopausal women. Doppler insonation should be performed to confirm high impedance flow and serum CA- 125 should be measured. If both these investigations are normal and the woman is asymptomatic, it is reasonable to follow the cyst with serial ultrasound scans to confirm no growth rather than proceed to laparoscopy. The differential diagnosis of a simple adnexal cyst includes:

i rregular clump of echoes due to clot. Repeat scans helpful to show change. • Ruptured cyst-typical history, irregularly-shaped cyst with surrounding fluid. • Torsion of cyst or ovary-heterogeneous enlarged ovary with or without a thick-walled cyst with internal echoes. Presence of colour flow within the ovary is said to indicate viability of the ovary, hence laparoscopy is worthwhile to try and preserve function. • Endomctriosis. • Acute / chronic tubo-ovarian abscess. • Dermoid cyst-complex mass with cystic and solid areas, fat and/or calcification.

• Other neoplastic ovarian tumours, benign and malignant. • Pedunculated fibroid differentiation from an ovarian mass depends on identification of the ovaries separately.

• Ectopic pregnancy-should always be considered in a patient of


Paraovarian cysts may reach up to 10 cm, usually recognisable by the fact they are close to, but can he separated from, the ovary by gentle pressure with the ultrasound probe (Fig. 34.34). • Endometriomas (chocolate cysts)-usually contain internal echoes and have a thick wall but may look entirely simple. • Hydrosalpinx-a small hydrosalpinx may mimic an ovarian cyst but can be distinguished by its rather elongated shape, its position around or on the surface of the ovary and the presence of incomplete septations due to inucosal folds (Fig. 34.35). e Neoplastic cysts-particularly benign cystadenomas and some borderline tumours. • Peritoneal cysts or fluid trapped around the ovary due to adhesions. These may be asymptomatic or cause cyclical pain (entrapped ovary s.vndrome). The patients usually give a history of complicated pelvic surgery or infection.

Complex adnexal masses can be due to complicated simple cysts; however, various inflammatory and neoplastic causes must be considered in addition to some nongynaecological causes. Complex adnexal masses

child-bearing age. Pregnancy test important.

• Other inflammatory masses-e.g. appendix or diverticular mass. • Other neoplastic masses-e.g. arising from the bowel or peritoneum (benign peritoneal mesothelioma).

Endeinetriosis is an incidental finding in up to 25% of laparoscopies. Symptoms are variable but the most common is dysmcnorrhoea. The majority (up to 90%) of endometriotic (chocolate) cysts contain diffuse internal echoes due to old blood. The echogenicity of these internal echoes varies from very low level, only discernible

scanning endovaginally, to moderately high, which may cause some confusion with a dermoid cyst (Fig. 34.36). The echoes may show gravity-dependent layering creating a fluid-fluid level (Fig. 34.37). The wall thickness of the cysts varies and highly reflective foci or flecks of calcification may be seen within the wall. Septations, creati ng multilocular cysts, are common, the various locules containing echoes of differing densities, indicating haemorrhage of different ages (Fig. 34.38). Very large endometriotic cysts occasionally occur and may mimic a solid mass; however, compression of the mass with the probe will usually demonstrate the mass is deformable and the i nternal echoes move very slowly. Deposits are most easily recognised on the ovary and in the broad ligament; however, endometriotic deposits do occur anywhere in the pelvis or indeed outside the abdomen. Rarely nodules occur on the bowel, on the pleura and in the soft tissues, particularly at the sites of scars (Figs 34.39, 34.40). Adhesions and diffuse small endometriotic deposits cannot be

Previous Page 1080

CONTENTS

Next Page


TAS. Endometriosis in the bladder (cursors). Arrowheads mark position of the uterus. Fig. 34.40

Fig. 34.37

TAS. Bilateral endometriomas. Note the fluid level on the left

and the irregularly thickened wall.

TAS. Endometriosis. Complex ovarian mass with internal septations and echoes of varying density. Differential diagnosis must include a malignant tumour. Fig. 34.38

TAS. Acute pelvic infection with a thick-walled tubo-ovarian abscess (arrow) and free pus in the pouch of Douglas (arrowhead). Fig. 34.41

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ Fig. 34.42

Fig. 34.39

Huge pleural effusion in a young girl. Aspiration revealed heavily blood-stained fluid with multiple macrophages typical of pleural endometriosis.

visualised with ultrasound so a normal ultrasound examination does not exclude endometriosis. Endometriosis normally resolves after the menopause but may be reactivated if the patient is taking HRT, hence endometriosis should still he considered as part of the differential diagnosis of a complex cyst in post menopausal patients taking HRT.

EVS. Large thin-walled chronic hydrosalpinx.

Pelvic in flamnxtforv disease is becoming increasingly common as a cause of adnexal masses, both in the acute and chronic phases. The ultrasound in acute infection may show free fluid (pus) in association with a complex adnexal mass, which comprises the ovary and thickened surrounding tube (Fig. 34.41). Doppler insonation shows low impedance how due to a surrounding inflammatory reaction. In more chronic disease the ovary may he more easily definable (Fig. 34.35) with a thin-walled hydrosalpinx adjacent to the ovary. The hydrosalpinx may contain internal echoes due to either blood or pus (Fig. 34.42) and the ovary may look like a polycystic ovary

Previous Page

CONTENTS


1081

• I ncreased number of episodes of ovulation, for example: -following treatment with ovulation induction agents -the nulliparous state. Some protection is conferred by multiparity, breast-feeding and use of the contraceptive pill. Epithelial ovarian tumours Eighty-five per cent of malignant ovarian tumours are epithelial in origin and the commonest epithelial carcinoma (60-80%) is a serous cvstadenocarcinoma.

Serous tumours are predominantly cystic masses. They may show wall thickening and nodularity, internal solid areas and septations. Malignant tumours tend to have more nodularity and solid areas than their benign counterparts (Figs 34.44-34.46). Mutinous c.vstadenocarcirunnas are also large predominantly cystic masses but tend to be multiloculated with multiple thick i nternal septations and diffuse internal echoes due to their high Fig. 34.43 TAS. Complex pelvic mass behind the uterus (arrowheads) due to peritoneal mesothelioma. Note multiple internal septations and how the mass conforms to the shape of the pelvis. because of follicular fluid trapped by surface adhesions. Rarely patients present with right hypochondrial pain due to perihepatitis (Curtis-Fitz-Hugh syndrome). Adhesions around the liver have been described on ultrasound but it is rare to see any abnormality around the liver, and indeed the pelvic ultrasound may also he normal in even quite severe pelvic inflammatory disease. Benign peritonea/ mesotheiionut is a slow-growing rare tumour that is difficult to treat because of its propensity to local recurrence (Fig. 34.43).

TAS. Adnexal cyst with one solid area and some fine internal echoes suggestive of a serous cystadenocarcinoma. Histology confirmed a borderline malignant tumour. Fig. 34.44

Ovarian tumours are classified into three main types, according to their cells of origin: epithelial (60-70%), sex cord stromal (5-10%) and germ cell (15-20%) tumours. In addition, approximately 5% of significant ovarian tumours are metastatic in origin. Primary t umours can be associated with the production of various hormones, including oestrogens, progestogens and androgens. Calcification is seen in cystadenomas and cystadenocarcinomas, fibromas and dermoid cysts/tcratomas. In addition to classification as benign or malignant, some ovarian tumours are classified as borderline malignant, indicating that they have a hcuer prognosis, with a low risk of local recurrence and even lower risk of mcta.static disease.


Malignant ovarian tumours Carcinoma of the ovary is responsible for about 5000 deaths/annum in the UK: 80% of tumours occur in women over 50 years of age. Presenting symptoms (pain, abdominal distension, vaginal bleeding, bowel and urinary dysfunction) usually occur late in the disease with two-thirds of patients having spread outside the pelvis at the time of diagnosis. This late presentation is responsible for the overall high mortality rate of approximately 70% at 5 years. Metastatic spread occurs most commonly to the peritoneum, with multiple peritoneal nodules. ()mental thickening and ascites. Lymphatic spread to the para-aortic nodes and liver metastases are also seen. Risk factors for development of ovarian carcinoma include:

TAS. Malignant adnexal cyst with internal echoes and irregul arly thickened wall (arrowhead). Fig. 34.45

• Family history of ovarian, breast, endometrial or colorectal carcinoma. Women with a pathogenic mutation in the BRCA1 or BRCA2 genes have a lifetime risk of 40c%o or 25 /. respectively.

TAS. Solid tumour mass (white arrowheads) surrounding the posterior aspect of the uterus (black arrows). Fig. 34.46

Previous Page 1082

CONTENTS

Next Page


EVS. Ovarian fibroma. Homogeneous solid mass (arrows) arising from the ovary (arrowheads). Fig. 34.49

Fig. 34.47

TAS. Benign mutinous cystadenoma showing the typical

multiloculated appearance-impossible to differentiate from a malignant tumour. mucin content (Fig. 34.47). Benign mucinous cystadenomas similarly contain thick irregular scptations such that, in the absence of ascitcs or lymphadenopathy, it is impossible to distinguish a benign from a malignant mucinous tumour. However, in spite of their worrying appearance. only 10%- oh mucinous tumours are actually malignant. Between 60 and 70% of serous tumours and 5 and 10 (/( , of mucinous tumours are bilateral. Other rarer epithelial tumours include endometrioid carcinomas (these are associated with endometrial thickening due to endometrial hyperplasia or carcinoma in approximately 20-30%:), clear cell carcinomas and Brenner's tumours. Brenner'.s tumours are invariably benign, large at presentation and may be associated with a mucinous cystadenoma or dermoid cyst.

TAS. Typical dermoid with a floating echogenic area with acoustic shadowing due to fat, with or without calcification. . Fig. 34.50

Sex cord tumours This group of tumours includes Jibroma.s and the hormone-secreting tumours such as thecomas, granulo.sa cell tumours and Sertoli cell tumours (urrhenuhlastumas). They are usually benign solid tumours. Patients with functioning tumours often present with the symptoms due to the excess hormone production, e.g. post menopausal bleeding (Fig. 34.48). Fibromas are benign slow-growing tumours which when large may be associated with ascites and pleural effusions, a condition known as Meigs'syndrome . Fibromas can be heavily calcified (Fig. 34.49).


Germ cell tumours Dermoid cysts (or benign cystic tercrtomas) are the commonest tumours in this group; 95% are benignparticularly in patients aged between 20 and 50 years. Dermoid

TAS. Dermoid cyst in a pregnant patient. Note the echogenic nodule (arrows) and dense acoustic shadowing (arrowheads). Fig. 34.51

Fig. 34.48 TAS. Solid ovarian mass with a thickened endometrium (arrowheads) in a postmenopausal patient. Histology revealed a benign functioning thecoma.

cysts are typically complex adnexal masses with variable amounts of cystic and solid areas. They typically show areas of markedly i ncreased reflectivity and acoustic shadowing due to fat, calcification or teeth. Fat commonly floats at the top of the cyst. obscuri ng deeper structures and the true extent of the mass (Figs 34.5034.52). The echogenic nature of the cyst can also make it difficult to differentiate from bowel (Fig. 34.53), hence the size of a dermoid cyst may be underestimated with ultrasound. Approximately 25% are discovered incidentally. Management of asymptomatic dermoid cysts has changed over the last few years. Whereas previously all would have been removed, it is now considered acceptable not to operate on small (less than 5 cm), inciden-

Previous Page

CONTENTS


1083

TAS. Bilateral adnexal masses due to ovarian metastases. Note predominantly cystic mass on the right and partly solid mass on the left. Fig. 34.54

EVS. Solid-appearing dermoid cyst. Note the thick septum and two nodules (arrows and arrowheads) casting shadows. Fig. 34.52

Fig. 34.55

TAS.

Metastatic ovarian carcinoma causing omental thicken-

i ng (arrows).

EVS. Echogenic dermoid cyst (arrows). Note how the mass mimics a loop of bowel. The remainder of the ovary (arrowheads) is seen. Fig. 34.53

tally discovered dermoid cysts providing the ultrasound findings are typical and there is no growth on follow-up scans. Between 10 and 1 5% of dermoid cysts are bilateral.


Malignant germ cell tumours (dvsgernrinonuas, immature tera-

tomas) occur predominantly in young women (mean age of approxi mately 20 years). They arc usually large solid tumours but typically only stage I at presentation. They are associated with raised levels of various tumour markers, e.g. hCG, AFP, CA-I25.

Ovarian metastases ( Kruckenherg's tumours) Ovarian secondary tumours most commonly arise from primary tumours of the stomach, colon, pancreas or breast. They may be solid, cystic or complex ovarian masses, frequently bilateral and usually associated with ascites. Secondary tumours are less likely to be multiJ ocular than primary ovarian tumours but otherwise there are no specific distinguishing features (Fig. 34.54).

Differentiation of benign from malignant masses A considerable amount has been written attempting to use ultrasound to differentiate benign from malignant adnexal masses with accuracy rates ranging from 50 to 98%! Clearly the presence of metastatic disease (Figs 34.55, 34.56) indicates malignancy but in earlier disease it is difficult to be sure. Various scoring systems based on morphology and colour and spectral Doppler have been devised. A recent paper comparing the different schemes was able to show that a combination of both morphology and Doppler indices is more accurate than either used alone but there is no agreement as to which Doppler index is best and at which level the threshold should be set to distinguish

TAS. Metastatic ovarian carcinoma showing serosal tumour (arrowheads) around a loop of bowel (arrow). Fig. 34.56

between high and low impedance flow. Current research is investigating the use of ultrasound contrast media in association with 3D ultrasound in the hope that assessment of the pattern of vessels within a mass will help, but to date this is unproven. Features suggestive of malignancy are:

• Hypoechoic solid area within the mass (hig hly echogenic solid areas due to fat or calcification are typical of dermoids).

• Thick (more than 3 mm) nodular septations. • Size of mass greater than 7 cm, although very large but simple cysts are usually benign cystadenomas.

• Central rather than peripheral vascularity. • RI less than 0.6 (Fig. 34.57). RI greater than 0.8 is suggestive of benign disease but there is an indeterminate range of 0.6-0.8;

CONTENTS

Previous Page

1084

Next Page


Ultrasound has a crucial diagnostic and therapeutic role in the management of infertility. Its main uses are:

• to confirm normal pelvic anatomy • to assess ovarian morphology • to look for pelvic pathology such as endomet riosis and hydrosalpingcs

• cycle monitoring -to confirm ovulation in natural cycles -to monitor the response to ovulation induction agents, such as clomifene and Pcrgonal, and try and prevent ovarian hyperstimulation syndrome -to time hCG injections prior to assisted conception techniques such as IVF and donor insemination -to confirm development of the typical midcycle endometrium, as this has a bearing on conception rates • to guide for procedures such as cyst aspiration and oocyte collection.

TAS. Large malignant germ cell tumour in a 25-year-ofd. Doppler shows typical low impedance flow with RI of 0.5. Fig. 34.57

I n some centres this initial ultrasound examination is combined (hysterocontrastsonography) to confirm tubal with HyCoSy patency. Ovarian hvpcrstimululion svndrunte (DHSS) results in very enlarged ovaries (up to 10 cm in length) with multiple follicles. It is associated with ascites and pleural effusions and, when severe. may l ead to hypovolaemia, disseminated intravascular coagulation, venous thrombosis and even death. Mild forms are common and usually self-limiting. HvCoSv (hystcrocorttrustsonography ) is performed by inserting a small balloon catheter through the cervix into the uterine cavity. An endovaginal scan is performed while the ultrasound contrast agent Echovist is injected via the catheter. Tubal patency is confirmed when contrast is seen to flow along both fallopian tubes and around the ovary (Figs 34.59, 34.60). The technique is quite difficult to l earn but in experienced hands accuracy rates of 80-90c/( for tubal patency can he achieved. Its obvious advantage over conventional hysterosalpingography is the lack of ionising radiation. Early reports claimed it was also less painful but this has not been substantiated and it has the disadvantage of not showing detailed tuba) anatomy. Therefore its precise role is still to be determined. It is probably justified as a screening test for tubal patency in patients with a low probability of tuba[ disease.

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ EVS. Ovarian cyst with nodule in a 65-year-ofd. Doppler shows low impedance fl ow (RI 0.50) suggestive of a malignant tumour. Histology revealed a benign cystadenofibroma with a Brenner tumour. No evidence of malignancy. Fig. 34.58

however, low impedance [low can be seen with benign disease ( Fig. 34.58) and high impedance [low with malignant disease.

Screening for ovarian cancer The 5 year survival for stage I disease is over 80%, hence there has been considerable interest in screening for early tumour. Serum CA- 125 measurement, transabdominal and endovaginal ultrasound have all been investigated but found unreliable. CA-125 measurements are normal in up to 50c/ of stage I tumours, and abnormally high results are found in healthy controls and patients with endometriosis, cirrhosis and other abdominal malignancies. The difficulties in distinguishing benign from malignant masses on ultrasound also lead to a considerable number of false positives and unnecessary laparoscopies. Current recommendations are therefore that whole population screening is not justified. However, most authors agree there is benefit in screening patients known to be at increased risk of the disease, particularly those thought to have hereditary ovarian cancer.

HyCoSy. Contrast (Echovist) is seen outlining the cavity and entering the fallopian tube (arrows). Fig. 34.59

Previous Page

CONTENTS


1085

I VUs, barium studies, cystograms and sinograms are all occasionally necessary in the investigation of gynaecological patients but have been largely superseded by ultrasound, CT and MRI. Cystograms and barium enemas are used to demonstrate vaginal fistula usually due to surgery or radiotherapy for malignant disease: the investigation of cyclical rectal bleeding and pain may require a barium enema to exclude the short smooth stricture typical of endometriosis and barium studies of the small and large bowel can be used to demonstrate the typical serosal metastases of ovarian carcinoma (Fig 34.62).

Fig. 34.60

HyCoSy.

Echovist outlines a fibroid polyp (arrows) in the

uterine cavity.

The role of the plain radiograph in current gynaecological practi e s very limited. However, it is clearly important to be able to recognise features of gynaecological disease on plain films and it is still necessary to perform a plain abdominal X-ray to look for an IUCD that cannot be found with ultrasound. Abdominal radiographs may also be requested to assess bowel dilatation in postoperative patients and to exclude bowel obstruction as a cause of abdominal distension in patients with advanced ovarian cancer. Rarely direct invasion by pelvic tumours may cause bone destruction. Causes of pelvic calcification visible on a plain abdominal X-ray include:

• fibroids-typically typically coarse popcorn-type calcification • dermoid cystscommones tnest ovarian mass to calcify (teeth commonest • • • •

Hysterosalpingography (HSG) remains important in the investigation of infertility. It is an accurate means of assessing the uterine cavity and tuba) patency but has a low sensitivity for the diagnosis of pelvis adhesions so cannot replace laparoscopy. The main cut rent i ndications for HSG are infertility and recurrent miscarriage. Rare i ndications include checking the efficacy of tubal sterilisation and assessment of the tubes prior to attempted reversal of sterilisation. Technique The procedure is performed in the first half of the menstrual cycle following cessation of bleeding. The patient is asked to refrain from unprotected sexual intercourse from the date of her period until after the investigation to be certain there is no risk of pregnancy. Examination in the second half of the cycle is avoided because the thickened secretory-phase endometrium i ncreases the risk of venous intravasation, and may cause a falsepositive diagnosis of cornual occlusion. Routine use of antibiotics i s controversial but a recent consensus statement from the Royal College of Obstetrics and Gynaecology recommends that all nonpregnant women under 35 years of age undergoing uterine instrumentation should receive prophylactic antibiotics or be screened for relevant organisms. The suggested antibiotic regimen colliprises metronidazole I g rectally at the time of the procedure plus doxycycline 100 mg twice daily for 7 days. Numerous different types of cannula are available. All possess some means of preventing reflux of contrast through the cervix and i deally should allow traction on the uterus. Once the cannula is in place, water-soluhle contrast medium is injected slowly under fluoroscopic control until the uterine cavity is distended, the tubes filled and contrast is seen to spill freely from the distal ends of the tubes. Spot films should be taken during the early filling phase to ensure small filling defects are not obliterated by contrast, during early tuba] filling before the isthmic portions are obscured by contrast, and after complete filling of the tubes to demonstrate free peritoneal spill (Fig. 34.63). Additional oblique views help to demonstrate the position of the uterus and any fibroids. It is important that the uterine cavity is visualised en face. This is usually achieved by traction on the cervix, but if the uterus is retroverted it may be more effective to push the cervix so the uterine fundus tips back into the pelvis and is seen upside-down. Nowadays with modern fluoroscopic units it is also possible to angle the tube rather than manipulate the uterus. I nadequate distension of the uterus (due to cervical reflux) and tubal spasm can give rise to a false-positive diagnosis of cornual occlusion. Intravenous hyoscine butylbromide (Buscopan) or


and/or a fat-fluid level pathognomonic; Fig. 34.61) other ovarian masses-cystadenomas /carcinomas, fibromas p seudomyxoma peritonei-from rupture of a mutinous tumour fallopian tube calcification-rare, should suggest tuberculosis uterine, i.e. endometrial ossification from chronic endometritis.

Fig. 34.61 Dermoid cyst. Note calcification and teeth with a fat-fluid level (arrow).

Previous Page

1086

CONTENTS

Next Page


Fig. 34.62 Barium enemas. (A) Serosal metastases from ovarian carcinoma. (B) Short smooth stricture due to endometriosis (arrowheads). Note the puckering of the serosal surface due to adhesions (arrow).


Fig. 34.63 HSG. Normal cavity. Both tubes visible with regular mucosal folds and free peritoneal spill. Note how the contrast flows around loops of bowel (arrows).

glucagon have been suggested as treatment for this but a World Health Organization (WHO) study failed to confirm they were of any value. It is probably more important to avoid rough manipulation of the cervix and allow time for spasm to relax. Gentle traction on the uterus and change in position of the patient can also help. If there is doubt about the appearances distally, delayed films will help distinguish contrast flowing into a large hydrosalpinx from contrast spilling into the peritoneum and loculated spill. Complications of HSG include:

• Pain-due to uterotuhal distension or peritoneal spill. Minimise by slow injection of contrast and the use of isosmolar contrast agents. • infection-rare, but more frequent in patients with a past history of pelvic intlamiiiatory disease and hydrosalpinges.

Fig. 34.64 HSG. Venous intravasation. The myometrial plexus is shown with drainage into the ovarian veins (arrowheads). Peritoneal spill is also seen.

• Vasovagal reactions-usually from manipulation of the cervix or i nflation of an occlusion balloon in the cervical canal. of no clinical significance but can make i nterpretation of the images difficult. It occurs more corm only i n the presence of fibroids or tubal obstruction (Fig. 34.64). • Allergic reaction to contrast media-very rare.

• Venous intavasation

Congenital uterine abnormalities The uterus develops by fusion of the paired mullerian duct systems. Complete or partial failure of fusion is estimated to occur in 3-4% of the general population. The range of resulting abnormalities is shown in Fig. 34.65.

Previous Page

CONTENTS


Fig. 34.67

1087

HSG. Bifid uterine cavity. Impossible to be sure if this is bicor-

nuate or septate.

Fig. 34.65

I mportant congenital abnormalities of the uterus.

Minor degrees of abnormality are of no clinical significance; howeyer, there is an increased incidence of recurrent miscarriage in patients with a septate uterus, so differentiation from a bicornuate uterus is important. On HSG an angle of >90 0 between the horns is suggestive of a bicornuate uterus but definite differentiation from a septate uterus requires further investigation to look for a fundal


notch. Truly unicornuate uteri arc rare, so if an apparently unicornuate uterus is demonstrated on HSG care should be taken to look for a rudimentary horn or second cervix (Figs 34.66-34.68).

Fig. 34.68

HSG. Unicornuate uterus. No evidence of a rudimentary horn.

Other uterine abnormalities Filling defects in the uterus are

caused by:

• fibroids (Figs 34.69-34.71) • polyps / endometrial hyperplasia • intrauterine adhesions, caused by dilatation and curettage ( D&C), tuberculosis or following exposure to diethyloestradiol ( DES) • pregnancy.

HSG. Uterus bicornis bicollis. Note the completely separate cervical canals and uterine horns, both of which have patent tubes. Fig. 34.66

The effect of fibroids on an HSG depends on their position within the uterus. Subserosal fibroids may cause displacement of the cavity but are otherwise undetectable; mural fibroids enlarge the cavity and may or may not cause distortion; submucous fibroids appear as polypoid filling defects within the uterine cavity, indistinguishable from endometrial polyps. Early-filling films are necessary to demonstrate small fibroids and oblique views are helpful in confirming their exact location.

Previous Page

1088

Fig. 34.69

CONTENTS


HSG. Filling defect due to submucous fibroid.

Fig. 34.71

Fig. 34.72

HSG. Polypoid endometrium causing multiple filling defects (arrows) only seen on the early filling film.


HSG. Cavity and right fallopian tube being distorted by large mural fibroid. Note small calcified fibroid on the left (arrows). Fig. 34.70

Next Page

HSG. Cavity enlarged by fibroids.

Multiple small filling defects causing an irregular lobulated outline to the uterine cavity are also seen with endometrial hyperplasia, for example in patients with polycystic ovary syndrome (Fig. 34.72). Intrauterine synechiae/adhesions cause linear or irregularly-shaped filling defects that are not obliterated by increasing amounts of contrast (Fig. 34.73). They are associated with recurrent miscarriage and.

Fig. 34.73

HSG. Linear filling defects in the uterine cavity due to adhesions.

i f extensive, they can completely obliterate the uterine cavity, causing Asherman's syndrome (Fig. 34.74). A sherman's syndrome is amenorrhoea due to intrauterine synechiae, usually caused by dilatation and curettage for postpartum haemorrhage or retained products of conception. Rarely it follows a normal pregnancy. Treatment comprises hysteroscopic resection of the adhesions and insertion of an intrauterine device to separate the walls of the cavity. I ntrauterine adhesions and small irregularly-shaped cavities are also seen in patients with chronic endometritis due to tuberculosis (Fig. 34.75). Genital Tuberculosis primarily affects the fallopian tubes and 50r// of patients with tubal disease will have a uterine abnormality. Tubal tuberculosis leads to a rigid abnormal tube with occlusion in the isthmus. The ends are frequently clubbed and there are diverticula-like projections from the tubal surface. Tubal (and very rarely ovarian) calcification can be seen. A small irregular T-shaped cavity, with constrictions around the body, is also associated with c.sposure to DES, a drug that was used from 1940 to 1960 to treat recurrent miscarriage. It resulted in a range of genital abnormalities in the daughters of treated mothers and is associated with increased incidence of subfertility, ectopic pregnancy and pregnancy loss, as well as an increased incidence of clear cell carcinoma of the vagina.

Previous Page

CONTENTS


Fig. 34.74

HSG. Severe Asherman's syndrome with complete oblitera-

tion of the uterine cavity.

1089

Fig. 34.76 HSG. Enlarged uterus with multiple diverticular-like projections of contrast into the myometrium typical of adenomyosis.


HSG. Normal right tube but a large left hydrosalpinx. Note the mucosal folds on the left have been obliterated and there is no distal spill. Fig. 34.77

Fig. 34.75

HSG. Tuberculous endometritis leading to an irregular T-

shaped uterine cavity.

A denomvosis (endometriosis interna) is a rare cause of subfertility: it can be recognised by diverticula-like projections of contrast from the cavity into the myometrium, usually in association with an enlarged uterus (Fig. 34.76).

Tubal disease Pelvic inflammatory disease is the most common cause of both distal and proximal tLibal occlusion. Preservation of mucosal folds within a hydrosalpinx is said to be associated with a good response to tubal surgery; however, nowadays patients are frequently referred directly for IVF rather than being considered for tubal surgery. Peritubal adhesions cannot be identified reliably on HSG but their presence can be inferred if contrast remains l oculated around the tube instead of spreading freely in the peritoneum, and if the tube looks angulated or distorted. Delayed i mages may be of value in determining this (Fig. 34.77). Other causes of tubal occlusion are endometriosis, postabortal or postpuerperal infection and tuberculosis. Salpingitis isthrnica nodosa (Fig. 34.78) is characterised by multiple diverticula-like collections of contrast projecting from the tubal lumen. It is usually due to pelvic inflammatory disease or endometriosis and is associated with an increased incidence of subfertility and ectopic pregnancy.

HSG. Salpingitis isthmica nodosa (arrows). The right tube is patent but the left tube is very irregular and beaded and terminates in a hydrosalpinx (arrowhead). Fig. 34.78

Cornual polyps are occasionally seen as tiny filling defects at the cornua but they rarely cause obstruction and are of questionable clinical significance.

Previous Page

1090

CONTENTS

Next Page


Selective fallopian tube catheterisation and recanalisation

Approximately 50% of proximal tuba] occlusions have been shown to be due to spasm, amorphous debris and fine adhesions, rather than to a true histological occlusion, and these are amenable to treatment by selective fallopian tube catheterisation.

400 ChM before tubal filling has been associated with infertility and a poor response to tuba) recanalisation.

Following a conventional HSG a 5F catheter is manipulated into the origin of the tube. Contrast is injected to confirm tubal occlusion and then a guide-wire (0.35 in) is used to probe the tube and hopefully dislodge the obstruction. Success rates for tuba] recanalisation of 70-80%%r have been achieved, with subsequent pregnancy rates of 1 0-45%. This compares favourably with success rates achieved by I VF. The technique is popular in the USA but has been slow to gain acceptance in the UK, most gynaecologists preferring to send their patients with tuba] disease directly for IVF (Figs 34.79, 34.80). Selective tubal catheterisation also allows measurement of tubal filling pressures. A pressure transducer is connected between the catheter and an injection pump. Once the catheter tip is positioned i n the cornea, contrast is injected at a constant rate and the backpressure monitored until tuba] filling is achieved. Normal tubes fill with pressures less than 200 cm H2O. A pressure rise of greater than

Fig. 34.79

The role of CT in the evaluation of gynaecological diseases in the pelvis has declined since the advent of endovaginal scanning and MR1. As a general rule, benign disease should be investigated initially by ultrasound and then MRI, rather than CT, which is used to solve specific problems. Staging of malignant disease requires CT or MRI, depending on the site of the primary tumour. MRI is superior to CT for staging cervical and uterine carcinoma, particularly with respect to local disease, but CT still has a role in ovarian carcinoma because of its ability to detect peritoneal deposits. Currently CT and MRI have similar capabilities for detecting lymphadenopathy, although the use of different imaging planes and development of specific contrast suggest that MRI will eventually prove to be more accurate. However, CT is frequently used as an i maging modality in patients with non-specific lower abdominal symptoms such as pain, or to determine the site of origin of a mass, so it is clearly necessary to be aware of the CT appearances of gynaecological conditions. CT of the pelvis is performed following opacification of the small bowel with oral contrast (given at least an hour before the scan) and with a moderately full bladder. Some institutions also use a vaginal tampon and rectal contrast. Intravenous contrast enhancement helps to distinguish lymphadenopathy from pelvic vessels, so i s mandatory when CT is performed as a staging investigation. CT anatomy of the pelvis The vagina is identified as a thin rectangular structure with a brightly enhancing mucosa. The ceryix and uterus appear as soft-tissue masses, only distinguishable from each other by their shape, i.e. the cervix is round anti the uterus oval or triangular in cross-section. The endometrium is not easily distinguished from the myometrium but distension of the cavity by fluid is discernible, particularly following enhancement with


HSG. Bilateral cornual occusions.

i ntravenous contrast. The ovaries can usually be seen in adult premenopausal patients as soft-tissue masses postcrolateral to the uterus; however, their position is variable and their precise appearance depends on whether or not there are cysts/follicles present. The atrophic ovaries of post menopausal patients are frequently indistinguishable from surrounding structures. Uterine fibroids Fibroids typically cause an enlarged lobulated

uterus. They are usually isocchoic with the myometrium and show similar enhancement following intravenous contrast administration; however, they may contain calcification (up to 10%) or areas of reduced attenuation due to degeneration or necrosis. High-attenuation areas are also seen due to haemorrhage. Pedunculated fibroids may be difficult to distinguish from an adnexal mass and submucous fibroids may expand the cavity, mimicking an endometrial carcinoma. There are no specific features to differentiate fibroids from adenomyomas or other rare myometrial tumours, such as metastases or leiomyosarcomas (Figs 34.81, 34.82).

Selective salpingography. 5F catheter and wire manipulated into the right uterine cornua. Subsequent injection of contrast shows tubal patency with free peritoneal spill. Fig. 34.80

Endometrial carcinoma Endometrial carcinoma typically causes a hypodensc, irregular mass expanding the uterine cavity, some-

Previous Page

CONTENTS

Next Page

CT. Small adnexal cyst (arrow) in postmenopausal patient. EVS confirmed a small simple cyst. Fig. 34.83

Fig. 34.81 CT. Bulky uterus with low-density areas due to fibroids. One small fleck of calcification.

CT. Bulky mildly heterogeneous uterus with posterior displacement of the cavity due to adenomyosis rather than fibroids. Fig. 34.82

CT. Complex mass in the pelvis typical of a dermoid cyst (arrows). The mass is of mixed attenuation but contains a large amount of fat. It has a calcified rim and a dense area of calcification (arrowheads) i nferolaterally due to a tooth. Fig. 34.84


ti mes associated with blood, fluid or pus within the cavity. CT is good at determining the extent of extrauterine disease but cannot easily differentiate stage I from stage 2 disease, so formal staging

ureters. Short-axis measurements of greater than 0.8 cm in the

is far better performed with MRI.

Cervical carcinoma is suggested by the presence of an enlarged irregular cervix. The tumour enhances less than the surrounding normal cervical stroma, forming a relatively hypodense area following intravenous contrast. Fluid may be seen within a distended uterine cavity if the cervical canal is obstructed by tumour. Parametrial invasion is demonstrated by loss of clarity of the cervical margins, with an eccentric soft-tissue mass and stranding into the paracervical fat. However, minor soft-tissue stranding should be interpreted with caution as inflammatory changes and oedema following dilatation and curettage or cone biopsy can give similar appearances. Loss of the fat planes between the cervix and the ureters, rectum and/or bladder is indicative of advanced disease (i.e. stage 3 or greater). There is also sonic correlation between cervical size and prognosis-a maximum anteroposterior depth greater than 4 cm is associated with a significantly higher incidence of nodal metastases and a poor prognosis. Evaluation of a patient with cervical carcinoma must also include assessment of the iliac and para-aortic lymph nodes and both kidneys because of the propensity of the tumour to invade the Cervical carcinoma

common iliac nodes is said to indicate a high likelihood of metastasta disease and this is particularly true if the nodes are relatively hypoechoic following contrast enhancement, i.e. they possess the same enhancement characteristics as the primary tumour. Adnexal masses Functional ovarian cysts are visible as thinwalled low-attenuation masses within the ovaries (Fig. 34.83). Pain can be due to rupture, haemorrhage or torsion. On CT haemorrhage may be recognised by the presence of high-attenuation fluid within a cyst. Acute torsion causes severe pain and most patients proceed to laparoscopy without CT; however, subacute torsion can be difficult to diagnose clinically. CT findings include deviation of the uterus to the side of the torsion and engorgement of adjacent blood vessels around a non-enhancing mass (the enlarged ischaemic ovary). Apparently simple functional cysts should be further evaluated with ultrasound to confirm their benign and transient nature. The cause of complex (i.e. partly cystic and partly solid) adnexal masses is difficult to determine on CT. A dermoid cyst can be diagnosed with confidence if fat and calcification are seen within it (Fig. 34.84) (approximately 90%) but otherwise it is difficult to distinguish neoplastic from inflammatory masses. The same criteria

Previous Page

1092

CONTENTS

Next Page

A TEXTBOOK OF RADIOLOGY AND IMAGING i noperable metastatic disease, e.g. liver metastases, for whom surgery may have little to offer.

Angiography Arteriography as a purely diagnostic procedure is

rarely indicated nowadays. However, embolisation of both uterine arteries can be performed safely with a high success rate and low complication rate, hence it can be used to treat intractable bleedi ng, such as may occur following childbirth, from pelvic tumours, following radiotherapy, following trauma and from the very rare

Fig. 34.85

CT. Bilateral cystic adnexal masses. Note the tiny gas bubble (arrow) seen in one of the masses. Laparoscopy confirmed bilateral hydrosalpinges with chronic infection.

are applied to adncxal masses on CT as for ultrasound, i.e. a complex mass with thick irregular septations and wall nodularity is strongly suggestive of a carcinoma. Rarely gas is seen within a mass and indicates a pelvic abscess (Fig. 34.85). Calcification may be seen in a pcdunculated fibroid or a solid ovarian tumour such as a fibroma. Endometriotic cysts have a very variable appearance on CT, with areas of high attenuation within them due to haemorrhage; however, both sensitivity and specificity for endometriosis are low on CT and quite extensive peritoneal implants can be missed.

CT does, however, retain a role in staging of ovarian carcinoma, l argely because of its ability to detect peritoneal and serosal deposits in addition to liver and nodal metastases. Typically there is diffuse thickening of the omentum-so-called omental cake with nodular thickening of the peritoneum and serosal surface of the bowel (Fig. 34.86). Most ovarian carcinomas are already advanced at the time of presentation but surgery is still helpful to reduce tumour bulk. Preoperative staging of ovarian carcinoma is therefore performed to assess tumour hulk and detect those patients with

uterine arteriovenous malformations. A rteriovenous malformations are suspected from a history of recurrent extremely heavy bleeds which usually follow a miscarriage or uterine intervention, such as dilatation and curettage or evacuation of retained products of conception. Colour Doppler ultrasound demonstrates an area of abnormal vascularity in the uterus with high flow rates and low i mpedance flow on spectral Doppler (Fig. 34.87). More recently transcatheter embolisation of the uterine arteries has been advocated as a means of treating large fibroids in patients who wish to avoid hysterectomy. Embolisation of both uterine arteri es is performed using polyvinyl alcohol (PVA) particles, with or without gel foam. Antibiotic cover is given, with sedation and analgesia as required. Results to date show a reduction in uterine sire of 40-70% and a reduction in abnormal menstrual bleeding in 8090% of patients. The complication rate is low, the most common complication being pain due to ischaemic necrosis. Postprocedural infection occurs in 1-2%. There are reports of successful pregnancy following embolisation but amenorrhoea, thought to be due to ovarian failure, has also been reported, so the technique should be considered very carefully in patients who are trying to conceive.


Percutaneous aspiration and drainage Follicular aspiration for

oocyte retrieval is routinely performed with ultrasound guidance but there is reluctance to aspirate cysts because of fear of seeding malignant cells in to the peritoneal cavity. However, cysts that are

Patient with a history of recurrent very heavy vaginal bleedi ng. Selective internal iliac artery injection shows an abnormal stellate coll ection of vessels on the right. Embolisation was performed, with good symptomatic relief. Fig. 34.87

Fig. 34.86

CT. Patient with ovarian carcinoma, peritoneal deposits (white

arrowheads), para-aortic lymphadenopathy (black arrowheads) and ' omental cake' (black arrows).

Previous Page

CONTENTS

Next Page

likely to be benign can be safely aspirated-either transabdominally across a distended bladder or endovaginally. This is effective treatment for pain caused by functional cysts or adhesion-related peritoneal cysts, with a risk of recurrence of 15-20%. Similarly patients with large tubo-ovarian abscesses may benefit from aspiration or drainage of the pus, the preferred route dependingon the location of the abscess. Ultrasound guidance can also help to guide drainage of an obstructed uterus.

Jeremy P. R. Jenkins MRI is an important technique in the evaluation of pelvic pathology due to its ability to obtain images with a high soft-tissue contrast resolution and discrimination in multiple planes. It is now the primary technique of choice in the staging of pelvic malignancy, with the exception of staging ovarian malignancy, where CT is the preferred technique. Advantages of MRI in the pelvis have been discussed in Chapter 3I .

Both sagittal and transverse imaging planes are required when assessing the uterine body, cervix and vagina. T,-weighted scans display the characteristic zonal anatomy, with three distinct areas within the uterine body ( Fig. 34.88). There is a hyperintense central zone representing the endometrimn combined with secretions in the endomctrial canal, an outer area of intermediate signal due to myometrium, and a low-signal functional zone between, from a layer of compressed myometrium. Changes in size and signal inten-

Fig. 34.89 Normal zonal anatomy of the cervix on a transverse T,-weighted spin-echo (TSE 3500/100) image. b = bladder; s = cervical stroma; straight arrow = cervical mucus and epithelium.

sity of the three zones in the uterus occur during the menstrual cycle, with an increase in volume and signal of the myometrium in the secretory phase. Following intravenous gadolinium-chelate the endometrium and myometrium enhance, with the functional zone remaining low-signal on T,-weighted images. cervix The is a cylindrical-shaped structure measuring 2-4 cm in l ength, connecting with the body of the uterus at the isthmus. The l evel of the isthmus is approximately at the peritoneal reflection on the bladder. The cervix has two distinct layers: a hyperintense central zone representing cervical mucus and epithelium, with an outer zone of low signal, similar to the uterine functional zone, due to the fibrostromal wall (Fig. 34.89). A further peripheral layer of i ntermediate signal may be seen continuous with the myomeu i um. The paramerrium has an intermediate signal on T,-weighted images, with increase in signal on T-weighted scans. After intravenous gadolinium-chelate the compact cervical stroma retains its low signal, with enhancement of the paracervical tissue and inner cervical epithelium on T,-weighted images. There are numerous glands lining the cervical canal and the ducts of these glands can become blocked, producing retention (nabothian) cysts (Fig. 34.90). These are commonly seen on MRI of the female pelvis. The vagina can be identified as a high-signal central zone of mucus and epithelium surrounded by a low-signal muscular wall (Fig. 34.88). The vagina cm he divided into three regions: an upper third is characterised by the lateral vaginal fornices. a middle third i s at the level of the bladder base, and a lower third is at the level of the urethra. There is a high-signal venous plexus surrounding the cervix and vagina, best seen on transverse T,-weighted images ( Fig. 34.90). Intrauterine contraceptive devices (IUCDs) are safely imaged with MRI, with no adverse heating or torque effects demonstrated. All IUCDs show a signal void, the extent depending on the type of device in situ, on all pulse sequences (Fig. 34.91). A contraceptive diaphragm will, however, produce a signilleant signal artefact.

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ Fig. 34.88 Normal uterus on a sagittal T 2 -weighted spin-echo (TSE 3500/100) image with normal zonal anatomy of central high-signal endometrium (e), the junctional zone (j) and the outer myometrium (m). urine-filled bladder.

Previous Page

1094

CONTENTS

Next Page


Fig. 34.90

Nabothian cysts (curved arrow) within the cervix on a fatsuppressed transverse T 2 - weighted spin-echo (FSE 3500/100) i mage.


Fig. 34.91 An IUCD (arrow) within the uterus (u) on a transverse fatsuppressed T 2- weighted spin-echo (FSE 3500/100) image. f = free fluid in the sacral cul de sac.

The upper two-thirds of the vagina, cervix, uterus and fallopian tubes develop from fusion and descent of paired rudllerian ducts. The lower third of the vagina is derived from the urogenital sinus. Partial or complete failure of the ducts to fuse results in a spectrum of complex abnormalities. MRI is the technique of choice in assessment and evaluation of these congenital lesions. MRI can demonstrate unicornuate, bicornuate and septate uteri, and uterine didelphys ( Fig. 34.92). A uterine didelphys is one in which two separate uteri and cervices are visualized. A septate uterus is one in which the uterine septum fails to resorb, which results in failure of correct placental implantation and subsequent miscarriage. MRI is able, unlike other imaging techniques, to differentiate a septate from a bicornuate uterus. The importance of making this distinction is that the surgical approach for treating the two anomalies is dif-

Fig. 34.92 (A,B) Uterine didelphys showing two separate uteri and cervices (arrows) are demonstrated on transverse T 2 - weighted spin-echo (FSE 3500/100) i mages. Note the normal left ovary (o).

ferent: a bicornuate uterus requires abdominal surgery, whereas the septate uterus can be repaired hysteroscopically. MRI, as with other i maging methods, can be used to assess any renal tract abnormalities coexistent with these miillerian duct anomalies.

Leiomyoma This is the most common solid uterine tumour, hcing single or multiple. These tumours are composed of smooth muscle with varying amounts of fibrous tissue, and occur in 20-30% of premenopausal women. They are located in the submucosal, intramural and subserosal spaces of the uterus. Rarely they can occur along the broad ligament or be entirely separate f ront the uterus. Submucosal tumours project into the endometrial

Previous Page

CONTENTS


Multiple leiomyomas (I) on (A) sagittal and (B) T2 -weightec i mages with a large cervical intramural leiomy oma (c) and smaller intramural tumours (e). There is a degenerating serosa eiomyoma (d) and a smaller serosal leiomyoma (s) adjacent to a loculatec cystic collection due to an associated hydrosalpinx (h). b = bladder. Fig. 34.93

1095

Fig. 34.94 Endometrioma deposit (arrow) within the bladder wall on (A) coronal T,-weighted (SE 650/25) and (B) coronal T 2 -weighted spin-echo (TSE 3500/100) i mages. Note the areas of high signal within the superficial margin of the deposit due to haemorrhage in (A), and the low signal on the T 2 - weighted in (B). b = bladder; u = uterus.


spin-echo (FSE 3500/100)

cavity, and intramural lesions arise within the myometrium. Subserosal leiomyomas occur along the serosal surface of the uterus. As the tumours are oestrogen-dependent they can grow rapidly dduringgpregnancy, and tend to regress following g g the gp y' menopause. MRI provides an accurate assessment of the site, size and number of uterine leiomyomas, with lesions as small as 3 mm diameter being detected. Leiomyomas are classified according to their position-submucosal, intramural, subserosal or cervical (Fig. 34.93). Non-degenerating leiomyomas have a characteristic uniform signal intensity, being indistinguishable from myometrium onT1-weighted i mages, with a lower signal on T,-weighted scans ( Fig. 34.93A). Occasionally calcification within these tumours produces a low signal on all pulse sequences. Degenerating tumours show a variable and non-specific signal appearance with an i ntermediate-high signal on T,-weighted and a high signal on T2weighted images (Fig. 34.93B). Malignant transformation cannot be differentiated from benign degenerating tumours. The use of intravenous gadolinium-chelate does not improve the detection rate orcharacterisation of leiomyomas. While the assessment of leiomyomas is usually by clinical examination and ultrasound, these techniques can be limited in the

presence of a retroverted or displaced uterus. In addition, difficulty can be encountered in discriniinatin90%, which reduces to 50-60% or less for those with bulky disease. This invasive cervical

carcinoma is better demonstrated on MRI than CT and appears as an area of high signal contrasted against the low-signal cervical stronta on T,-weighted images. The presence of a low-signal stromal ring around the high-signal tumour is good evidence (93-100% sensitivity) of a confined tumour. The absence of a lowsignal stromal hand, however, is not certain evidence of parametrial spread due to limitation in spatial resolution using current MRI

techniques. The use of high-resolution MRI, as indicated above, is likely to overcome this limitation. The cervix may enlarge, leading

to obstruction of the endometrial canal with distension of the uterus from retained secretions (Fig. 34.99). I n FIGO stage II disease the tumour extends beyond the cervix. In stage IIA the tumour extends into the upper two-thirds of the vagina but not into the parametrium (Fig. 34.100), and stage IIB

disease extends into the parametrium but not to the pelvic side-wall (Fig. 34.10I ). This distinction is critical, as most institutions treat l esions above stage IIA disease by radiation therapy. The reported accuracy of MRI in the demonstration of parametrial or vaginal spread is approximately 70-9014. On MRI T,-weighted images

extension into the parametrium is identified by the high-signal

Fig. 34.98 Carcinoma of the cervix (stage 1132) showing exophytic tumour (arrows) within the vaginal canal on sagittal T2 -weighted spin-echo (TSE 3500/100) i mage. Note the intact low-signal vaginal wall. b = bladder.

tumour breaching the low-signal cervical stromal wall on transverse, or transverse-oblique sections parallelling the short axis of the cervix. False-positive results on MRI are due to surrounding

Previous Page

109 8

CONTENTS

Next Page


Bulky exophytic cervical carcinoma confined within the cervix on a sagittal T2 - weighted spin-echo (TSE 5041/132) i mage. The lowsignal vaginal wall remains intact apart from an area of tumour infiltration posteriorly (arrow). Within the tumour in the anterior fornix there is an area of necrosis. b = bladder. (Courtesy of Dr R. J. Johnson, Christie Hospital.) Fig. 34.100

Fig. 34.102 Bulky carcinoma of the cervix extending into the body of the uterus, lower third of the vagina (small arrows) and bladder (open arrow) on a sagittal T2 - weighted spin-echo (TSE 5041/132) i mage. (Courtesy of Dr R. J. Johnson, Christie Hospital.)


Tumour (t) infiltrating the parametrium and left iliacus muscle (i), with left-sided involved lymph nodes (n) and a right ovarian metastasis (o) on a transverse T 2 - weighted spin-echo (TSE 3500/100) i mage. b = bladder. Fig. 34.103

Carcinoma of the cervix (stage I I B) showing tumour (t) within the parametrium (arrows) on transverse T 2- weighted spin-echo (ESE 3500/100) i mage. Note the loss of the normal low signal from the cervical stroma. b = bladder. Fig. 34.101

oedema, vascular parametrium or an inflammatory reaction to the tumour. Vaginal extension is indicated on MRI by high-signal tumour replacing the normal low-signal vaginal wall (Fig. 34.100). Overstaging of disease occurs particularly from large exophytic tumours in the region of the anterior fornix. As direct inspection of vaginal infiltration can be easily performed, this assessment on MRI is not critical. In FIGO stage III disease there is extension into the lower third of the vagina (stage IIIA) (Fig. 34.102), or to the pelvic side-wall with or without hydronephrosis, or a non-functioning kidney due to no known

other cause (stage IIIB) (Fig. 34.103). MRI criteria for pelvic sidewall invasion includes tumour within I cm of the muscles of the pelvic wall, vascular encasement or high-signal tumour replacement of low-signal adjacent muscles (levator am, piriformis, obturator internus) (Figs 34.103, 34.104). Overstaging of tumour can again occur due to surrounding oedema or inflammatory change. I n stage IV disease the tumour extends outside the reproductive tract, with tumour involvement of the mucosa of the rectum or bladder (stage IVA) (Fig. 34.105), or disease outside the true pelvis or distant metastases (stage IVB). Sagittal and transverse T2-weighted scans allow assessment of tumour infiltration into the lower uterine segment, bladder, rectum and vagina, where high-signal tumour replaces the normal low-signal structures (Figs 34.106, 34.107). The use of dynamic contrast enhancement allows a clearer assessment of

Previous Page

CONTENTS

Next Page

Fig. 34.106 Carcinoma of the cervix (t) infiltrating through the bladder wall (arrow) on a transverse T2-weighted spin-echo (TSE 3500/100) i mage. b = bladder.

tumour extension into the bladder or rectal wall, compared with T_,-weighted images. Although /sinph node status is not part of the FIGO staging, the presence and extent of lymphadenopathy has important implications in treatment and prognosis. The presence or absence of pelvic lymphadenopathy is assessed on T,-weighted images, with nodes greater than 7-10 mm in diameter being considered abnormal (see

will increase the accuracy in the detection of nodal disease irrespective of size or anatomical distribution. The treatment and prognosis of invasive cervical carcinoma is dependent on tumour volume, extent of disease, histological grade, depth of invasion and vascular and lymphatic involvement. Following histological diagnosis accurate staging is important in order to optimise treatment-demonstration of parametrial extension often precludes surgery and requires radiotherapy (Fig. 34.108). Clinical staging is often inaccurate, particularly in those with more advanced disease-two-thirds of these patients may be incorrectly staged. Imprecision in clinical staging is due to errors in assessment of size of endophytic tumours, uterine body extension, infiltration of parametria and pelvic side-wall and lymphadenopathy. X-ray, CT and ultrasound (including transrectal ultrasound) are used for staging and assessment of tumour volume but are limited due to poor tissue discrimination and difficulty in delineating adjacent organ involvement. MRI has an overall accuracy range in staging of approximately 78-92`7, with improved accuracy in more advanced tumours. The accuracy of MRI for demonstrating extent of tumour invasion of the pelvic side-wall is over 90%, with a similar value for detecting bladder and rectal wall i nvolvement. Good concordance has been achieved with tumour volume measurements on MRI when compared with data obtained by histological review. The advantages that MRI has over other i maging techniques include an improved tissue resolution and discrimination combined with a multiplanar facility, particularly with the use of phased-array coils, fast acquisition sequences with facility for fat suppression, and dynamic contrast enhancement methods. An advantage of dynamic contrast enhancement is in the assessment of neoangiogenesis of the tumour, which has a direct relationship to tumour growth rate and necrosis, but is not routinely used in staging tumours. T-weighted images remain superior in the evaluation of parametrial spread as the normal parametrium is well vascul arised. The detailed imaging technique is outlined in the Royal College of Radiologists' booklet ` A Guide to the Practical Use of

Ch. 31) (Fig. 34.103). Similar signal intensity appearances are, however, obtained from hyperplastic and nmetastatic nodes. A new MRI lymphographic contrast is now available using ultasmall iron oxide particles (USIOP), which accumulates in normal nodes, producing a signal void while sparing metastatic nodes, which retain their abnormal signal (see Chapters 2, 59). The use of this agent

MRI in Oncology'. MRI is indicated in patients with tumours greater than 2 cm diameter on clinical examination, when the tumour is primarily in the endocervical canal, or if it is of an infiltrative type. MRI can be of particular value in assessing pregnant patients with invasive cervical carcinoma and in detecting concomitant uterine disease, e.g. Iciomy-

Cervical carcinoma (t) extending into the parametrium producing left hydronephrosis (straight arrow) and extending posteriorly through the perirectal fascia into the rectal mucosa (curved arrow) on a transverse T 2 - weighted spin-echo (TSE 5041/132) i mage. b = bladder (Courtesy of Dr R.J. Johnson, Christie Hospital.) Fig. 34.104


Recurrent carcinoma of the cervix (t) infiltrating into the parametrium and right levator ani on a transverse T 2- weighted spin-echo (TSE 5041/132) i mage. Note the low-signal vaginal wall (arrow). b = bladder, r =rectum. (Courtesy of Dr R. J. Johnson, Christie Hospital.) Fig. 34.105

Previous Page

CONTENTS

Next Page

Fig. 34.107 Extensive recurrent cervical carcinoma (t) following a hysterectomy on (A) selected transverse T 2 -weighted spin-echo (TSE 3500/100) i mages, and (B) coronal T 2- weighted spin-echo (TSE 3500/100) fat-suppressed image. There is a large tumour recurrence infiltrating through the lower two-thirds of the vagina into the pelvic floor and in the bladder wall (open arrow), with separate tumour nodules in the right ischial rectal fossa (straight arrow) and left parametrium. There is bilateral inguinal lymphadenopathy (small arrows). Note the mass of higher signal than tumour in the left adnexa from the native ovary. b = bladder.


Fig. 34.108 Carcinoma of the cervix (straight arrows) on sagittal T 2- weighted spin echo (SE 1500/80) (A) before treatment, (B) 6 weeks, and (C) 6 months after radiotherapy. Note the rapid reduction in size of the tumour between (A) and (B). The small area of high signal in the cervix in (C) is due to either residual tumour or post-treatment change. Note the low-signal area in the uterus due to a non-degenerating leiomyoma (curved arrow), and the high mucosal signal in the posterior wall of the bladder (b) from radiotherapy change in (B) and (C).

omas (Fig. 34.108). MRI can be used in the diagnosis of recurrent disease and in aiding the distinction of tumour from post-treatment change (Fig. 34.109). Tumour recurrence i s indicated by the presence of a soft-tissue mass which, on T,-weighted images, exhibits a high signal compared with muscle and fat. Improved conspicuity of disease can be achieved using fat-suppressed sequences. Although longstanding post-treatment fibrosis can be of low signal on all pulse sequences, signal appearances otherwise may overlap with those obtained from tumour. Less than 6 months after radiotherapy both recurrent tumour and fibrosis exhibit hyperintensity on T, -weighted

i mages, making distinction difficult. Six to 12 months post-radiotherapy the radiation fibrosis becomes lower in signal. The use of dynamic contrast enhancement with integration of signal intensity-time curves may allow separation of post-treatment fibrosis from recurrent tumour-the tumour neoangiogencsis allows a more rapid contrast uptake than in radiation fibrosis. Significant overlap in measured T, values has been demonstrated between normal cervical tissue and tumour, although a reduction in tumour T, has been noted in cervical carcinomas following radiotherapy.

Previous Page

CONTENTS

Extensive radiation change involving the bladder, vagina, rectum and bowel loops on (A) sagittal T2 - weighted (SE 1 500/80) and (B) transverse T,-weighted (SE 800/40) i mages. In (A) the bladder (b) has a thickened wall with a high-signal-intensity mucosa around the posterior wall, The uterus (u) is enlarged and the vagina (arrows), rectosigmoid (r), and adjacent small bowel loops (I) show thickened walls with high signal. No evidence of recurrence of cervical carcinoma, which was confirmed on histological review. The high signal from the sacrum and L5 vertebra is due to radiation-induced fatty infiltration of the marrow spaces. The area of signal void within the vagina in (A) i s due to a tampon in situ. Fig. 34.109

Next Page

Endometrial carcinoma is a common gynaecological malignancy, being the most prevalent invasive malignancy of the female genital tract in the USA. In 1998, 36 100 new cases of endometrial carcinoma were diagnosed, with approximately a sixth of patients dying from their disease. There has been a threefold increase in the incidence of endometrial carcinoma over the last 30 years, with a strong link to long-term oestrogen exposure without opposed progesterone. Risk factors include obesity, diabetes mellitus, hypertension, multiparity, late-onset menopause, polycystic ovaries, and the l ong-term use of tamoxifen for the treatment of breast cancer. In the UK this tumour is second in prevalence to ovarian malignancy. Endometrial carcinoma may be localised or diffuse and mainly occurs in postmenopausal women. Over 90% present with postmenopausal bleeding. Approximately 90%% of tumours are well differentiated adenocarcinomas arising within the uterine epithelium. Localised tumours are polypoidal with a superficial attachment to the endometrium, whereas diffuse l esions infiltrate the entire endometrium and invade the myometrium, spreading beyond the uterus and cervix to involve adjacent organs. The depth of infiltration of the myometrium relates to the presence of nodal metastases. Only a few per cent of patients have nodal involvement with superficial invasion, increasing to approximately 40 1Z( for deep myometrial infiltration. The detailed imaging technique is outlined in the Royal College of Radiologists' booklet A Guide to the Practical Use of MRI in Oncology. On MRI endometrial carcinoma shows a signal intensity appearance similar to normal endometrium, which can cause difficulty in defining small lesions. Large lesions expand the endometrial cavity (Figs 34.110, 34.111 ) and can have a low signal on T-weighted images. Widening or signal heterogeneity on T,weighted images within the endometrial canal may be the only abnormal finding in early-stage disease (Fig. 34.116). Blood clot, adenomatous hyperplasia and degenerating submucosal leiomyoma can produce similar changes, making histological review essential. The most reliable criterion for the diagnosis of myometrial invasion is disruption of the junctional zone. Difficulty can occur in this


Fig. 34.110

Stage IC endometrial carcinoma (e) on a sagittal T 2 3500/100) image. b = bladder. -weightdspnco(TSE

Fig. 34.111 Endometrial carcinoma (e) on a sagittal T2 -weighted spinecho (TSE 3500/100) i mage with associated lymph node involvement (demonstrated on other sections) making this a stage I I I C tumour. b = bladder.

CONTENTS

Previous Page

1102


Box 34.2

FIGO/TNM staging for uterine corpus TNM

FIGO IA

IB IC IIA II B III A

a

IIIB III C IVA

IVB

Next Page

Tumour confined to endometrium I nvasion < half depth of myometrium I nvasion > half depth of myometrium Extension to cervix, endocervical glands only I nvasion of cervical stroma I nvasion of uterine serosa/adnexae/+ve peritoneal cytology Vaginal involvement (direct spread or metastasis) Pelvic and/or para-aortic lymph node metastases Invasion of bladder/bowel Distant metastases, including extrapelvic nodal disease

TI Tl

a b

TI c

T2a T2b T3a T3b N1 T4

M1

The vast majority of small ovarian cysts visualised on MRI are ovarimc,lillicles i n various stages of development. In women of reproductive years a cyst > 1 cm is usually a corpus hrtecun cyrst,buton occasions will he a follicular cyst. These simple ovarian cysts are thinwalled and tend to have a low-intermediate signal on T,-weighted i mages and a high signal on T,-weighted scans (Fig. 34.1 12). Simple ovarian cysts are not uncommon in post menopausal women. The ovaries lie in the adnexa lateral to the uterus, maintaining a constant relationship with the pelvic ureter. The ureter courses posterior or lateral to the ovary, this relationship remaining constant irrespective of the position of an ovarian mass in the pelvis. Thus an ovarian mass will displace the ureter posteriorly or laterally

assessment as the junctional zone may not he visible in some postmenopausal women. Following intravenous gadolinium-chclate the endometrial carcinoma enhances, increasing the contrast difference between tumour and normal endometrium and improving the conspicuity of smaller lesions. In addition, viable tumour can be differentiated from non-enhancing necrosis and blood clot. Dynamic contrast enhancement with integration of signal intensity-time curves allows separation of viable from necrotic tumour. Staging of endomctnal carcinoma is important in defining appropriate treatment, which is primarily surgical with or without pelvic radiotherapy (Box 34.2). Clinical staging is inaccurate: one-fifth of patients are understaged. As with the other imaging techniques, MRI cannot provide a histological diagnosis and is not indicated unless a positive histology has been obtained. The overall reported accuracy for MRI in staging is 85-92/( ; it has particular advantage in defining the depth of myometrial invasion and in demonstrating extrauterine extension. MRI is indicated in those patients in whom physical examination is difficult and there is a clinical suspicion of advanced disease, those who are unsuitable for surgical staging, or if the tumour is of high grade.


Bilateral cystic ovaries (o) in a 25-year-old on a coronal fatsaturation T2 -weighted spin-echo (TSE 3500/100) image showing a lowsignal haemorrhagic right ovarian cyst (large arrow-probably a corpus l uteum cyst) with some surrounding free intraperitoneal fluid (small arrows). r - rectum; u - uterus. Fig. 34.112

The normal-sized ovaries are best demonstrated on transverse or coronal scans, and can he identified in 96% of women of reproductive age. They measure 1.5-3 cm in diameter and have a variable signal on T I - and T,-weighted images. The premenopausal ovary shows a low-intermediate signal, similar to muscle, on T,-weighted images. This appearance is altered if there is haemorrhage present. On T weighted scans the ovary is usually of low signal, but it can be of high signal in some individuals. The cause of this high signal is unknown but is probably due to a looser vascular and connective tissue in the medulla of the ovary. A low-signal rim, in keeping with fibrous cortical tissue, can be observed on the T,-weighted images. On high-resol ution T,-weighted imaging, numerous small peripheral cysts (follicles) are seen (Fig. 34.112), with a more intermediate-high signal from the central stroma of the ovary. This appearance must be distinguished from that of polycystic ovarian disease. The postmenopausal ovary demonstrates a low signal on T,-weighted images, with few if any peripheral follicular cysts. On T,-weighted images, difficulty can occur in identifying the ovaries separate from adjacent bowel and uterus. The use of bowel -specifie oral contrast agents can he of help in this regard. After intravenous gadolinium-chclate the normal ovaries enhance, allowing improved detection of non-enhanci ng follicular cysts.

Multiloculated thin-walled haemorrhagic benign ovarian cysts (c) showing fluid-fluid levels on a transverse T 2 - weighted spin-echo (SE 2000/120) image. There is a coincidental uterine leiomyoma (I). Fig. 34.113

Previous Page

CONTENTS



1103

Fig. 34.115 Dermoid cyst (arrows) showing a unilocular mass with a nodule within high-signal fat on (A ) coronal T 1 - weighted spin-echo (TSE 700/12) i mage and (B) sagittal T 2 - weighted spin-echo (TSE 3500/100) i mage. A fat-suppressed sequence (not shown) was also performed to confirm the fat contents of the cyst. b = bladder.

Fig. 34.114 Large left adnexal mass (d) due to an ovarian dermoid displacing the uterus which contains a coincident subserosal leiomyoma (I) on placing corona) T,-weighted (SE 720/25) and (B) transverse T2 - weighted spinecho (TSE 2000/80) images. The dermoid cyst has solid and cystic components with some areas of intermediate-high signal noted in (A).

(Fig .34.117B), in contradistinction to pelvic lymphadenopathy which lies on the pelvic side-wall lateral to the ureter.

Benign ovarian cysts These tend to be small (.v /3 and asphyxiating thoracic dysplasia-all postaxial-and with Holt-O ram syndrome and Fanconi's anaemia (preaxial).

Syndactyly Occurs in A pert's syndrome, Fanconi's anaemia, Laurence-Moon-Biedl syndrome, trisomy 13 and trisomy /8. ( For a fuller list consult specialist texts; see 'References and suggestions for further reading'.)

This is a very important condition because success in its treatment depends upon early recognition. The incidence of hip instability at birth is 5-10/1000, and of frank dislocation 1-1.5/1000. One-third of cases of congenital dislocation of the hip are detected late, but these often have a family history of the disease. Instability and dislocation is usually unilateral (L : R = I I : I )

I

F1g.35.3 Trauma to the distal radius has resulted in partial fusion of the epiphyseal plate and partial growth arrest. Avulsion of the ulnar styloid occurred at the same time. people. In syndromes they are usually proximodistal and occur in A pert'ssyndrome, dvschondrosteosis, chondroectodermal dysplasia, Holt-O rain syndrome and Turner's syndrome. Similar appearances may follow trauma, infection and rheumatoid disease.

but both hips may be involved (unilateral : bilateral = 11:4). Females are more commonly affected (F : M = 5 : I). Sixty per cent of affected children are first horn. These children are far more likely to have been breech presentations (breech : vertex = 6 : I ), possibly because the abnormal lie does not permit reduction in utero. Children born by caesarean section are thus also more likely to have associated instability and dislocation. A family or twin history is also common; a subsequent child has a 6% risk of involvement. Congenital dislocation of the hip is more commonly found in the winter (winter : summer = 1.5 : I ), possibly because of tight swaddling of children in a position of dislocation in winter.

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ Absolute indications

Polydactyly is more common in black-skinned people. It may he postaxial (ulnar) and may range from a minor ossicle to complete duplication of the little linger. Preaxial l esions (radial) range from minor partial duplication of the thumb distal phalanx to complete thumb duplication. On occasion the hand may be duplicated. Syndactyly may be associated with polydactyly.

If any one of these is present, the hips should he examined by i maging: I. 2. 3. 4.

Family history of congenital dislocation of hip Neonatal hip instability Limb shortening Limitation of hip abduction in flexion.

Relative indications

If any two of these are present, the hips should be examined by i maging:

I. 2. 3. 4. 5.

Ilg, 35.4 Chondroectodermal dysplasia (Ellis-van Creveld syndrome). Polysyndactyly is demonstrated. In addition, the phalanges are abnormal in morphology.

Breech presentation First-born child Caesarean section Other congenital abnormalities Excessive fetal moulding.

I maging Ultrasound (Ruth Green)

Ultrasound is now the accepted method of primary investigation of suspected developmental dysplasia of

CONTENTS

Previous Page

1110

Next Page


Fig. 35.5 (A) Coronal scan of a 3-month-old baby's hip. The ilium (II) is parallel to the transducer. The femoral head (F) is round and of speckled low echogenicity with a central bright echogenic ossific centre. The bony acetabulum (Ac), cartilaginous labrum (Lb), triradiate cartilage (tr) and ischium (Is) should be seen if in the correct plane (gluteal muscle = GI). (B) Coronal scan of developmental dysplasia of the hip. Note the shallow acetabulum (Ac) and displacement of the femoral head (F) (double-headed arrow). (C) Coronal scan of normal hip showing Graf angles (a = bony acetabular roof angle and /3 = the cartilaginous acetabular roof angle).

the hip. The image is obtained with the child on his/ her side with the hip slightly flexed, the transducer is parallel to the ilium and the image obtained is in the coronal plane (Fig. 35.5A). The measurement of Graf angles may be made in the assessment of hip dysplasia to determine the degree of dysplasia (Fig. 35.58) and plan management (Graf I984a, b). The baseline is drawn as a straight-line extension from the ilium, a line drawn from the triradate cartilage to the point where the bony acetabulum meets the baseline represents the alpha angle. The line drawn from the lateral aspect of labrum to the baseline where the lateral bony margin of the acetabulum meets the baseline represents the beta angle, a Table 35.1

measurement of the cartilaginous development (Fig. 35.5C). A si mplified Graf classification is demonstrated in Table 35.1. Plain film radiology The advent of neonatal hip ultrasound.

especially if it can be carried out as a screening test and again at 3 months, has limited the use of conventional radiology. The indications for conventional radiographs given below (taken from Catterall's work (personal communication, 1990)) of course apply also to ultrasound. At birth, neither femoral head is ossified but, occasionally, a notch above the acetabulum may be present (Fig. 35.6) and even in

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ Simplified Graf classification

Type

Description

Alpha angle

Beta angle

Comment

Treatment

I

Good cartilaginous and osseous roofing of femoral head. Normal contact and centring of femoral head

>>-60

3 months delayed maturity

Follow-up and consider (abduction) orthosis

Illa

Cartilagenous roof pushed upward. Femoral head pushed cranially subluxing. No structural change of cartilage

3.0 per 1000 Prevalent screen >0.4 to >0.9 per 1000 I ncident screen >0.5 to >1.0 per 1000

Prevalent screen >3.6 per 1000 I ncident screen >4.0 per 1000

>0.75

2.0 per 1000 I ncident screen >2.2 per 1000

The rate of invasive cancers detected in eligible women b. The rate of cancers detected which are i n situ carcinoma c. Standardised detection ratio (SDR)

3. To maximise the number of small invasive cancers

The rate of invasive cancers l ess than 15 mm in diameter detected in eligible women i nvited and screened

Prevalent screen >1.5 per 1000 I ncident screen >1.65 per 1000

4. To achieve optimum i mage quality

a. High contrast spatial resolution b. Minimal detectable contrast (approx.) 5-6 mm detail 0.5 mm detail c. Standard film density

>10p/mm

5. To limit radiation

Mean glandular dose per film to standard breast using a grid

some angiomatous malformations) may cause only minor changes. These water-soluble non-ionic contrast media, which are licensed for intrathecal use (e.g. lohexol), are virtually non-neurotoxic, do not cause epileptic fits and are apparently devoid of arachnoid toxicity. They can he used without withdrawing other drugs which the patient may be taking, and anticonvulsant cover is unnecessary. The only relafive contraindication is an unexpectedly severe reaction to a previous i njection of the contrast medium by any route. Minor adverse reactions, which occur in about 25% of patients and are relatively more

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ Fig. 54.4 Myelogram and CT myelography in a patient with rheumatoid arthritis and seemingly mild anterior atlanto-axial subluxation and severe spinal cord damage. (A) Sagittal image reformatted from multiple axial slices (no contrast). (B and C) Lateral views of the myelogram in flexion and extension. (D) Axial CT images of the myelogram showing the flattened damaged spinal cord.

Previous Page 1646

CONTENTS

Next Page

A TEXTBOOK OF RADIOLOGY AND IMAGING the radiologist's hands. Lumbar puncture is used unless there is a positive contraindication. The latter include: I. Manifest or suspected infection of the tissues overlying the l umbar theca. 2. Suspected obliteration or marked stenosis of the lumbar theca due to previous surgery, arachnoiditis, congenital abnormality or recent failure of lumbar puncture under screen control. 3. Suspected difficulty with manipulation of contrast medium to the cervical region, as in patients with a severe kyphoscoliosis or deformities of the neck, or where the head-down position induces dyspnoea, nausea, vomiting or vertigo. Because contrast media are heavier than CSF, it can he induced to form pools under the influence of gravity by placing the patient in such a posture that the region of interest is dependent. The lumbar region i s most simply examined by introducing a relatively large volume of iohexol, containing 240 mg I/ml, with the feet dependent. The puncture is made away from the site of particular interest, usually high in the lumbar region with the patient prone and the back slightly flexed over a firm pad. During injection the foot of the table is at least 20° dependent, and ideally a sufficient quantity of the contrast medium (as estimated by fluoroscopy) is introduced to fill the theca to the level of the second

Fig. 54.5 Myelogram with non-ionic contrast medium. (A) AP projection of cervical region in prone position. The spinal cord is symmetrically situated in the opacified subarachnoid space. The nerve roots are shown descending through the root sheaths to the intervertebral foramina. An anterior radiculomedullary artery ascends to the anterior spinal artery. The periphery of the left sixth cervical nerve root sheath is obstructed by a small lateral disc protrusion. The myelogram is otherwise normal. (B) Lateral projection, supine position, dorsal region. The spinal cord tends to take the shorter route toward the convexity of the thoracic curve in the anterior part of the subarachnoid space, which is therefore wider behind the spinal cord. The descending nerve roots and denticulate ligaments are faintly visible.

lumbar body. Erect A P, lateral and oblique films are taken in all cases and supplemented with prone, horizontal decubitus, and supine films as necessary for diagnosis. When pathology relevant to the patient's symptoms is evident on the erect films, the examination is discontinued, but when this is not the case the study should be extended as high as the mid-dorsal level. For examinations of the dorsal region, i ohexol containing 240 mg I/mI is injected with the patient in the lateral decubitus position, with a firm pad beneath the head and neck to tilt the vertex toward the ceiling, and with the head of the table dependent so as to pool the contrast medium in the concavity of a dorsal curve. The injection is made under screen control, continuing until the region of interest has been adequately filled. After lateral films have been obtained, the needle is removed and the patient placed supine for AP and oblique projections. For examination of the cervical region the patient is prone, with the neck extended by a soft pad under the chin and the lumbar lordosis flattened by a pillow beneath the abdomen. The contrast medium may be directly pooled in the cervical lordosis in patients with a suitably flat back, tilting the head of the table down while about 10 ml iohexol, containing up to 300 mg I/ml, is injected. Alternatively, the injection is carried out so that the contrast medium forms a pool in the lumbar region, and the head of the table is smoothly lowered to transfer the pool to the cervical theca under screen control after the injection has been completed. The latter method is usually adopted for total myclography, in which the lumbar region is examined first, then the cervical, and finally the dorsal region, by turning the patient supine, taking care to elevate the vertex in all manoeuvres to avoid intracranial spill. Should obstruction to the upward passage of contrast medium be diagnosed while the needle is still in the lumbar theca, it can generally be overcome by further injection of iohexol or normal saline while the head of the column lies against the obstruction. The patient may be aware of some discomfort as the intrathecal pressure is increased. but this is usually tolerable, especially if it is explained that the necessity for a further puncture in the cervical region may be avoided.


frequent in women, include headache, nausea, vomiting, backache and radicular pain, but these are generally short lived. Although the empirically recommended dose is 3g I equivalent, this can be cautiously increased to 4.2g I if necessary for complete examination of l arge suharachnoid spaces without increasing the incidence of adverse effects. It should be noted that many of the newer watersoluble contrast media now used in angiography are not licensed for intrathecal injection: only those that are should be used (Fig. 54.5). Myelographic technique Defects of blood clotting should he reversed if possible to reduce any tendency to sub- or extradural haemorrhage. General anaesthesia may be necessary in young chitdren. Otherwise, unless the patient is exceptionally nervous, myelograins arc performed without premeditation or special preparation, apart from an explanation of the procedure. The meal preceding the i nvestigation is light, but a plentiful fluid intake is encouraged. A 22-gauge needle is preferred. After a CSF sample has been taken, the contrast medium is injected with fluoroscopic control, always via a connecting tube to avoid disturbing the needle or irradiating

Previous Page

CONTENTS

Next Page NEURORADIOLOGY OF THE SPINE

For cervical studies AP and lateral films arc made and supplemented with oblique projections as necessary, especially if nerve root compression is suspected. Obstruction due to cervical disc protrusion or spondylosis may be overcome by tilting the patient enough to flex the neck while obtaining sufficient lateral bowing of the head to avoid tipping the contrast medium irretrievably into the intracranial compartment. If the problem is not fully elucidated in the prone position, the patient is turned supine and the neck flexed. This is particularly necessary for examination of the cisterna magna, cerebellar tonsils, and lower posterior fossa. The neck flexion relieves the buckling of the ligamenta flava present in extension, and shows the degree of mobility of the cervical spine and the effect of these factors on canal diameter; it is thus useful in elucidating the components contributing to a canal stenosis. Cervical puncture is best performed between the first and second vertebra in the posterior third of the canal, under fluoroscopic control. Suspicion of an upper cervical or foramen magnum mass lesion contraindicates cervical puncture; suspected mid or lower cervical obstruction is a relative contraindication to cervical puncture as the primary approach. For examination of the cervical region, the contrast medium is introduced with the patient prone and the neck extended; for the dorsal or lumbar region it is introduced with the patient supine and the neck flexed. After myelography the patient is encouraged to resume gentle activity and to eat and drink normally. Bed rest is avoided, unless

i ndicated for other reasons, since it increases the incidence of minor reactions; these are treated symptomatically. Technical errors Extra-arachnoid injection of non-ionic contrast medium is not a significant complication but it should not be mistaken for pathology. Extradural contrast medium extends along nerve roots and muscle planes away from the spinal theca (Fig. 54.6). Subdural contrast medium tends to remain localised and to flow slowly under gravity, but it may extend throughout the subdural space and simulate a swollen cord (Fig. 54.7). Since the subdural space does not normally contain fluid, the contrast medium will not layer or pool as it does in the subarachnoid space. Exceptionally, enough CSF may be present in the subdural space to allow pooling; this occurs most frequently following multiple lumbar punctures. Leakage of CSF from the punctured theca, or of blood from a blood vessel, may compress the spinal theca and simulate a mass; the site of such `needling artefacts' at the level of the injection should suggest the correct explanation. Computed myelography (CM). The whole spine now can be studied rapidly using modern CT with fast multislice and spiral facilities. In-plain resolution and slice thickness and spacing can be tailored to the clinical situation. Reformatting in longitudinal plains i s also very helpful. Once the contrast is distributed evenly the examination of all regions can be carried out with the patient supine. CTM is sometimes used to supplement MRI when it is necessary to evaluate some arachnoid cysts or diverticula, or identify the level or presence of CSF leaks, such as into the pleural cavity after spinal surgery. CTM may also he necessary in primary i nvestigation of clinical myelopathy when MRI is unavailable,

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ Fig. 54.6 Extradural injection of contrast medium. Myodil globules have extended throughout the lumbar and sacral epidural space, through the i ntervertebral foramina and along the course of the sacral plexuses.

1647

Fig. 54.7 Subdural contrast medium. (A) AP projection. (B) Lateral proj ection. (C) Oblique projection. The appearances superficially simulating an enlarged spinal cord are due to the contrast medium outlining the outer border of the arachnoid membrane. Note that the inner border of the contrast column appears lobulated and that its upper border does not form a fluid level with unopacified CSF.

Previous Page 1648

CONTENTS

Next Page


Fig. 54.8 Spinal angiogram, selective injection. Normal study. The arteria radicularis magna arises from the left tenth intercostals artery. It fills the anterior spinal artery and there is retrograde filling of other anterior radiculomedullary vessels. The posterolateral arteries of the spinal cord are also filled through the cruciate anastomosis at the conus.

Fig. 54.9 Dural arteriovenous fistula, left second lumbar angiagram. AP and lateral projections. (A and B) Large arrow, the fistula; small arrows, the draining vein. c. Showing the anatomy of the radiculomedullary arteries prior to surgical procedures in which they may be at risk, as for example in the mid and lower thoracic region in disc and scoliosis surgery.


i mpossible (due to skeletal deformities, excessive obesity or claustrophobia), impractical (life-support systems) or contraindicated ( pacemakers or other powered implants, some intracranial aneurysm clips and mechanical heart valves).

5. Spinal angiography using digital subtraction

Spinal angiography has limited but important continuing applications (Fig. 54.8). It is essential for the preoperative or pre-embolisation elucidation of arteriovenous fistulas and angiomatous malformations involving the spinal cord, or shunting into the coronal venous plexuses which drains cord substance. Aneurysms are much less frequently associated with spinal than with intracranial angiomatous malformations, but they too can be reliably shown by angiography (Fig. 54.9). Angiography is also used for: a. Confirmation of haemangioblastomas suspected clinically in von Hippel-Lindau disease, or from appearances on MRI or even

myelography. However, MRI before and after gadolinium enhancement has replaced angiography in most cases. r b. The preoperative investigation of hypervascular lesions of the vertebrae, including angiomas, some aneurysmal bone cysts, giant cell tumours and chordonws, which may be amenable to preoperative embolisation.

Spinal angiography requires injection of non-ionic contrast medium i nto the orifices of the individual segmental vessels supplying the appropriate radiculopial and radiculomedullary arteries, and is performed using a tapered-tip 4 French catheter, generally introduced through a femoral artery. A simple cobra curve similar to that used for visceral angiography is appropriate for examination of all the vessels; in some cases catheterisation of the lumbar and sacral vessels is facilitated after the curve has been reversed by forming a l oop in a renal or mesenteric artery. A complete study therefore requires a very large number of individual injections of contrast medium, and very careful technique. General anaesthesia is usually recommended because of the length of the procedure and discomfort caused by injecting small somatic arteries. A urinary catheter may be necessary to avoid a distended opacified bladder obscuring pathology. In the presence of pre-existing clinical myelography, at l east some neurological deterioration should be expected in up to 1 5 1Y( of patients.

6. Radioisotope studies

Radionuclide hone scanning continues to be useful for: a. Diagnosis of spinal metastases from a known or suspected malignancy

Previous Page

CONTENTS


1649

h. Occasionally for the detection of inflammatory processes and primurv bone tumours, especially osteoid osteoma and osteohlastoma c. Elucidation of localised pain. with normal plain X-ray. I ntrathecal injection of radionuclides may very occasionally still he useful in localising or confirming a CSF leak, especially a suharachnoid-pleural fistula, which sometimes complicates t horacic disc surgery.

7. Ultrasound This method has limited applications for: a. Elucidation of the contents of meningoceles and meningomyeloceles b. Detection of intramedullary cysts after laminectomy c. Measurement of spinal diameters d. Detection of conus position and spinal cord tethering in infants

8. Discography Discography has been completely replaced by CT and MRI for diagnosis of disc disease causing neurological symptoms. Lumbar discography is still applied prior to discolysis and pcrcutaneous extraction. In orthopaedic radiology it is often used for assessing discogenic pain, for which many authorities consider the results to be of questionable significance. For details of techniques, see Ghelman 1988. Water-soluble contrast media suitable for i ntrathecal use should only he used.

The spinal canal The spinal canal is hounded: (i) anteriorly, by the vertebral bodies and i ntervertebral discs, hacked by the posterior longitudinal ligament: and (ii) posterolatcrally. by the pedicles and laminae lined by the ligamenta (lava. Though the cross-sectional shape of the canal, which varies widely between individuals, is best shown by CT, some dimensions are visible on plain films. Of particular significance is the minimum diameter, which gives an indication of the amount of space for the spinal cord and/or nerve roots: it is the sagittal diameter i n the cervical and lumbar regions, and measurements below 12 mm and 1 4 mm, respectively, are considered as potentially significant developmental narrowing. This may become critical if encroached on further, for example by the minor spondylosis commonly accompanying ageing. The sagittal and interpedicular diameters arc vir-

Fig. 54.10 Cystic astrocytoma, cervical spine lateral projection. The spinal canal is expanded. The posterior borders are concave. The line of the conjoined laminae bordering the posterior margin of the spinal canal is flattened and its length increased. pedicles. These measurements are useful for reference when dilfuse narrowing or widening is suspected (Fig. 54.10). The normal intervertebral foramina are oval or boot shaped and fairly symmetrical in the absence of scoliosis. Masses extending i nto the foramina cause hone erosion, with thinning of the adjacent l amina and pedicles and rounding of the foramina; such masses are mainly neurogenic tumours. Other tumours, and, in the cervical region, kinked vertebral arteries, also less commonly erode foramina. All mass lesions eroding hone are large and easily shown by MRI. However failure to recognise erosion on plain X-rays may result in an important delay in diagnosis (Fig. 54.11).


tually equal in the thoracic region. Interpedicular distance increases at the level of the cervical and lumbar enlargements and the i ncrease continues down the lumbar spine in normal individuals. I nt aspinal masses growing by expansion tend to erode bone. The cortex may he thinned or destroyed if a mass is expanding quickly but tends to he reconstituted when expansion is slow. Enlargement of the spinal canal is usually associated with abnormal flattening or even medial concavity of the vertebral surfaces bordering the canal. Local expansion is generally evident from comparison with adjacent levels, a step of 3 mm or greater being abnormal. Published tables of sagittal ( Ullrich et al 1980) and interpedicular diameters at every level arc available. The former are made from the midpoint of the dense cortical line of the posterior border of the vertebral body to that of the conjoined laminae, which is always visible unless spina hifida is present. The latter is between the most medial points of the cortices of the

The spinal cord and nerve roots

The spinal cord descends from the medulla oblongata, commencing at about the level of the foramen magnum, and terminates at the cones medullaris, which lies between the lower border of the 1 2th thoracic and the upper border of the third lumbar vertebra. It is approximately cylindrical, with slight enlargements in the lower cervical and thoracolumbar regions, corresponding to the cells supplying cervical and lumbosacral innervations. On anteroposterior projections it lies symmetrically in the subarachnoid space: on l ateral views it tends to deviate toward the concavity of the normal curves. The nerve roots pass laterally from the anterolateral and posterolateral margins of the cord at each segment. Normal differential growth of the axial skeleton and spinal cord causes relative cranial displacement of the segmental innervation by about one segment in the lower cervical region, two in the lower thoracic and three in the upper lumbar, so that the obliquity of desceit of the nerve roots increases progressively from cervical to sacral levels. The roots generally appear straight with the spine slightly flexed and in neutral position, but they may be tortuous or redundant when the spine is extended. Especially when the canal is relatively narrow, this appearance may he exaggerated, and has occa-

Previous Page 1650

CONTENTS

Next Page


Fig. 54.11 Neurofibromatosis. (A,B) Lumbar spine X-ray. The L3/L4 intervertebral foramen is enlarged, with erosion of the inferior border of the third l umbar pedicle and the adjacent part of the posterior surface of the vertebral body. (C) Myelogram. A neurofibroma on the fourth lumbar nerve root is partly intradural, causing a well-defined filling defect within the theca at the L3 level, and partly extradural, displacing the theca medially away from the pedicle and intervertebral foramen. There is another intradural tumour at T11 /T12 level which deviates the termination of the spinal cord and nerve roots toward the left and causes complete obstruction. Histology showed this tumour to be a neurofibrosarcoma.


sionally been confused with tortuous vessels. The anatomy of at least some of the roots can almost always be resolved, and distinction can be further aided by a film made in flexion to stretch the roots. The cerebellar tonsils and biventral lobules lie posterior and l ateral to the lower medulla at or just above the foramen magnum. A tonsil extending below the inferior margin of the foramen and

Each nerve root carries a sheath of arachnoid, extending toward the point where it penetrates the dura passing through the epidural fat to gain the appropriate intervertebral foramen. The normal sheath looks like a rose thorn, and deformity or occlusion of the sheath is an important myelographic sign of compression by an extradural mass. The dura is closely applied to the arachnoid, but tearing of the

below the hiventral lobule on a coronal section is termed ectopic. Slightly more marked degrees of ectopia are evident on sagittal studies also. Minor degrees of ectopia are frequent and commonly asymptomatic, but ectopia may produce symptoms by partly obstructing the CSF pathways and/or compressing the medulla or upper cord, or be associated with syringomyelia.

arachnoid by any type of trauma, including surgery or puncture, will allow entry of CSF into the potential subdural space; this can compress the subarachnoid space and increase the chances of a subdural injection of myelographic contrast medium. The epidural space contains a variable quantity of fat which outlines the nerve roots and epidural venous plexuses, facilitating visu-

The meninges The subarachnoid space containing the CSF usually extends down to the second segment of the sacrum, but it may terminate at any l evel between the fifth lumbar and fouth sacral segments. Thin arachnoid suspending ligaments, the denticulate ligaments, extend from the lateral borders of the spinal cord at each level between the foramen magnum and LI, and posterior septa also divide the subarachnoid space incompletely. These structures are commonly shown on myelography and the latter may cause pocketing of CSF, which is not abnormal.

alisation by CT and MRI. When the epidural space is narrow, the contours of the spinal canal are accurately reflected by myelography and normal annuli may cause slight impressions on it. A wide epidural space may obscure disc protrusions on myelograms; although they are clearly shown on CT and/or MRI. Blood vessels The anterior spinal artery is shown by spinal angiography, but may be visible on myelograms as a linear defect approximating to the line of the anterior median sulcus of the spinal cord; the major anterior radiculomedullary arteries may also be shown, passing superi-

Previous Page

CONTENTS


orly to join the anterior spinal artery at an acute angle, forming a hairpin loop. The upper four cervical segments are supplied by the descending anterior spinal artery formed by anastomosing branches from the i ntracranial segments of the vertebral arteries. The lower half of the cervical cord and the upper two segments of the dorsal cord are supplied by two to four anterior radicular vessels arising from the vertebral, deep cervical and/or superior intercostal arteries and very rarely from the ascending cervical artery The anterior spinal artery in the mid-thoracic region is usually supplied by a single anterior radicular branch from one of the adja cent intercostal arteries. The lower thoracic, lumbar and sacral region is supplied mainly by the artery ofA damkiewicz, which may arise at any level between the eighth dorsal and the fourth lumbar artery on either side, but most commonly arises from the left tenth intercostal artery. When i ts origin is relatively high or low, there is usually a supplementary supplying vessel to the anterior spinal artery. Small vessels accom_ P toy the lumbar and sacral nerve roots and form an anastomosis ing the lower cord if the artery of Adamkiewicr is occluded: it may also permit emboli introduced into lumbar vessels to cause i schaemia of the spinal cord. The anterior spinal artery is not a continuous vessel in many patients: the most important watershed is at the periphery of the Adamkiewicz Supply (Fig. 54.9). In addition, in all regions, a segmental radicular or radiculopial supply accompanies the posterior nerve roots; some of these give a medullary component to the peripheral part of the posterolateral aspects of the cord. From the anterior spinal artery, sulcocomnussm-ol arteries run through the anterior median sulcus, passing to each side to supply a hemicord. There are several arteries to each segment, with considerable overlap in their distribution, so that no nervous tissue is dependent on a single penetrating vessel. There is also a peripheral arterial anastomosis around the cord from which small arteries supply the outer parts of the white matter tracts. Veins of the corona/ perimedullary plexus may cause tortuous curvilinear shadows on the posterior surface of the spinal cordon supine myelography films as dark flow voids often apparently. notching the dorsal surface of the spinal cord on MRI when enlarged, and as small high-signal features on the pial surface on MRI after intravenous gadolinium even when normal. They are also shown on carefully subtracted venous phase spinal angiograms.

Skeletal abnormalities of the spine may be further elucidated by CT, and MRI when neural abnormalities are suspected.

The most severe forms of dysraphism are due to: 1. Adhesion between the endoderm and ectoderm of the embryonic disc prior to formation of the notochord in the third week of gestation. This causes the various manifestations of the split notochord/ syndrome, which include dorsal enteric fistula, cystic remains of such a fistula (enteric duplications, foregut, gastric, enterogenous, neurenteric and spinal enteric cysts), and anterior or combined spina bifida. The intraspinal cysts are intradural and usually extratnedullary i n location and typically anterior or anterolateral to the spinal cord; they are unilocular, tense, and may cause local hone erosion and enlargement of the spinal canal as well as spinal cord compression. 2. Derangements of the normal sequential closure of the neural tube and its separation from the dermis and failure of retrogressive filum terminale. diferntaohculemas,wihnorlyfmteaudhconsmlri.TpotenaycblfsupDisturbance of closure of the neural tube produces myelomeningocele or meningocele. Mesoderm prematurely interposed before the neural tube is closed or the caudal cell mass fully regressed, which in normal development would have formed the primative meninx, is induced to form lipoma between the gaping margins of the tube. The lipoma extends through the overlying local defect in the meninges (lipomyelomeningodysplasia Fig. 54.12), or simply an intramedullary lipoma. A little fat in the filum terminale is a frequent i nsignificant finding with intact overlying meninges. Failure of separation of the neurogenic from the dermogenic ectoderm nay he manifest as a dermal sinus, epidermoid or dermoid t umour .


When kvphosis or scoliosis i s due to developmental vertebral anomalies such as dorsal or lateral hemivertebra, to bone destruction by i nflammation or neoplasm, or to fracture or dislocation, the causative lesion is generally evident on the plain X-ray. The short kyphosis of neurofibromatosis, which generally involves about five vertebrae, and the thoracolumbar kyphos is of the mucopolysaccharidoses (particularly Morquio's disease) are recognised from the clinical and/or more diffuse radiological bony abnormalities. Also kyphosis and scoliosis may he associated with spinal tumours and syringohydromyelia; in these conditions enlargement of the spinal canal is commonly, but not necessarily, evident.

1651

Fig. 54.12 Lipomyelomeningodysplasia. Sagittal T,-weighted image. There is high signal from the epidural fat and from the lipomatous mass. The latter is subcutaneous, within the lumbosacral spina bifida, and extends through the dura and then superiorly to blend with the posterior border of the spinal cord which extends down to the lumbosacral region. The dark band extending centrally within the intradural part of the lipoma as far as the cord is presumed to be a fibrous septum.

Previous Page 1652

CONTENTS

Next Page


Plain X-ravs reveal the degree of spina hifida. With overt meningocele or meningomyelocele, the pedicles and lamina are splayed in addition to the failure of fusion. If the vertebral bodies are split, a posterior mediastinal mass suggests an enteric cyst; gas within the mass confirms a connection to the gut. Increased width of the spinal canal with retained medial convexity of the pedicles occurs with diastemaiomvelia. This is a dysraphic state of unknown cmbryogenesis, but is probably initiated by an accessory neurenteric canal or neuroentodermal adhesions. Narrowing of a disc space at the same level and interlaminar fusion are supportive evidence, and a bony spicule dividing the c anal is

confirmatory; in most cases, however, the division is cartilaginous or fibrous (Fig. 54.13), and may be absent, especially when the cervical region is involved. Bony malformations are well shown by CT, and diastematomyelia, aberrant lamina, and spina bifida are best assessed by using it. However MRI will generally allow adequate analysis of the spinal deformities (Fig. 54.14) and is the technique of choice for dysraphism requiring full elucidation. It is however true that on plain CT the low density of fat can reveal the extent of lipomas both superficial and within the spinal canal, including the infiltration of the neural placode. Meningoceles and meningomyeloceles will be shown, but details of intrathecal structures, such as position of the c,)nus, tethering of the cord, presence of an embryonic tumour or cyst, and site of origin of nerve roots, will require elucidation by MRI. Cystic lesions without free communication with the subarachnoid space opacify only on CT sections taken late after myelography. MRI is the best technique to show the anatomy of the brainstem and cervical cord which is are usually affected by a Chiari malformation (Fig. 54.15) in the presence of meningomyelocele, but usually not in other forms of dysraphism. In caudal regression syndrome there may be absence of the caudal spine, which may involve a hemisacrum, the whole sacrum and/or a variable number of lumbar vertebrae. It is associated with neurogenic bladder and usually paralysis of some muscles innervated from sacral segments. The conus may he high, or the cord may end low in a nodular expansion. Low tethering or a lipomeningocele may be associated.


Diastematomyelia, lipoma. (A) Sagittal T,-weighted sequence. (B) Axial T 1 sequence. A posteriorly situated lipoma at L5 and S1 levels extends through the dura into the subcutaneous tissues. A band extends from the inferior margin of the posterior surface of the fifth lumbar vertebral body into the lipoma. The spinal cord is divided into two unequal parts, which extend one behind the other through the lumbar region and which both blend into the anterior border of the lipoma at L5 level, with the band passing between them. Fig. 54.14

Diastematomyelia. (A,B) Myelogram. The width of the spinal canal is increased in the lower thoracic region and it is divided by a bony spicule at T10 level. There is a long cleft of the spinal cord extending from the fourth thoracic to the second lumbar levels; the conus medullaris is at L3/4 disc level. (C) Computed myelogram. Contiguous axial sections of l ower thoracic region pass through the bony spur which is dividing the spinal cord and the subarachnoid space. Fig. 54.13

Previous Page

CONTENTS


1653

Fig. 54.15

Chiari malformation, syringomyelia. Midsagittal T1 -weighted section through brainstem and upper cervical region. The cerebellar tonsils are elongated and extend well below the level of the arch of the atlas. The medulla is elongated and the depressed dorsal column nuclei are below the tonsils. The medulla is compressed between the odontoid and the depressed cerebellar tonsils. There is a syrinx distending the spinal cord below the inferior border of C2 and there is four-ventricular hydrocephalus.

This may affect two or more vertebral bodies and occurs in about 0.6% of the population. It is distinguished from acquired fusion by the following features:

Congenital fusion of cervical vertebrae with spondylosis: myelogram. Congenital fusion of the fourth, fifth and sixth cervical vertebrae with degenerative changes at discs above and below. Osteophytes at C3/4 level compress the spinal cord and nerve root sheaths. Osteophytes at C6/7 l evel compress the root sheaths only. Fig. 54.16

and quantify the deformity. These however may he abnormal in the presence of other insignificant anomalies, such as a congenitally short clivus and therefore a prominent slope of the foramen magnum. A tlanto-occipital assimilation (Fig. 55.17) may he associated with instability at the atlanto-axial joints, with forward slip of the atlas on flexion. This may result in quadriparesis, sometimes accompanied by lower cranial nerve palsies and cerebellar dysfunction, which may be slowly progressive or related to mild trauma. Fusion of C3 and C2 vertebrae is often associated.


1. The fused segments seem generally taller than adjacent unfused bodies and are reduced in AP diameter, with a tendency to ` wasp waist' at the level of the fusion (Fig. 54.16). The vertebral bodies tend to retain their fetal morphology on axial sections. 2. The arches are commonly also fused and the intervertebral foramina are round. 3. Angulation, which is common in postinflammatory fusion, is absent because there is no bone destruction. 4. Association with spina hifida is not infrequent. Though any part of the spine may be affected, the cervical region is most com-

ciated with failure of segmentation. Early or accentuated spondylosis tends to affect the intervertebral discs adjoining the fused segment. Plat vbasia, i n which the angle between the central part of the anterior fossa and the clivus (basal angle) is increased to 142° or more, has little if any clinical significance. Basilar invagination, i n which the normal slope of the posterior fossa down to the foramen magnum is reversed, may cause neuraxial compression or cranial nerve dysfunction. This deformity may he congenital or acquired in conditions such as rheumatoid arthritis which destroy (basilar erosion) or soften (basilar impression) hone. The latter include Paget's disease, osteogenesis imperfecta and osteomalacia. Several measurements have been advocated to recognise

y malformations arc not associated with malformations Somebnyivld.ThcasynromefKlip-Fshrtnck oftheCNS.Tsincludreobty,whairlndSpegsour-ia mentation of basiocciput or dens, ossicles near the foramen magnum, ponticles and bipartite facets of the atlas and paracondylar and epitransverse processes), clefts of the arch of the atlas, lateral hemivertebrae, butterfly vertebrae, absent pedicle or lamina, persistent ncurocentral synchondrosis, cleft and hypoplastic pedicles, and cleft lamina. Rarely, absence or hypoplasia of a pedicle is a feature of the mesodermal dysplasia of neurofibromatosis. A bsent pedicle can be mistaken for hone destruction in patients with pain or malignancy. In plain films, however, the margins of the defect arc well corticated and there is generally evidence of more extensive vertebra malformation. In the cervical region this includes absence of the ipsilateral superior articular facet and retrolocation of the lateral mass of the vertebra posterior to those of the other vertebrae. The adjacent parts of the vertebral bodies and the pedi-

Previous Page

1654

CONTENTS

Next Page


Atlanto-occipital assimilation: reformatted CT sections in coronal (A), sagittal (B,C) through condyles and axial (D) planes. The left occipital condyle and lateral mass of atlas are small, relatively dense and completely fused. The joint space between the right occipital condyle and l ateral mass is incomplete. The posterior arch of the atlas is closely applied but not fused to the skull bone. Fig 54.17


Absent thoracic pedicle: CT. The left pedicle of the seventh thoracic vertebra is almost completely absent. The bone bordering the defect is corticated and the articulation with the corresponding rib is anomalous. The bone defect is filled by fat continuous with the epidural fat outli ning the outer margin of the dura, which is in a normal position. Fig. 54.18

cles above and below the malformation are not eroded, which further excludes a neurolibroma or kinked vertebral artery. In the t horacic region the ipsilateral posterolateral border of the body of the affected vertebral is prolonged beyond the normal site of the ncurocentral synchondrosis and forms an anomalous articulation with the rib. CT confirms the nature of the defect, showing the anatomical features to advantage (Fig. 54.18). The position of the pedicle is filled in by fat which overlies the normal theca; there is no soft-tissue mass, lack of definition, or irregularity of the bony margins, as would be usual with inflammatory and neoplastic processes. With aplasia of a lumbar pedicle, there is abnormal orientation of the articular process and thickening of the contralateral neural arch. Such thickening is also present, though usually less conspicuous, in the other regions of the spine. It may be associated with sufficient increase in bone density to simulate a focal sclerosing lesion such as osteoid osteoma, but the presence of the congenital anomaly is of itself an adequate cause of the increase in bone density. Dural ec/asia (Fig. 54.19) may occur as part of the mesodermal dysplasia of neurofibromatosis; it is commonest in the thoracic region, and may be sufficiently pronounced to form a lateral mettingocele. It also may complicate ankylosing spondylitis, where diffuse ectasia or pocketed localised outpouchings (sometimes contaming adherent spinal roots) may be seen, and among the grossest

Previous Page

CONTENTS


1655

Fig. 54.19 (A,B) Dural ectasia in neurofibromatosis: myelogram. The subarachnoid space is markedly expanded in the lower half of the thoracic and upper lumbar regions. The corresponding part of the spinal canal is markedly expanded in this region and there is erosion of the posterior borders of the vertebral bodies and of the pedicles.

examples have been seen in connective tissue disorders like Ehlers-Danlos. These cause well-defined paravertebral masses associated with widened intervertebral foramina, which are usually asymptomatic incidental findings but may cause pain. They are of CSF signal intensity on MRI, and the anatomy is best shown by this technique. They are also of CSF density on CT, and though they opacify freely on myclography, the latter is rarely necessary for diagnosis.


Plain X-rays are first obtained, and in most cases provide adequate information regarding the presence of major fractures and dislocations. Careful attention to the anatomical alignment will reveal some abnormalities in all cases with displacement, though these may be subtle, as in rotatory dislocation with interlocking facets. A l ong bone' type fracture should be suspected if trauma occurs in ankylosing spondylitis. The degree of compromise of the spinal canal is well elucidated by MRI and CT (Fig. 54.20); fractures of neural arches and vertical fractures of vertebral bodies are better shown by CT but horizontal fractures may only be adequately visualised after reformatting of axial data. Foreign bodies, herniated intervertebral discs, displaced bone and haematoma arc shown and the degree of encroachment into the spinal canal is evident. Contusion and oedema may obscure softtissue planes and confuse the distinction between spinal cord swelling and extrinsic compression.

Fig. 54.20 Comminuted fracture of thoracic spine: fractures of the ninth and eighth thoracic vertebrae with anterior subluxation of the eighth. Axial sections through the ninth (A) and the eighth (B) thoracic vertebra, and reformatted sagittal (C) and coronal (D) sections. There is a comminuted fracture with disruption of the body and neural arch of the ninth thoracic vertebra. A fragment has been separated from the posterior inferior margin of the body of the eighth thoracic vertebra and there is anterior subluxation of the rest of the vertebra. The spinal canal is markedly narrowed and the spinal cord is compressed by displaced bone fragments. The subarachnoid space is opacified by a blood clot at T9 and T10 levels. (E,F) MRI of cervical spine. Whiplash injury in road traffic accident 6 months previously. Sagittal sections: T,-weighted (E); T2 weighted (F). The spinal canal is narrowed due to spondylosis. A small focus of aging haematoma is shown in the spinal cord at CT, possibly an unrelated cavernoma rather than an evolving haemorrhagic contusion.

Previous Page 1656

CONTENTS

Next Page


MRI is particulary valuable in showing the extent of spinal cord damage by contusion, haemorrhage or disruption, and/or the degree of compression by displaced bone or haematoma. In penetrating i njuries, the presence of ferromagnetic foreign bodies near the spinal cord must first be excluded by plain X-rays. In the absence of MRI, iohcxol myelography and/or CT myelography is performed: (i) if there is a partial cord dysfunction which is not improving in the absence of detectable cord compression; and (ii) if there is significant persisting cord dysfunction after reduction of a dislocation to less than one-third encroachment in the spinal canal.

Traction on spinal nerve roots causes tearing of the arachnoid and dura of the root sheath, allowing leakage of CSF both along the course of the torn nerve root and along the epidural space. The fluid collections are surrounded by compressed areolar tissue and fibrosis, forming pseudomeningoceles. These are most common in the cervical region in association with brachial plexus avulsion, where they may encroach on the apical pleura and be visible as apical masses on the chest X-rays. Such meningoceles are well shown by MRI and CT and are outlined on myelography. In large meningoceles, dilution of contrast medium may detract from the detail of conventional myelography: the opacification is very adequate for CTM. Nerve roots may be torn without an associated meningocele, and some roots may he intact when a meningocele is present. Exact details of anatomy showing these features can be obtained from myelograms and CTM and from MRI. An extradural pseudomeningocele may displace the cord hut virtually never causes signs of cord compression and is not an indication for surgery.

Fig. 54.21 Old fracture of odontoid peg with non-union and anterior subluxation. (A) Reformatted sagittal section. (B-E) Axial sections. B is at the l evel of the upper border of the anterior arch of the atlas, C at the level of the lower border of displaced odontoid, D through the upper border of the posterior arch of the second cervical vertebra, and E 3mm below D. The theca is impressed by the posterior margin of the lower fragment of the second cervical vertebra, and the spinal cord is deformed and compressed between it and the posterior arch of the first cervical vertebra. Arrows mark anterior arch of atlas, odontoid fragment, lower fragment of second cervical vertebra; arrowhead = posterior arch of second cervical vertebra.


This may follow penetrating or blunt trauma and may be suspected when persistent pleural effusion follows the injury or spinal surgery, especially corto-transversectomy in approaching a thoracic i ntervertebral disc; occasionally pneumocephalus occurs. The leak may be confirmed by conventional myelography or CTM.

The maximum normal distance between the cortical margins of the arch of the atlas and the odontoid is 2.5 mm in adults and 5 mm in children. A tlanto-axial subluxation and dislocation may cause acute or slowly progressive quadriparesis. It may be secondary to:

1. Abnormalities of the odontoid process

These are due to: (i) fracture or (ii) congenital anomaly (os odontoideum). Fractures of the dens are relatively common, accounting for about l-2% of all fractures of the spine. They occur through the base of the dens, most commonly due to a hyperflexion injury, which generally also fractures the posterior arches of the atlas, allowing enlargement of the ring of the atlas to take place during anterior subluxation or dislocation without cord compression. Hyperextension fractures are much less frequent, and are also

usually associated with disruption of the posterior arch. Isolated dens fractures without subluxation may be difficult to detect in the acute stage even with tomography. Failure of union leads to pseudoarthrosis with an irregular horizontal corticated linear defect at the base of the dens; this predisposes to anterior subluxation (Fig. 54.21) with relatively minor trauma. The os odontoideuni is a separate ossicle which becomes significant when it carries the whole of the articular surface, with the anterior arch of the atlas being then liable to subluxation with relatively minor trauma. In such cases the odontoid itself is relatively short, and often maintains its usual superiorly convex corticated upper border. The articulation between the os odontoideum and the dens is above the usual site of the fracture through the dens. However it is recognised now that most if not all are either old fractures with remodelling, or aberrent ossification of the cartilagenous dens, as in some of the inherited connective tissue disorders, most especially Morquio's disease and spondyloepiphyseal dysplasias. This anomaly may also be isolated or associated with others, including atlanto-axial assimilation. 2. Hypoplasia of the dens Hypoplasia of the dens occasionally complicates complex fusion anomalies of the upper cervical spine, but when isolated it usually

Previous Page

CONTENTS

Next Page

NEURORADIOLOGY OF THE SPINE

165 7

i s a misdiagnosis because the separated dens is either small, misplaced or ankylosed to other structures, or remains unossified ( Stevens et al 1994).

Most commonly rheumatoid arthritis but occasionally ankylosing spondylitis, tuberculous or pyogenic arthritis present with these symptoms. A tlanto-axial subluxation occurs in over 6% of patients with rheumatoid arthritis. The atlas is usually displaced forward, with variable elements of rotatory and lateral displacement. The dens may be subtly eroded at its base or markedly waisted, and eventually be reduced to an irregular spike of bone. Erosion of the lateral joints is accompanied by descent of the skull with pseudobasilar impression. The lower joints of the cervical spine may also be affected by the rheumatoid process, and pannus from any of the joints may encroach on the spinal canal. Limb joint affliction may cause difficulty with clinical assessment of neurologi cal disability and of the site of neuraxial compression. Neuroradiological assessment of rheumatoid patients using conventional myclography may be both difficult and painful. Ideally such cases are evaluated by MRI (Fig. 54.22). The site, nature, and degree of compression from bone or pannus are usually elucidated sufficiently for planning. The degree of instability is still probably best assessed by X-ray films in flexion and extension of the neck. CT myelography making reformatted thin sections from just above the foramen magnum down to the lower cervical region (Fig. 55.23) also is satisfactory. Plain CT on bone windows is useful to fully evaluate the integrity of the bones for treatment planning.

4. Presumed laxity of the joint capsules and ligaments This may occur in association with: (i) pharyngitis or tonsillitis, particularly i n children; (ii) atlanto-occipital assimilation; (iii) mucopolysaccharidoses, in which incomplete ossification of the dens and os odontoideum are a frequent feature.


Rheumatoid arthritis with atlanto-axial subluxation. (A,B) CT axial sections at level of the atlas. (C) Midsagittal reformatted section. There is erosion of the odontoid peg and anterior subluxation of the atlas; the spinal cord is compressed by the odontoid peg and posterior arch of the atlas. Fig. 54.23

5. Spondylolisthesis

Rheumatoid arthritis. Sagittal MRI, T 2 - weighted contrast. Severe erosion of the odontoid, vertical atlanto-axial subluxation and degenerate subaxial disease with spinal cord damage are shown. Fig. 54.22

Spondylolysis is a descriptive term referring to defects in the region of the pars intra-articularis of a vertebra, most frequently at the fifth l umbar level. Such defects, which used to be considered as congenital abnormalities, are now thought to be almost always the result of unhealed stress fractures. Spondylolisthesis indicates anterior or posterior slipping of a vertebra from any cause. It is commonly associated with separation of spondylolyses with anterior slipping of the superior vertebra, which may, but commonly does not, cause narrowing of the spinal canal. The fractures in the pars intra-articularis are commonly visible on the routine lateral film of the lumbar spine but are better demonstrated on oblique projections or more exactly by CT. The degree of spondylolisthesis is graded by the fractional shift of vertebral body on the lateral projection but minor slips are shown on oblique projections. On the latter the axes of the apophyseal joints are normally situated progressively slightly more anteriorly as the lumbar spine is descended; this situation is reversed at the level below a forward

Previous Page 1 658

CONTENTS

Next Page


Fig. 54.25 Degenerative lumbar canal stenosis: T 2 - weighted sagittal section of lumbar spine. There is low signal from all the discs, indicating ageing, but the decline in intensity is greater in the lower lumbar region due to additional disc degeneration with dehydration. The lower lumbar discs are narrowed and there is anterior subluxation of two vertebral bodies, with minor posterior protrusion of the annuli impressing the anterior surface of the theca. There is more prominent impression of the posterior surface of the theca due to infolding of the posterior ligaments associated with apophyseal osteoarthritis. The combination is causing degenerative canal stenosis.

Fig. 54.24

Spondylolisthesis of the fourth lumbar vertebra at CT. (A,a) At the level of the pedicles and superior articular facets. (B,b) Through fractured pars intra-articularis. Images are made at window setting appropriate to show both soft tissue and bone. Note that the forward slip of the body of L4 has elongated the sagittal diameter of the spinal canal; there is no nerve root compression.

In all types of spondylolisthesis the intervertebral disc at the level of the slip projects posterior to the plane of the displaced body; this should not be misinterpreted as significant disc prolapse or protrusion unless there is considerable encroachment on the epidural fat between the disc and the theca or root sheath.

slip. On CT the fracture in the pars is at the level of the pedicles; it has irregular margins and adjacent sclerosis (Fig. 54.24), features which should serve to distinguish it from the apophyseal joints which cross the plane of, and are imaged at the level of, the intervertebral discs. The degree of slip and neural compromise is also shown on MRI. However, lack of signal from sclerotic but intact bone may occasionally simulate a break. Stress fractures may heal after spondylolisthesis has occurred, and the partes interarticulares are then elongated and sometimes sclerotic. Spondylolisthesis can also occur as a result of:

Acute epidural haematoma This condition may present with local and/or radicular pain, usually soon followed by signs of spinal cord or root compression. There may he a history of trauma, antecedent lumbar puncture, anticoagulation, aspirin therapy, hypertension or a blood-clotting disorder. It may also rarely be due to an angiomatous malformation. CT may be diagnostic if an extradural mass with attenuation of around 60 HU is shown. MRI is the elective study: intensity patterns will suggest the diagnosis and the degree of narrowing of the theca and neuroaxial compression will be shown. Myelography classically reveals a long extradural compression, generally from the posterior aspect of the spinal canal (Fig. 54.26).

I. Trauma. 2. Pathological fractures or bone softening, as for example in Paget's disease or osteogenesis imperfecta. In such cases the abnormality causing the slip may affect partes interarticulares or pedicles. 3. Congenital dysplasia of the superior articular facets of the sacrum. In this case severe forward slipping is often associated with considerable deformity of the upper border of the sacrum and of the inferior margin of the fifth lumbar vertebra. Since the whole of the fifth lumbar vertebra is displaced anteriorly, it is easily distinguished from the common type. 4. Disc degeneration and osteoarthritis of the apophyseal joints. This is a common cause of lateral recess stenosis and clinically may stimulate disc prolapse (Fig. 54.25).

Such appearances on plain X-rays are generally due to neoplasm or i nflammation. Inflammatory lesions tend to extend through the i ntervertebral discs (Fig. 54.27) and the contiguous bone, and to form paravertebral swellings which spread along soft-tissue planes. Less often, apophyseal joints are involved. On MRI the inflammatory exudate causes increase in T, and T, throughout the affected region. The disc returns high signal on T2-weighted sequences, with diagnostic loss of the low-signal cleft, which is a feature of the normal annulus. The vertebral end-plates become indistinct and lost on T, weighted images early in the process. In tuberculosis, a low-


Previous Page

CONTENTS


1659

Pyogenic infection, L5/S1 disc. (A) Axial section at level of L5; (B) at level of S1. (C) Axial reformatted section. The bone is destroyed adjacent to the disc space, leaving a large cavity with adjacent sclerosis around the disc. There is no intraspinal extension. Fig. 54.27

Fig. 54.26

Acute intraspinal haemorrhage. Sagittal (A) and axial (B) MRI with T,-weighted contrast. The extensive posteriorly located haematoma is well shown. This one probably is subdural in location, and was spontaneous.

Staphylococcus and Streplococcus are implicated but many other organisms, including A ctinomvces, Blasturttvccs, Coccidio ides, typhus and Brucellu, also cause spinal osteomyelitis. Brucella tends to cause extensive sclerosis. 4. The tendency of tuberculosis to be multifocal in the spine. 5. The tendency of tuberculosis to cause deformity, especially gibbus formation.


density abscess expanding psoas or other muscles may also be shown. Malignant neoplastic destruction tends to respect the intervertebral discs, and soft-tissue masses tend to be confined by the periosteum. Paravertebral swellings may he evident on plain films. In the cervical region, widening of the retropharyngeal space (normally under 0.75 cm) or retrotracheal space (normally up to 1.5 cnm in children and 2.25 cm in adults) may be caused; in the thoracic region the paramediastinal pleura is displaced laterally. They are better shown by CT or MRI (Fig. 54.28), which are valuable complementary studies, particularly in the lumbar region. Distinction between tuberculous and pyogenic inflammation may be suggested by: 1. The presence of associated lesions in other organs, especially the lung, urinary tract and other hones in tuberculosis. 2. The tendency of untreated tuberculosis to induce less bone formation, to produce cold abscesses extending along muscle planes and beneath the anterior spinal ligament, eroding contiguous bone. 3. The tendency for pyogenic organisms to induce sclerotic bone reaction within the vertebral bodies and under the elevated periost eam. The various organisms cause generally similar reactions, which require bacteriological study for distinction. Commonly

Fig. 54.28 Neurofibroma: axial section, cervical region, T 1 -weighted sequence. There is a slightly lobulated mass grossly enlarging the left intervertebral foramen and eroding the adjacent bone. An intraspinal component of the mass displaces the theca and spinal cord toward the right and compresses them. A paravertebral component forms a mass displacing the deep cervical muscles. The extent of this dumb-bell neurofibroma is evident from this single study.

Previous Page 1660

CONTENTS

Next Page


Hydatid cysts These are responsible for less than 1.0% of inflammatory spinal lesions but important because their radiological appearance is suggestive. In bone, hydatids spread slowly through intertrabecular spaces by exogenous vesiculation, passing through the periosteum into the soft tissues. In the spine, the thoracic region is most commonly affected, and the disease tends to involve adjacent vertebrae and ribs and to spare the intervertebral discs. The cysts cause bubble-like round or lobulated circumscribed lytic lesions in the bones, with virtually no sclerotic reaction, and adjacent soft-tissue masses which tend to be extensive and cause extradural compression of the spinal theca.

Metastases The commonest tumours affecting the vertebrae are metastatic. Those from the breast and prostate and neuroblastoma frequently i nduce sclerotic bone reactions (Fig. 54.29), also a feature of lymphoma and histiocytosis X. Sclerotic vertebra plana in children is characteristically due to the cosinophil granulomatous form of histiocytosis X. Most metastases, including those from the commonest primary neoplasm, a bronchial carcinoma, are lytic; multiple myeloma, Burkitt's lymphosarcoma and Ewing's sarcoma cause similar destructive lesions, though the latter may cause periosteal reaction and bone expansion. In common with most tumours, metastases cause increase in T, (low signal) and T z (high signal) at M7R1. Melanoma is an exception; melanin is paramagnetic and may cause l ow signal on T2- weighted sequences. The bone lesions and softtissue components, including any thecal compression, are well demonstrated on MRI.

Fig. 54.29 Metastasis CT. Sclerotic metastases involving the right side of the sacrum were shown by plain films but are seen more clearly by axial CT. This also shows bone destruction around the right sacral foramen and a vertical pathological fracture through the body of the sacrum. It should be noted that, although focal neoplastic changes are well shown, diffuse infiltration of marrow may be more difficult to recognise until T, and T z i ncreases are pronounced. Following deep X-ray therapy, T, is decreased, probably due to fat replacement of haemopoietic tissue. Haemosiderin deposits after multiple transfusions of blood cause low signal on T z-weighted sequences due to susceptibility changes.

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ Fig. 54.30 (A) Osteoblastoma of cervical spine. The right pedicle and lateral mass of the sixth cervical vertebra are enlarged and sclerotic, with an irregular central low-density nidus. (B) Osteoid osteoma. CT scan of mid-thoracic spine. There is sclerosis with focal expansion involving the left lamina. The tumour encroaches on the epidural space and contains a small nidus which is of low density.

Previous Page

CONTENTS


166 1

Sacral chordoma. (A) A fairly well defined region of destruction i nvolving the bodies and left lateral masses of the lowest two segments of the sacrum. (B,C) CT sections through the fourth and fifth segments of the sacrum. A mass of density slightly lower than that of muscle is expanding and destroying the bone of the bodies and left lateral masses of the fourth and fifth segments of the sacrum. (Contrast in loops of small bowel). Fig. 54.31


Osteoblastoma and osteoid osteoma In the spine these tumours involve the neural arch; they may present with local or referred pain as well as cord compression. Sclerosis and expansion of the affected part of the neural arch may be better demonstrated with CT than plain X-rays (Fig. 54.30); a low-density nidus with or without a sclerotic centre is generally evident on CT.

Spinal chordomas (Fig. 54.31) These primary tumours arise from notochord remnants, but usually become manifest in middle age; they are commoner in males. The majority (50%) involve the sacrococcygeal region, and only 1 5% arise in the remainder of the vertebral column, the rest originating i n the clivus. The tumour causes hone lysis, which tends to involve several segments of the spine and to be associated with expansion. Calcification or residual bone fragments within the tumours, and some sclerotic reaction are not infrequently present. CT and MRI are particularly valuable in delineating the extent of the sacral chordomas, which may be entirely below the level of the spinal theca; higher tumours cause non-specific extradural compression.

Fig. 54.32 Haemangioma of bone, lateral lumbar spine. The secondary trabeculae of the body of the second lumbar vertebra are thickened; the vertical striate appearance is typical of a haemangioma.

Previous Page 1662

CONTENTS

Next Page


Fig. 54.33 Paget's disease. (A) Lateral thoracic spine. The body and neural arch of the twelfth thoracic vertebra are enlarged and the bone texture is abnormal. (B,C) CT axial sections, lower dorsal vertebra. Note abnormal irregular bone texture throughout the vertebral body, neural arch and appendages; the whole vertebra is slightly enlarged.

Vertebral haemangiomas These usually asymptomatic tumours are diagnosed by observing the characteristic increase of prominence of the secondary bone traheculae of the vertebral body causing a striate or honeycomb pattern (Fig. 54.32). Haemangiomas are commonly associated with regions of fatty change in marrow, causing lowering off, and high signal on T,-weighted sequences. T, is increased, with high signal on T,-weighted sequences: regions of signal void can occur occasionally if there is high flow. Uncommonly, expansion of the affected vertebra may cause a paravertebral swelling and symptomatic extradural compression shown on MRI or myclography. The extent of the abnormality is well shown on MRI or CT. Angiography and cmbolisation of the feeding arteries may reduce the morbidity of laminectomy.

some spinal root and spinal cord compression in up to 20% of asymptomatic older subjects. Therefore it is usually asymptomatic. Despite a considerable literature linking specific appearances to specific clinical features, it should be recollected that evidence for such linkages is purely circumstantial. Entrapment neuropathies due to extruded disc material and some cases of spondylotic myelopathy arc clear exceptions however. Imaging is prioritised appropriately to demonstrate involvement of neural tissue. Disc prolapse Extrusion of the softer material from within an intervertebral disc into or through a posterior or posterolateral radial tear in the annulus fibrosis is a common cause of neural involvement_ This takes the form of a focal broad-based bulge in the margin of the annulus, or a focal mass extending upwards or downwards in the anterior epidural space. Far lateral protrusions or extruded fragments involve the intervertebral foramena, not the spinal canal. They arc commonest in the lumbar spine. These lesions usually are well shown by CT and MRI. Disc material is denser than CSF in the thecal sac and is clearly visible against epidural fat on CT (Figs 54.34-54.36). There is a danger that a very large extrusion, which fills the spinal canal, will be overl ooked on CT, because the disc material may closely resemble the thecal sac, and the density difference between it and CSF may not be appreciated. On MRI extruded fragments often are brighter than parent disc substance in T,-weighted images, and may enhance after intravenous gadolinium; sometimes they are heavily calcified, which is usual in the thoracic region. MRI is more reliable than CT at demonstrating soft disc protrusions in the cervical spine, where there is less epidural fat. Myelography or CTM should be required rarely in modern departments. Only CT or MRI will demonstrate far lateral protrusions. Plain X-rays commonly are normal or show non-specific changes, such as reduction of disc space or changes in vertebral alignment due to muscle spasm. A high proportion (70%) of extruded, soft-tissue fragments, usually yielding a higher signal than the parent disc on MRI, disappear spontaneously, usually within 3 months (Fig. 54.36).


Paget's disease of bone This commonly involves vertebrae, usually affecting the neural arch and appendages in addition to the body (Fig. 54.33). The coarse trabecular structure and expansion of the whole vertebra are characteristic features, and the body often loses height due to softening. Encroachment on the spinal canal is an infrequent cause of compression of the spinal cord or cauda equina; myelography shows non-specific extradural compression, whereas MRI shows the bone thickening in addition.

Spondylosis is a useful general term for this process caused by wear and tear, though it is most often used when describing the cervical spine. The process involves the intervertebral discs, vertebral bodies and facet joints usually all together but to varying relative extent. The process becomes of particular relevance to neuroradiology when the spinal cord or roots become affected. It is the commonest cause of entrapment spinal neuropathy and of neurological disability due to spinal cord disease. It is extremely common, being virtually universally present in the over sixties, and it is associated with at least

Previous Page

CONTENTS


1663

Fig. 54.34 Far lateral lumbar disc prolapse. (A,B) CT axial sections. (A) Through L3/4 disc space, showing a soft-tissue mass contiguous with the right posterolateral aspect of the disc, encroaching into the intervertebral foramen and extending lateral to it. (B) Section 6mm higher than A, and at level of the lower part of the body of the third lumbar vertebra. The disc prolapse forms a soft-tissue mass in the path of the emerging right L3 nerve root, which is not visible due to the absence of epidural fat. The left L3 root (arrow) is clearly shown. (C,D) MRI L4/5 l ateral disc prolapse. Sagittal T 1 -weighted sections. Encroachment of prolapsed disc substance into the L4/5 i ntervertebral foramen obliterates the epidural fat around the emergi ng nerve root.

Degenerative change Loss of normal tension or volume in the nucleus pulposus tends to cause circumferential bulging of the annulus fibrosus and tension on the periosteum, resulting in osteophytosis and marginal sclerosis. These changes are most frequent in the lower cervical and lower l umbar regions and are accompanied by disc narrowing and altered stresses on the apophyseal joints, resulting in osteoarthritis. They are revealed by plain X-rays and CT. Reactive changes of three types occur in the adjacent vertebral bodies, and are well shown by MRI (Modic et al 1988). Degeneration in ligaments can lead to fibrocartilaginous metaplasia, calcification, ossification or myxomatous degeneration, all of which can result in diffuse thickening or a focal mass which may compress neural tissue. This process can involve the posterior longitudinal ligament, the cruciform ligament at the craniocervical junction, the ligamenta flava and the capsular ligaments of the facet joints. It includes entities such as ossification of the posterior longitudinal ligament (OPLL) (Fig. 54.37), retroodontoid pseudo-tumour, ossification of the ligamentum flavum; synovial cysts can also be accommodated in this category. Accelerated degenerative changes are seen in such conditions as ochronosis and Charcot spine. Destructive discovertebral lesions due to non-specific degenerative processes may be seen at one or more levels in ankylosing spondyltis, rheumatoid arthritis, chronic renal failure (dialysis) or as isolated phenomenon. The process is distinguishable from infection on MRI by absent or non-uniform high signal in the involved disc, and irregularity or fragmentation of the vertebral end-plates rather than destruction.

Plain X-rays show most of the features of degeneration. If the measured sagittal diameter of the spinal canal in the cervical region is less than 10 mm, the spinal cord will be compressed. Deformation of spinal and intervertebral canals is best shown in axial sections on CT or MRI, the latter permitting direct visualisation of the effects on neural structures. As in the case of disc prol apse, myelography and CT myclography should only rarely be required in modern departments. With the exception of focal masses directly compressing spinal roots, clinicoradiological correlation generally is poor, even with MRI. Bony stenoses of intervertebral foramena only rarely are focal enough to suggest the probability of surgical cure. Compression or moulding of the spinal cord by osteophytes in cervical spondylosis also correlates poorly with the presence and severity of clinical myelopathy. However, MRi often presents direct evidence that compression is damaging cord substance, in the form of focal signal changes in the spinal cord at or minimally below the relevant intervertebral level. The signal change is deep within the cord, often si mulating accentuation of the central grey matter, unlike plaques of demyelination for example, which usually extend to the pial surface.


Postoperative changes Recurrent myelopathy and/or radiculopathy after surgery may be discogenic or reactive. Distinction is usually evident on CT and/or MRI, with either a typical mass continuous with, and having char-

Previous Page 1664

CONTENTS

Next Page


Fig. 54.35 Extruded (sequestrated) disc protrusion. (A) Sagittal T 2- weighted section of lumbar spine. The L4/5 disc space is narrowed and the signal returned from the nucleus is decreased. There is a slightly lobulated extradural mass behind the L4/5 disc and upper half of the body of the fifth lumbar vertebra which is compressing the spinal theca. The signal returned from it is similar to that of the normal L3/4 nucleus. It was an extruded fragment removed at surgery. Such fragments commonly give higher signal than the damaged disc from which they originate. (B-E) Sequestrated lumbosacral disc prolapse. T,-weighted sections lumbar spine: B,C sagittal; D,E axial. The disc fragment extends behind the upper part of the right side of the body of the sacrum. It displaces the first sacral nerve root posteriorly and erodes the sacral body.


Fig. 54.36 Extruded fragment from a degenerate L3/L4 i ntervertebral disc. (A,B) Sagittal and axial MRI with T2 -weight d con ras . A l ge migratory extruded disc fragment ascends on the right behind the L3 vertebral body. (C,D) Sagittal and axial MRI 8 weeks later shows spontaneous resolution of the extrusion.

actcristics of, disc substance in the former, or a contracting lesion Spinal stenosis extending around the affected part of the theca and/or nerve root Although applicable in all regions, this term is most often used and continuing into the soft tissue in the latter. On T,-weighted when describing changes in the lumbar spine. It is a feature of conMRI, disc usually returns higher signal than scar, in which signal ditions such as achondroplasia and acromegaly, where even mild decreases with ageing. Recent scar enhances immediately but old degenerative changes produce severe multilevel stenoses, especially scar much less and more slowly. Disc fragments usually do not in the lower lumbar region. enhance but may do so minimally and slowly. GadoliniumThe term stenosis should be used with caution, because it has a enhanced MRI appears to be more accurate than enhanced CT in tendency to become a diagnosis when it may be inappropriate. this assessment. Clinical correlation only becomes good when the lumbar spinal

Previous Page

CONTENTS


Fig. 54.37 Ossification of the posterior longitudinal ligament. Sagittal (A) and axial (B) MRI with T 2 - weighted contrast. The axial image is between the intervertebral discs and shows the large low-signal thickened ligament l ooking somewhat like a mushroom.

1665

Fig. 54.39 Canal stenosis. CT sections through lumbar spinal canal. Note the short pedicles (bottom right image) and medially placed apophyseal j oints. The combination causes marked reduction of the sagittal diameter of the lateral parts of the spinal canal, which is further compromised by osteoarthritic changes in the apophyseal joints.


Fig. 54.38 Lumbar spinal canal stenosis with entrapment of the cauda equina. Sagittal MRI with T 2 - weighted contrast showing stenosis of the spinal canal at L4-L5; CSF signal is excluded at the level of the stenosis, and there is redundant coiling of many of the intradural spinal roots above.

Fig. 54.40 Syringomyelia. Parasagittal T,-weighted sequence. The lowsignal subarachnoid CSF outlines the expanded spinal cord and also the i ntramedullary cyst, which shows a typical lobulated appearance. One of the cerebellar tonsils is outlined with its lower pole extending to the level of the arch of the atlas.

canal is very narrow: cross-sectional area even less than 110 mm'-

veins (Fig. 54.38). The lateral part of the spinal canal, sometimes

on CT (Ullrich et al 1980), or narrow enough to eliminate CSF

referred to as the lateral recess, often is the most affected, usually

signal on T,-weighted MRI; an additional sign of entrapment of the

mainly due to hypertrophy of the facet joints (Fig. 54.39). However,

cauda equina is redundant coiling of intradural roots above the

with the exception of S I the spinal roots are within the thecal sac as

slenotic level on MRI, occasionally mistaken for dilated intradural

they cross the disc and simply are displaced medially rather than

Previous Page 1666

CONTENTS

Next Page


Intradural lipoma. T1 -weighted MRI. Sagittal section through posterior fossa and upper cervical region. The posterior fossa and foramen magnum are normal. There is a large mass returning high signal and typical of fat. It lies within the subarachnoid space posteriorly, enlarging the spinal canal and displacing the spinal cord anteriorly. The spinal cord is markedly thinned and the cord substance appears to be almost completely replaced by fat over most of the extent of the tumour. The appearances are typical of a large i ntradural and partially intramedullary lipoma. These tumours typically extrude from the posterior surface of the cord between the dorsal columns. Fig. 54.42

end ymoma. Sag ittal (A) and axial (B) MRI with T2-weighted contrast shows an extensive tumour filling much of the lumbosacral spinal canal and cavitated spinal cord above. Fi Fig. 54.41

Ep

entrapped. On conventional myelograms the thecal sac assumes a waisted appearance at such levels.

i ncluding the extradural part within the sacrum, and cause enlargewent (sometimes massive) of the spinal canal similar to that seen in giant neurofibromas. Such tumours may blend superiorly with the expanded lumbodorsal cord or simulate an intradural extramedullary mass, which may be amenable to almost complete surgical removal (Fig. 54.41).


Syringomyelia This describes a predominantly glial lined fluid-filled cavity or cavities filled within the spinal cord usually centred on the central canal but extending into the dorsal columns through the white commissure. They may be cylindrical and extend through most of the spinal cord, or fusiform where only a localised number of segments are involved. They may enlarge the spinal cord, or be collapsed with normal or small cord size. About 80% of the extensive type are associated with a Chiari I malformation. Most of the rest are idiopathic. A small number are post-traumatic, or postarachnoiditic, and occasional reports in the past have linked them to a variety of other conditions though with less clarity. Identical cavities commonly extend through otherwise uninvolved above and below intramedullary tumours, especially haemangioblastoma (Fig. 54.40). The cavities are well shown by MRI, together with their causative lesions; intravenous gadolinium may be necessary to exclude tumours in idiopathic cases. Plain CT may show them, but generally CTM is required, with imaging early and late (6 or 10 hours) after contrast administration. Ependymomas These generally very slowly growing gliomas comprise about twothirds of all intramedullary tumours, being especially frequent in the dorsolumbar region. They may involve the filum terminale,

Astrocytomas These tumours account for about 30% of cord gliomas, but are relatively more common in children. They may cause focal or extensive expansion of the cord, sometimes continuing into the brainstem, and like ependymomas may have a considerable cystic component. Expansion of the spinal cord tends to he less marked than with ependymomas.

Multiple sclerosis. An axial T 2 -weighted MRI at C6 shows a typical plaque in the right lateral column, reaching to the pial surface of the spinal cord. Fig. 54.43

Previous Page

CONTENTS


1667

Other types of intramedullary glioma and sarcoma occur, but cannot generally be suspected prior to histological evidence. An exception is melanoma, which may give low signal due to susceptibility changes on T 7 - weighted MRI sequences related to the presence of melanin.

Lipoma and dermoid These may be associated with dysraphic features and CT may show l ow density within the fat; T i - weighted MRI reflects high signal which is pathognomonic in the clinical context (Fig. 54.42). Dermoids are often extramedullary and in about 20% a dorsal dermal sinus is demonstrable by MRI.

Haemangioblastoma This constitutes about 5% of intramedullary tumours, but it is potentially curable by surgery and can be diagnosed by spinal angiography. This tumour should be suspected in patients: (i) with previous manifestations or a family history of von Hippel-Lindau syndrome, especially cerebellar tumours or retinal angiomas; (ii) with extensive serpiginous defects in association with intramedullary swelling, especially in the absence of obstruction of the subarachnoid space. In our experience about a third of such cases will have haemangioblastomas. The intramedullary swelling is commonly extensive, due to cyst formation. The tumour nodule reflects low signal on T,- and high on T2- weighted sequences and enhances markedly with gadolinium. If MRI is available angiograms usually arc not necessary. The haemangioblastoma opacifies densely at angiography, usually as a homogeneous nodule, but sometimes as a ring density with central necrotic or cystic change. In contrast to angiomatous malformation, it tends to retain the contrast medium into the venous phase, but there may also be arteriovenous shunting through it (see Chapter 57).

I nflammatory lesions

Myelogram neurofibroma. (A) AP projection. (B) Oblique proj ection. The tumour is intradural on the right side at C4/5 level and its margins are clearly outlined by the contrast medium in the subarachnoid space which is widened around the tumour. The spinal cord is compressed and displaced by the mass. Fig. 54.44

dorsolateral regions most commonly, extending right to the pial surface. Mild swelling is associated with acute lesions. The lesion of acute transverse myelitis usually is far more extensive and involves most of the cord cross-section, with more swelling progressing to atrophy over time. Far more rarely intramedullary abscesses (usually associated with dermal sinuses), tuberculous or sarcoid granulomas, HTLV- I or other viral lesion, and schisostomiasis of the lumbar enlargement are encountered (Fig. 54.5).


The spinal cord frequently is involved in multiple sclerosis (MS) (Fig. 54.43). Plaques generally are limited in extent and involve the

Fig. 54.45 Neurofibromatosis with multiple tumours. Sagittal T,-weighted MRI sections. (A) Posterior fossa and cervical region. (B) Thoracic region. (C) Cervical region after gadolinium enhancement. (D) Thoracic region after gadolinium enhancement. The patient has neurofibromatosis and there are multiple tumours. Some are intradural extramedullary schwannomas. A good example is shown anteriorly at C2 level. The upper border of this tumour is outlined against the cerebrospinal fluid, but its posterior border blends with the spinal cord which gives signal of similar intensity. The extent of this tumour i s evident on the enhanced scan. Some of the tumours are extradural. An example is shown posteriorly in the mid-thoracic region B; the upper and lower borders are outlined against the extradural fat behind the spinal cord. The anterior margin is only slightly denser than the cord substance. After injecting gadolinium (D), the tumour is enhanced. Its anterior border, which is compressing the dura and the cord, is more evident but the intensity of signal from the enhanced tumour is similar to that of the epidural fat, making the upper and lower limits more difficult to define. An intramedullary tumour is also present, expanding the cord in the lower dorsal region. The margins are not distinguished from cord substance on the plain scan B but are evident with ring enhancement after gadolinium (arrow in D).

Previous Page 1668

CONTENTS

Next Page


Meningioma. (A,B) Myelogram. A partially calcified mass forms a slightly irregular defect within the theca on the left side posteriorly in the upper thoracic region and compresses the spinal cord. (C,D) Computed myelogram. The calcification within the tumour is mainly peripheral and is well shown on D, where the subarachnoid space is almost totally occluded. On C, the irregular medial margin of the tumour is shown, together with the flattened spinal cord, which is displaced anteriorly and towards the right. (E-G) Cervical meningioma-MRI: T 2 -weighted E; T,-weighted F; T,-weighted after gadolinium G. The tumour has a broad base on the dura anterior to the spinal cord which is displaced posteriorly and compressed. The tumour returns high signal on T 2 -low on T,-weighted, and enhances markedly. Fig. 54.46


All these lesions are well shown only by MRI. Some can be difficult to distinguish from intramedullary tumours. Enhancement of the meninges after intravenous gadolinium can be helpful to indicate inflammation in some of the rarer conditions though not MS or acute transverse myelitis.

always visible on MRI. More acute lesions may be associated with swelling and enhancement after intravenous gadolinium administration. Striking signal change confined to the dorsal columns may be seen as a consequence of selective destruction of primary sensory neurones in the dorsal root ganglia in Sjdgren's syndrome. The spinal cord may also be involved in Bchcet's disease.

I nfarction

Acute arterial and venous infarcts usually involve long segments of the spinal cord substance usually intensified centrally. Infarcts are shown only by MRI, as extensive signal change with mild swelling progressing to atrophy. Acute infarcts usually are more extensive even than acute transverse myelitis but ambiguity can exist in clinical contexts. They are seen most commonly after thoraco-abdominal aortic aneurysm repair and thrombosis of dural arteriovenous fistulas and their draining veins.

These are outlined in the subarachnoid space and tend to widen the space around them by acting as wedges, displacing and flattening the spinal cord and/or nerve roots, and tending to cause local expansion of the spinal canal. Such masses are likely to be benign tumours, usually neurolibromas or meningiomas (Figs 54.44-54.46).

Meningiomas

Other systemic disorders

Subacute combined degeneration of the spinal cord in congenital and acquired vitamin B,, deficiency or dysmetabolic states results i n extensive lesions in the dorsal and lateral columns, often hut not

These are especially frequent in the thoracic segment of the spinal canal (over 80%) and in the high cervical-foramen magnum region. The majority (over 80%) occur in middle-aged women. The tumours may be partly calcified and are generally entirely intradural

Previous Page

CONTENTS


(over 90%); extradural involvement should suggest the possibility of malignancy. Meningioma tends to reflect MRI signal in much the same way as neural substance, but is clearly defined by marked enhancement after gadolinium injection. Neurofibromas (schwannomas)

These usually arise from a posterior nerve root; about two-thirds are entirely intradural, the rest being either entirely extradural or having an extradural component which sometimes enlarges an intervertebral foramen. They tend to occur in younger patients than do meningiomas, and have no predilection for either sex. They usually reflect high signal on T,-weighted and low on T,-weighted MRI sequences and enhance with gadolinium. Neurofibromas are often relatively larger than meningiomas by the time they present with cord or root compression, especially those in the lumbar canal or with a paraspinal extension, which may be massive tumours causing extensive bone erosion. Spinal tumours very uncommonly present with subarachnoid haemorrhage or raised intracranial pressure; ncurofibroma is most frequently implicated and ependymoma less often. Multiple intradural masses

These may be: (i) neurofibromas, or less frequently meningiomas, which may occur in combination with each other and with gliomas i n patients with neurofibromatosis; (ii) metastases from a malignant neoplasm of central nervous tissue, most commonly a medulloblastoma; or (iii) metastases from a systemic primary, especially carcinoma of the breast or bronchus, which occur most frequently on the cauda equina. Suspected meningeal spread of malignancy may require intravenous contrast for visualisation even on MRI.

These displace the theca away from the bony margins of the spinal canal and tend to compress the theca and nerve root sheaths. Only by considering the history and observing any local bone and/or soft-tissue changes may the traumatic, neoplastic, inflammatory or other nature of the lesion be determined. A disc protrusion is generally suspected from the acute or recurrent history of pain and the position of the extradural impression, which is often, but by no means always, anteriorly placed in the spinal canal and in part opposite a disc. A previously established or currently evident extraspinal neoplasm, especially in the presence of single or multiple foci of abnormal bone or radionuclide uptake, is strong evidence that an extradural mass is metastatic. However, radiological evidence of bone change is sometimes absent and the possibility of complicati ng haematoma or infection should be considered if the patient is receiving chemotherapy. Direct extension of retroperitoneal tumours, especially lymphomas or neuroblastomas, is generally evident on CT and/or MRI. Rarely, advanced retroperitoneal fibrosis affects the extradural space and compromises neural function. Extradural abscesses are usually localised extensions from disease of the spine; they may complicate septicaemia or be metastases from a distant known focus of infection, or apparently occur in isol ation. They usually spread in the relatively unconfined segment of the extradural space in the thoracic region dorsal to the spinal cord, forming a long, smooth swelling, deviating the theca anteriorly,

1669

and compressing the cord. The lumbar and cervical regions are l ess frequently affected, but any suspicion of the diagnosis is a contraindication to spinal puncture in the involved segment. Unusual extradural masses include: 1. Extramedullary haemopoiesis. This occurs in severe anaemias, especially thalassaemia, and both extensive paravertebral masses and bone changes secondary to marrow hypertrophy will he evident. 2. Increase in epidural fat in patients on prolonged steroid therapy. The nature of the compression is evident on CT and/or MRI, and is rarely if ever significant.

l. pachymeningitis, which previously occurred as a complication of syphilis, but is usually now an exceedingly rare fibrosis of the dura of unknown aetiology. 2. Neoplastic infiltration by carcinoma or lymphoma. 3. Deposition of mucopolysaccharides, when metabolism of these substances is abnormal due to any one of a group of inherited enzyme defects. Typical clinical and biochemical abnormalities have usually been established by the time that cord compression occurs. 4. Extension of intracranial thickening of various causes.

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ Spinal dural arteriovenous fistula with intradural drainage. Fig. 54.47 Sagittal (A) and coronal (B) MRI with T 2 - weighted contrast in the cervical region showing markedly enlarged intradural veins.

Previous Page

1670

CONTENTS

Next Page


Fig. 54.48 Dural arteriovenous fistula, left second lumbar angiogram. (A) AP projection. (B) Lateral projection. The fistula is on the left side of the dura at L2 level and it drains superiorly through a vein ascending along the posterolateral aspect of the subarachnoid space to enter the posterior coronal venous system at T12 level. Veins then pass around the left side of the cord to fill the anterior coronal plexus also. Large arrow = fistula; small arrows = draining vein.

Fig. 54.49 AVM of the spinal cord. (A) Left vertebral angiogram. (B) Left sixth intercostal angiogram. The AVM is at the cervico-thoracic junction. It i s supplied (A) from above by the enlarged anterior spinal artery descending from the cervical region and (B) by an enlarged tortuous vessel, presumably a posterior spinal artery, ascending along the posterolateral aspect of the cord.


All these conditions cause concentric narrowing of the subarachnoid space which may give rise to difficulty with spinal puncture. They are well shown by MRI, but may requir intravenous contrast.

An arteriovenous malformation or fistula draining into the spinal venous plexuses generally causes prominent dilatation of the veins, which become abnormally tortuous extending over several spinal segments and sometimes throughout the entire spinal canal. They are usually evident on T 2 - weighted MRI as low-signal linear structures. High signal within the spinal cord, usually throughout its l ower half or so down to the conus medullaris, usually is present also, due to oedema and/or infarction. Myelography will show the abnormal vessels; sometimes minor cord swelling is also present ( Fig. 54.47). Spinal angiography is necessary to localise the arteriovenous shunt prior to surgery or embolisation. The majority of spinal arteriovenous fistulas occur in middle-aged men and present with progressive paraparesis, usually with incontinence, often with pain; sometimes the disability increases with exercise. These fistulas are usually in the dura (Fig. 54.48), generally in the dorsal or lumbar

regions, but they may he at any level from the sacral region to the foramen magnum. They are often related to a nerve root and are supplied by one, or occasionally two or three, radiculopial arteries, and drain through one or two veins to the coronal plexus. These fistulas are eminently suitable for embolisation procedures or surgery. Angiomas of the spinal cord itself (Fig. 54.49) have a wider age range, not infrequently presenting in childhood, sometimes acutely with haemorrhagc. They involve any region of the spinal cord, most frequently the cervical, and tend to have multiple feeding arteries, often including the anterior spinal artery.

Cyst-like dilatations of nerve root sheaths are a common normal variation, especially in the lumbar and sacral regions. Such root cysts and another anatomical variation, the conjoined root sheath, containing the nerve roots for two or more adjacent levels, have assumed significance because on plain CT they may superficially simulate an extruded disc fragment; the latter is, however, always more dense than CSF. The term 'arachnoid cyst' refers to two separate conditions: 1. Intrtulural arachnoid cyst: l oculated and expanding collections of CSF, usually within or adjacent to the septum posticum,

Previous Page

CONTENTS


which communicate with the free subarachnoid space through narrow necks. 2. Extradural arachnoid cysts, which are herniations of the arachnoid through defects in the dura. They may remain within the spinal canal, expanding it, or extend through intervertebral foramina, causing erosion of the pedicles and root pain.

1671

Infiltration of the meninges with chronic inflammatory cells and fibrosis most commonly follows chemical damage to the meninges, most often in the past by iophendylate (Myodil, Pantopaque), once widely used for myelography, or ionic myelographic contrast media, but occasionally by drugs used in therapy, including steroids and

Cysts at the level of the spinal cord are most frequently situated posteriorly in the dorsal region, and cause backache and cord compression, sometimes varying with posture. A particular variation is the intrasacral arachnoid cyst, which may cause gross expansion of the sacral spinal canal with marked thinning of the hone. Such

spinal anaesthetics; other causes are subarachnoid haemorrhage and i nfections. It used to be thought that it occurred very commonly after l umbar disc surgery, but it is now known that the cause was the contrast media used in the preoperative myelogram (Fitt & Stevens 1 995). Modern myelographic media do not cause arachnoiditis, and cysts may extend through into the pelvis and cause anterior disarachnoiditis has almost disappeared as a complication of extradural placement of the rectum. They are evident on MRI as masses spinal surgery. As well as the usual pyogenic organisms and tubercugiving signals of either similar or slightly greater intensity than the l osis, in which it generally occurs in association with extradural subarachnoid CSF. They displace adjacent nerve roots and seginflammation, sarcoidosis, schistosomiasis, cysticercosis, cryptococments of the spinal cord, which is often deformed by pressure from cal infection and coccidioidomycosis have all been implicated. The the cyst. condition is associated with a variable amount of demyelination, The intradural arachnoid cysts form long masses on myeloaxonal and nerve cell degeneration. graphy, usually posterior to the spinal cord, causing partial or comArachnoiditis is best diagnosed by MRI, high-resolution axial plete obstruction in the dorsal region. The size of the cyst may be T2- weighted fast spin-echo acquisitions usually being the most shown to vary with posture. Though the cyst communicates with helpful. Chronic adhesive arachnoiditis results in adhesions bethe subarachnoid space, opacification may be slow; the cyst may be tween intradural roots and with the wall of the thecal sac, producing revealed only by subsequent CTM or sometimes by introducing mass-like clumps, or incorporation of roots into the thecal wall additional contrast medium by cervical puncture. creating an empty thecal sac. The causal part of the thecal sac Extradural arachnoid cysts are also generally large posteriorly usually is involved. It can be very extensive, however, with dissituated masses, displacing the theca anteriorly or anterolaterally. tortion and cavitation of the spinal cord. The organised adhesion The neck of the cyst is usually situated superiorly, and cyst filling, may calcify or even ossify, producing arachnoiditis ossificans. which may he free or very delayed, may be best achieved in the Clinico-radiological correlation generally is poor. supine position. Where communication is free the cyst may be seen at myelography to fill and empty with posture; slow-filling cysts may be confirmed by CTM. The cysts are usually well shown by MRI and may be suspected Fitt, C. J., Stevens, J. M. (1995) Post-operative arachnoiditis diagnosed by when different flow conditions are present in the cyst and the free high resolution fast spin echo MRI of the lumbar spine. Neuroradiology, subarachnoid space. Damage to the spinal cord by pulsatile expan37,139-145. Ghelman, B. (1988) Discography. In: Kricun, M. (Ed), Saunders. W. B. (eds) sion at the closed end of the cyst may be evident as myelopathy on I maging Modalities in Spinal Disorders, pp. 538-556. Philadelphia. T,-weighted sequences. In both types of cyst, the diagnosis also is Hueftle, M. G., Modic, M. T., Ross, J. S., et al (1988) Lumbar spine: evident on CTM. The density of the fluid within the cyst may be postoperative MR imaging with Gd-DTPA. Radiology, 167, 817-824. si milar to or less than that in the subarachnoid space, but the entire Lacatelli, B., Laurena, R., Ballard, P., Mark, A . S. (1999), MRI in vitamin B,; deficiency myclopathy. Canadian Neurological Sciences, 26, 60-63. extramedullary lesion will be homogeneously opacified, and the Modic M. T., Steinberg P. M., Ross J. S., et al (1988) Degenerative disc membrane between it and the subarachnoid space may be visible. disease: assessment of changes in vertebral bone marrow with MRI. Multiple intradural arachnoid diverticula, which may erode the Radiology, 166, 1 97-199. margins of the lumbar canal or sacrum and be associated with Ross, J. S., Masaryk, T. J., Modic, M. T., et al (1987) Magnetic resonance i maging of lumbar arachnoiditis. A merican Journal of Neuroradiologv, 8. symptoms of radiculopathy, are an uncommon accompaniment of 885-892. ankylosing spondylitis. Herniation of the spinal cord through a Stark, D. D., Hendrick, R. E., Hahn, P. F., Ferrucci, J. T. (1987) Motion dural defect may simulate an arachnoid cyst. In this rare condition artifact reduction with fast spin-echo imaging. Radiology, 164, 183-191. the hernia is usually anteriorly and the spinal canal may be widened Stevens, J. M. (1995) I maging of the spinal cord. A review. Journal of locally with the spinal cord hernia extruding into the focal expanNeurology, Neurosurgery and Psychiatry, 58, 403-408. Stevens, J. M., Kling Chong, W ., Barber, C. et al (1994) A new appraisal of sion. This is sometimes directly visualised by MRI, but more often abnormalites of the odontoid process associated with atlanto-axial only ventral angulation of the spinal cord is shown, sometimes subluxation and neurological disability. Brain, 117, 133-148. wrongly diagnosed as posterior compression by an intradural arachUllrich, C. G., Binet, E. F., Sanecki, M. G., Kieffer, S. A . (1980) noid cyst. The èmpty' suharachnoid space may be mistaken for a Quantitative assessment of the lumbar spinal canal by computed tomography. Radiology, 134, 1 37-143. cyst on superficial analysis.


Previous Page

CONTENTS

Next Page

.



.

Previous Page

CONTENTS

Next Page

55

The methods currently used include: 1. Magnetic resonance angiography (MRA) 2. CT angiography 3. Sonographic vascular imaging 4. Intravenous digital subtraction angiography (IV DSA) 5. Direct intra-arterial angiography.

Direct intra-arterial angiography using standard film has for many years provided the gold standard for angiography but is now largely replaced by minimally or non-invasive techniques in well-equipped units. Some workers, however, still prefer direct angiography for special purposes such as the investigation of subarachnoid haemorrhage. Direct angiography also remains essential for interventional work (see Ch. 56). The different shades of meaning attached to the term `noni nvasive' have been described elsewhere in this text (see Ch. 12, p. 319, and Ch. 15, p. 416). Some procedures so referred to are better described as minimally or mildly invasive, since intravenous

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ Fig. 55.1 (A) MRA showing all four neck vessels in AP view. (B) Segmental MRA study of right carotid and vertebral in lateral view.

Fig. 55.2 MRAs. (A) Normal left carotid and vertebral. (B) Stenosis of right common carotid bifurcation. The stenosed channel is not outlined owing to low flow.

1673

Previous Page 1674

CONTENTS

Next Page


(I.V.) constrast and not direct intra-arterial contrast is used. These procedures include contrast enhanced magnetic resonance (CEMR), I.V. DSA and I.V. contrast enhanced spiral CT. The procedures may also be described as without radiation hazard (MRI and ultrasound) or with radiation hazard (DSA and CT).

This technique has expanded greatly in recent years (see Ch. 15 for more details). Two basic strategies are used for imaging, namely ti me of flight effects (TOF) and phase contrast (PC). These can be acquired either multislice (two-dimensional) or three-dimensionally (3D). TOE is most frequently used in imaging the cerebral circulation. Background suppression is improved by utilising magnetisation transfer (MTC), in the form of a broad-spectrum prepulse. Fixed and `walking' presaturation slabs are also helpful to obtain selectively images of either arteries or veins. Image quality in 3D techniques can be enhanced by reducing the size of the image volume, which permits better visualisation of slower flowing protons i n smaller arterial branches. Carotid and cerebral imaging is often best performed using an overlapping 3D volume technique, known as multiple overlapping thin-slice arteriography (MOTSA). Enhanced visualisation is also achieved with interactive vascular i maging (IV!), which permits segmentation of individual vessels or segments of vessels for multidirectional maximum intensity pixel ( MIP) displays. Intravenous enhancement with gadolinium is now widely used in MRA. The technique most widely used is threedimensional contrast-enhanced MRA (3D CE MRA ). It is necessary to acquire a time-resolved sequence so that images from the arterial phase are acquired before venous enhancement occurs. Image sets are acquired at approximately 5 s intervals. Acquisition planes generally are coronal or sagittal rather than transverse. Between 20 and 25 ml of contrast is injected into a peripheral rather than central vein. Digital subtraction of precontrast images and other postprocessing may be used to improve the display and MIP projections and 3D surface renderings are produced. The technique is most commonly used to display the aortic arch and origins of the great cerebral vessels. A customised neurovascular neck coil also is a great advantage to ensure good regional coverage.

Fig. 55.3 MRAs. (A) Right internal carotid thrombosis. (B) Severe right i nternal carotid stenosis. Owing to low flow the proximal segment does not outline. Perfusion imaging is also possible with both contrast-enhanced MRA and CT angiography. After an intravenous contrast bolus, a series of images are acquired over time, often of just a single slice, from which, with appropriate computer-based processing, parameters such as regional time to peak contrast, relative regional cerebral blood flow and blood volume can be calculated and presented as colour coded maps. Spatial resolution is generally better than in nuclear medicine-based techniques. These techniques are widely

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ used, most often in research settings.

2. CT angiography

(Figs 55.5,55.6; see also Ch. 15)

Spiral CT incorporates a slip ring coupling between the rotating and stationary portions of the gantry to allow for continuous tube rotation and data acquisition as the patient is transported through the

Fig. 55.4 MRAs. (A) Normal internal carotid bifurcation and proximal segments of the anterior and middle cerebral arteries in lateral view. Lateral (B) and rotated lateral (C) views of a posterior communicating aneurysm.

Previous Page

CONTENTS

Next Page

VASCULAR IMAGING IN NEURORADIOLOGY

Fig. 55.5 Spiral CT of the neck vessels showing sequestrated study of right carotid bifurcation. (A) In the AP view the stenosis is obscured by a calcified plaque. (B) The image is computer rotated through 180° to show the stenosed lumen. (C) 3D CT reconstruction. (Courtesy of Dr A. AIKutoubi.)

1675

gantry. The dataset produced can be reconstructed at arbitrary intervals to generate overlapping sections, and the rapid acquisition time enables an entire vascular territory to he imaged within 30 s. Imaging the cerebral circulation requires the injection of approximately 90 ml of contrast medium (300 mg/ml iodine) through a cannula or catheter inserted into a peripheral arm vein, over 20-30 s. The timing of the spiral CT imaging depends on circulation time and usually requires a test dose to monitor opacification of the vessels of interest, as in CE MRA. Collimator width and table speed have important consequences on the quality of the results. As usual there is a trade-off between the volume imaged and spatial resolution, and collimator width and signal-to-noise ratio. I mages can be viewed as axial sections of any prescribed thickness. They can also he presented, like magnetic resonance angiograms, as MIP projections or as threshold shaded surface displays (SSDs). Segmentation of the arterial system from other dense structures such as bone is necessary and can he very timeconsuming, but semiautomated methods are now on line; image editing is a major part of this type of angiography. 3D CT angiography in neuroradiology is used mainly to image the extracranial carotid and vertebral arteries, and to detect and display intracranial aneurysms. It generally is used when MRA is not an option.

This is widely used for imaging the extracranial carotid vessels and can also be used to measure blood velocity in the middle cerebral and other intracranial arteries. Grey scale imaging is used to visualise the vessels and to document the state of the vessel wall. Doppler spectral analysis is used to determine blood flow velocity. Colour flow mapping is also used to enable a rapid survey of arterial segments to find sites where further spectral analysis may be beneficial. The technical details of vascular sonography are given in Chapter 15. Considerable operator skill and time are required.

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ Fig. 55.6 3D CT reconstruction of circle of Willis and adjacent vessels showing bilateral aneurysms. (A) AP view. (B) Computer-rotated view. (Courtesy of Dr A. Al-Kutoubi.)

Digital techniques have completely replaced film-screen combinations. Computed fluoroscopy enables rapid manipulation of multiple images with subtraction of the bones and other tissues from the opacified blood vessels. Adequate arterial studies may be obtained with relatively low intra-arterial concentrations of contrast medium. Although enjoying widespread clinical use in the early 1980s, i nitial enthusiasm for intravenous DSA soon became tempered with caution as limitations became apparent. The great advantage of intravenous techniques is that it is very safe to the patient. The procedure does not require hospitalisation and is performed on an outpatient basis. It is therefore cost-effective. It generally requires placement of a catheter via an arm vein, into the superior vena cava or left atrium. Contrast resolution is not as good as with direct arteriography. For intracranial lesions all four vessels are opacified at the same ti me, and in lateral views there is superimposition of the vessels of the two sides. With uncooperative patients, movement can degrade the images to an unacceptable level. Large boluses of contrast agent are used for intravenous injections so that multiple projections require quite large total doses. These drawbacks led to some modification of the early enthusiasm for

Previous Page 1676

CONTENTS

Next Page


the intravenous technique, and many units switched to the use of digital techniques with the more dangerous intra-arterial i njections. The main uses of intravenous DSA are: 1. For arch aortography for neck vessel study (Fig. 55.7). 2. For selected intracranial studies, e.g. confirmation of large aneurysms, angiomas, fistulas, or vascular tumours suggested by CT, or to show the relationships of the major arteries to a l arge pituitary adenoma, or to show involvement of a major venous sinus (Fig. 55.8). 3. For some postoperative and follow-up studies, particularly after aneurysm clipping and treatment of stenoses. 4. For dural sinus phlebograpy. The intravenous technique ensures that all vascular territories are opacified together, allowing the generation of full-contrast venograms of good temporal resolution (Fig. 55.8). For the reasons stated above, intravenous studies of the extracranial neck vessels are suboptimal in a proportion of patients. This proportion was stated to be 15-20% in the early studies, but with

Fig. 55.8 DSA study showing the cortical veins, sagittal and lateral sinuses. Note large vein of Trolard and smaller vein of Labbe.

i mprovements of apparatus and technique should now be less than 1 0%: however, it may be necessary to use more than one projection to ensure this, with consequent increase in total contrast medium dose.

This was originally performed by percutaneous puncture of the common carotid artery in the neck. This technique became obsolete with the increasing use of percutaneous catheterisation. Direct carotid angiography is now generally practised by the introduction of catheters into the lumen of the common carotid artery from a percutaneous transfemoral approach. This technique permits selective catheterisation of the internal and external carotid arteries; however, caution should be observed in passing catheters into the internal carotid artery in middle-aged and elderly patients, particularly if there i s any suspicion of atheromatous internal carotid artery stenosis. About 10 ml of contrast medium are usually injected within 1 -2 s into the common carotid artery. A slightly smaller quantity, about 8 ml, is adequate for selective injection of the internal carotid artery. Selective injection of the external carotid artery can be accomplished with a similar amount of contrast medium (8 nil). In the case of the external carotid artery the injection can be made more slowly, taking 2-4 s. Carotid angiography can be carried out either under local anaesthesia or under general anaesthesia. In cases where more than one artery is to be injected at one session, or if the patient is likely to be very uncooperative or nervous, general anaesthesia is more commonly used. However, the advent of DSA and low-osmolarity contrast media has further reduced the need for general anaesthesia. As with other forms of arteriography, it is important to obtain rapid serial films as the contrast medium passes through the circulation. Many types of apparatus were available for achieving this, but today most imaging is performed by DSA. Different `programmes' can be used and tailored to the individual case. A standard series could consist of seven images taken at I per second for 7 seconds and timed so that arterial, capillary and venous phases of the angiogram are all covered. With angiomas, arteriovenous fistulas and vascular tumours the early images would be taken at two or three frames per second to allow for the rapid arteriovenous shunting.

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ Fig. 55.7 The neck vessels shown by DSA following an intravenous injection of contrast medium. The internal carotid and vertebral origins are normal.

Previous Page

CONTENTS

Next Page

VASCULAR IMAGING I N NEURORADIOLOGY

1677

Vertebral angiography For catheter demonstration of the carotid or vertebral artery, several alternative catheters are available, and different workers have preference for different types. We prefer the Mani (5F) headhunter catheters, and even smaller catheters (4F and 3F) are now available. The smaller catheters have the added advantage of being less liable to form clot or to damage arterial walls. Normally the left vertebral artery is easier to catheterise than the right, but if one side proves difficult the other can usually he entered. It is normally only necessary to catheterise one vertebral artery, as a forced injection will fill the contralateral vertebral artery by reflux as well as filling the basilar artery. Sometimes the catheter can be passed into the subclavian artery but will not enter the vertebral artery. In these cases an injection can be made into the subclavian artery and an indirect vertebral arteriogram obtained. In patients whose vessels are too tortuous for the suhclavian arteries to be catheterised from below, or in whom diseased iliacs prevent passage of a catheter, success can still be obtained by transaxillary catheterisation. In about 5% of cases the l eft vertebral artery arises direct from the aortic arch. If no vertebral artery can be found arising from the left subclavian this condition should be suspected, the catheter withdrawn into the arch, and an attempt made to manipulate it into the anomalous vertebral. If this fails, the right vertebral should be attempted. For direct injection into the vertebral artery we use only 6-8 nil of contrast medium. Low-osmolarity contrast media are now mandatory. As with carotid angiography, improved detail can he obtained by the use of magnification with a fine-focus tube and by the use of subtraction films. Where the first part of the subclavian artery is injected, about 1 5 ml of contrast medium are injected in about 2 s. It may help to obtain better filling of the vertebral artery in these cases if the appropriate brachial artery is occluded at the time of injection. Arch aortography It is still contended by some workers that cerebral symptoms may arise from stenosis or thrombosis of the origins of the internal carotid arteries or of the vertebral arteries. It has also been claimed that intrathoracic lesions of the innominate, left common carotid and left subclavian arteries can be a cause of cerebral symptoms. I njection of contrast medium into the aortic arch enables all the Great vessels in the thorax and the vertebral arteries in the neck, together with the carotid bifurcations, to be demonstrated. The important features of arch aortograms are illustrated in Figure 55.9. It is usual to take rapid serial films with the supine patient rotated about 45 ° to the right and the head turned to the right l ateral position. This normally avoids superimposition of the carotid and vertebral arteries and shows the carotid bifurcations clearly. DSA by intravenous injection has also been widely used for arch aortography. The main indications for direct intra-arterial angiography in wellequipped departments are now limited as follows:

Fig. 55.9 Arch aortogram in right posterior oblique position. There is a congenital anomaly, in that the right subclavian artery arises distally from the aortic arch (arrowheads). Its origin is superimposed on the left subclavian origin. Both carotid bifurcations are well shown.


I. To elucidate the arterial supply and venous drainage of arteriovenous shunts for treatment planning, not simply for diagnosis. 2. Prior to endovascular therapeutic procedures such as preoperative embolisation of tumours, treatment of aneurysms or arterial stenoses, etc. 3. Investigation of definite intracranial subarachnoid haemorrhage ( may not always he necessary).

4. Very occasionally as an investigation of last resort. Most often this will he in context where less invasive imaging has been unsatisfactory. Spinal angiography This procedure involved selective angiography and a very careful radiographic technique. The arterial supply to the spinal cord is illustrated in Figure 55.10. The main irdication today is in interventional techniques (see Ch. 56).

The general complications of angiography have been discussed in Chapter 15. Intra-arterial techniques carry the particular risk of i nducing temporary or permanent brain damage. The risks are highest in the elderly and those with pre-existing vascular disease. I n the past the commonest causes of brain damage were the use of toxic contrast agents, and local damage to the carotid or vertebral arteries. Nowadays, brain damage attributable to the contrast medium is virtually non-existent, and local trauma is also relatively i nfrequent. The commonest cause of damage is embolism from a clot forming in or around the catheter tip. The risk of permanent

Previous Page

1678

CONTENTS

Next Page


Fig. 55.10 Normal blood supply to the spinal cord as seen (A) in anterior view and (B) in lateral view. (C) Blood flow currents in the longitudinal spinal arteries. (After DiChiro & Wener 1973.)

disability or death has been variably estimated over the years, but two of the most recent assessments have put this at 1.2% and 5.2% (1995 and 1994). The best results are obtained where the procedures are carried out by experienced workers only. It must also be stressed that several recent studies from reputable centres have i ndicated that intra-arterial DSA is no safer than conventional film-screen techniques and, even more importantly, that arch angiography carries the same or even slightly high , r stroke risk

than selective arteriography. These serious complication rates make i t obvious that cerebral angiography should only be practised by experienced personnel who are fully aware of these hazards. Moreover, intra-arterial techniques should only be resorted to when absolutely essential on clinical grounds, and when the information required for patient management is not obtainable by less invasive techniques. Despite a few early reports to the contrary, which were not confirmed by later studies, intravenous DSA carries no risk of stroke, but carries a small risk of usually insignificant systemic side-effects, the most frequent being angina (0.6%). Despite the earlier claims to the contrary, it is most likely that the systemic side-effects are no more frequent in intravenous DSA than in a double-dose contrast-enhanced CT or intravenous urogram. The same is probably true for 3D CT angiography. Carotid sonography and MRA carry no risk of side-effects. Selective spinal angiography carries the risk of damage to the spinal cord and must be practised with caution. Only small doses of low-osmolar contrast medium should be injected into the posterior intercostal and lumbar arteries. In patients with arteriovenous fistulas, worsening myelopathy has been reported in up to 20% of cases. Spinal cord damage can also result from nonneuroradiological angiography, such as bronchial, thyroid and parathyroid angiography.

The features of the normal internal carotid arteriogram are illustrated in Figures 55.11 and 55.12. The normal circulation time from i njection of the contrast medium into the internal carotid artery to its disappearance from the veins of the brain varies with the individual and averages 5-7 s. In certain pathological conditions it may of course be appreciably slower. With angiomatous malformations or arteriovenous fistulas it may he extremely rapid. It is customary to divide the angiogram into four phases. The first phase lasts 1-2.5 s and is referred to as the arterial phase. Films taken during this period show the arterial tree. The second or capillary phase lasts about I s or less but is rarely clearly defined on the X-ray film, as


Fig. 55.11

(A) Normal internal carotid arteriogram-lateral view. (B) Diagram to illustrate (A). 1 . Anterior cerebral artery; 2. frontopolar artery; 3. pericallosal artery; 4. middle cerebral artery and its branches; 5. anterior choroidal artery; 6. posterior communicating artery; 7. posterior cerebral artery; 8. callosomarginal artery; 9. ophthalmic artery.

Previous Page

CONTENTS

Next Page


1679

Fig. 55.12 (A) Normal internal carotid arteriogram-AP view. (B) Diagram to illustrate (A). 1. Anterior cerebral artery; 2. middle cerebral artery and its branches; 3. posterior cerebral artery; 4. anterior choroidal artery; 5. lenticulostriate arteries; 6. sylvian point. there is usually some late arterial or early venous filling superi mposed. The third and fourth phases are characterised by venous filling and last 4-5 s. They include the early and late phlebograms, respectively. The early phlebogram outlines the superficial veins of the hemisphere, and the late phlebogram shows the deep veins. The blood flow through the internal carotid artery is considerably greater than through the external carotid artery. Thus, following i njection of the common carotid artery, the intracranial vessels are usually visible in the lateral films a second or so before the branches of the external carotid become superimposed. The experienced worker has no difficulty in differentiating branches of the external from branches of the internal carotid, as the distribution of these vessels follows a standard pattern. The middle meningeal artery and its branches which stem from the internal maxillary branch of the external carotid also fill slightly later than the internal carotid branches. Early filling of the middle meningeal artery together with internal carotid branches can occur if the vessel is hypertrophied to supply a meningioma or angioma. This point is further commented on below. The cervical internal carotid artery has no branches, but occasionally the ascending pharyngeal or other branches of the external carotid may arise from it, as may the rare proatlantal intersegmental artery and the hypoglossal artery (Figs 55.33, 55.34). Sometimes the internal carotid artery shows a prominent loop as it lies lateral to the oropharynx, and this is known as the 'tonsillar loop' (Fig. 55.13). It can be regarded as a normal variant. The normal internal carotid artery also forms a loop as it lies in the l ateral wall of the cavernous sinus, which is usually referred to as the ` carotid siphon'. The suprasellar portion of the internal carotid artery just before its bifurcation is normally inclined laterally (from below upward) as seen in the anteroposterior view. In this view the major branches of the anterior and middle cerebral arteries pass, respectively, medially and laterally. These two vessels, together

with the termination of the internal carotid, resemble the letter T when seen in this projection (Fig. 55.12). The major branches of the internal carotid artery all arise above the cavernous sinus, but there are some minor branches arising in its intrapetrosal and precavernous segments. The former include the caroticotympanic and


Previous Page 1680

CONTENTS

Next Page


Fig. 55.14 Right ophthalmic artery. (A) In lateral view. (B) In axial view. 1. Supraorbital branch; 2. main artery with ethmoidal branches; 3. lacrimal branch; 4. central retinal branch.

pterygoid (vidian) arteries, which are rarely recognisable at angiography. The precavernous and intracavernous branches include: I . The important meningohypophyseal trunk (Fig. 55.51), which gives rise to: a. the tentorial artery b. the dorsal meningeal artery c. the inferior hypophyseal artery The tentorial artery may he enlarged and easily recognisable with tentorial meningiomas. 2. The inferior cavernous artery, which supplies the wall of the cavenous sinus and its contents and anastomoses with the middle meningeal artery. 3. The capsular artery. Under normal circumstances these arteries are only recognisable in high-quality subtraction films with magnification, as are the tiny superior hypophyseal arteries which arise above the sella. The ophthalmic artery arises from the internal carotid just above the sella and medial to the anterior clinoid process. It passes forward to the optic canal beneath the optic nerve and then into the orbit. A good quality film will show its terminal branches outlining the posterior aspect of the globe (Fig. 55.14). Its terminal supraorbital branch supplies a small cutaneous area above the medial aspect of the orbit. The artery of the falx arises from the anterior ethmoidal branch of the ophthalmic artery and passes through the cribrifonm plate to supply the anterior part of the falx. It may be markedly hypertrophied to supply meningiomas or arteriovenous malformations. The posterior communicating and posterior cerebral arteries fill i n only about a third of common carotid arteriograms. However, they fill in a higher proportion of cases following selective injection of the internal carotid artery. In other cases, the dominant blood supply to the posterior cerebral artery is through the vertebrobasilar circulation. Sometimes the origin of the posterior communicating artery, as seen in lateral view, is slightly expanded. This appearance, known as the ìnfundibulum' of the posterior communicating artery, has been regarded as a normal variant. On the other hand, there is some pathological evidence that it may be associated with a defect in the vessel wall and predispose to aneurysm formation. The anterior choroidal artery is small but is readily identifiable. Its normal position in both lateral and anteroposterior view is illustrated in Figure 55.15. Displacements of this vessel were of considerable help in identifying and localising mass lesions.


Fig. 55.15 Anterior choroidal artery (arrowed). (A) In AP projection. (B) In lateral projection. The normal anterior choroidal artery arises from the posterior

aspect of the internal carotid just distal to the origin of the posterior communicating artery, but rarely it arises from the middle cerebral or posterior cerebral artery. The artery is directed backward and medially to the medial aspect of the anterior part of the temporal l obe. It then passes round the uncus and turns laterally and backward into the choroidal fissure to enter the temporal horn. There it supplies the choroid plexus, which is sometimes seen as an illdefined blush of contrast. As with the posterior communicating artery, an infundibulum is occasionally seen at the origin of the artery from the internal carotid artery. The lenticulostriate arteries arise in two groups of 2-4 tiny arteri es, the medial and the lateral. Normally they stem from the upper surface of the middle cerebral artery trunk and pass directly up through the anterior perforated substance into the Basel ganglia and i nternal capsule. Some of them may also arise from the internal carotid bifurcation or the origin of the anterior cerebral. They are

Previous Page

CONTENTS

Next Page


1681

readily recognised in good-quality anteroposterior films, but are usually obscured by larger overlying vessels in the lateral view. The normal disposition of these vessels is illustrated in Figure 55.16. They pass upward and medially for a short distance and then laterally in an arc that is concave inward. Masses or haematomas in the region of the external capsule displace these vessels medially, while masses in the basal ganglia or internal capsule distort or separate them. The anterior cerebral artery (Figs 55.1 1, 55.12) passes medially from the bifurcation of the internal carotid to reach the midline. This small segment of the artery lies above the optic chiasm or optic nerve and is usually convex upward. It is best seen in the anteroposterior projection. Near the midline it is joined by the anterior communicating artery to its fellow on the opposite side. Beyond the anterior communicating artery the anterior cerebral turns forward and upward in the interhemispheric fissure. It passes around the anterior aspect of the corpus callosum and turns backward along its upper surface. It continues as the pericallosal artery to the back end of the corpus callosum. In lateral view the segment between the anterior communicating artery and the gems of the corpus callosum usually has a gentle concavity downward. This is variable in degree and should not be mistaken for displacement by a mass. A few tiny perforating branches are given off from the proximal horizontal segment of the anterior cerebral artery, but these are variable and difficult to identify on angiograms. The frontopolar branch is generally the first major branch of the anterior cerebral and arises proximal to the knee or bend of the vessel around the corpus callosum. This vessel passes forward and upward towards the anterior pole of the frontal lobe and is best seen i n lateral view. It divides into 2-3 branches which pass to the superomedial margin of the hemisphere and over on to the convexity. The next major branch of the anterior cerebral is the callosomarginal. This usually arises near the gems and passes backward and upward, giving off' anterior, middle and posterior internal frontal branches. It terminates in the paracentral branch around the paracentral lobule. The branches just described are variable and may arise directly from the anterior cerebral. The callosomarginal artery lies in the callosomarginal sulcus for part of its course, and while in the sulcus it may be lateral to the in idline. This should not be mistaken for a true displacement by a mass. The terminal branches of the artery, like those of the frontopolar artery, reach the superomedial border of the hemisphere and pass over it to the convexity. Here they anastomose with terminal ascending branches of the middle cerebral artery. The pericallosal artery represents the continuation of the anterior cerebral after it has given off the major branches just described. Normally it is fairly closely applied to the upper surface of the corpus callosum and terminates in a precuneal or posterior callosal branch. The proximal segment of the anterior cerebral artery is occasionally hypoplastic, forming one of the many variations of the circle of Willis. In these cases the major part of the vessel fills from the opposite side through the anterior communicating artery. Such congenital hypoplasia should be distinguished from spasm associated with subarachnoid haemorrhage. The middle cerebral artery passes laterally and slightly forward from the bifurcation of the internal carotid artery. This segment is best seen in the AP view and is horizontal or convex upward. It gives rise to the lenticulostriate arteries, which have just been

Fig. 55.16 Lenticulostriate arteries in AP projection. They are difficult to identify in lateral projection because of the superimposed middle cerebral vessels. The anterior temporal and orbitofrontal branches are also shown passing downward. described, and also to anterior temporal and orbitofrontal branches (Fig. 55.16). These can arise from a common trunk. About 2 cm from its origin it reaches the insula (or island of Red), where it usually bifurcates or trifurcates. The main branches pass backward over the surface of the insula and are hidden by the opercula. These branches lie deep in the Sylvian fissure and deep to the external surface of the brain. The branches loop downward on the undersurface of the inferior parietal operculum and then pass out through the Sylvian fissure to reach the surface of the hemisphere. These


Fig. 55.17 Relationship of the middle cerebral artery and its branches to the insula in a sagittal brain section. (A) Insula. (B) Middle cerebral artery.

Fig. 55.18 Relationship of middle cerebral vessels to skull vault in the AP projection. Arrow points to the angiographic sylvian point which represents the posterior limit of the insula. Distance A, from the skull vault to the lateral aspect of the i nsula, varies from 20 to 30 mm. Distance B, from the sylvian., point to the skull vault, measures 30-40 mm. (After Taveras 1969.)

Previous Page 1682

CONTENTS

Next Page


downward loops can be identified in the lateral angiograms, and the margins of the triangular insula can thus be defined. The surface of the insula is also identified in the same way on the AP film (Fig. 55.17). The appearance of these vessels and their relationship to the skull vault in the frontal projection were of considerable i mportance in detecting displacement by tumours and other masses. Taveras refers to the posterior limit of the sylvian triangle as the angiographic sylvian point' and shows how this can he identified in the frontal projection. The last artery to emerge from the insula occupies at its medial bend the position of the angiographic sylvian point (Fig. 55.18). The sylvian vessels and sylvian point should appear symmetric on the two sides. As the branches of the middle cerebral artery emerge from the sylvian fissure they turn, in the case of the anterior branches, upward, and in the case of the posterior branch, backward. The anterior or ascending branches pass upward in the prerolandic areas to supply the lateral surface of the hemisphere in the frontal and parietal regions. These arteries are rather variable and tortuous. The posterior branches are more regular and three major arteries can usually be identified. From above downward these are the posterior parietal, angular and posterior temporal arteries.

The branches of the middle cerebral artery supply the whole of the convexity of the hemisphere except a 2.5 cm band near the sagittal fissure, and the occipital lobe.

The superficial cerebral veins arc very variable in position and distribution. Most of the veins of the hemisphere run upward and backward to end in the superior longitudinal sinus (Figs 55.8, 55.19). They usually enter the sinus against the direction of the blood flow, which is from before backward. Two large veins are named by the anatomists. The first, the vein of Trolard, i s a large vein which passes upward and backward over the hemisphere to enter the superior sagittal sinus in the parietal region. The second is the vein of Labbe. This large vein passes horizontally across the temporal region to enter the lateral sinus. In most angiograms one or other of these veins can be recognised but it is unusual for both veins to be present in the same angiogratn (Fig. 55.8). The smaller middle cerebral veins run forward in the sylvian fissure and then to the spheroidal ridge to end in the sphenoparietal sinus, which drains to the cavernous sinus. The deep cerebral veins fill slightly later than the superficial veins just described. They are of considerable importance in angiographic diagnosis because their position is much more constant than that of the superficial veins. The internal cerebral veins lie one on each side of the midline. They commence just behind the foramen of Monro and pass backward in the roof of the third ventricle, lying in the tela choroidea. Two smaller veins drain into each internal cerebral vein at its commencement. These are the septal vein, which runs on the medial surface of the frontal horn, and the striothalamic vein, which runs in the floor of the lateral ventricle to reach its lateral wall (Fig. 55.20). In the lateral view the junction of the striothalamic vein with the septal vein forms the origin of the internal cerebral vein and is known as the `venous angle'. This provides, in most cases, a recognisable landmark, as it normally lies just behind the foramen of Monro; however, the anatomy is not constant and in some cases the striothalamic vein may enter the internal cerebral vein more posteriorly. At its posterior end the internal cerebral vein unites with its fellow on the opposite side and enters the vein of Galen. This is constant in position and is easily identified. Just proximal to the vein of Galen other small tributary veins may enter the internal cerebral veins. The auricular vein drains the area of the trigone. The basal vein arises anteriorly above the sella and passes round the midbrain to enter the back end of the internal cerebral vein. Less commonly, it drains directly into the vein of Galen. Ventricular size can be assessed in the lateral view, as the septal, striothalamic and auricular veins all commence in the walls of the ventricles. Thus the shape of the ventricle can be roughly outlined by drawing in a line around the origins of these vessels. These small veins drain the subependymal veins of the cerebral white


matter. Normally the latter are very tiny and are rarely recognisable i n a routine angiogram. In certain pathological conditions, however,

Fig. 55.19 (A) Phlebogram showing superficial and deep veins. (B) Diagram to illustrate (A) 1. Internal cerebral vein; 2. vein of Galen . 3. striothalamic vein; 4. septal vein; 5. venous angle; 6. inferior sagittal sinus; 7. straight sinus; 8. superficial cortical veins; 9. basal vein; 10. lateral sinus.

these veins may become hypertrophied. Thus, with highly malignant cerebral gliomas where there is rapid arteriovenous shunting through large tumour vessels, and with angiotnatous malformations, the subependymal veins may be quite large and readily visible. The vein of Galen i s a short thick vein which curves upward and backward behind the splenium of the corpus callosum. Here it joins

Previous Page

CONTENTS

Next Page


1683

Fig. 55.21 External carotid artery branches. 1 . Superior thyroid artery; 2. lingual artery; 3. facial artery; 4. ascending pharyngeal artery; 5. occipital artery; 6. posterior auricular artery; 7. internal maxillary artery; 8. superficial temporal artery.

Fig. 55.20 (A) Deep veins shown at late phlebography. (B) Diagram of l ate phlebogram (lateral view) showing relationship to the ventricle. 1. Internal cerebral vein; 2. vein of Galen; 3. striothalamic vein; 4. vein of the septum pellucidum; 5. venous angle; 6. atrial vein; 7. subependymal veins; 8. straight sinus; 9. basal vein.

occasionally all three arteries, may arise in common, as may the ascending pharyngeal and occipital arteries, which arise from the posterolateral aspect of the external carotid. In about 15% of cases the superior thyroid arises from the common carotid near its bifurcation, and rarely the occipital or ascending pharyngeal, or both, may arise from the internal carotid. The posterior auricular artery is a small branch which arises just before the main terminal branches, the superficial temporal and i nternal maxillary. The large internal maxillary artery gives rise to the important middle meningeal artery as well as supplying the nasal fossa, palate, mandible and infraorbital region. The external carotid artery normally arises at the level of the C3 vertebra from the bifurcation of the internal carotid artery. Selective injection of the external carotid has been practised in the past for the better demonstration of intracranial tumours (such as meningiomas) deriving most of their blood supply from the middle meningeal artery. It was also useful for demonstrating the contribution of the external carotid artery to the blood supply of other intracranial tumours, such as acoustic neurinomas and glomus jugulare tumours, and for the demonstration of ducal arteriovenous fistulas supplied mainly or entirely by the external carotid artery. With the greater use of embolisation techniques for the treatment of tumours and angiomatous malformations, superselective angiography of the external carotid artery is being increasingly practised. The technique has been used for the demonstration and embolisation of angiomas of the face, lips and tongue. It has also been used in the i nvestigation and treatment of tumours. These include meningiomas, glomus jugulare tumours and nasopharyngeal angiofibromas. Another indication is the demonstration and treatment of dural arteriovenous fistula.


the inferior sagittal sinus to form the straight sinus which passes downward in the apex of the tentorium to the torcular herophili (Fig. 55.20). The straight sinus usually drains into the left lateral sinus, while the right lateral sinus usually drains the superior sagittal sinus.

The detailed angiographic anatomy of the smaller branches of the external carotid artery is described in great detail in the excellent monograph of Djindjian and Merland (1978). The named major branches of the external carotid artery may be listed as follows (Fig. 55.21): I. Superior thyroid artery 2. Lingual artery 3. Facial artery 4. Ascending pharyngeal artery 5. Occipital artery 6. Posterior auricular artery 7. Internal maxillary artery 8. Superficial temporal artery.

The first three branches arise from the anteromedial aspect of the proximal segment of the artery. The upper or lower of these two, or

Previous Page 1684

CONTENTS

Next Page


The vertebral artery arises from the subclavian artery at the root of the neck. It passes backward to enter the transverse process of C6. Occasionally it may enter a transverse process at a higher level. In its cervical course the vertebral artery gives off muscular branches that supply the paraspinal muscles. It also gives rise to tiny spinal or radicular branches. The muscular branches anastomose with branches of the occipital artery and with the ascending pharyngeal artery. These anastomoses assume some importance in cases of carotid occlusion or stenosis of the vertebral origins and in embolisation procedures.

The spinal branches are tiny and are rarely recognised on routine angiograms. They supply the meninges and may anastomose with the anterior and posterior spinal arteries which supply the cord. The normal vertebrobasilar intracranial circulation is illustrated in Figures 55.22 and 55.23. The termination of the vertebral artery in its i ntracranial portion gives rise to several important vessels. These are the posterior inferior cerebellar artery, the anterior and posterior spinal arteries, and small anterior and posterior meningeal arteries. The anterior and posterior spinal arteries are tiny and difficult to i dentify, although the anterior spinal artery is usually visible in a good-quality vertebral arteriogram. It is seen passing downward i nto the spinal canal as a very fine vessel directly anterior to the cord. Like most small vessels in the posterior fossa, it is best i dentified in good subtraction films. The posterior meningeal artery, when identified, is seen as a near midline vessel passing upward and just anterior to the occipital bone in the lateral view. The posterior inferior cerebellar artery usually arises from the terminal segment of the vertebral artery. With selective catheter vertebral angiography it is usually possible to fill by reflux the termination of the contralateral vertebral artery. Thus both posterior inferior cerebellar arteries can be shown from the injection of a single vertebral artery (Fig. 55.23). The normal anatomy of the posterior inferior cerebellar artery is illustrated in Figures 55.24 and 55.25. This vessel was of considerable importance in angiographic diagnosis in the posterior fossa. There is a wide variation in its course and distribution but a fairly typical pattern is followed in most cases. The point of origin of the artery may be from the vertebral artery below the foramen magnum or as high as the junction of vertebral and basilar arteries. Sometimes it arises from the basilar or in common with the anterior inferior cerebellar artery. Normally it arises from the vertebral artery just above the foremen magnum. The first part of the artery, as seen in lateral view, loops round the medulla and then downward to curve round the lower margin of the tonsil. It then passes up anterior and medial to the tonsil to reach the roof of the fourth ventricle in the region of the inferior medullary velum. The downward loop just described usually marks the lower limit of the tonsil and has therefore been used as an index of tonsillar herniation. However, this sign should he treated with caution since the loop may reach below the foramen magnum in normal patients as an anomaly. In these latter cases the loop tends to be narrower with a `hairpin' appearance. Just below the apex of the fourth ventricle the posterior inferior cerebellar artery, as seen in lateral view, loops slightly backward again before dividing into its major branches. This important landmark is known as the 'choroidal loop', as it supplies choroidal branches to the fourth ventricle and marks out the inferior roof of the fourth ventricle. The vermis branch, which lies near the midline, forms a flat loop convex downward with a local exaggeration in the region of the pyramid. Its terminal portion curves round the tuber in the posterior cerebellar notch. Normally the vermis branch, as seen in the axial view, does not cross the midline; displacement across the midline can be regarded as evidence of a mass on the ipsilateral side of the posterior fossa. The other major branch of the posterior inferior cerebellar artery is the tonsillohemispheric. This arises at the same point as the vermin branch but runs farther downward in lateral view along the posterior margin of the tonsil. It gives off anterior or tonsillar branches and posterior or hemispheric branches which curve down and back around the under aspect of the cerebellar hemisphere.


Fig. 55.22 (A) Normal vertebral arteriogram-lateral view. (B) Diagram to illustrate (A). 1. Vertebral artery; 2. posterior inferior cerebellar artery; 3. basilar artery; 4. superior cerebellar arteries; 5. posterior cerebellar artery; 6. posterior choroidal arteries; 7. thalamoperforate arteries; 8. posterior temporal artery; 9. internal occipital artery.

Previous Page

CONTENTS

Next Page

VASCULAR IMAGING I N NEURORADIOLOGY

1685

Fig. 55.24

Posterior inferior cerebellar artery. (A) Lateral view. Asterisks mark the apices of the caudal and cranial loops. The apex of the cranial l oop is closely related to the nodulus (N) and roof of the fourth ventricle (dotted). The upper or vermis branch runs near the midline around the i nferior vermis (including uvula-U, pyramid-P, tuber-T). The lower or tonsillohemispheric branch runs near the posterior margin of the tonsil giving off anterior or tonsillar branches and posterior or hemispheric branches. The tonsil lies between 2 and 3. (B) Anteroposterior view, in halfaxial projection. The segments marked in the lateral view are identified by the same numbers and letters. Segment 3 (the tonsillohemispheric branch) may be difficult to identify in this view. The apex of the cranial loop (*N) lies usually within 2 mm of the midline and the terminal portion of the artery (T) returns to the midline. (Reproduced from Wolf, B.S., et al (1962) A merican journal of Roentgenology, 87, 322-337). The appearance just described and illustrated can be regarded as the standard configuration of the posterior inferior cerebellar artery, but it should be appreciated that there are many variations from this standard pattern. The basilar artery is formed by the junction of the two vertebral arteries just above the foramen magnum. It passes upward directly behind the clivus in the lateral view and it terminates behind or just above the tip of the dorsum sellae. In the anterior or Towne's projection the basilar artery lies in the midline. However, displacement from the midline, or lateral kinking, is quite common in the middle-aged and elderly, particularly in hypertensive patients, and such kinking does not necessarily imply displacement by a mass. In elderly and hypertensive patients the basilar artery may also be elongated and kinked, as seen in lateral view. Sometimes it terminates well above the dorsum sellae and these cases of a high basilar termination may cause an indent in the floor of the third ventricle. Lateral kinking of the basilar artery may also be seen in young patients. In these patients the kinking is always to the side away from a large dominant vertebral artery, and has no pathological significance. The basilar artery has numerous small branches which supply the pons and are difficult or impossible to visualise in an angiogram. There are, however, larger branches which can be recognised in many cases. These include the paired anterior inferior cerehellar arteries. These two vessels usually arise within a centimetre of the origin of the basilar artery (Fig. 55.23). In lateral view they are obscured by the mastoids and petrous hones, and are best identified i n the Towne's view or in the direct anteroposterior view. Subtraction films will usually show them well, and clear of overlying bone. They extend directly laterally to a 'floccular loop', and supply branches to the internal auditory meatus and also to the inferior surface of the cerebellum. Here they anastomose with branches of the posterior inferior cerebellar artery.


Fig. 55.23 (B) Diagram

(A) Vertebral arteriogram-AP view (subtraction print). to illustrate (A). 1. Vertebral artery; 2. posterior inferior cerebellar artery (PICA); 3. basilar artery; 4. superior cerebellar artery; 5. posterior cerebral artery; 8. posterior temporal artery; 9. i nternal occipital artery; 10. anterior inferior cerebellar artery (AICA). The right PICA arises from the right AICA in this case.

Previous Page 1686

CONTENTS

Next Page


Fig. 55.25 Relationship of normal arteries in the posterior fossa to the brainstem. 1. Posterior inferior cerebellar artery; 2. superior cerebellar artery; 3. medial posterior choroidal artery. (After Huang & Wolf 1970.) The superior cerebellar arteries arise just before the termination of the basilar artery. These paired arteries curve round the midbrain to reach the superior surface of the cerebellum, where they divided into several branches. In a lateral film they lie below the posterior cerebral arteries and are seen to pass over the surface of the cerebcllum. In this view they are partially superimposed on the posterior temporal branch of the posterior cerebral artery, as this lies on the other side of the slanting tentorium, The posterior cerebral arteries are the terminal branches of the basilar artery. Both posterior cerebral arteries fill readily through the basilar in nearly 90% of cases. In most of the remaining cases only one posterior cerebral artery fills well. Very rarely neither posterior cerebral artery is filled. In these latter cases there is a dominant supply from the carotid artery through large posterior communicating arteri es. The posterior cerebral arteries curve around the cerebral peduncles to reach the dorsal aspect of midbrain. Here they pass through the tentorium to reach the undersurface of the temporal lobes. Each posterior cerebral has two main branches, a posterior temporal, which supplies the undersurface of the posterior part of the temporal lobe, and an internal occipital branch, which passes to the medial aspect of the occipital lobe. This divides into terminal branches named the calcarine and parielo-occipital arteries. The thalamoperforating arteries arise from the proximal segment of the posterior cerebral artery and lie close to the midline. They are easily seen in the lateral view, where they appear to be 1-3 in number and show as fine vertical vessels passing up into the thalamus. They are difficult or impossible to define in the AP view.

artery. It passes round the midbrain together with, and usually obscured by, the posterior cerebral artery (Fig. 55.25). In the capillary phase the choroidal plexus is outlined as a welldefined blush of contrast. This is a normal appearance and should not be mistaken for a pathological circulation. A small splenium branch of the posterior cerebral artery also supplies the pial plexus on and behind the splenium of the corpus callosum, which may thus he outlined on the vertebral angiograrn. The posterior communicating artery i s occasionally outlined, but it is unusual to see good antegrade filling of the carotid system from the posterior communicating artery in a normal patient. However, the phenomenon can be demonstrated in certain pathological conditions. In one case we have demonstrated filling of the whole of the cerebral circulation from the vertebrobasilar system in a patient with bilateral internal carotid thromboses. Retrograde filling has also been demonstrated following compression of the carotids in the neck and simultaneous injection of the vertebral artery.


Posterior choroidal arteries Each posterior cerebral artery gives off a medial posterior choroidal artery and also two or more lateral posterior choroidal arteries. The medial posterior choroidal artery arises first, just lateral to the bifurcation of the basilar

The veins of the posterior fossil

The precentral cerebellar vein is small but easy to recognise in the lateral phlebogram (Fig. 55.26). It passes in the midline over the superior surface of the cerebellum and lies dorsal to the midbrain. If the aqueduct and fourth ventricle are pressed backward the precentral cerebellar vein will be pressed backward with them. A mass in the upper vermin will displace the vein forward and upward. A mass i n the pineal region will press the precentral cerebellar vein downward. The precentral cerebellar vein drains into the great vein of Galen at its posterior end, where it joins the straight sinus. In the frontal view the vein is difficult to recognise because of the other l arger overlying midline veins. Occasionally its origin can be seen in the AP view as an inverted `Y' draining up from the fourth ventricle.

Previous Page

CONTENTS

Next Page


1687

Fig. 55.26 Normal veins of the posterior fossa and their relationship to the normal structures. 1. Inferior vermian vein; 2. precentral cerebellar vein; 3. posterior mesencephalic vein; 4. anterior pontomesencephalic vein; 5. petrosal vein; 6. lateral mesencephalic vein; 7. brachial vein; 8. vein of the lateral recess of fourth ventricle; 9. transverse pontine vein; 10. peduncular vein; 11. retrotonsillar veins. (After Huang & Wolf 1970.)

The anterior pontoinesencephalic vein runs on the anterior surface of the pons, where it forms an irregular concave curve. It marks out the posterior wall of the pontine cistern and the anterior surface of the pons. Below, it communicates with the right and left petrosal vein. Above, it drains into the posterior inesencephalic vein. The latter is superimposed on the basal vein in lateral view and may drain into it. The lateral mesencephalic vein also communicates with the posterior mesencephalic or basal vein. It passes downward and forward at an angle of about 80° in the lateral view and connects with the superior petrosal sinus, via the brachial and petrosal veins. The superior vermian vein passes along the superior vermis to terminate in the great vein of Galen near the termination of the precentral cerebellar vein. It may join the precentral vein, or the basal vein, or he continuous with the inferior vermian vein. Again it is difficult to identify in the anterior view, although easily seen in the l ateral view. The paired i nferior vermian veins lie near the midline. Both commence with a superior and inferior retrotonsillar vein outlining the hack of the tonsil and drain up the inferior vermis. They usually end in the straight sinus but can join the superior vermian vein. In l ateral view they are easily recognised as they lie as much as 1 cm from the occipital bone and mark the anterior margin of the cisterna magna. The petrosal vein lies just above and lateral to the internal audi-

The brachial vein drains into the petrosal vein and is best seen in the Towne's view where it lies at an angle of 45 ° to the midline. The lateral mesencephalic vein may connect it with the posterior mesencephalic vein.


tory meatus. It is fed by the brachial vein and drains into the superior petrosal sinus. It is best seen in the Towne's view and was of considerable importance in the assessment of angle tumours, as acoustic neurinomas or other angle tumours usually elevate or obliterate the petrosal vein.

The normal aortic arch and great vessels are illustrated in Figure 55.27A. This arrangement is present in the majority of people but anomalies are seen in about a quarter of the population. In our own angiographic material the variation that was seen most often was a common origin for the innominate and the left common carotid artery (Fig. 55.27B). Obvious cases arc recognised readily but it is sometimes difficult, or even impossible, to distinguish minor degrees, even in the oblique position. because of overlapping of vessels. About 20% of our white patients had this anomaly, but it appears to be even more common among blacks, being present in 36% of reported autopsies. The remaining anomalies show a fairly constant incidence and are also of importance. Thus the left vertebral arises from the aortic arch in 5% of patients, so its origin can only be seen in these cases by arch aortography and not by subclavian catheterisation (Fig. 55.27D). An aberrant right suhclavian artery is present in I -2% of patients. This is also important because it may he a cause for right transaxillary catheterisation failing to enter the ascending aorta (Fig. 55.27E). In a few of these cases the right and left common carotid arteries arise from the aorta by a common trunk (truncus bicaroticus).

Previous Page

CONTENTS

Anomalies of the aortic arch branches. (A) Normal. (B) J oint origin of innominate and left common carotid. (C) Left common carotid arises from innominate. (D) Left vertebral arises from arch. (E) Anomalous right subclavian. (From Sutton, D., Davies, E.R. (1966) Clinical Radiology,

Next Page

Fig. 55.27

17, 330-345.)

Fig. 55.29 Anomalies of the circle of Willis. The posterior communicating artery is most frequently involved and combined lesions are common. (a) Hypoplastic anterior communicating (3%); (b) hypoplastic proximal segment of anterior cerebral (2%); (c) hypoplastic posterior communicating (22%); (d) carotid origin of posterior cerebral (15%). (After Alpers et al 1959.)

The posterior communicating artery may also he small or hypoplastic. Such anomalies of the circle of Willis can be very i mportant when stenosis or obstruction of a major extracranial vessel occurs, as it normally provides a collateral circulation. Congenital hypoplasia of the intracranial part of the internal carotid arteries has been described and in such cases a collateral circulation may develop from the primitive vascular `rete' or network present in the embryo at the base of the skull. Congenital anomalies of the brain such as agenesis of the corpus callosum may also be associated with arterial anomalies. In this condition the anterior cerebral arteries may be represented by a large common trunk in the midline, or two arteries may be present. In either case the anterior cerebral artery pursues an abnormal course. It passes more vertically upward, and as there is no corpus callosum the pericallosal branch is abnormal in course and situation. The common trunk anterior cerebral artery (azvgos anterior cerebral artery) may also be seen as an isolated anomaly. The common carotid artery usually bifurcates at the level of C3 or C4 just below the angle of the jaw. In some 15% of cases, however, it may bifurcate at lower levels, although rarely below the C6-7 disc space. In about 2% of cases the bifurcation lies at or about the C2


Fig. 55.28

Some rare anomalies of the great vessels (see text).

Apart from the commoner anomalies illustrated in Figure 56.27, many other rarer anomalies have been described. These include (Fig. 55.28): I . Origin of all four great vessels independently from the arch (A) 2. Both common carotids arise together (truncus bicaroticus) between the right and left subclavians (B) 3. Origin of the left common carotid from the left subclavian (C) 4. Origin of both common carotids from the left subclavian (D) 5. The left vertebral arises from the arch, and the left common carotid from the innominate (a variant of anomalous left vertebral from the arch already described) (E).

Congenital anomalies of the intracranial circulation are relatively uncommon, with the exception of anomalies of the circle of Willis (Fig. 55.29). The first part of the anterior cerebral artery may be small or hypoplastic on one side. In these cases both anterior cerehral arteries fill from the other side with the aid of the anterior communicating artery.

level. Anomalies of the internal and external carotid arteries are less common, but are recorded. They include: 1. External carotid artery: a. Absence of the external carotid with its territory supplied by collaterals b. Branches of the external carotid arising from the internal carotid c. Lateral position of the external carotid instead of normal medial position relative to the internal carotid at its origin. 2. Internal carotid artery: a. Aplasia of the internal carotid artery b. Hypoplasia of the internal carotid artery

Previous Page 1690

CONTENTS

Next Page


Acoustic artery (arrow) connecting basilar artery to internal carotid. This had been ligated in the neck for an aneurysm arising from its suprasellar segment. A vertebral arteriogram shows it still filling from the basilar (intra-arterial DSA study). Fig. 55.32

Origin of the vertebral artery from the carotid in the neck. This is an example of the proatlantal intersegmental artery. Lateral view. Fig. 55.34

other branches of the external carotid may involve it, with resulting unilateral blindness. Equally rare is a middle meningeal artery arising from the stapedial artery. This is an anomalous branch of the petrous segment of the internal carotid which passes through the middle ear and t( amen.


Vascular lesions will be considered under the following headings: I. Aneurysm 2. Angiomatous malformation 3. Arteriovenous fistula 4. Intracerebral haematoma 5. Embolus 6. Vascular stenosis and thrombosis.

In this chapter we are concerned with the findings and appearances at angiography. However, these lesions are also considered in some detail in Chapter 58, where the emphasis is mainly on the findings at scanning by CT or MR.

Fig. 55.33

Hypoglossal artery connecting carotid and basilar arteries.

Lateral view.

A nomalous middle meningeal artery: the middle meningeal artery or its anterior branch may occasionally arise from the ophthalmic artery. The converse, an ophthalmic artery arising from the middle meningeal, may also occur. Although very rare, the anomaly i s of considerable importance, as embolisation of the maxillary or

Most intracranial aneurysms are berry aneurysms. Fusiform aneurysms and vascular ectasias can also occur, as can dissecting aneurysms (Fig. 55.35). Aneurysms may present clinically in three different ways. By far the commonest mode of presentation, and that seen in over 90% of the cases encountered, is by subarachnoid haemorrhage. The second mode of presentation is by pressure upon cranial nerves,

Previous Page

CONTENTS

Next Page


Fig. 55.35 Vertebral arteriogram showing dissecting aneurysm of a posterior interior cerebellar artery (MR study).

A Trauma B Spontaneous 1. Aneurysm 2. Angioma 3. Atheroma 4. Tumour 5. Blood dyscrasia or anticoagulants particularly around the circle of Willis and lateral to the cavernous sinus and giving rise to oculomotor pareses. The third cranial nerve i s the one most commonly involved. Third, in a few cases, the aneurysm may reach a very large size and simulate an intracranial tumour. Symptomless aneurysms are occasionally encountered as unexpected findings at cerebral CT or MRI.

advantage of being able to prescribe retrospectively an almost unlimited range of projections, free of vascular overlap, can make these forms of angiography preferable to more conventional intraarterial studies; they may enable the neck of the aneurysm to be shown better (Fig. 55.4) and may also be better at distinguishing aneurysms from vascular loops; these can sometimes pose a diagnostic problem when only a limited number of projections are available. The non- or minimally invasive techniques are the most appropriate to use if screening for intracranial aneurysms is considered useful in high-risk populations (such as patients with polycystic disease of the kidneys) or with patients who have suffered severe sudden headache that could have been due to subarachnoid haemorrhages (thunderclap headache). These noni nvasive techniques are also the most appropriate for investigating patients with cranial nerve palsies possibly due to aneurysms (Fig. 55.4). Intra-arterial angiography should no longer be used for these indications, assuming that these non-invasive techniques are available. Finally, an increasing number of patients presenting with subarachnoid haemorrhage or painful IIIrd cranial nerve palsies are now proceeding to surgery on the basis of MR or CT angiography only. When the brain CT findings localise the source of bleeding, direct angiography may occasionally be li mited, if clinically advisable, to a single carotid angiogram or vertebral angiogram, thus saving the patient from the more extensive four-vessel studies. The latter are, however, usually requested by the neurosurgeon, if the patient is considered fit, as aneurysms may be multiple. Direct angiography

These required injection of both internal carotids and either both vertebrals, or one vertebral with good reflux down the other, thus demonstrating all relevant intracranial vessels. For the carotid angiogram, lateral, AP and oblique films are usually required; it may also be necessary to take an oblique transorbital view, and occasionally an axial view. The aim is to show the aneurysm clearly and the relationship of its neck to the vessel of origin. For the vertebral angiogram lateral and Towne's views arc required with good subtraction films. A lateral oblique view, to show the origins of the posterior inferior cerebellar arteries clearly, may also be helpful. Aneurysms are demonstrated in the majority of cases (55%) and angiomas in a significant minority (10%). Only 15-20% of cases are negative after direct angiography. Repeat angiography after 6 weeks is still advocated by some units in angiographically negative cases, and an aneurysm is then recognised a further 2% or so. Follow-up of patients with negative findings indicates that they have a much better prognosis than patients in whom an aneurysm is shown. Cerebral aneurysms are multiple in between 5 and 15% of cases. Such multiple aneurysms can be well shown by MR (Fig. 55.36). In such cases the clinical features or CT scan may help to decide the source of the haemorrhage, or local vascular spasm may point to the aneurysm responsible. Following a recent bleed, the vessels in the region of the ruptured aneurysm often become spastic and narrowed. Sometimes the spasm is more widespread, involving most of the intracranial arteries. This spasm is readily identified on the arteriograms and is a factor contributing to the clinical symptoms. Other features which may help to identify the bleeding aneurysm, when multiple aneurysms arc present, are the size of the Four-vessel studies


This is discussed in detail in Chapter 58. I n the past, suspected cases of SAH were investigated by direct angiography and often required four-vessel studies. Today, the primary investigations of choice are CT or MRI. These will often provide diagnostic evidence by showing blood in the basal cisterns or sulci or an intercerebral haematoma (see Ch. 57), but some cases may be equivocal or negative. In cases where surgery or intervention is being considered, cerebral angiography will be required. Non-invasive angiography by MRI or spiral CT is increasingly favoured but in patients considered likely to benefit from surgery or where non-invasive findings are negative or equivocal, direct angiography is still required by many neurosurgeons. It is also required where endovascular treatment by the radiologist is being considered or undertaken. Non-invasive angiography

The distinction between mildly noninvasive procedures (IV DSA and IV CE spiral CT) and noni nvasive methods (MRI and ultrasound) has been explained above, as has the distinction between those with and without radiation hazards. MR angiography and 3D CT angiography have been shown to detect convincingly all aneurysms greater than 2-3 mm in size and some aneurysms smaller than this. The

1691

Previous Page 1692

CONTENTS

Next Page

A TEXTBOOK OF RADIOLOGY AND IMAGING fill from one carotid yet have ruptured into the contralateral frontal l obe. Natural history Aneurysmal subarachnoid haemorrhage is still a devastating disease: at least 50% of patients reaching hospital die or are left disabled. The cause is rebleeding in about 15% and delayed ischaemic neurological deficit in most of the rest. Exclusion of the aneurysm by operative clip or endovascular techniques reduces but does not eliminate the risk of rebleeding, and medical therapies aim to reduce the risk of ischaemia. Although the risk of rebleeding is still often stated to be 50% in the first 6 months, untreated this undoubtedly is an overestimate. Most rebleeds occur within 24 h and the risk steadily declines, perhaps to about 2% per year, or even less, by around 6 months postictus. The risk of bleeding of unruptured aneurysms, formerly given in about 2% per year, is now known to be less than 0.1 %, basilar tip aneurysms having the highest risk (about 0.8% per year). It is probable that a subarachnoid haemorrhage is a sudden incident lasting only a short time. It is followed by a drop in blood pressure, with sealing off of the point of haemorrhage by clot. It is therefore extremely unlikely that haemorrhage will be shown at angiography by contrast extravasation. However, like other investigators, we have, on isolated occasions, seen evidence that the aneurysm was actually bleeding at the time of injection. It should be emphasised that this is excessively rare and we have only seen it three times in a review of 1000 cases of subarachnoid haemorrhage examined by direct angiography. Site of aneurysms Aneurysms giving rise to subarachnoid haemorrhage may occur anywhere in the cerebral circulation but they are most frequently around the circle of Willis and in one or two other sites. Thus, in one large series, supratentorial aneurysms occurred most frequently at: 1. The origin of the posterior communicating artery (27%) 2. The junction of anterior communicating and anterior cerebral artery (27%) 3. The bifurcation or trifurcation of the main middle cerebral trunk (20%) 4. The terminal segment of the internal carotid artery (6%).


Fig. 55.36 Multiple aneurysms. MR angiography without contrast injection. (A) MIP projection (contrast reversal); (B) a 3D shaded surface display; (C) direct left carotid angiogram shows anterior communicating aneurysm and posterior communicating aneurysm.

aneurysm and the presence of local haematoma. In the majority of cases it is the largest aneurysm which bleeds. Ruptured aneurysms may haemorrhage into the brain and produce large intracerebral haematomas (Fig. 55.37). These may act as mass lesions and produce displacement of superficial vessels and even shift of the anterior cerebral artery across the midline. In these cases, and in the absence of a CT scan, the diagnosis of haeniatoma will be based on the history and the demonstration of a source for the haemorrhage, such as an aneurysm or angioma. CT scanning will of course show an intracerebral haematoma clearly prior to angiography. Such intracranial haematomas may themselves be the cause of symptoms and require surgical treatment by evacuation. Sometimes an anterior communicating aneurysm may

Aneurysms of the peripheral vessels are relatively uncommon, although occasionally encountered. In the posterior fossa, aneurysms causing suharachnoid haemorrhage and shown by vertebral arteriography have been found, mainly: I . At the basilar termination 2 At the junction of posterior communicating and posterior cerebral arteries 3. At the origin of the posterior inferior cerebellar artery from the vertebral e artery (Fig. 55.35) 4 ve At verteb the bra l robasilar junction. They have also been shown at the origin of other smaller vessels from the basilar artery and very rarely on more peripheral vessels. Aneurysms-with the exception of the large atheromatous aneurysms, to be described later usually arise at points of bifurcation of cerebral arteries. There is evidence of a defect in the muscle coat at this point and this has given rise to the theory of congenital origin of these lesions. However, there is no doubt that other factors, such as age, atheroma and hypertension, are most important in their aetiology.

Previous Page

CONTENTS

Next Page


1693

Fig. 55.37 (A) CT scan showing blood in subarachnoid space and large haematoma in left temporal lobe, indicating ruptured aneurysm at middle cerebral bifurcation. (B) Lateral view showing displacement of middle cerebral vessels by large haematoma in the left temporal lobe from ruptured i rregular aneurysm at left middle cerebral bifurcation (arrows).


The vast majority of cases present in middle-aged or elderly patients and there is a higher incidence among hypertensives. This probably accounts for the recorded associations with coarctation, polycystic kidneys and fibromuscular hyperplasia. Subdural haematomas, usually shallow, may also result from ruptured aneurysms and should be carefully looked for when i nspecting the angiograms or preliminary axial scan. Many neurosurgeons, when they consider internal carotid ligation, feel that it is important to have cross-compression angiographic studies in order to determine whether there is free flow across the anterior communicating artery. Such cross-compression studies are made by compressing the carotid on one side while injecting the other. With a freely patent anterior communicating artery, both anterior cerebrals and both middle cerebrals will fill; however, a negative result can be unreliable.

We have noted above that some cases of aneurysm may present by pressure upon cranial nerves leading to the orbit. The commonest nerve to be involved is the Illrd nerve, but the IVth, Vth, and Vith nerves can also be affected. These lesions are usually produced by

Fig. 55.38 Posterior communicating aneurysm shown in lateral view by arterial DSA study.

Previous Page

1694

CONTENTS

Next Page


Fig. 55.39 Posterior communicating aneurysm shown in lateral view (MR study). aneurysms lying in the cavernous sinus or arising from the origin of the posterior communicating artery (Figs 55.4. 55.38, 55.39). Intracavernous aneurysms are more common in women and are bilateral in 1 0% of cases (Figs 55.40, 55.41). Occasionally a suprasellar aneurysm may involve the optic chiasm or an optic nerve. The primary investigation for suspected pressure producing aneurysms are non-invasive (see later).

Large supratentorial aneurysms presenting clinically as tumours are relatively rare. The true diagnosis may not be suspected before CT or angiography unless marginal calcification is present before CT or MRI, unless calcification is present on simple X-ray. Large athcromatous aneurysms of the basilar artery are also rare, but when they do occur they may also simulate tumours. Vertebral angiography would demonstrate an aneurysm as a surprise finding i n such cases unless a prior CT scan has suggested its presence. Basilar ectasia also due to atheroma can produce a dilated, elongated and kinked basilar artery with similar clinical symptoms. Again, CT or MRI will diagnose these without direct angiography. Occasionally l arge aneurysms of the internal carotid artery may bulge forward into the orbit. Such aneurysms will erode the sphenoitial fissure and may


Fig. 55.41 Bilateral intracavernous aneurysms. MR angiography. (A) MIP projection. (B) 3D shaded surface display. This study was made after intravenous contrast injection: contrast-enhanced 3D MR angiography.

even produce proptosis. In these cases differential diagnosis at simple X-ray is from a secondary deposit or the rare erosive type of meningioma. However, the erosion produced by an aneurysm is fairly characteristic, and usually involves the lateral margin of the optic foramen or anterior clinoid and jugum (Figs 55.42-55.44).

Fig. 55.40 Large left intracavernous aneurysm projecting medially and laterally. There is a smaller right intracavernous aneurysm (intravenous DSA study).

These lesions are of congenital origin and arc to he differentiated from the simple arteriovenous fistula which is usually due to trauma. Although they are generally accepted as being of congenital origin, the majority of cases present in early adult life. It is probable that they increase in size with age and that this increase proceeds more rapidly once adult blood pressure has been established. The communication between the arteries and veins lies over the surface of the brain, but if large in size it extends down into the brain substance in the manner of an inverted cone. In the silent areas of the brain they may be completely symptomless. Should they lie, however, in such sensitive areas as the motor cortex or the occipital

Previous Page

CONTENTS

Next Page



1695

Fig. 55.42 Huge intra- and suprasellar aneurysm of right internal carotid. (A) Lateral view. (B,C) CT cuts through anterior clinoids and suprasellar cisterns, respectively, show bony erosions (arrow) and rounded mass of aneurysm (arrow). (D) Post-enhancement CT shows the aneurysm clearly. cortex, they will produce appropriate symptoms and physical signs. In the former case jacksonian epilepsy is likely to result; in the latter case cortical field defects may be demonstrated. Apart from the physical signs, these lesions may give rise to headaches, usually of a migrainous character. A small but significant proportion of cases present by rupture of the angioma and subarachnoid haemorrhage. Simple X-ray of the skull in these cases has been discussed above (see Ch. 53), but CT or MRI are the usual first investigations and are often diagnostic (see Ch. 58), as is intravenous DSA (Figs 55.45, 55.46). However, arteriography will still be required for accurate delineation of the lesion for neurosurgical purposes or for intervention by the radiologist (see Ch. 56). Angiomatous malformations vary in size from small lesions difficult to demonstrate at angiography to large and spectacular vascular masses (Fig. 56.46). It is a characteristic angiographic feature that the drainage veins fill in the normal arterial phase. This is

because the blood is passing direct from arteries to veins with no i ntervening capillary circulation. Depending on the size of the arteriovenous shunt, there will be compensatory hypertrophy of the afferent vessels supplying it, and also increase in size of the drainage veins. With large angiomas blood appears to be diverted away from the remaining normal cerebral vessels. Indeed, the lesion has been likened to a `sponge' or parasite in this respect. This diversion of blood may possibly give rise to cerebral ischaemia and account for some of the symptoms. In cases where an angioma has ruptured, an intracerebral haematoma may result, with displacement of vessels as by any other intracranial mass. Such intracranial haematomas are well shown by CT or MRI. With small angiomas which have ruptured, the angioma may clot and it may be difficult or impossible to demonstrate the lesion at angiography. However, the clinical history of a subarachnoid haemorrhage with radiological evidence of an intracercbral

Previous Page 1696

CONTENTS

Next Page


Fig. 55.43 Giant terminal right-internal carotid aneurysm. (A) Coronal MRI (B,C) Contrast-enhanced 3D angiogram B is a source image showing the neck of the aneurysm C is a shaded surface display.


Fig. 55.44 Large basilar aneurysm presenting as a suspected pontine tumour. (A) Lateral and (B) axial MR. (C,D) Large terminal basilar aneurysm partially thromhosed; AP and lateral vertebral arteriogram.

haematoma at a peripheral site may cause suspicion. Sometimes the internal carotid, so that the external carotid artery should he some evidence of a lesion may remain-for example, delayed shown as well as the internal carotid. In one large series 80% of filling in a peripherally clotted vein. supratentorial angiomas were supplied purely by pial vessels, 1 5% Planning treatment of angiomas requires full demonstration of all by both pial and dural vessels, and 5% were purely dural in blood feeding arteries and drainage veins. In some cases this will require supply (see Dural AV Malformations, below). Infratentorial bilateral carotid angiography; in others, vertebral angiography may angiomas arc not uncommon and constituted 20% of the cases in also be required, even though the lesion is supratentorial. Some the series just quoted; among these infratentorial angiomas, dural angiomas are fed by nieningeal vessels as well as by branches of lesions were relatively more frequent (33%).

Previous Page

CONTENTS

Next Page


1697

Cerebellar arteriovenous malformation. (A,B) MR angiography without contrast injection. A 3D shaded surface display. Fig. 55.46

opening directly into the vein. Heart failure or irreversible brain damage may occur, and also a form of hydrocephalus in which ventricular shunting may he detrimental and is to be avoided.

Fig. 55.45

Parasagittal angioma shown by intravenous DSA.

The above discussion concerns the common artcriovenous trial formation (AVM). Other types of angioma include cavernous angionut , venous angioma and capillary is langiec tusiu. These are all less typical or negative at angiography but are shown by CT or MRI. The cavernous angioma is not shown: the venous angioma may show a small collection of venules with a single drainage vein but can easily be missed, capillary telangiectasia may show a small non-specific capillary blush. Aneurysm of the vein of Galen (Figs 55.47, 55.48) This is in fact another manifestation of cerebral angioma, the dilated vein being secondary to AV shunting through an angioma, usually in the brainstem or medial temporal lobes. Some cases usually presenting in infancy, however, seem to he due to a congenital malformation of the venous system; the hugely dilated central venous structure is a persistent embryonic vein known as the median vein of the prosencephalon. The AV shunt is from choroid d arteries

Spinal angiomas With the increasing use of spinal angiography (see also Figs 56.10 and 56.58) spinal angiomas have been demonstrated in increasing number and treated by surgery or embolisation (see Ch. 56). Spinal angiography is of little help in the diagnosis of spinal tumours except in spinal hacmangioblastoma, where a characteristic appearance similar to that seen in the posterior fossa will be shown (see Ch. 54).

This may he due to trauma causing a direct communication between an artery and a vein. It may also he due to rupture of the wall of an aneurysm or diseased artery into an adjacent or adherent vein. Intracranial arteriovenous fistula is rare and is most frequently seen in the cavernous sinus. Most of the cavernous sinus cases encountered are due to trauma, but a significant proportion, particul arly in female patients, are due to the second of the mechanisms mentioned, i.e. spontaneous rupture of a diseased arterial wall. The condition may he very difficult to demonstrate by angiography. If there is a large shunt through the fistula it will best he shown


(A) Large angiomatous malformation in the region of the anterior end of the corpus callosum. (B) Drainage is mainly by a hypertrophied i nternal cerebral vein to the vein of Galen and straight sinus. Fig. 55.47

Previous Page 1698

CONTENTS

Next Page

A TEXTBOOK OF RADIOLOGY AND IMAGING 3. The basal vein - the straight sinus 4. The convexity veins -* various sinuses. As already noted, the venous outflow may pass to the contralateral as well as the ipsilateral veins. Subtraction films are essential in elucidating the complex venous anatomy. Vertebral AV fistula has been encountered in the past, mainly as a complication of vertebral angiography by needle puncture. It has also been encountered following closed trauma. So-called `congenital' cases have also been described, although birth trauma cannot be excluded in these cases. Successful treatment of both caroticocavernous and vertebral fistulas has been effected by detachable balloon catheters. This is described in Chapter 56.

Fig. 55.48 Vertebral angiography demonstrates a vein of Galen aneurysm, secondary to an angioma. by using larger doses of contrast medium and higher concentrations than usual. Figure 55.49 illustrates a case of spontaneous AV fistula. In this case there is a cross-flow of contrast medium to the cavernous sinus on the opposite side. The contralateral internal carotid artery appears as a filling defect in the opacified sinus. With cavernous sinus fistula, several different venous outflow pathways may occur. These include: 1. The superior ophthalmic vein -- the facial veins; or the inferior ophthalmic vein -- pterygoid plexus ---> the facial veins 2. The superior petrosal sinus --- > the sigmoid sinus

It is now well known that many apparent spontaneous caroticocavernous fistulas are dural shunts between meningeal branches of the internal or external carotid artery and dural veins near the cavenous sinus. Differentiation of these dural shunts from direct internal caroticocavernous sinus fistula is important both for prognosis and treatment. Clinically the signs of an arteriovenous fistula are less marked with a dural shunt, but they include proptosis, chemosis and elevated intraocular pressure. The shunt is one of low pressure and low flow and needs first-class subtraction films for its demonstration. Selective angiography of both internal and external carotid arteries will be required, as the lesion may be fed by meningeal branches of the internal or external carotid artery, and can have multiple feeders. From the external carotid artery, either the middle meningeal or the distal internal maxillary branches may supply the fistula. From the internal carotid, the meningohypophyscal trunk may be the source of supply (Fig. 55.50). This small vessel arises from the internal carotid just before it enters the cavernous sinus. Bilateral injections will be necessary, as once a fistula is established it can be fed from both meningohypophyseals. Similar dural arteriovenous malformations can occur at other sites,


Fig. 55.49 AV fistula in the cavernous sinus (caroticocavernous fistula). (A) Lateral view showing contrast entering cavernous and superior petrosal sinus (arrow). (B) AP view showing contrast medium entering left cavernous sinus (arrow) and crossing over to fill the right cavernous sinus (arrow). The right carotid siphon appears as a defect in the opacified sinus.

Previous Page

CONTENTS

Next Page


but they are commonest at the base of the skull and in the occipitomastoid area. Figure 55.51 shows diagrammatically the meningeal vessels from which these lesions can arise. Dural arteriovenous malformations probably arise as a fistula which progressively recruits secondary arterial feeders from the existing microcirculation or perhaps via flow-induced angioneogenesis. They drain most often via dural veins to the extracranial veins. However, some drain intradurally into the cerebral veins, where they may be responsible for intracranial haemorrhage or i ncreased intracerebral venous pressure and interstitial fluid pres-

1699

sure. These lesions may change over time due to thrombosis of draining venous pathways, and although spontaneous complete cure may occur, this is unusual. Lesions draining extracranially may suddenly begin to drain intracranially, and clinical changes such as loss of tinnitus do not necessarily mean spontaneous cure. In a review of 28 cases it was reported that 14 lesions were basal, mostly in the middle fossa and involving the cavernous sinus, 1 1 were occipitomastoid, one was on the convexity, and one at the free edge of the tentorium. Lesions that are entirely fed by branches of the external carotid are eminently suitable for treatment by emboli-

Fig. 55.50 (A) Dural AV fistula supplied by meningohypophyseal trunk (arrow). (B) The drainage is into the superior ophthalmic vein (arrows). Subtraction prints.


Previous Page

CONTENTS

Next Page


1689

c. Separate origins of the internal and external carotid arteries-on the right side from the innominate and subclavain, on the left side from the aortic arch. A rare but important variant of the internal carotid is an anomalous intrapelrosal course. Normally the vessel kinks medially and forward after entering the carotid canal and lies close to the inferior aspect of the tympanic cavity. In this anomaly it buckles laterally into the tympanic cavity. This is best seen in the AP or oblique views and the vessel is narrowed and uneven. Clinically this can give rise to pulsating tinnitus and a reddish pulsatile `tumour' behind the tympanic membrane. The correct diagnosis can he confirmed by angiography, and a potentially fatal operation can be avoided. Size of arteries

The vertebral arteries vary markedly in size. On the one hand the vertebral artery may be almost as large as an internal carotid artery, on the other hand it may have a lumen as small as I or 2 mm and may appear very tiny on the arteriograms. Such small vertebrals (usually the right) may terminate in the posterior interior cerebellar artery, and not join the basilar artery. The average vertebral artery lies somewhere between these extremes, and on the X-ray film its lumen measures 3-4 mm in diameter. In the same individual the two vertebral arteries can be very different in size, although in approximately 30% of patients the two vertebral arteries are equal. In the remaining 70% of patients, about 40% have a left vertebral larger than the right and about 30% have a right vertebral larger than the left. The basilar artery can also vary greatly in diameter, being quite small when both posterior cerebrals are supplied from the internal carotids. Anomalous vertebral arteries An important anomaly, the origin of the left vertebral artery direct from the aortic arch between the l eft common carotid and the left subclavian arteries, is described above. Left vertebral catheterisation from the femoral artery in these cases is more difficult, and from the axillary artery it is i mpossible. There are a number of other but much rarer anomalous origins for a vertebral artery. Thus it may arise from the i nnominate artery, from the inferior thyroid artery, or even from the common carotid artery in the neck. Another very rare anomaly i s the origin of the right vertebral from the distal part of the aortic arch. Such a vessel must pass behind the oesophagus like an anomalous subclavian artery in order to reach the right side of the neck. The terminal segment of one vertebral beyond the origin of the posterior inferior cerehellar artery may be hypoplastic. Fenestration of the terminal segment has also been described, as has fenestration of the basilar artery.

Fig. 55.30 p l e of

exam

Anomalous carotid-basilar anastomosis (arrow). This is an the trigeminal artery. Lateral view.

hypoglossal artery, so called because they accompany the corresponding cranial nerves (Fig. 55.31). The acoustic artery connects the internal carotid artery with the basilar by an artery accompanying the VIIth and VIIlth nerves through the internal auditory meatus (Fig. 55.32). It is excessively rare, and special projections, including a basal view, may be necessary before a case can be regarded as authenticated. The hypoglossal artery, which passes through the anterior condyl oid foramen at the base of the skull after arising from the internal carotid in the neck, is less rare, although still less common than the trigeminal artery (Fig. 55.33). The proatlantal intersegmental artery i s also very rare (Fig. 55.34). The essential feature is that the artery arises from the internal carotid and joins the horizontal part of the vertebral artery as it lies on the atlas.


Communications between the carotid and vertebrobasilar

system Normal communications exist between the muscular branches of the vertebral artery and muscular branches of the ascending pharyngeal and occipital arteries. Large abnormal communications between the internal carotid and vertebrobasilar systems may also occur. The most frequent of these (about I in 500 cases) is between the internal carotid artery as it enters the cavernous sinus and the termination of the basilar artery. This artery passes round the base of the sella and is termed the trigeminal artery, because of its close relationship to the trigeminal nerve (Fig. 55.30). Anastomosis at a more proximal level is less frequent. Named communications are the acoustic artery and the

Anomalous communications between the carotid and vertebrobasilar systems. (a) Trigeminal artery; (b) acoustic artery; (c) hypoglossal artery; (d) proatlantal intersegmental artery. Fig. 55.31

Previous Page 1700

CONTENTS

Next Page


Fig. 55.52 Dural AV fistula supplied by the occipital artery and draining i nto the lateral sinus and internal jugular vein. Superselective occipital artery i njection (subtraction print). This case was successfully treated by percutaneous embolisation. cation (Fig. 55.52). This is a particularly valuable treatment of these l esions, as direct surgery is difficult and carries a high rate of complications and recurrence.

I ntracranial haematomas may be classified as extradural, subdural and intracerebral. Arteriography can diagnose these lesions but has long been superseded by non-invasive methods, i.e. CT or MR] and, i n the case of infants, ultrasound. Diagnosis is discussed in detail in Chapter 58.

Atheromatous stenosis and thrombosis of the internal carotid arteries before they enter the cranium can be a cause of cerebral symptoms. Thus the internal carotid artery may be stenosed or thrombosed just distal to its origin from the common carotid artery (Fig. 55.53). Less commonly, atheromatous stenosis is seen in the carotid siphon. Originally it was thought that simple stenosis produced symptoms but later it was suggested that platelet emboli from ulcerating plaques were responsible (see below). However, it is now accepted that only severe stenosis (i.e. more than 70%) is a major risk factor for stroke and an indication for surgery. I n the same way, but less frequently, the vertebral artery may be stenosed or thrombosed by atheroma near its origin from the subclavian artery (Figs 55.54, 55.55). It has also been claimed that the vertebral artery can be compressed in its cervical portion by pressure from osteophytes, and, further, that rotation of the neck may compress the vertebral artery in some elderly patients. The clinical significance of vertebral lesions is debatable, as there are still no reliable controlled studies relating them to vertebrobasilar i nsufficiency, and the value of surgery for vertebral stenosis remains controversial. Finally, stenosis or thrombosis of the innominate, common carotid and subclavian arteries may occur in the thorax, and was claimed by some to affect the cerebral blood supply; however. smooth non-ulcerating lesions in these vessels are unlikely to produce cerebral symptoms. Multiple lesions are well shown by i ntravenous DSA (Fig. 55.56). Collateral circulation I n the case of unilateral internal carotid thrombosis, a collateral circulation may develop from several sources. Blood may reach the affected hemisphere through the circle of Willis, i.e. by flow across the anterior communicating artery from the opposite side, and by retrograde flow along the posterior communicating artery from the basilar to the internal carotid artery. In addition, anastomoses usually open up between the terminal branches of the maxillary artery and those of the ophthalmic artery. This permits blood from the external carotid artery to reach the ophthalmic artery and flow retrogradely into the cerebral circulation. Anastomoses may also develop between the muscular branches of the vertebral artery in the neck and branches of the external carotid artery, such as the occipital artery (Fig. 55.57). The usually excellent collateral circulation may explain why symptoms are often absent, or are variable and unpredictable in patients with carotid stenosis and thrombosis. Thus we have encountered symptom-free patients with bilateral carotid thrombosis at one end of the spectrum and patients with symptoms and unilateral stenosis at the other. Eventual thrombosis may result in a minor or major stroke, or paradoxically it may cure the symptoms.


Cerebral embolus most commonly arises from clot in a fibrillating auricle. It is therefore most frequently seen as a complication of mitral stenosis. Like embolus elsewhere, it is sometimes seen following coronary infarction and clot formation in the left ventricle. It may also result from paradoxical embolus, or from dislodgement of clot in an aortic aneurysm. Small emboli can also result from clot forming on the surface of atheroniatous plaques in the aorta and great vessels. Catheter clot embolus is a complication of arteriography and is discussed in Chapter 15. A cerebral vessel occluded by embolus is readily shown by arteriography, provided it is sufficiently large. In the writers' experience the middle cerebral artery is the commonest site for a major cerebral embolus. This is not surprising as the middle cerebral artery takes the major blood flow after bifurcation of the i nternal carotid artery. Embolus of smaller peripheral vessels may require careful study of rapid serial angiographic films before the lesion is recognised. A major embolus is now readily diagnosed by CT (see Ch. 58) or by MRI, but the tiny emboli which occur in transient ischaemic attacks (TIAs) cannot be so identified.

Transient ischaemic attacks Symptoms in cases of internal carotid stenosis may be due to small platelet emboli detached from the surface of ulcerating atheromatous plaques. These can

give rise to the so-called transient ischaemic attacks (TIAs) with focal symptoms, such as hemiparesis, hemiparaesthesia or monocular amblyopia, which recover fairly rapidly. If untreated, carotid occlusion may eventually ensue. Rather surprisingly, as already noted, this can result in clinical improvement with spontaneous cure of the TIAs. On the other hand, thrombosis may result in stroke, which may be minor, major or even fatal. Whether this i s due to a major embolus or to extension of the thrombosis is

Previous Page

CONTENTS

Next Page


170 1

Fig. 55.53 (A) Stenosis of the internal carotid arteries. (B) Thrombosis of the internal carotid artery in the neck.


Fig. 55.54 Bilateral stenoses of the i nternal carotid artery sinuses. (A) MR angiography using the neurovascular coil (3D contrast-enhanced MIP projection) showing the extensive coverage achieved. (B) Higher resolution 3D contrast-enhanced image of the left carotid artery shown as a shaded surface display. The other arteries have been excluded from the image. (C) Left carotid and vertebral arteries. MR angiography contrast-enhanced 3D shaded surface display. Mild smooth narrowing of the left carotid sinus is shown. Neurovascular coil, other areteries are excluded from the display.

Previous Page 170 2

CONTENTS

Next Page


Fig. 55.55 Arch aortogram showing occlusion of right common carotid artery and left internal carotid artery and severe stenosis of the left subclavian artery. The right vertebral is also occluded. (Subtraction print.)

Fig. 55.57 Possible collateral pathways in cerebrovascular insufficiency. (A,B) Posterior communicating and anterior communicating arteries of the circle of Willis. (C) Communication between maxillary and ophthalmic artery. (D) Communication between muscular branches of the vertebral and the occipital artery. (E) Communication between the vertebral and ascending cervical arteries. (F) Communications between the inferior thyroid and superior thyroid arteries of the same and opposite sides. (G) The vertebral artery providing retrograde flow to the subclavian (subclavian steal). (H) Pial anastomoses between terminal branches of the anterior cerebral, middle cerebral and posterior cerebral arteries. (J) Pial anastomoses in the posterior fossa. (K) Meningeal anastomoses and rete mirabile. (From Sutton, D., Davies, E.R. (1966) Clinical Radiology, 17, 330-345.)

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ Fig. 55.56 Four-vessel neck studies by intravenous DSA in three different patients. (A) Left carotid stenosis. (B) Bilateral carotid stenosis. (C.) Right common carotid thrombosis.

Previous Page

CONTENTS

Next Page


170 3

difficult to determine. Other factors that may increase the risk of stroke include heart disease with poor cardiac output, hypertension or hypotension. Also relevant is the presence or absence of previous brain infarcts or other lesions. As there was no reliable method of predicting the outcome, surgery was generally accepted as the treatment of choice and internal carotid endarterectomy, an operation pioneered at St Mary's Hospital, London, has been widely practised. The value of the operation has also given rise to much controversy and to controlled trials in Europe and North America (see below).

An interesting feature is that the contrast medium, as it passes slowly up the cervical carotid, sinks to the posterior aspect of the vessel, producing a `layering' effect. This is better appreciated if the radiograph is viewed in the brow-up position in which it was taken. It has been shown experimentally in animals that spontaneous respiration invariably ceases when intracranial pressure becomes equal to the diastolic blood pressure. The parallel between this observation and the clinical fact that most of the patients with pseudo-occlusion arc in need of artificial respiration is clear.

The vertebral circulation may be seriously affected by thrombosis of the left Subclavian artery in its first part and proximal to the left vertebral origin. When this occurs, the l eft vertebral artery acts as a collateral to the left Subclavian artery. Blood flows up the right vertebral artery and then retrogradely down the left vertebral artery to the left subclavian artery, where blood pressure has been lowered by the proximal subclavian block. This siphonage of blood from the vertebrobasilar system to the arum has been christened the 'subclavian steal' phenomenon. It has been claimed that the diversion of blood from the brain may be considerable and result in cerebral symptoms. The lesion is well demonstrated by arch aortography. Less cornmonly, siphonage of blood down the right vertebral artery may occur if there is an innominate or proximal right subclavian block. In the same way `carotid steal' may result from occlusion of a common carotid or innominate artery. In addition to the vertebral artery acting as a collateral to the arm, a further collateral circulation may be derived from the external carotid artery and its branches to the thyroid axis. Collaterals have also been demonstrated passing from the external carotid to the muscular branches of the vertebral and this provides a further contribution to the vertebral blood flow. Other sources of collateral supply are through the intercostals to the scapular and internal mammary arteries.

Multicentre trials in Europe and North America have established that severe, and only severe, atheromatous stenosis (greater than 70%) of the proximal part of the internal carotid artery is a clear and independent risk factor for stroke, and only i nfrequently is it a necessary cause. They have also demonstrated that correcting severe stenosis by carotid endarterectomy significantly reduces stroke incidence in this group, but the number of strokes thereby prevented in follow-up periods of 2-5 years is actually quite small, amounting to no more than I or 2% of all strokes. Little or no controlled data exist about the significance of vertebral artery stenosis or thrombosis in the presence of symptoms of vertebral vascular insufficiency, and none on the efficacy of operative treatment. Only a very small number of patients (about 2%) have more proximal stenosis, near the origins of the common carotid or subclavian arteries. These are usually smooth non-ulcerating lesions, and it is doubtful if they ever result in cerebral symptoms. It is also our experience that ` subclavian steal' is unlikely to produce cerebral symptoms and is usually a chance finding, I n assessing the stroke risk in an individual patient with a carotid stenosis, it is necessary to consider the lesion not as an isolated phenomenon, but in relation to the state of the cerebral circulation as a whole and the status of the brain itself. For example, cerebral CT or MRI may show diffuse ischaemic brain damage in white matter and basal ganglia indicative of severe intracerebral vascular disease, or cortical infarcts in multiple vascular territories. Either of these findings would suggest that an isolated carotid stenosis is of l ess significance as a risk factor in that particular case.

Subclavian steal

Surgical results


Thrombosis of the internal carotid artery

This may result from

several causes:

I . I n the past, needle injection of the common carotid artery, particularly in the hands of an unskilled worker, was sometimes accidentally made into the arterial wall. If the injection was subi ntimal, an appearance resulted which superficially resembled thrombosis. 2. Thrombosis of the internal carotid artery has sometimes been wrongly diagnosed when the tip of a needle, cannula or catheter has been at the bifurcation of the common carotid artery and the main injection has entered the external carotid artery. As a result, the internal carotid artery, although patent, was not visualised. In all cases where internal carotid thrombosis is suspected, the catheter tip at test doses must be demonstrated to be below the point of bifurcation of the common carotid artery. 3. Another condition in which apparent thrombosis of the internal carotid artery may he seen is grossly raised intracranial pressure in a comatose and moribund patient. In these cases cerebral blood flow may be so reduced that, even on delayed serial films, the contrast medium passing through the internal carotid artery has still not progressed beyond the cavernous sinus. We have seen several such cases where autopsy has shown the vessels to be patent intracranially. When this phenomenon is seen it implies that cerebral perfusion is at an extremely slow rate and the prognosis is invariably had.

Imaging The most complete visualisation of the cerebral vascul ature is routinely obtained by IV DSA (Fig. 55.56) but this now can also be achieved by MRA or invasively by arch aortography or a combination of arch and selective carotid arteriography. In general, however, the requirement of visualising all the cerebral circulation (including cervical and intracranial vessels) diminishes in importance if carotid stenosis is regarded as only a potentially remedial risk factor and not necessarily a cause of stroke. Sonography, MRA and 3D CT are usually directed only at the carotid bifurcations, but the latter two techniques now may be extended quite easily to include great vessel origins and intracranial vessels by performing further runs. For MRA this is time consuming and involves changing coils, and for 3D CT angiography it may need to he done on separate days because of concern about contrast volume and solute load. What is actually done still varies widely and continues to be driven by strongly held but often ill-founded views. This is an area where radiologists need to be more than technicians, and should be aware of the issues i nvolved. Readers are urged to study carefully the reports of the European and North American Symptomatic Carotid Stenosis trials, and recent studies of the accuracy of intra-arterial angiogra-

Previous Page

1704

CONTENTS

Next Page

A TEXTBOOK OF RADIOLOGY AND IMAGING Dissecting aneurysm This may also follow direct trauma to the

neck and usually involves the high cervical segment. Diagnosis is readily made on direct angiography if the false channel fills with contrast medium, but should also be suspected when it does not and there is merely narrowing of a segment of the artery in the same high cervical position (Fig. 55.58B). Spontaneous dissection may also occur in this area, with hypertension as a predisposi ng cause. Clinically there may be sudden neck or head pain and frequently a partial or complete Horner's syndrome. As noted, angiography rarely shows the false lumen, but the true lumen may be narrowed, irregular or eccentric. MRl is nor, the method of choice to make a diagnosis by using high-resolution fast spinecho sequence to show the dissected artery in the axial plane (see Fig. 58.13A): MR angiography is not required, and may even appear normal, as it should be noted that some carotid dissections may resoly .4pidly Some dissections have an undetectable effect on the side of the true lumen. Intra-arterial angiography should be avoi/ed. Intracranial thromboses Thrombosis of intracranial vessels

Fig. 55.58 (A) Fibromuscular hyperplasia producing beaded appearance of the internal carotid. (B) Dissecting aneurysm of internal carotid artery. The channel is narrowed and irregular.

develops either from atheroma or, less commonly, from ruptured aneurysms. Atheromatous stcnosis of the internal carotid artery occurs intracranially, either in or just above the cavernous sinus. Of the other major intracranial vessels, the middle cerebral artery is most frequently involved (Fig. 55.59). Thrombosis of small peripheral vessels is easily missed; if it is suspected, rapid serial films may be necessary to demonstrate the l esion clearly. With intracerchral thrombosis a collateral circulation may he demonstrated, filling the occluded vessel retrogradely. Intracranial thrombosis of the internal carotid artery may also occur from atheroma in the cavernous sinus or just hefore the termination of the artery. Basilar thrombosis is usually fatal but can occur with survival of t he patient, and cases have been diagnosed by vertebral angiography. A syndrome consisting of multiple progressive intracranial arteri al occlusions in children, originally thought to occur only in


phy both in estimating stenosis and in predicting stroke risk by assessing plaque morphology, all of which are cited at the end of this chapter. Most importantly, it must be appreciated that if the stroke risk of preoperative investigation rises above about 2% in appropriately operated patients, the statistical significance of the

beneficial effect of carotid endarterectomy may be wiped out. If patients eventually shown not to be in the group that benefits from carotid endarterectomy are also exposed to this level of risk, more strokes will be caused than prevented. Many centres have now opted to investigate patients with suspected carotid stenosis exclusively non-invasively, using a combination of either sonography and MRA, 3D CT or IV DSA, and they rarely if ever perform i ntra-arterial studies, even in operated cases. A persistent indication for intra-arterial studies, however, is when non- or less invasive techniques have suggested but not proven an occlusion, because very occasionally this turns out to be an operable very tight stenosis (Fig. 55.1). Fibromuscular hyperplasia The internal carotid artery is

occasionally involved, usually in its upper cervical segment ( Fig. 55.58), although this rarely causes significant symptoms. There may be involvement of the renal arteries and associated hypertension.

Traumatic thrombosis of the internal carotid This may occur

after severe direct trauma to the neck.

Fig. 55.59

Middle cerebral artery thrombosis (arrow).

Previous Page

CONTENTS

Next Page

VASCULAR IMAGING IN NEURORADIOLOGY Japanese ( moyamoya disease), has since been reported in American and British cases. The disease results in rapidly progressive stenoses of the major vessels of the circle of Willis, leading to occlusion and multiple tiny basal collateral vessels. The process stops once occlusion occurs, so that survival is possible if collaterals are adequate. The distal arterial vessels are not involved and the aetiology is unknown. The role of CT and MRI in the diagnosis of infarcts and TIAs is discussed in Chapter 58.

1705

Cortical venous thrombosis, apart from local absence of venous filling and presence of surrounding collaterals, may show evidence of a local mass due to infarction and oedema. Obstruction of venous sinuses may also occur with tumours. Thus meningiomas frequently involve the sagittal and sometimes the lateral sinus, and glomus jugulare tumours may invade the internal jugular vein and sigmoid sinus.

Angiography is sometimes requested to help support a clinical diagnosis of intracranial arteritis. Because some of the changes reported in these cases are subtle or involve only small branches, highresolution infra-arterial studies are often recommended. However, even the finding of multiple small vessel occlusions or irregularities

Thrombosis of cerebral veins and dural sinuses is not uncommon, Clinically, the commonest cause is infection, either local or distant. Meningitis or encephalitis may be complicated by venous thrombosis, often fatal, and infection may spread from the paranasal sinuses or middle ear and mastoid. Thrombosis may also occur as a corn- is non-specific and can be simulated by multiple emboli or vasospasm, and the role of angiography in this context is at best lication of pregnancy (about I case per 2500a births). Rare causes controversial. In a recent review, it was concluded that CT and MRI i nclude nclude skull trauma and craniotomy, cardiac disease with right of the brain provided better supportive evidence for arteritis than heart failure, oral contraception, dehydration and blood disorders, angiography, and that angiography was not indicated when these The symptoms and signs vary with the localisation and extent of were normal. the lesion and with associated brain oedema, infarction, haemorrhage or infection. They include hemiplegia, papilloedema and epilepsy, general or focal, and drowsiness or coma. Superior sagittal sinus Thrombosis of the anterior part is usually asymptomatic because of excellent pial collaterals. In any case, absence of the anterior part of the sagittal sinus is a common congenital finding when the ascending frontal vein courses parallel to the midline to join the sagittal sinus more posteriorly. Thrumbosis of the more posterior part of the sinus is more serious and produces symptoms of the type just described. The thrombosis may extend into the cortical veins and produce areas of infarction. Transverse and sigmoid sinuses Thrombosis is frequently septic, arising from the middle ear or mastoid. Congenital absence of one transverse sinus may occur and should not be considered as thrombosis without supportive evidence.

For many years angiography was the most important method of localising and characterising cerebral WMOLIl'S. With the advent of CT in the 1970s and MRI in the 1980s its value has steadily diminished, and CT and MRI are now the primary investigations of choice. Angiography for tumour investigation is now an ancillary method for use in rare cases or for special indications. Thus, some neurosurgeons require angiography in some gliomas already diagnosed by CT or MRI, to assess the vascularity and probable degree of malignancy prior to surgery. Angiography may also he required to assess some meningiomas for similar reasons, and to assess vascular relationships or possible sinus involvement. Occasionally it is done for preoperative embolisation of large meningiomas via their external carotid feeding vessels. Vertebral angiography is still requested for haemangioblastomas of the posterior fossa, as it remains the most accurate method of defining small nodules, which can easily be missed by CT or MRI. Angiography is also useful for the demonstration of glomus .jugulare and glolnus tymlzatticum tumours and for the embolisation of the former. These indications are discussed in more detail in Chapters 56 and 57.


Cavernous sinus Involvement of the cavernous sinus is associated with exophthalmos, chemosis, oedema of the eyelids and oculomotor palsies.

Radiological appearances Acute dural sinus thrombosis usually i s obvious on unenhanced CT or MRI: the thrombosed sinuses are distended by high-density or high-signal blood clot. Once recanali sation has begun to occur, it can be difficult to diagnose, whatever modality is used. CT or MRI also demonstrate associated parenchyma) damage. The thrombosis can be well shown by serial angiography of the internal carotid following arterial DSA (Fig. 55.8), but intravenous DSA has the advantage of displaying all the venous sinuses with uniform simultaneous opacification. MR venography is also now widely used for this purpose, although non-visualisation of a transverse sinus is not unusual in normal individuals and turbulent flow effects can simulate intral uminal clot.

On intra-arterial angiography, thrombosis of the sagittal sinus usually slows the circulation, failure of a section of the sinus to outline, and dilated collateral venous channels bypassing the obstruction or draining to other areas. The bridging veins may be abnormal and `corkscrew' in appearance. The lateral sinuses are best seen in the frontal projection and partial occlusions identified.

Alpers, B. J., Berry. R. G., Paddison, R. M. (1959) Anatomical studies of the circle of Willis i n normal brain. Archires ofNenrology and Psychiatry. 81, 409-418.

Black, W. C. (1994) Editorial. Intracranial aneurysm in adult polycystic kidney disease: is screening with MR angiography indicated'.' Radiology, 191, 18-20. Buelmke, D. A., Chambers, T. P. (1995) Editorial. Spiral CT angiography: an alternative to conventional angiography. Radiology, 195, 317-319.

Debrun. G., Legre, J.. Kasbarian. M., et al (1978) Endovascular occlusion of vertebral fistulae by detachable balloons with conservation of the vertebral blood flow. Radiologv. 130, 1 41-147. Di Chiro, G., Wener, L. (1973) Angiography of the spinal cord. Journal of Neurosucgerv, 39, 1 -29.

Djindjian. R.. Merland, J. J. (1978) Superselectire Angiography of the Evternol Carotid Artec-v. Berlin: Springer.

Previous Page

CONTENTS

Eastcott, H. H. G., Pickering, F. W., Rob, C. G. (1954) Reconstruction of internal carotid artery in a patient with intermittent attacks of hemiplegia. Lancet, ii, 994-995. European Carotid Surgery Trialist Collaboration Group (1991) MRI European Carotid Surgery Trial. Interim results (1991) of symptomatic patients with severe (70%-99%) or with mild (0%-29%) carotid stenosis. Lancet, 337, 1235-1243. Executive Committee for the Asymptomatic Carotid Atherosclerosis Study (1995) Endarterectomy for asymptomatic stenosis. JA MA , 272, 1 421-1428. Harris, G. K., Tran, D. D., Sickels, W. J., Cornell, S. H., Yuh, W. T. L. (1994) Diagnosing intracranial vasculitis: the roles of MR and angiography. A merican Journal of Neuroradiology, 15, 317-330. Huang, Y. P., Wolf, B. S. (1970) Differential diagnosis of4th ventricle tumours from brain stem tumours in angiography. Neuroradiology, 1, 4. Katz, D. A., Marks, M. P., Napel, S. A., Bracci, P. M., Roberts, S. L. (1995) Circle of Willis: evaluation with spiral CT angiography, MR angiography, and conventional angiography. Radiology, 195, 445-449. Leclerc, X., Gauvrit, J. Y., Nicol, L., Provo, J. P. (1999) Contrast enhanced MR angiography of the craniocervical vessels. A review. Neuroradiology, 41,867-874. Lie, T. A. (1968) Congenital A nomalies of the Carotid A rteries. Amsterdam: Excerpta Medica. National Institute of Neurological Disorders and Stroke (1995) Carotid endarterectomy in patients with asymptomatic internal carotid artery stenosis. Clinical advisory. Journal of the Neurological Sciences, 129, 76-77. Newton, T. H., Potts, G. G. (eds) (1974) Radiology of the Skull and Brain, Vol. 2 A ngiographv. St. Louis: Mosby.

Next Page

North American Symptomatic Carotid Endarterectomy Trial Collaboration (1991) Beneficial effects of carotid endarterectomy in symptomatic patients with high grade stenosis. New England Journal of Medicine, 325, 445-453. Osborn, A. G. (1980) Introduction to Cerebral A ngiography. Hagerstown: Harper Row. Peevers, F. L. M., Kroger, R. (1979) Dural and direct cavernous sinus fistula. A merican Journal of Roentgenology, 132, 599-606. Prince, M. R., Grist, T. M., Debatin, J. F. (1998) Contrast MRA ngiography. London: Springer. Rothwell, P. M., Salinas, R., Ferrando, L. A., Warlow, C. (1995) Does the angiographic appearance of a carotid stenosis predict the risk of stroke independently of the degree of stenosis. Clinical Radiology, 50, 830-833. Stevens, J. M. (1995) Editorial. Imaging carotid stenosis. Clinical Radiology, 50,821-822. Stevens, J. M., Barker, C. J., Kerslake, R., Broz, M., Barter, S. (1991) Extended use of cranial CT in the evaluation of patients with stroke and transient ischaemic attacks. Neuroradiologv, 33, 200-206. Taveras, J. M. (1969) Multiple progressive intracranial arterial occlusions: a syndrome of children and young adults. A merican Journal of Roentgenology, 106, 235-268. Vanninen, R., Manninen, H., Koivisto, K., Tulla, H., Partanen, K., Puvanen, M. (1994) Carotid stenosis by digital subtraction angiography: reproducibility of the European Carotid Surgery trial and the North American Symptomatic Carotid Endarterectomy trial measurement methods and visual interpretation. A merican Journal of Neuroradiology, 15,1635-1641. Weber et al. A symptomatic A neurvsms.


Previous Page

CONTENTS

Next Page

56 This chapter is meant to provide an overview of interventional neuroradiology. Although some general comments have been made within each section, full details of specific procedures and techniques should be obtained from dedicated texts and articles. These procedures should not be performed by occasional operators: occlusion of the main hepatic artery that interferes with the arterial supply to 750 g of liver will probably have no adverse effect and may not he noticed by the patient, while occlusion of a vessel depriving I g of brain of its arterial supply may leave a patient paralysed, aphasic or dead. The procedures outlined below should only he performed by experienced, well-trained operators with detailed knowledge of' the relevant anatomy, anatomical variants, equipment and pathophysiology. Even in competent hands, interventional neuroradiology is not without risks; the incidence of irreversible neurological deficit from diagnostic neuroangiography is generally accepted to he in the range of 0.3-3.0°k, for interventional neuroangiography it is an order of magnitude higher. A neuroradiologist with ongoing handson experience and in-depth knowledge minimises the risks to the patient. The decision to proceed with an interventional neuroradiological procedure necessitates a consideration of the relative risks and benefits of this modality compared to conventional surgery or other techniques such as radiosurgery. Clearly, before any procedure is undertaken, it must offer an advantage over the natural history of the condition to be treated. Careful consideration must be given to the specific risks of' the procedure and these must be explicitly discussed with the patient to obtain informed consent. Interventional neuroradiology cannot exist in isolation and must be viewed as one component of' a multidisciplinary team. Close interaction with anaesthetic, ncurosurgical, neurological, radiotherapy and intensive care specialties is essential. Iligh-quality biplane angiography, i deally with 3D/rotational angiography, MRI and CT arc required.

Generally, a guiding catheter is used to ensure stable access and provide a route for injection of contrast for road-mapping. Microcathctcrs, either flow guided or over the wire. are usually used to reach the target. The anatomy of the vessels used for access to the target and the target itself must be carefully defined by high-quality angiography. Potentially hazardous anatomical variants should be sought. The optimal endovaseular route to reach the target is not necessarily the technically easiest one, minimising the chance of damage to normal tissues should always take precedence. The possibility that flow patterns will change in the course of the procedure must always he kept in mind and close attention should be given to the appearance of' previously non-visualised collateral pathways. Angiography mist he repeated as often as is necessary to monitor the effects of treatment and to look for unwanted effects or alterations in flow. Angiography at the end of the procedure demonstrates the effectiveness of the technique and defines any adverse effects on normal vessels. Heparin is often administered to decrease the risk of thrombosis during the procedure. Haemostasis at the puncture site may therefore not be easily obtained by simple manual pressure. particularly i f a large calibre sheath has been used. Reversal of anticoagulation or the use of puncture site closure devices (such as Angioseal°° or Pcrclose") can he considered in such cases. Endovascular procedures can he broadly divided into two categories: those that aim to restore a narrowed or obliterated cndovascular lumen supplying normal tissues and those that aim to exclude abnormal blood vessels from the circulation.


The femoral approach is almost exclusively used. A sheath of' appropriate size is always placed. Patients may have multiple diagnostic and therapeutic procedures and care should he taken to nmininuse trauma at puncture sites and to avoid frequent re-puncture at the same site.

Recanalisation can be achieved using mechanical devices such as dilatation balloons and stents or by using drugs with a thrombolytic or vasodilatatory effect.

Following the result of randomised trials, intravenous thrombolysis with recombinant tissue plasminogen activator (rTPA) has now heel) approved for the treatment of acute ischacmic stroke, within

1707

Previous Page

CONTENTS

3 hours of onset. The potential for clot recanalisation with intravenous thrombolysis is markedly dependent on the site of the occlusion. It is more successful in distal middle cerebral artery occlusions than in recanalisation of larger vessels such as the internal carotid artery (ICA) or proximal (M1) segment of the middle cerebral artery. Intra-arterial thrombolysis (Fig. 56. 1) achieves a higher local concentration of the thrombolytic agent in the vicinity of the clot, which should theoretically be more efficient, but it is more complex to achieve and therefore takes longer, increasing the delay between onset and treatment. The question remains whether a more efficient l ysis can prolong the therapeutic time window beyond 3 hours. A randomised trial (PROACT, Prolyse in Acute Cerebral Thromboembolism trial) has shown an overall benefit from intra-arterial thrombolysis in acute middle cerebral artery occlusion within 6 hours after the onset of stroke, despite a greater rate of haemorrhagic complication. In this trial 9 mg of recombinant pro-urokinase had been infused through a single end-hole microcatheter placed in the proximal third of the middle cerebral artery (MCA) thrombus over a period of 2 hours. Unfortunately urokinase has since been withdrawn from the market, however, the trial provided a proof of principle' that intra-arterial thrombolysis works and alternative pharmaceutical agents such as streptokinase and rTPA remain available for intra-arterial thrombolysis. Apart from the randomised PROACT trial there are numerous reports of successful intra-arterial throm-

Next Page

bolysis, not only in the MCA but also in internal carotid artery and vertebro-basilar occlusions. Success of recanalisation is reduced in distal internal carotid artery or T-junction occlusion. A retrospective analysis of acute basilar artery occlusions showed a considerable reduction in mortality with thrombolysis within 48 hours. Young patients with short-segment basilar occlusions benefited most from thrombolytic treatment and had better chances of recovery. It has been shown that ischaemic changes on diffusion-weighted MRI can be reversed with intro-arterial thrombolysis in certain cases. In the future it is likely that MR perfusion and diffusion i maging will play a much bigger role in selecting patients for thrombolysis, independent of rigid time-to-treatment windows. New developments in the endovascular treatment of acute stroke include the use of mechanical devices for extraction of acute thrombus. These have the advantage of reducing the treatment time (with a corresponding prolongation of the therapeutic window) and of a theoretical reduction in haemorrhagic complications associated with thrombolytic agents.

Cerebral vasospasm represents a significant cause of morbidity and mortality in patients with subarachnoid haemorrhage. Early treatment of ruptured aneurysms together with refinement of endovascular (see below) and surgical techniques have led to a reduction


Fig. 56.1 I ntra-arterial thrombolysis. Left internal carotid injection (A) in a patient who suffered an acute stroke during a coronary angiogram showing an abrupt change in calibre of the horizontal ( M1) portion of the left middle cerebral artery with occlusion of some M2 branches. Angiogram through microcatheter (B) that had been advanced into the acute middle cerebral artery clot. Superselective angiogram during thrombolysis (C) with urokinase shows re-opening of previously occluded MCA branches. Left internal carotid injection following the procedure (D) shows successful lysis of the acute l eft MCA thrombus. (Courtesy of Dr James Jackson.)

Previous Page

CONTENTS

Next Page

I NTERVENTIONAL NEURORADIOLOGY of morbidity and mortality from aneurysm re-rupture, but management of cerebral vasospasm remains still a challenge. Approximately one-third of patients with subarachnoid haemorrhage develop symptomatic vasospasm with delayed ischaemic deficit, of which 30% will die and a further 30% will have a permanent deficit. Medical treatment consists of hypertension, hypervolaemia and haemodilution (triple H therapy). Patients who fail to respond to medical therapy may benefit from endovascular treatment, which falls into two categories: mechanical dilatation of stenotic segments (balloon dilatation) and pharmacological relaxation of the contracted vessels with intra-arterial papaverine infusion. I ntracranial balloon angioplasty is performed with a small, compliant balloon attached to a microcatheter (Fig. 56.2). It is only suitable for treatment of the proximal segments of intracranial vessels, as dilatation of smaller branches carries a high risk of vessel rupture. Balloon dilatation has a more permanent effect than the infusion of papaverine through a microcatheter, which often l eads only to a transient increase of the vessel lumen. Papaverine may, however, be useful to facilitate access for endovascular treatment of a ruptured aneurysm if this is impeded by vasospasm. It can also be used to treat vasospasm in distal branches of cerebral arteries that are not suitable for balloon dilatation. The complications of both techniques include vessel dissection and rupture, vessel thrombosis or embolus and rupture of an unsecured aneurysm. Early endovascular treatment has a higher success rate. Treatment of an obviously infarcted territory should be avoided.

The European Carotid Surgery Trial (ECST) and the North American Symptomatic Carotid Endarterectomy Trial (NASCET) demonstrated that carotid endarterectomy significantly reduced the risk of recurrent stroke in patients with an over 70% carotid artery stenosis. Subsequently percutaneous transluminal angioplasty (PTA) and stenting have emerged as alternatives to surgery in these

1709

patients. A large randomised multicentre trial, the Carotid and Vertebral Artery Transluminal Angioplasty Study (CAVATAS) showed an almost identical 30-day outcome of mortality and major stroke between surgical and endovascular treatment, however, endovascular treatment was safer in terms of minor morbidity such as cranial nerve palsies and haematoma. Three-year follow-up data suggested no difference in outcome between the two treatments, suggesting that both are equally effective in preventing stroke. There has been a considerable evolution in endovascular techniques during and since the CAVATAS study. The majority of the patients in the endovascular arm of CAVATAS were treated with balloon angioplasty alone and only 22% were treated with stents. In the meantime primary stenting has become the endovascular technique of choice for carotid artery stenosis, replacing PTA. The majority of major strokes after carotid PTA are the result of dissection of the carotid artery at the time of balloon inflation with subsequent thrombosis. Dissection and occlusion of the carotid artery are less likely to occur with stenting because the stent maintains laminar flow across the stenosis and seals the site of dissection, preventing a free intimal flap. There is also evidence from endovascular intervention in the coronary circulation of superior outcomes with stenting compared with balloon angioplasty. Carotid stenting (Fig. 56.3) is generally performed under local anaesthesia. The patient should receive antiplatelet therapy (ideally a combination of aspirin and clopidogrel prior to and following the procedure) to minimise the risk of thromboembolic complications. Following preliminary angiography, a guiding catheter is placed in the common carotid artery, the stenosis is crossed with a soft tip guide-wire and a self-expandable stent is placed across the stenosis, normally bridging the carotid bifurcation. The self-expandable stent allows differential expansion of its distal and proximal portions adapting to the different diameters of the common and internal carotid arteries. Additional dilatation of the stenosis with a balloon catheter placed within the stent is usually performed. Recent developments in carotid stenting include protection devices such as filters and occlusion balloons, which can be placed


Fig. 56.2 Angioplasty of vasospasm. Right ICA angiogram following clipping of a right MCA aneurysm, which had bled, demonstrates severe vasospasm of the M1 segment of the right MCA (A). An unsubtracted image taken during balloon dilatation (B) demonstrates the inflated angioplasty balloon and aneurysms clip. Postangioplasty right ICA angiogram (C) demonstrating a normal calibre of the M1 segment.

Previous Page 1710

CONTENTS

Next Page

A TEXTBOOK OF RADIOLOGY AND IMAGING devices suggest that intracranial balloon dilatation and stenting have reached an acceptable safety level. Data on long-term patency, which may be compromised by a proliferative reaction of the vessel i ntima, are not yet available.

A number of embolic agents are available and new agents are being developed. A broad distinction can be made between mechanical devices (such as coils and balloons), particles and liquid embolic agents. Detachable balloons are one of the oldest mechanical devices and can be either of silicone or latex. They are used to occlude large vessels or large fistulas such as a carotico-cavernous fistula. One of the risks of this technique is the accidental detachment of a balloon i n an unwanted position, which can lead to obstruction of normal cerebral vessels. Systems with non-detachable balloons are avail-

Fig. 56.3 Carotid stenting. The preceding common carotid artery angiogram (A) demonstrates a severe stenosis at the origin of the internal carotid artery. An unsubtracted image during the procedure (B) shows a self-expandable stent placed across the stenosis with a residual waist at level of stenosis. The postprocedure right CCA angiogram (C) demonstrates successful treatment of the stenosis. distal to the carotid stenosis prior to stent deployment to catch any embolic material released during stenting. There is currently an active debate whether these protection devices should be used routinely, as they prolong and complicate the procedure and can, in rare cases, cause dissection of the distal internal carotid artery. It remains to be seen whether the benefits of theses devices outweigh these drawbacks. Larger published carotid stenting series suggest an acceptable complication rate and medium-term results comparable to endarterectomy. Prospective trials randomising surgical treatment versus carotid stenting are underway on both sides of the Atlantic.

able for temporary test occlusions. Guglielmi detachable coils (GDC) are currently the standard devices for embolisation of cerebral aneurysms. They are made of platinum and attached to a stainless steel delivery wire. The coils have a circular memory and are available in many different diameters and lengths. Detachment of the platinum coil from the delivery i s achieved by electrolysis following application of an electrical current once the coil is deemed to he in a safe position. The advantape of such a system is that the coil remains firmly attached to the delivery wire until the current is applied, which allows the retrieval of a coil in an inappropriate position or of inappropriate size. Various other detachable coils have recently become available, using similar principles with various detachment systems and coil configurations. Embolisation with particles leads to a less permanent occlusion than embolisation with liquid agents. Particles are therefore frequently used for preoperative embolisation. Polyvinyl alcohol (PVA) particles are currently the most frequently used particulate embolic agent and are supplied in sizes ranging from 50 to 1000 µm. PVA particles are diluted in radiopaque contrast medium as a suspension. Small particles (50-I50 µm) are more effective in inducing tumour necrosis but carry a higher risk of cranial nerve or skin damage. Gel foam sponge is prepared from pork-skin gelatine and is a pliable material that can be cut to appropriately sized particles. It is used for occlusion of more proximal vessels, frequently following embolisation of distal vessels with PVA. It has no significant sclerosing effect and recanalises within approximately 6 weeks. Trisacryl gelatine microspheres (Embospheres) are a new non-absorbable embolic agent. They tend not to aggregate, arc deformable and appear to penetrate deeper into the vascular system than PVA particles. There also preliminary evidence that Embospheres have a longer occlusive effect. Dehydrated (98%) alcohol is a potent liquid embolic agent. It has a sclerosing effect on cells it comes in contact with by virtue of causing dehydration and denaturation of peptides. Its main application is the percutaneous treatment of haemangiomatous malformations in the head and neck. It should be used with extreme care. Acrylate glues such as N-butyl-cyanoacrylate (NBCA) are liquid embolic agents, mainly used for embolisation of cerebral AVMs and dural fistulas. Prior to injection, they are diluted in a lipophylic contrast agent (Lipiodol) or made radiopaque with Tantalum powder. On contact with an ionic solution, such as blood, NBCA solidifies


Intracranial arteriosclerotic disease represents a significant cause of ischaemic strokes. It is more frequently seen in African-American or Asian ethic groups compared to Caucasians. The WarfarinAspirin Symptomatic Intracranial Disease Study showed that the annual stroke risk remains as high as 10% under therapy with is aspirin in patients with a greater than 50% stenosis of the MCA or distal ICA. Warfarin therapy reduced the annual stroke risk to 3.6% but was associated with a greater risk of haemorrhagic complications. The surgical treatment option consists of an extracraniali ntracranial bypass operation, which has not been shown to confer any additional benefit over aspirin therapy. The possibility of treati ng these patients with endovascular techniques such as balloon angioplasty and stenting represents an important development. This is a technique in evolution and dedicated intracranial stents have only recently become available, after the initial use of coronary stents in the intracranial circulation. Preliminary data using recent

Previous Page

CONTENTS

Next Page

I NTERVENTIONAL NEURORADIOLOGY rapidly, causing permanent vascular occlusion. The microcatheter through which it has been delivered has to be swiftly withdrawn to avoid the risk of gluing the catheter in-situ. It is a very effective but dangerous embolic agent, which is not easy to handle, and should only be used by experienced operators. Onyx is a new non-adhesive liquid embolic agent which consists of a mixture of ethyl-vinyl alcohol polymer (EVOH), dimethyl sulfoxide (DMSO) as a solvent and tantalum to render it radiopaque. In contrast to NBCA, it solidifies slowly, minimises the danger of in situ gluing of a microcatheter. Onyx has been used for cerebral arteriovenous malformations (AVMS) and giant cerebral aneurysms in which other forms of endovascular or surgical treatment are difficult.

Test occlusion of the internal carotid or vertebral arteries is performed to assess the ability of the cerebral circulation to tolerate permanent sacrifice of such a vessel. Sacrifice of the internal carotid or vertebral artery may be contemplated in the context of surgery for skull base and cervical neoplasms encasing these vessels, or for the treatment of giant aneurysms, where surgical or endovascular treatment with preservation of the parent artery is not possible. Testing the clinical tolerance of an occlusion is the most important part of the procedure, which has therefore to be performed under local anaesthesia. A temporary (non-detachable) occlusion balloon, fixed at the tip of a microcatheter, is advanced to the point of the anticipated permanent occlusion and then inflated for approximately 30 minutes with careful monitoring of the patient's neurological status, which is best undertaken by a separate operator. In the event of neurological deterioration, the balloon has to be deflated immediately to prevent a permanent deficit. In addition to clinical criteria, an injection of the contralateral ICA during inflation of the test occlusion balloon is made to assess cross flow through the circle of Willis and to detect any delay in the transit of contrast medium on the side of the occlusion. This is best done by comparing the appearance of cerebral veins in the cerebral hemispheres. A delay in venous opacification of 2 seconds or more indicates that the occlusion is unlikely to be tolerated. Test occlusion

1711

can also be combined with a number of simultaneously performed adjunctive tests, such as pharmacologically induced hypotension, transcranial Doppler, SPELT or Xenon CT. An added benefit from these procedures has, however, not been convincingly demonstrated at present. Complications of test occlusion include arterial dissection (symptomatic and asymptomatic) and transient or permanent neurological deficit. Routine use of heparin during the test occlusion limits the incidence and severity of these complications.

Trauma can result in a tear in the cavernous segment of the internal carotid artery, leading to a defect in the arterial wall with blood shunting directly into the cavernous sinus surrounding this arterial segment (direct carotico-cavernous fistula). From the cavernous sinus arterialised blood drains to reverse the flow in the ophthalmic veins, cerebral veins or the petrosal venous sinuses. Patients present with symptoms of venous hypertension, sometimes months after the trauma. Increased pressure in the ophthalmic veins typically causes proptosis, scleral injection and conjunctival haemorrhages. Increased cerebral venous pressure can cause headaches, seizures and venous haemorrhages. The goal of treatment is the occlusion of the arteriovenous fistula with preservation of the lumen of the internal carotid artery. This is best achieved from an intra-arterial approach using detachable balloons (Fig. 56.4). A microcatheter with a deflated attached balloon is passed from the carotid artery through the defect in its wall into the cavernous sinus. The balloon is then gradually inflated until the fistula is occluded from the venous side, preserving patency of the ICA and flow to the ipsilateral intracranial vessels. If this is achieved the balloon is detached by a gentle pulling action. Insertion of more than one balloon may be necessary in large fistulas. In experienced hands a successful occlusion can be obtained in up to 90% with this method. This is usually followed by rapid improvement of clinical symptoms. In the remaining cases alternative treatment options have to be considered. These include coil embolisation either transarterially, or using a transvenous approach via the inferior petrosal or ophthalmic


Fig. 56.4 Balloon embolisation of carotico-cavernous fistula. Internal carotid artery angiogram (A) demonstrating a traumatic carotico-cavernous fistula shunting into the ophthalmic veins and petrosal venous sinuses. Note the absence of opacification of the ipsilateral intracranial vessels due to steal through the shunt. Unsubtracted image (B) at the end of the procedure shows an inflated balloon, filled with contrast medium, within cavernous sinus. The postprocedure internal carotid angiogram (C) demonstrates complete closure of the fistula with normal anterograde opacification of the ipsilateral intracranial vessels.

Previous Page 1712

CONTENTS

Next Page


veins. Sacrifice of the carotid artery, which should always be preceded by a test occlusion, represents the last treatment option. Risks of this procedure include stroke following arterial occlusion after unintentional balloon detachment and cranial nerve palsies, usually transient, due to pressure from the balloon or coils.

The most common cause of epistaxis is the spontaneous or idiopathic form, frequently related to smoking and hypertension. Other causes of epistaxis include tumours (juvenile angiofibroma in male adolescents and malignant tumours in adults), trauma (caroticocavernous fistulas, traumatic false aneurysm or iatrogenic following endoscopic sinus surgery) and Osler-Weber-Rendu disease. Most cases of spontaneous epistaxis occur in the anterior septal area of the nose (Kiesselbach's plexus), a location that is easily accessible for cautery and packing. Posterior epistaxis is often more difficult to treat, requiring posterior nasal packing and endoscopic cauterisation. Endovascular embolisation is indicated if the bleedi ng is refractory to this initial treatment. Blood supply to the nasal fossa derives from the internal carotid artery (ethmoidal branches of the ophthalmic artery) and external carotid artery (maxillary artery and facial artery). The main blood supply derives from the sphenopalatine artery, a branch of the maxillary artery, which can be superselectively catheterised with microcatheters. Refractory idiopathic epistaxis is typically embolised with polyvinyl alcohol particles. This has an 87% success rate after embolisation of internal maxillary branches alone and a 97% success rate with additional facial artery embolisation. Potential complications include stroke and blindness as a result of inadvertent embolisation of ICA branches. It is therefore extremely important to check for the presence of dangerous external to internal carotid artery anastomosis before and during particle embolisation. It should be noted that in the setting of haemorrhage in acute trauma, particle embolisation is not a suitable technique for the treatment of bleeding due to rupture of a traumatic false aneurysm or vessel transection, as the volume of fluid used to deliver the par-

ticles may result in re-rupture, precipitating fresh haemorrhag Coil or glue embolisation should be considered in this setting.

Endovascular techniques are becoming more widely accepted and evolving technology is allowing endovascular treatment of arterial aneurysms that would have previously required surgery, or would have been regarded as untreatable. Published evidence suggests that endovascular techniques have comparable rates of immediate neurological complication to conventional surgery. while avoiding a craniotomy and surgical manipulation of brain. I n specific settings (aneurysms in the posterior fossa where surgical access is often difficult (Fig. 56.5), patients in poor medical condition), an endovascular approach is generally preferred. Some centres proceed directly to endovascular treatment immediately foll owing diagnostic angiography. Endovascular treatment also has an advantage in acute subarachnoid haemorrhage (SAH) where two aneurysms are found and it remains unclear which one has ruptured; both aneurysms can be treated acutely in the same sitting from an endovascular approach (Fig. 56.6) whereas surgical treatment might require two separate craniotomies. A large multicentre trial comparing endovascular and conventional surgical treatment of intracranial aneurysms is currently underway and it seems likely that in the future, endovascular techniques will play an increasingly important role in the management of intracranial arterial aneurysms. Because these techniques are relatively new, their long-term efficacy has not been established. Careful follow-up is therefore essential. Follow-up catheter or MR angiography has already demonstrated recurrence of aneurysms, especially at the basilar and carotid terminations. Currently, the most widely used technique is coil embolisation. This is achieved by careful placement of a microcatheter within the aneurysm lumen, through which detachable coils are deployed, provided they could be suitably positioned. If a satisfactory position cannot be obtained, the coils are retrievable before detachment. Thermally, mechanically and electrolytically detachable coils are


Fig. 56.5 Coil embolisation of superior cerebellar artery aneurysm. Left vertebral angiogram (A) demonstrating right superior cerebellar artery aneurysm. Note the caudal fusion and the origin of the perforating vessels off the more rostrally located P1 segment. Despite the broad neck and involvement of the superior cerebellar artery origin, it was decided to attempt endovascular treatment because of the posterior fossa location. A 3D GDC coil has been deployed (B) in good position, forming a basket across the aneurysm neck, avoiding the region of the superior cerebellar artery origin and the laterally proj ecting nipple, presumed to be the site of rupture. Final angiogram (C) demonstrating complete exclusion of the aneurysm without compromise of the adjacent vessels.

Previous Page

CONTENTS

Next Page

I NTERVENTIONAL NEURORADIOLOGY

1713

Fig. 56.6 Coil embolisation of bilateral intracranial aneurysms. Left ICA angiogram (A) demonstrating a terminal carotid aneurysm. Note the nipple pointing medially, an angiographic sign often seen in acutely ruptured aneurysms. Right ICA angiogram (B) demonstrating a posterior communicating segment aneurysm. CT (not shown) had revealed the greatest density of subarachnoid blood to lie in this region. Because of the disparity between the angiographic and CT findings, it was not certain which of the two aneurysms had bled. It was therefore decided to treat both in a single sitting, despite the risk of intervention in two different arterial territories. Bilateral surgical intervention was not considered to be a safe option. Left ICA angiogram (C) demonstrates the terminal carotid aneurysm to be excluded by coils. Right ICA angiogram (D) demonstrates the posterior communicating segment aneurysm to be excluded by coils. Unsubtracted image (E) shows the position of the coils.


Fig. 56.7 Coil embolisation of of wide-necked aneurysm with balloon remodelling. Subtracted and unsubtracted left internal carotid angiography before (A,B) and after (C,D) balloon remodelling assisted GDC coiling. The large supraclinoid aneurysm has a very wide neck (a). A remodelling balloon was placed across the neck of the aneurysm, seen on the unsubtracted image (B). Also note the microcatheter tip placed within the aneurysm lumen. The remodelling balloon was inflated for short periods during the coil deployment to prevent coil prolapse into the internal carotid artery. This enabled complete exclusion of the aneurysm, without compromise of lumen of the internal carotid artery (C). The dense coil mass is seen on the final unsubtracted image (D).

available, in a wide variety of shapes and sizes. Embolisation of

point where they are suitable for widespread clinical use, though

aneurysms using fluid materials and exclusion of the aneurysm l umen with stents are techniques that have not yet evolved to the

they are performed in some highly specialised centres and may be i mportant in the future.

Previous Page

1714

CONTENTS

Next Page

A TEXTBOOK OF RADIOLOGY AND IMAGING aneurysms. Aneurysms with wide necks that would previously have been regarded as difficult to treat are increasingly amenable to endovascular treatment since the development of techniques preventing the prolapse of coils through the aneurysm neck, such as balloon remodelling (Fig. 56.7), stenting devices to bridge the aneurysm neck. Potential complications of endovascular treatment include aneurysm perforation or rupture (Fig. 56.8) (by guide-wire, microcatheter or coil), inadvertent occlusion of the parent or branch arteries by coils (Fig. 56.9) and thromboetnbolic events. Where preservation of patency of the parent vessel cannot be assured (fusiform or serpentine aneurysms, wide-necked aneurysms, false aneurysms and arterial dissection), sacrifice of the parent vessel may be considered, usually preceded by test occlusion. This can be achieved with either coils or detachable balloons, depending on the vessel configuration and indication for treatment.

Fig. 56.8 Rupture during coil embolisation of intracranial aneurysms. Left vertebral angiogram (A) demonstrating extravasation of contrast from ruptured basilar tip aneurysms during coil placement. Note coils within aneurysm and extravasated contrast around the basilar artery in the i nterpeduncular and pre-pontine cisterns. Left vertebral angiogram (B) demonstrating cessation of extravasation after reversal of heparin with protamine and further packing of the aneurysm, to exclusion. The patient suffered no adverse consequences. Irrespective of the technique used, the aim is total exclusion of the aneurysm lumen with preservation of the parent artery and its branches. This is most reliably achieved in narrow-necked saccular

Cerebral arteriovenous malformations arc less common than cerebral aneurysms, with a prevalence of 0.2-0.5%. Most patients with cerebral AVMs develop symptoms before the age of 40 years. which is in contrast to cerebral aneurysms where only a quarter of patients present by the age of 40. Cerebral AVMs are an important cause of cerebral haemorrhage in young adults. Approximately 60% of patients with cerebral AVMs present with haemorrhage; epilepsy is the second commonest presentation, followed by nonhacmorrhagic neurological deficit and headache. AVMs are presumed to have usually pursued an indolent course before they present. They may occur in any location in the brain,


Fig. 56.9 Arterial occlusion due to coil embolisation of intracranial aneurysms. Left ICA angiogram (A) demonstrating coils within an aneurysm arising at the right side of the large anterior communicating artery and the origin of the right A2 segment. Note the good filling of the A2 segment. Left ICA angiogram (B) demonstrating occlusion of the A2 segment after deployment of a subsequent coil. The coil was successfully retrieved before detachment, but the vessel remained occluded. Fortunately, the patient rapidly developed collateral flow from the posterior cerebral artery and suffered no adverse consequences.

Previous Page

CONTENTS

Next Page


1715

either supra- or infratentorial, from the cortical surface to the cular treatment as well as for the other treatment modalities. It may ependymal surface, or within the choroid plexus or ventricular require several staged embolisation procedures and complete AVM system. Arterial supply and venous drainage may therefore be from obliteration is often not possible. Endovascular treatment can be any combination of superficial cortical vessels or deep perforating combined with other forms of treatment. It is sometimes performed vessels. Meningeal vessels may also contribute to supply, usually to reduce the size of the AVM prior to radiosurgery, as the severity after previous haemorrhage or surgery, but occasionally spontaand incidence of side effects, such as radiation necrosis and radianeously. tion induce white matter disease, depend directly on the volume of AVMs can be treated by endovascular embolisation, micro- irradiated brain tissue. The role of partial AVM embolisation is actively debated and surgery, radiosurgery or a combination of these methods. There is remains controversial. It can be effective in improving the control considerable uncertainty and lack of uniformity of opinion regarding the indications for, and strategies of, AVM treatment. The of epileptic seizures and in reducing the risk of haemorrhage by tarchoice of treatment is often governed by local expertise and the geted embolisation of specific components such as false aneurysms. Partial embolisation can also bring relief of symptoms related to goal of treatment may vary in individual patients depending on the venous hypertension and the elimination of external carotid artery clinical presentation, AVM location and size, and age of the patient. supply is often effective in alleviating headaches. The perceived risks and benefits of treatment should be weighed up for each specific case and discussed by a multidisciplinary team. A commonly acknowledged aim is the prevention of cerebral haemEndovascular treatment of vein of Galen aneurysmal orrhage or re-haemorrhage (which is associated with higher mortalmalformations (VGAM) i ty and morbidity than an initial haemorrhage). The mean annual VGAM is characterised by arteriovenous shunts at the choroidal level, risk of haemorrhage from a cerebral AVM is probably 2-4%, but within the subarachnoid space, in the region of the vein of Galen. may be considerably increased in the presence of certain patient They must be distinguished from cerebral arteriovenous shunts drainfactors and angiographic features. The former include older age and i ng into the vein of Galen and producing dilatation of it as these l esions have entirely different natural histories and prognoses. initial presentation with haemorrhage and the latter include false aneurysms, venous ectasias or stenoses, a single draining vein, deep I n VGAM, there is associated aneurysmal dilatation and abnormal venous drainage and venous congestion. persistence of the median vein of the prosencephalon, an embryologiThe most commonly used endovascular approach is arterial cal midline venous structure and precursor of the vein of Galen. The embolisation using a permanent fluid agent such as NBCA, which malformation often drains into a persistent falcine sinus and there may is diluted and rendered radiopaque with Lipiodol or Tantalum be agenesis of the straight sinus and anomalies of other dural venous (Fig. 56.10). The degree of dilution governs the speed of polymerisinuses. sation and can be varied to suit the rapidity of arteriovenous shuntVGAM and dural sinus malformations are the only intracranial i ng and distance to be crossed be this embolic agent. More recently AVMS, demonstrated to have been present in the fetal period. They Onyx has also been successfully used for embolisation of cerebral arc more common in males. Most present in either the neonatal period AVMs. or in infancy. In general, the high flow arteriovenous shunt of VGAM Small AVMs can be treated by endovascular means, radiosurgery is initially well tolerated from a neurological point of view and early or, if in a superficial location, by microsurgery. Radical obliteration presentation is with cardiac failure or altered CSF dynamics and of medium-sized or large AVMs represents a challenge for endovasmacrocephaly.


Fig. 56.10 Vermian AVM embolisation. Pre-embolisation left vertebral angiogram (A) demonstrating vermian arteriovenous malformation supplied by vermian branches of the superior cerebellar arteries. Microcatheter NBCA injection (B) into arteriovenous malformation nidus. Postembolisation of both major pedicles, left vertebral angiography (C) demonstrates almost complete obliteration of the nidus.

Previous Page 1716

CONTENTS

Next Page


Fig. 56.11 Embolisation of vein of Galen aneurysmal malformation. Left lateral angiogram (A) demonstrating aneurysmal dilatation of the vein of Galen and rapid arteriovenous shunting into a persistent falcine sinus. A microcatheter was passed into one of the feeding pedicles (B) and NBCA rendered radiopaque with Tantalum was injected (C) through it. This achieved complete closure of the shunt (D). (Courtesy of Dr W. Taylor.)


If possible, treatment of the shunt in VGAM is deferred to allow for growth of the child, as intervention in neonates is technically difficult and hazardous. Criteria for intervention in infancy include progressive macrocephaly, developmental delay, seizures and reversible neurological deficit. Cardiac failure unresponsive to medical management is an indication for urgent embolisation in the neonatal period. An arterial approach using a permanent embolic agent (NBCA) is most often used (Fig. 56.11). Venous approaches have also been described. The initial aim is to reduce the level of arteriovenous shunting to a degree that the cardiac failure can be managed. Further staged treatment can be performed after maturation of the child. Total cure is often possible.

neuropsychiatric syndromes. Impediments to venous outflow result i n venous reflux and congestion, which predisposes to haemorrhage and cerebral parenchymal damage. Cerebral venous congestion is an indication for treatment (Figs. 56.12, 56.13). Endovascular techniques are the first-line intervention in DAYS. An arterial approach utilising a permanent embolic agent is often used. The goal should be occlusion of the site of arteriovenous shunting with preservation of patency of the dural venous sinuses. Care should be taken not to redirect venous flow toward veins that drain the brain by occlusion of alternative outflow pathways. A venous approach may also he used, though this often necessitates sacrifice of a segment of dural venous sinus at the zone of arteriovenous shunting. This may be justified when there are multiple sites of shunting, multiple sources of arterial supply or inaccessible arterial sources. Venous approaches can be particularly useful in the treatment of cavernous sinus dural fistulas.

Endovascular treatment of dural arteriovenous shunts

Dural arteriovenous shunts (DAVS) occur within the walls of dural venous sinuses or their tributaries. They may be related to previous episodes of sinus thrombosis. Depending on their location and severity of shunting, they may present with pulsatile tinnitus, bruit, ophthalmic symptoms (cavernous sinus DAYS), seizures, intracranial haemorrhage (Fig. 56.12) or decreased mental function and

Embolisation of vascular tumour of the head, neck and CNS prior to surgery can significantly reduce intraoperative blood loss, shorten the operative procedure time and facilitate the removal of these tumours. Hypervascular tumours which may benefit from preoperative embolisation include meningiomas, metastases, glomus

Previous Page

CONTENTS

Next Page



1717

Fig. 56.12 Embolisation of cortical vein dural arteriovenous shunt presenting with parenchymal haemorrhage. CT without contrast at presentation (A) demonstrating right parenchymal haematoma. Right external carotid angiogram (B) showing arteriovenous shunting into a cortical vein in the region of the haematoma. Microcatheter angiogram from a distal position within a frontal branch of the right middle meningeal artery (C) better demonstrating the dural arteriovenous shunt into a cortical vein. There were several additional branches to the dural shunt off the left middle meningeal artery (not shown). NBCA/ Lipiodol was injected from this position (D). This achieved complete closure with no arteriovenous shunting evident from either external carotid artery (E,F). The internal carotid injection (G) demonstrates preservation of the superior sagittal sinus. CT without contrast after embolisation (H) demonstrates the dense glue cast and the evolving haematoma.

jugulare tumours (Fig. 56.14), benign and malignant bone tumours and juvenile nasopharyngeal angiofibromas. Embolisation is usually performed via the endovascular route but can also be achieved by direct injection of embolic agent into the tumour (see below). The aim is a reduction of blood supply to the tumour and this is most efficient if the embolic agent can penetrate into the small vessels within the tumour, which is best achieved using parti-

cles (polyvinyl alcohol or Embosphereso) of a small size. The choice of particle size is a balance between safety and efficiency and depends on whether a highly selective catheter position could be achieved. Smaller particles penetrate deeper into the tumour but carry a higher risk of skin necrosis and cranial nerve palsies, risks they share with liquid agent such as glue or concentrate alcohol, which are occasionally used.

Previous Page 171 8

CONTENTS

Next Page


Fig. 56.13 Embolisation of dural arteriovenous shunt causing cortical venous congestion. CT with contrast (A) demonstrating dilated cortical and deep veins, including the septal vein. Right ECA angiogram (B) demonstrating a dural fistula of right transverse sinus supplied by meningeal and transosseous branches with retrograde cortical veinous reflux. Unsubtracted image (C) showing NBCA glue cast at the site of the fistula. Note how the glue has penetrated the venous side of the fistula, with the cast corresponding to the zone of shunting demonstrated in (B). Right ECA angiogram (D) demonstrating complete occlusion of the fistula. Note the decrease in calibre of the occipital artery on this follow-up angiogram. (Courtesy of Prof. P. Lasjaunias.)

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ Fig. 56.14 Embolisation of glomus jugulare tumour. External carotid angiogram (A) showing a hypervascular tumour at the level of the jugular foramen. Superselective catheterisation of the ascending pharyngeal artery (B) prior to embolisation demonstrates that this vessel provides the main vascular supply to the mass, a typical finding in glomus jugulare tumours. The external carotid artery angiogram following particle embolisation (C) shows almost complete obliteration of the vascular supply. There is also some vasospasm near the catheter tip.

Previous Page

CONTENTS

Next Page


171 9

Fig. 56.15 Paliative embolisation of malignant extra-axial tumour. A contrastenhanced T 1 -weighted MRI (A) demonstrates a markedly enhancing metastasis with extradural and extracranial extension (presumed to be from a known neuroendocrine primary). Superselective angiography of the left internal maxillary artery (B) confirms marked hypervascularity of the mass. Injection of the same vessel following particle embolisation (C) shows devascularisation of the tumour. A contrastenhanced MRI after one week (D) shows large non-enhancing necrotic areas within the tumour.


External carotid artery branches supply many of these tumours. Before and during the embolisation, a meticulous search for the presence of anastomoses from the external to the internal carotid or vertebral arteries is mandatory. It is important to realise that these collateral pathways can open during the procedure following alteration of the local haemodynamics. Anatomical variants such as ophthalmic supply from middle meningeal artery must be actively sought. Endovascular tumour embolisation may also be indicated as a palliative procedure, to reduce tumour size or associated symptoms in patients with lesions not amenable to surgery (Fig. 56.15). Intra-arterial chemotherapy is an established form of treatment i n head and neck tumours that fail to respond to conventional therapy. Chemotherapeutic agents can be infused following catheterisation of external carotid branched via the femoral route or through an infusion pump, implanted via surgical cutdown of the

Fig. 56.16 Spinal dural arteriovenous fistula. Spinal angiogram in a patient with progressive leg weakness and sphincter disturbance. Selective injection of right thoracic intercostal vessels shows a dural arteriovenous fistula shunting into perimedullary veins, which drain superiorly and inferiorly.

Previous Page

1720

CONTENTS

Next Page


Vertebral haemangioma biopsy and embolisation. T 2-weighted axial MRI (A) of a vertebral haemangioma causing cord compression. CT-guided core biopsy (B). The biopsy needle was placed through the pedicle of the vertebral body to decrease the chance of causing haemorrhage within vertebral canal. Pre-embolisation right intercostal artery angiogram (C) demonstrating the hypervascular haemangioma. Postembolisation angiogram (D) demonstrating almost complete devascularisation after embolisation with PVA particles. Fig. 56.17


superficial temporal artery. Local toxicity of this treatment includes alopecia, mucositis, skin necrosis and cranial nerve palsies.

shunts

a gradual or stepwise loss of power in the legs, walking difficulties and sphincter disturbance. The symptoms are due to increased (arterial) pressure in the veins draining the spinal cord leading to venous congestion and cord oedema, evident as increased T2signal intensity within the cord on MR images. Definitive diagnosis of spinal dural fistulas requires high-quality spinal angiography, which should be performed under general anaesthesia (Fig. 56.16). Spinal dural fistulas may be treated by embolisation or surgery. Embolisation is an option if the fistula arises from a purely radicular artery, which does not provide any arterial (either a radiculomedullary or radiculopial artery) supply to the cord. It involves the i njection of a permanent occlusive agent such as NBCA glue, which must reach and occlude the fistulous point. Surgery of these lesions is usually straightforward and consists of exposure and electrocoagulation of the fistula, which is located on the dura. The chief aim of treatment is to prevent further progression of symptoms. Recovery from existing symptoms is possible but can be slow and depends on the degree of established cord damage, which is related to the length of clinical history. Diagnosis of this relatively infrequent condition is unfortunately often delayed and patients have often had negative or inconclusive investigations for lumbar or

In principal, endovascular treatment of arteriovenous malformations of the spinal cord does not differ from the brain, however, because of the small size of the vessels of the spinal cord, its dependence on one channel (the anterior spinal arterial axis) and the catastrophic consequences of occlusion of the arterial supply to even a small volume of the cord, it is an especially unforgiving territory to embolise in. Complete eradication of an arteriovenous shunt can seldom be safely achieved by embolisation. As in the brain, a permanent fluid embolic agent such as NBCA is preferred. Whereas spinal cord AVMs are seen in young patients, spinal dural fistulas most commonly present in elderly male patients and are thought to be associated with venous thrombosis, like cranial dural fistulas. Clinical symptoms may have an insidious onset with

thoracic disc disease. It is therefore important to heighten the awareness for this disease and the available treatment options. Their diagnosis is not infrequently missed, sometimes even after (usually inadequate) spinal angiography, and it must be remembered that the definitive diagnosis of spinal dural fistulas requires high-quality spinal angiography.

Previous Page

CONTENTS

Next Page

I NTERVENTIONAL NEURORADIOLOGY Endovascular treatment of vertebral tumours

Embolisation for vertebral tumours is generally performed preoperatively, aiming to reduce intraoperative bleeding (Fig. 56.17 C,D), though occasionally it may result in resolution of symptoms (either pain or cord compression). Both benign (especially haemangioma) and malignant (usually metastatic) hypervascular lesions may benefit from endovascular treatment. High-quality diagnostic spinal angiography must he performed first. This should demonstrate the blood supply to and degree of vascularity of the lesion. The arterial supply to vertebral bodies is from the two adjacent metameric levels, and these should he exami ned bilaterally. The blood supply to the spinal cord must be carefully examined as inadvertent embolisation of a radiculopial or radiculomedullary artery may have terrible consequences. It will also be useful for the surgeon to know the location of any radiculomedullary artery in the region to be operated on. If the lesion is hypervascular and there is no radiculomedullary or radiculopial artery arising from the same level, embolisation can he performed safely. It should be remembered that patterns of flow may alter in the course of embolisation, and the possibility of a previously unopacified vessel contributing to the supply of the spinal cord should be excluded. Embolisation may be performed either through a catheter positioned in the segmental artery supplying the tumour, or more selectively thorough a microcatheter in the feeding pedicles of the tumour itself. Occlusion of the segmental artery distal to the origin of supply to the tumour may be useful to reduce passage of embolic material to vessels supplying the skin and body wall.

Fluoroscopy or cross-sectional imaging can be used to accurately place a biopsy needle and to avoid injury to adjacent vital structures, such as the spinal cord or major arteries (Fig. 56.17B). Biopsies of the spine are frequently performed under fluoroscopic or CT guidance. Imaging guided biopsies can also be performed on an interventional ('open') MR system, which allows multiplanar scanning and offers more flexibility in choosing the access route. Special MR-

1721

compatible non-ferromagnetic instruments are required. In addition to percutaneous procedures performed in the radiology department, i maging plays an important role in neurosurgical navigation. This can be based on preoperative acquisition of volumetric CT or MR scans using a stereolactic frame or on intraoperative MR, which offers `real-time' monitoring of neurosurgical procedures.

Percutaneous access under image guidance can also be used for minimally invasive treatment.

Low-flow vascular malformations of the head and neck, such as hacmangio-lymphangiomas and venous malformations, can be treated with direct injections of concentrated alcohol under fluoroscopic guidance. This procedure has to be proceeded by an angiographic run during the injection of a contrast medium through the percutaneously placed needle to delineate the vascular compartments and venous drainage route. Injection of concentrated alcohol should not be undertaken in the presence of intracranial venous drainage. Spinal tumours or intracranial extradural tumours can be directly punctured using image guidance and embolised with particles or glue through a percutaneous needle, preoperatively or for palliation. Direct percutaneous tumour - embolisation can he more efficient than embolisation of several feeding pedicles via the endovascular route hut should only be undertaken be an experienced operator, using biplane fluoroscopy for intracranial lesions.


The aim of percutaneous vertebroplasty is mechanical strengthening of weakened or partially destroyed vertebral body by the percutaneous injection of acrylic cement, a substance which has been used in surgery for decades. The first image-guided percutaneous

Fig. 56.18 Percutaneous vertebroplasty. Fluoroscopic images in frontal (A) and lateral (B) projections showing transpedicular course of cannulas prior to first injection of PMMA. Note marked osteopenia and loss of height of vertebral bodies in this patient with multiple myeloma. Subsequently cannulas were placed (C) in the other severely affected levels. A total of three levels were treated (D). The patient was able to cease morphine after the procedure. (Courtesy of Professor Juergen Reul.)

Previous Page

1722

CONTENTS

Next Page


vertebroplasty (PVP) was performed in France in 1 984, using the injection of polymethylmethacrylate (PMMA) into a C2 vertebra that had been partially destroyed by an aggressive haemangioma. Thereafter, PVP was used to treat vertebral compression fractures caused by metastatic or primary bone tumours (haemangiomas, giant cell tumours) and osteoporosis (primary or steroid induced). PVP is able to provide rapid and lasting relief from the pain associated with vertebral compression fractures and is increasingly used for that purpose. High-quality fluoroscopic equipment is a prerequisite for PVP. It is essential for accurate placement of the cannula and real-time visualisation of the cement injection (Fig. 56.18). CT has been described as an aid to fluoroscopy, but adds considerable complexity and is most useful in the cervical spine to avoid the carotid vessels and down to approximately T4 where lateral fluoroscopy may be impossible through the shoulders. Reported clinical complications following PVP include fever, i nfection, cement pulmonary embolism, radiculopathy and cord compression. Treatment of osteoporotic compression fractures has a l ow rate (1-2 (lo ) of mostly ncrn-neurological complications. The complication rate increases in patients treated for malignant compression fractures, because of a higher risk of cement leakage from vertebrae destroyed by tumour. A 3-6% rate of (mostly transient) radiculopathy has been reported in this group.

Photodynamic therapy is a minimally invasive treatment with great promise in malignant disease. It can be applied before or after chemotherapy and he repeated as necessary. It is based on the interaction of an intravenously injected photosensitising drug and l ocally delivered light of a specific wavelength, which activates the drug. Interaction of the activated photoscnsitiser with oxygen results in the production of a highly cytotoxic singlet oxygen. This

Bendszus, M., Klein R., Burger R., Metz M. W., Hofmann E., Solyrnosi L. (2000) Efficacy of trisacryl gelatin microspheres versus polyvinyl alcohol particle in the preoperative cmbolization of meningiomas. A merican Journal of Neuroradiology, 21, 255-261. Borden J. A.. Wu J. K., Shucart W. A. (1995) A proposed classification for spinal and cranial dural arteriovenous fistulous malformations and i mplications for treatment. Journal of Neurosurgerv, 82, 1 66-179. Brandt T., Kummer R. V., Mullcr-Kuppers M.. Hacks W. (1996) Thrombolytic therapy of acute basilar artery occlusion: variables affecting recanalyzation and outcome. Stroke, 27, 875-881. Cavatas Investigators (2001) Endovascular versus surgical treatment in patients with carotid stenosis in the Carotid and Vertebral Artery Transluminal Angioplasty Study (CAVATAS): a randomised trial. Lancet, 357(9270),1729-1737. Chimowitz M. I., Kokkinos J., Strong J.. et al (1995) The warfarin-aspirin symptomatic intracranial disease study. Neurology, 45. 1 488-1493. Davies M. A., Saleh J., Ter Brugge K., Willinsky R., Wallace M. C. (1997) The natural history and management of intracranial ducal arteriovenous fistulae (part 1: benign lesions). Interventional Neuroradiologv, 3(4). Davies M. A., Saleh J., Ter Brugge K., Willinsky R., Wallace M. C. (1997) The natural history and management of intracranial dural arteriovenous fistulae (part 2: aggressive lesions). Interventional Neurnrndiology, 3 (4). Dovey Z.. Misra M., Thornton J., Charbel F. T., Debrun G. N.. Ausman J. 1. (2001) Guglielmi detachable coiling for intracranial aneurysms. The story so far. A rchives of Neurology, 58. 559-564. Elliott J. P., Newell D. W.. Lain D. J., et al (1998) Comparison of balloon angioplasty and papaverine infusion for the treatment of vasospasm following aneurysmal subarachnoid hemorrhage. .lournal of Neurosurgery. 88,277-284. Furlan A., Higashida R., Weschler L. (1999) Infra- arterial prourokinase for acute ischemic stroke. The PROACT study: a randomized controlled trial. Journal of the A merican Medical A ssociation, 282, 2003-2011. Gobi n Y., Laurent A., Merienne L., et al (1996) Treatment of brain arteriovenous malformations by emholization and radiosurgcry../ournal of Neurosurgerv, 85, 1 9-28. Hopper C. (2000) Photodynamic therapy: a clinical reality in the treatment of cancer. Lancet, 1. 212-219. Johnson L H., Whittle 1. R., Besser M., Morgan M. K. (1987) Vein of Galen malformation: diagnosis and management. Neurosurgerv, 20, Lasjaunias P. (1997) A revised concept of the congenital nature of cerebral arteriovenous malformations. Interventional Neuroradiologc, 3 (4). Lasjaunias P. (1997) Vascular Diseases in Neonates, Infants and Children. Berlin: Springer-Verlag. Lasjaunias P., Berenstein A. (1987) Surgical Nearoongiograplnr, Vol. 2; Endovascular Treatment of Craniofacial Lesions. Berlin: Springer-Verlag. Lasjaunias P., Alvarez H., Rodesch G., Garcia-Monaco R., Burrows P., Taylor W. (1996) Aneurysmal malformations of the vein of galen. Follow up ol' 1 20 children treated between 1984-1994. Interventional Neurnrndiology, 2, 15-26. Lasjaunias P., Berenstcin A., TerBrugge K. G. (2001) Surgical Neuroangiogrop/iv, Vol. I ; Clinical V ascular A natomy and V ariations, 2nd edn. Berlin: Springer-Verlag. Massoud T. F., Hademenos G. J. (1996) Risk of intracranial arteriovenous malformation rupture due to venous drainage impairment. A theoretical analysis. Stroke, 27(6). Mathis J. M., Barr J. D., Belkoff S. M., Barr M. S., Jensen M. F., Deramond H. (2001) Percutaneous vertebroplasty: a developing standard of care for vertebral compression fractures. A merican .loutnal of Neurnrndiology. 22, 373-381. Rodesch G., Hurth M., Alvarez H., Lasjaunias P. (1997) Embolisation of spinal chord arteriovenous malformations with glue through the anterior spinal axis. Interventional Neuroradiologv, 3, 1 31-143. Roubin G. S.. New G.. Iyer S. S., et al (2001) Immediate and late clinical artery stentmg in patients with symptomatic and asymptomatic carotid artery syenosis. Circulation, 103, 532-545. TerBrugge K. G. (2001) Vein of Galen management in neonatal period. A merican Journal of'Neurorudiology, 22, 1 236-1237. Tong F. C., Cloft H. J., Dion J. E. (2000) Endovascular treatment of i ntracranial aneurysms with Guglielmi detachable coils: emphasis on new techniques. Journal of Clinical Neuroscience, 7, 244-253. Vitek J. J. (1991) Idiopathic intractable epistaxis: endovascular therapy. Radiology, 181, 1 1 3-116.


technique has been used as surface illumination of superficial

mucosal lesions in the past. Interstitial light delivery to solid tumours of the head and neck is greatly facilitated by imaging-

guided placement of needles through which optic fibres can be i nserted, providing a focused delivery of laser light (Fig. 56.19).

Fig. 56.19 Photodynamic therapy of carcinoma tongue. Sagittal (A) and axial (B) T i -weighted scans performed on an 'open' interventional MR scanner of a patient with a carcinoma of the base of tongue causing airway obstruction. MR-compatible guiding needles, which appear as low-signalintensity structures, were inserted into the tumour using MR guidance. Palliative treatment with PDT was carried out by inserting optic fibres through these needles.

Previous Page

CONTENTS

Next Page

57 Disorders of the brain will be considered in this and the following chapter under the following headings:

• • • • • • • • • • • •

Methods of investigation Some aspects of technique Congenital lesions Neoplasms Vascular lesions Trauma Inflammatory lesions Degenerative and metabolic lesions White matter disease Epilepsy Ultrasound Radioisotopes.


The following methods arc in current use for investigation of diseases of the brain: 1. 2. 3. 4. 5. 6.

transcranially for the circle of Willis (see Ch. 15). Ultrasound may also be used during operations to assist the neurosurgeon in defining tumour extent and cysts. Simple radiography has been used in the past as a cheap preliminary screening technique but its accuracy and specificity are so low compared with those of CT and MRI that most workers now dispense with its use. It is, however, still valuable in the assessment of skull trauma. With suspected cerebral lesions it is helpful in lesions i nvolving the skull vault and base and has been discussed in detail i n Chapter 53. A rteriography is no longer used for the diagnosis of tumours except in special circumstances as described below, but is still essential for the definitive diagnosis of most vascular lesions (Ch. 55). Its use has also extended into the field of interventional radiology (Ch. 56). As explained in Chapter 55, non-invasive or minimally invasive techniques for arteriography in the form of MRA and spiral CT are now preferred to direct artcriography and should be used where possible.

Simple radiography Computed tomography (CT) Magnetic resonance imaging (MRI) Ultrasound Angiography Radionuclides.

CT and MRI are now the techniques of choice for investigating most suspected lesions of the brain, and this and the following chapter are mainly devoted to their use in defining intracranial pathology. The use of radioisotopes as a method for the localisation of i ntracranial lesions is now obsolete, apart from a few special indications which are discussed in the section on radionuclides at the end of Chapter 58. Positron emission tomography (PET) is proving useful in certain as aspects of brain disease by showing differences in local brain metabolism but is expensive and is only available at special centres. It is also discussed below. Ultrasound is very useful in neonates and infants but not in older children and adults, since it requires the window of an open fontanelle to image the brain (see Ch. 58). However, duplex and Doppler ultrasound can be used to investigate the neck vessels and

Some aspects of CT and MRI techniques

Data presentation CT and MRI present data as an analogue display on a console. Presentation is in the form of a `grey scale' in which whiteness is proportional to the X-ray attenuation coefficient or nuclear magnetic resonance (NMR) signal intensity at each point i n the image. The range can be varied by changing the `window' ( W) so that tissues within a wide or narrow range of CT or MRI characteristics can be evaluated. The centre point or level (L) of the window can also be varied. Standardising W and L is of more i mportance in CT; the range of imaging options and signal variation that occurs in MRI mitigates against rigorous standardisation. Figure 57.1 shows the relative density on the Hounsfield scale of the normal body tissues at CT. This scale is an arbitrary one with air at -1000 units and water at 0 units as fixed points. The numerical value assigned to the attenuation coefficient bears a linear relationship to the electron density of the tissue concerned. Although tissue contrast in MRI depends mainly upon proton density and T, and Tz relaxation times, other contrast mechanisms come into play, and absolute values show considerable variation according to measurement method and equipment used. Soon after the introduction of CT brain scanning, Ambrose (1973) showed that pathological lesions in the Contrast enhancement

1723

Previous Page 1724

CONTENTS

Next Page

A TEXTBOOK OF RADIOLOGY AND IMAGING gadolinium (Gd 3 +). This is chelated to a large carrier molecule such as DTPA and injected intravenously as a dose of 1-2 mM/litre/kg body weight. Head positioning Most CT studies are performed with the head

supine and the plane of section at 10-25 ° to Reid's baseline. This is comfortable for the patient and reduces radiation to the sensitive lenses of the eye. The appearances in a set of normal axial scans are illustrated in Figure 57.2. Modern gantries are wide enough to permit coronal or nearcoronal sections to be obtained directly if the patient is able to cooperate by assuming the appropriate position. Sagittal or nearsagittal sections can also be obtained (Fig. 57.3). In most cases, however, these projections are obtained by computer reconstructions of the stacked axial slices (reformat). However similar facilities arc now available with the new generation of multislice spiral

Fig. 57.1 The Hounsfield CT scale. The full scale on the left extends over 2000 units. The expanded scale on the right extends over 200 units and i ncludes all body tissues. Head scans are usually done routinely at a window l evel (L) of 34-40 and a window (W ) covering 0-75.

CT scanners. The issue is simpler with MRI, where the scan plane is set electronically and multiplanar facilities are readily available. Figure 57.4 shows a set of normal MR scans in the usual axial plane, and Figure 57.5 a set in the coronal plane. MRI protocols Protocol choice for MRI is much wider than for

CT even in brain imaging, the longest established and in many ways simplest of all organs to image. Most modern units no l onger use conventional spin-echo sequences, but one or more of brain frequently accepted intravenous iodinated contrast media the fast techniques are now available. In practice, this generally such as those used in pyelography in a manner different from normal brain. This is due partly to the breakdown of the entails the use of gradient-recalled echo ( GRE) imaging for threedimensional (3D) or volumetric acquisitions, and fast spin-echo blood-brain barrier which occurs in brain tissue damaged by for multislice (2D) acquisitions. Also, fast spin-echo techniques disease and partly to the abnormal vascularity of some neoplasms are used mainly to demonstrate proton density and T2-weighted and other pathological lesions. Aneurysms and angiomas also contrast, T,-weighted contrast being best displayed by gradientenhance with contrast medium, mirroring its concentration in the recalled echo techniques, conventional spin-echo or inversion blood, as do vascular structures such as the choroid plexuses. recovery. Fast spin-echo acquisitions can be used to acquire dual Since grey matter contains considerably more blood than white echoes, generating proton-density and T,-weighted images from matter, the difference between grey and white matter is also each excitation as in conventional spin-echo techniques. However accentuated by contrast, and the falx and tentorium are also enhanced. Thus, some lesions not obvious on the simple scan this takes twice as long as collecting a single echo at a single effective TE, so that the advantage of collecting a single echo is could be revealed and many others enhanced by the contrast that the spatial resolution can be increased for the same acquisimedium. Different workers have varied the amount of contrast medium used. A dose equivalent to 30 g of iodine for an average tion time. In the head GRE techniques are used mainly for voltimetric acquisitions with T,-weighted contrast, usually with some adult and given as a bolus is widely used, though many workers form of tissue magnetisation preparation, of which the simplest is readily use up to 40 g. i nversion recovery (IR) prepared. Aneurysms and other vascular structures are shown best immediThe routine MR examination of the brain performed by most ately following injection when blood concentrations of iodine are units is still a sagittal scout image or multislice T,-weighted GRE highest. Leakage through the blood-brain barrier shows best some acquisition followed by an axial dual-echo fast spin-echo acquisi30 minutes later. In practice both types of lesion are adequately tion. Most units prefer dual to single echo for routine brain imaging shown in sections taken 10-15 minutes following injection. because signal from CSF is lower, enabling lesions close to the It is usual to inspect the non-enhanced scan first and then, if ventricles or in cerebral cortex to be identified more reliably. necessary, to proceed to injection of contrast medium followed Acquisition in the coronal plane is preferred in special situations, i mmediately by the enhanced scan. It is not necessary to use dictated clinically such as in patients with habitual epilepsy or enhancement in every case. Where the simple scan is normal and the dementia, or to help localise lesions shown by other sequences. clinical indications are vague it has been general experience that the post-enhancement scan is most unlikely to add further information. When specialised post-processing is required, such as volumetry, surface rendering or registration, magnetisation-prepared GRE The inert gas xenon has also been shown to be lipid soluble and volumetric imaging is preferred, usually in the coronal plane using to cross the blood-brain barrier after administration as an anaescontiguous slices of I or 1.5 mm thickness. High-resolution techthctic. However, its high cost and the fact that its administration niques (single echo) are used for imaging small structures such as often requires anaesthesia make it impractical for routine use as an the pituitary gland or internal auditory meatuses (Fig. 57.5). enhancing agent. Newer sequences are now becoming generally available, sonic of MRI, despite its greater soft-tissue contrast sensitivity, also them exploiting novel NMR contrast mechanisms. One, which makes widespread use of contrast enhancement. T, and T, relaxadepends on conventional T,-weighted contrast, is fluid-attenuated tion enhancement is provided by the strongly paramagnetic ion


Previous Page

CONTENTS

Next Page I NTRACRANIAL LESIONS (1)


1725

Fig. 57.2 Serial slices from below upward in an elderly patient with some degree of atrophy with diagrams to illustrate normal anatomy. (A) T = temporal lobe; CP = cerebellopontine angle; B = brainstem; 4 = fourth ventricle; C = cerebellum. (B) D = dorsum sellae; P = pontine cistern; Ba = basilar artery; TH = temporal horn. (C) SC = suprasellar cistern; H = hippocampus; B = brainstem; V = vermis. (D) Ped = peduncle; I = insula; 3V = third ventricle; CQ = corpora quadrigemina. (E) FH = frontal horn; SP = septum pellucidum; CN = caudate nucleus; T = thalamus; CP = choroid plexus; OH = occipital horn. (F) IHF = interhemispheric fissure; B = body of lateral ventricle. (G) Su = sulci; F = falx.

Previous Page 1726

CONTENTS

Next Page

A TEXTBOOK OF RADIOLOGY AND IMAGING been limited by long acquisition times. A proven but not yet widely available strategy for improving its efficiency uses sequential inter-

Diagram illustrating position of patient for coronal or nearcoronal sections (A, B), and sagittal or near-sagittal sections (C) at CT. Fig. 57.3

i nversion recovery (FLAIR) and its variations which produce heavily T,-wcighted CSF-nulled images by coupling an inversion pulse followed by a long inversion time to a long echo time. Although FLAIR has had great success in displaying lesions not seen with conventional T,-weighted spin-echo sequences, it has

l eaving and rapid acquisition with relaxation enhancement (RARE). This enables very long repetition times to be used (10 seconds or more). Magnetisation transfer contrast imaging is widely available but relatively little used in clinical practice. It emphasises contrast derived by the reduction in relaxed magnetisation induced by application of a broad-spectrum off-saturation prepulse which saturates a tightly bound water fraction not observed directly with MRL What the water is bound to is uncertain, although in biological tissues it may be mainly macromolecules and membranes. Diffusion weighted contrast i maging can be achieved by using strong diffusion weighting gradients, where contrast depends on the diffusion coefficient of water. Diffusion behaviour is anisotropic, especially in white matter, diffusion being greatest in the direction of the tracts and smallest transverse to them. Anisotrophy can be characterised from three orthogonal directions, or better by an elliptical tensor, and directionality expressed as a trace, or one of many anisotropy indices. Diffusion weighted imaging has been used mainly in the evaluation of stroke, and now has widespread clinical application. Perfusion imaging can be produced by analysing the signal change over time in regions of interest after the administration of a bolus of intravenous contrast medium, or by techniques such as blood oxygenation level dependent contrast imaging (BOLD) or arterial spin tagging. The latter are used in activation studies in


MRI axial sections showing normal anatomy. (A) T,-weighted section shows CSF black and clear differentiation between white and grey matter. (B-G) T2- weighted sections in another patient show CSF white and white matter dark while grey matter remains grey. Compare with and relate to diagrams in Fig. 57.2. Fig. 57.4

Previous Page

CONTENTS

Next Page

Fig. 57.5 MRI corona) sections from before backward. (A,B) The sella and cavernous sinuses with carotid siphons. The chiasm is seen in the suprasellar cistern. (B,C) Pontine cistern and basilar artery (black). (D,E) Cerebellopontine angle cisterns with high-signal CSF in the internal auditory meatus. The medulla and spinal cord are outlined in (F). The hippocampus is also well shown.

functional MRI, which is currently mainly a research rather than a clinical tool. CT cisternography This technique greatly enhances the contrast

between CSF and neural tissue enabling window and level settings to be arranged to minimise the effects of artefacts, especially i n the posterior fossa and around the skull base. It requires the i ntrathecal injection of a water-soluble myelographic contrast medium, usually only about half or less of the amount that would be used for a myelogram. Contrast medium is run into the head by gravity, and imaging using high-resolution techniques can be performed to produce great anatomical detail of the cisternal structures. The technique is most widely used for demonstrating the site of CSF leaks in CSF rhinorrhoea but its other applications, such as demonstrating small tumours in the cerebellopontine angle, have been rendered obsolete by MRI. Air eisternography was also once used with CT but is no longer needed.

smaller with maturity and at term appear as narrow slits. During the same period there is progressive differentiation between white and grey matter, and the sulci and convolutions become defined. Contrast between cortex and medulla increases as myelination progresses. Low attenuation of the white matter is a normal finding in neonates and usually resolves in the first 2 or 3 months. MR/ is also sensitive to the changes of myelination. This manifests as relative shortening of T, and T, relaxation times of white matter, thus increasing the relative contrast between white and grey matter on images as the infant brain matures. Myelination begins in the brainstem and extends into the internal capsules and optic radiations by 6 months of life, forceps major and minor by 1 year, and into the gyral convolutions as in the adult by a year and a half. Ultrasound is widely used for the examination of the neonate and infant brain. The appearances are discussed below (see Ch. 58).


Congenital lesions of the brain are listed in Box 57.1. Most are well demonstrated by CT or MRI, and are discussed in this chapter. The appearance of the normal neonate brain differs from those of older children and adults and is also described here.

The neonatal brain In CT i mages the density of the brain is dependent on the stage of maturation/At full term the cortex shows convolutions, the cerebral sulci are well defined, and the cortex and white matter are differentiated. In premature infants before 30 weeks the brain is homogeneously low in attenuation with the cortex appearing as a thin denser rind without sulci. The sylvian fissures are shallow and wide and the ventricles appear relatively large.- They become

Hydrocephalus Meningocele and encephalomeningocele Macrocephaly Hemimegalencephaly Chiari malformations Aqueduct stenosis Dandy-Walker syndrome Arachnoid cyst Ependymal cyst Septal agenesis Septo-optic dysplasia Cyst of the septum pellucidum Cyst of the cavum Vergae Agenesis of the corpus callosum Lipoma of the corpus callosum Hamartoma Coarctation of the frontal horn Porencephaly Schizencephaly Hydranencephaly Holoprosencephaly Disorders of neuronal migration Phakomatoses

Previous Page 1728

CONTENTS

Next Page


Hydrocephalus Congenital hydrocephalus in infants and young children may be classified, like adult hydrocephalus, as communicating and non-communicating. In the former there is free communication between the ventricles and the basal cisterns, with obstruction to the flow of CSF in the subarachnoid space or basal cisterns. This is due to meningeal irritation by haemorrhage, infection or trauma. All the ventricles are enlarged and the basal cisterns may be prominent. In some congenital cases the cause remains obscure. In non-communicating hydrocephalus flow is obstructed within the ventricular system, usually at third ventricular, aqueductal or fourth ventricular level. The ventricles arc dilated above the level of obstruction and are normal below it. Non-communicating hydrocephalus in infants and children is commonly due to aqueduct stenosis. Other congenital causes are Dandy-Walker syndrome, arachnoid cysts and, very rarely, neoplasms. In infants the diagnosis is confirmed by ultrasound (see Ch. 58), and a specific causative lesion such as Dandy-Walker or arachnoid cyst may be demonstrated or the level of obstruction identified. In older patients CT or MRI will usually identify the level and in most cases the cause. Rarely hydrocephalus is due to oversecretion of CSF, as in some cases of papilloma of the choroid plexus. Low-pressure hydrocephalus This is encountered in the elderly and may be difficult to distinguish from atrophy. The distinction is discussed below (see Ch. 58). At MRI the signal from CSF is modulated by pulsatile motion on moderately T2 - weighted images, being lowest near the choroid plexuses, in the aqueduct, along the floor of the fourth ventricle, and in the anterior cisterns of the posterior fossa where the velocity of CSF motion is greatest. In obstructive hydrocephalus this effect is diminished or absent while in communicating hydrocephalus it is often, but not always, accentuated, especially in the aqueduct. Interstitial oedema secondary to hydrocephalus is clearly shown by MRI on T 2- weighted images, appearing as well-demarcated bands of increased signal extending centrifugally from the subependymal tissues to the central white matter.

Hydrocephalus Subdural effusions Normal (sometimes familial) Migrational abnormalities (see below) Lipidoses Spongy degeneration Alexander's disease Tuberous sclerosis

In older children CT or MRI will be necessary to differentiate and to elucidate other pathological processes. These are discussed under their specific headings. Hemimegaleneephaly In this condition enlargement of the skull is unilateral and confined to the vault. The asymmetrical skull is secondary to, a congenital unilateral enlargement of the brain, and is associated with fits, hemiplegia and hemianopia. Pathologically there are migrational abnormalities with thickening of the grey matter and increase of the white matter and with abnormal shallow sulci on the affected side. The ventricle on this side is also enlarged. The brain lesions are identified by CT or by MRI. Chiari malformations These were originally described by Chiari i n 1891 and 1896, and were graded Types 1-4 according to degree of deformity. Types 3 and 4 have the major deformities with cerebellar hypoplasia and downward displacement of the brainstem and a high cervical or occipital encephalocele.

Type 2 This also presents in neonates or infants, virtually always with a thoracolu mbar myelomeningocele. The associated brain deformities consist of caudal herniation of the medulla and vermis with a caudally displaced and elongated fourth ventricle. A backward kink may be seen at the caudally displaced junction of cervical cord and medulla. Hydrocephalus may he due to associated aqueduct stenosis or to occlusion of the ambient cisterns. There are also various other malformations of the brain. These i nclude hypoplasia or absence of the falx and tentorium with interdigitation of the cerebral hemispheres; absence of the corpus callosum with forward pointing of the frontal horns; deformity of the midbrain (dorsal beaking); enlargement of the cerehellar vermin and forward protrusion of the cerebellar hemispheres as pointed projections around the brainstem; gyral malformations, and other lesions. Many of these brain deformities can be identified by CT or MRI, and some by neonatal ultrasound. Simple X-ray may show a l acunar skull and forward bowing of the petrous bones and clivus.


Meningocele and eneephalomeningoeele These lesions are usually obvious clinically though, as noted in Chapter 53, small lesions of the vault or larger lesions involving the skull base and re sopharynx can be misdiagnosed, with potentially disastrous esultse The skull defects visible at simple X-ray have already been described. The nature and extent of the cerebral deformity, particul arly with basal meningocele, can be clearly defined by CT or MRI before surgical intervention. At MRt the anatomy is best shown on T,-weighted images.

Type 1 This is the least obvious clinically and may not be diagnosed until adult life. It consists of tonsillar herniation through the Maeroeephaly (megalencephaly) Pathologically this is defined as foramen magnum with or without varying degrees of elongation of a brain weighing more than 1800 g. Such brains may be struc- the medulla oblongata and fourth ventricle. turally normal and occur in children from large-headed families, Syringohydromyelia (see Ch. 54) This is associated with the but more frequently there are migrational abnormalities, a diffuse condition in up to 70% of cases. The patients do not present cliniglial proliferation or a pathological process. The latter include cerebral lipidoses, spongy degeneration of the white matter and cally until adult life when the symptoms and signs of syringomyelia Alexander's disease. Large brains may also be seen in tuberous develop; less typically, symptoms suggesting involvement of the l ower cranial or cervical nerves arc seen, and very rarely patients sclerosis (Box 57.2). In infants or children presenting with large heads the primary present with hydrocephalus. Bony anomalies of the craniocervical junction are present in differentiation is from hydrocephalus, which is the commonest cause. Ultrasound in infants will readily confirm hydrocephalus or around 1 5% of cases, particularly assimilation of the atlas. Basilar invagination and dysplasias of the vault are other rare associations. bilateral subdurals (see below).

Previous Page

CONTENTS


1729

I maging The diagnosis is best confirmed by MRI, which also shows the syrinx non-invasively (Fig. 54.4C). Caution is required when interpreting sagittal images of the craniovertebral junction as partial volume effects result in visualisation of 3-4 mm of cerebellar tissue apparently l ying below the foramen magnum in I 5-20% of normal individuals. Coronal imaging is helpful in equivocal cases. Aqueduct stenosis In this condition there is a congenital stenosis

or obstruction of the aqueduct which results in hydrocephalus i nvolving the third and lateral ventricles. While this usually presents in infancy or childhood, the condition is occasionally diagnosed in adults. In all cases it is necessary to exclude obstructive hydrocephalus due to a local tumour or ependymitis. There are also recognised associations with neurofibromatosis and Chiari Type 2 malformation. I maging The diagnosis can be made by CT but is more elegantly shown by sagittal MRI sections through the aqueduct (Fig. 57.6). Hydrocephalus is present, involving the lateral and third ventricles. The latter is often markedly dilated and its suprapineal recess can be large and rounded, simulating an arachnoid cyst. Anteriorly the third ventricle may extend down into the sella, which is often flattened. The mouth of the aqueduct is dilated like a bell or trump et, while the lower aqueduct and fourth ventricle appear normal. There is no evidence of a mass lesion in the brainstem or posterior fossa.

Fig. 57.7 Axial T2 - (A) and sagittal T,- (B) weighted images in a patient with the Dandy-Walker syndrome showing an enlarged posterior fossa with a high tentorium, and a large fluid-filled fourth ventricle-cisterna magna complex in association with vermian hypoplasia.

cisterna magna, or a gliomatous cyst is normally easy, with the distinguishpresence of a normal or small fourth ventricle as a key distinguishmg feature. The dysplastic cerebellum in Dandy-Walker syndrome also helps to distinguish the condition.

Dandy-Walker syndrome I n this condition there is a more or

l ess complete membranous obstruction to the foramina of Magendie and Luschka, which causes cystic dilatation of the fourth ventricle. The lesion may occur on its own or in association with other congenital anomalies such as meningocele or defects in the corpus callosum. The cerebellum is hypoplastic and in severe cases the hemispheres are vestigial and the vermis undetectable. The grossly dilated fourth ventricle extends upward to and above the tentorium and also backward to the occipital bone, which is thinned and bulging, expanding the posterior fossa. There is also marked hydrocephalus with dilatation of all the ventricles.


I maging CT or MRI shows the anatomical features described and permits an immediate diagnosis (Figs 57.7, 57.8). In the neonate or small infant, diagnosis may he made by ultrasound (Fig. 58.86). Differential diagnosis from an arachnoid cyst, a giant

Aqueduct stenosis shown by MRI. Sagittal midline section. (Courtesy of Dr Gordon Thomson.) Fig. 57.6

Fig. 57.8 (A-D), Dandy-Walker syndrome shown by CT.

Previous Page

1730

CONTENTS

Next Page


Fig. 57.9 J oubert's syndrome at the level of the ponto-mesencephalic junction showing the ' molar tooth appearance' with superior cerebellar peduncles clearly seen on the coronal T,weighted image (A) due to absence of vermian folia. In this other patient with Joubert's syndrome the superior cerebellar peduncles are typically very horizontally orientated on the sagittal T,-weighted image (B). (C) Axial T2W, view of patient with J oubett's syndrome showing molar tooth appearances.

Dandy-Walker variant This term is applied to cases where l esions are less marked and the posterior fossa is not enlarged, though the fourth ventricle is dilated and the vermis dysgenetic.

Fig. 57.10 Arachnoid cyst (arrows) involving pineal region and quadrigeminal cistern and compressing back of third ventricle, producing marked hydrocephalus.

Fig. 57.11 Arachnoid cyst in left anterior temporal (A) and Sylvian (B) region. It expands the overlying skull but produces minimal brain displacement.


Joubert's syndrome This condition is characterised clinically by

ataxia, mental retardation, episodic hyperpnoea, and abnormal eye movements, and is due to total aplasia of the cerebellar vermis. This is easily shown by CT or MRI where, in addition to absence of the vermis, the fourth ventricle appears large and triangular with the apex pointing backward at its midportion, and large and bat's wing' at a higher level (Fig. 57.9). Giant cisterna magna In this condition the cisterna magna

appears abnormally large or dilated and the posterior fossa may be enlarged. The fourth ventricle, cerebellum and vermis, however, are seen to be normal.

Arachnoid cyst Arachnoid cysts are intracranial but extra-

cerebral. They may be found anywhere around the brain, but are seen particularly in the middle and posterior cranial fossae, in the suprasellar region, and behind the third ventricle. They contain clear fluid indistinguishable from CSF and are lined by arachnoid tissue. Electron microscopic studies suggest they are due to congenital splitting of the arachnoid membrane. The clear CSF content helps to distinguish them from post-traumatic lesions such as subdural hygromas and from postinflammatory loculations. Arachnoid cysts are usually unilocular, but septated cysts may occur.

They can present in adults but more usually they present in infants or children, particularly if large enough to produce a mass effect or obstructive hydrocephalus. The latter is most frequently seen with posterior fossa or quadrigeminal cistern cysts (Fig. 57.10). There may be hypoplasia of underlying cerebral tissue such as the temporal pole, considered by some to represent a coexistent congenital malformation and by others to be secondary to the mass effect (Fig. 57.11). Symptomless cases are sometimes found coincidentally at CT or MRI examination of adults for other lesions. In neonates or small infants the cysts may be diagnosed by ultrasound, but in older children or adults they are diagnosed by CT or MRI (Fig. 57.12). They usually lie in one of the classic sites mentioned above, are of CSF density, and on CT have no enhancing capsule or adjacent calcification. With suprasellar arachnoid cysts, differential diagnosis is from a Rathke's cleft cyst or craniopharyngioma, and occasionally from a grossly dilated third ventricle. In the quadrigeminal cistern a dilated suprapineal recess or even a cystic pineal tumour may need consideration. Ependymal cyst These very rare cysts are lined by a thin wall of

ependyma and can be intra- or paraventricular. The latter can resemble a hydatid cyst at CT and show no marginal enhancement.

Previous Page

CONTENTS


1731

Fig. 57.12 MRI study of left anterior temporal arachnoid cyst, proton density. (Courtesy of Dr Gordon Thomson and Bristol MRI Centre.)

Large intravcntricular ependymal cysts can resemble unilateral hydrocephalus since they usually occur in the lateral ventricle and the wall is difficult to define at CT. Intrathecal or intraventricular contrast agent has been used to establish a diagnosis, by outlining the cyst within the opacified ventricle. Cyst of the septum pellucidum The double septum or fifth ven-

tricle is due to the abnormal persistence of the fetal cavum septi pellucidi. Ultrasound will show that this is present in over onethird of neonates. The incidence falls rapidly through childhood, but nevertheless the condition persists in 1-2% of adults, when it

Fig. 57.14 Cyst of the septum pellucidum (arrow in A) continuous (B) with cyst of the cavum Vergae (arrow). (C) Velum interpositum (arrows) containing enhanced internal cerebral veins.

midline. A similar situation can arise in severe hydrocephalus to]will be seen as a chance finding at imaging by CT or MRI lowing rupture of the septum. Both conditions are readily identi(Figs 57.13A, 57.14A). fied at CT or MRI. Cyst of the cavum Vergae This represents the so-called `sixth ventricle' and is a backward extension of the septal cyst, though much less commonly seen. Anatomically it lies beneath the posterior part of the corpus callosum with the velum interpositum below (Fig. 57.13B). On axial CT sections the eavum Vergae appears as a rectangular structure continuous with a septal cyst anteriorly, both being of CSF density (Fig. 57.14B). It should not be confused with the normal velum interpositum (Fig. 57.13), which is also of CSF density and lies above the third ventricle. The latter is frequently seen at axial CT, but is triangular in shape with the apex lying anteriorly, and contains the enhancing internal cerebral veins (Fig. 57.14C). Fig. 57.15 shows a cyst of the velum i nterpositum.

Septo-optic dysplasia ( De Morsier's disease) This is charac-

tensed by absence of the septum pellucidum and hypoplasia of the anterior optic pathways. There is squaring of the frontal horns with a large chiasmatic cistern, and the child has retarded development and poor vision. The anatomical features may be demonstrated by CT or by MRI.


Agencies of the corpus callosum This is usually partial but

sometimes complete. It may occur as an isolated finding, or in association with other malformations.

Septal agenesis Septal agenesis can occur as an isolated lesion

giving rise to a single ventricle communicating across the

Fig. 57.13 (A) Cavum septi pellucidi: C = corpus callosum; F = fornix; CS = cavum septi pellucidi. (B) Cavum septi pellucidi and cavum Vergae: VI = velum interpositum; CV = cavum Vergae; 3 = third ventricle.

Fig. 57.15 Axial T 2 - (A) and sagittal T,- (B) weighted images showing a cyst of the velum interpositum which extends into the trigone of the left lateral ventricle.

Previous Page 173 2

CONTENTS

Next Page


Fig. 57.16 (A) Complete agenesis of the corpus callosum shown by midline sagittal MRI section. (B) Partial agenesis of the corpus callosum. (C) CT shows lipoma of the corpus callosum with marginal calcification.

At CT complete agenesis is characterised by wide separation of

Partial agenesis usually involves the posterior part, and the anterior part may remain normal. MRI shows similar features on axial or

the medial borders of the lateral ventricles, with a high third ven-

coronal sections, but a midline sagittal section (Fig. 57.16) shows

tricle extending up between them and also enlargement of the occipital horns.


Fig. 57.17 Sagittal (A) and coronal (B) T,-weighted images showing an extensive lipoma of a dysplastic corpus callosum. Low signal within the lipoma on the T 2 axial image (C) is likely to represent calcification. There is enlargement of the posterior horns of the lateral ventricles (colpocephaly).

Fig. 57.18 (A,B) Midline bifrontal lobulated lipoma with marginal calcification. Shown by CT. (C) Hamartoma of the left temporal lobe. Coronal MRI study shows a small rounded lesion mainly of low signal similar to cortex.

Previous Page

CONTENTS

Next Page

the lesion most clearly. The third ventricle extends upward into and above the position of the absent corpus callosum. Radial configuration of the medial cerebral sulci extending through the absent call osal body region is also seen in the sagittal images. The axons which fail to cross form symmetrical longitudinal bundles along the medial surfaces of the hemispheres and impress the medial parts of the ventricles. Thus the frontal horns and bodies are laterally displaced and small. The atria and occipital horns are large and rounded due to lack of the normal forceps major impressions. A lipoma may sometimes he found at the site of the absent corpus callosum (see below). I ntracranial lipomas are rare t umours which occur mainly in relation to the corpus callosum, but can also occur elsewhere including the suprasellar and pineal areas. They may lie above a normal corpus callosum but are associated with partial or complete agenesis in 40% of cases. There is often marginal calcification where they merge with adjacent cortical tissue this can he of a characteristic brackets' type which occasionally permits diagnosis from a simple PA X-ray film ( Fig. 53.34). A large lesion may also produce an area of increased radiolucency between the brackets. The pericallosal arteries are usually incorporated in the tumour and may be fused (azygos or single anterior cerebral artery). Their typical situation together with their fatty density makes the lipomas readily identifiable at CT or MRI. Adjacent calcification is best seen at CT (Figs 57.17, 57. I8A,B), but this may he minimal or absent, particularly with small lesions. The single anterior cerebral artery may he visible at MRI on T-weighted images. Sometimes streaks of fat may be evident in an otherwise normal-looking corpus callosum. Lipoma of the corpus callosum

Hamartomas are congenital benign tumours containing an overgrowth of mature cells and tissues normally found i n the affected part. Some pathologists include glial lesions of the Hamartomas

Fig. 57.19 Porencephaly (A) T 2 - weighted axial study. (B) T,-weighted coronal study. Large cavity in left hemisphere.

brain in addition to dermoids, epidermoids and nasal gliomas. These are discussed in the section on neoplasnms. Hamartoma of the tuber cinereum is a specific condition associated with precocious puberty. At CT or MRI it appears as a rounded suprasellar mass of brain density and occupying the interpeduncular cistern. Hamartoma and allied lesions also occur in the temporal lobe (Fig. 57.18C) where it may present with epilepsy and contain calcification visible at simple X-ray or CT. I n this rare condition the frontal horns fail to develop normally and the lateral and medial walls appear adherent or apposed due to ependymal fusion. The rest of the ventricular system is normal. Its importance lies in recognising its congenital nature when seen as a chance finding at imaging, and Ventricular coarctation

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ Fig. 57.20 (A) CT shows minor degree of schizencephaly presenting as lateral slit lined by cortex. Axial T 2 - weighted image (B) of a patient with bilateral schizencephaly. The clefts radiate out from the lateral ventricles to the surface of the hemispheres and are lined by grey matter. The cleft on the right is closed-lipped whereas the one on the left is open-lipped and this is evident on the more superior axial T,-weighted image (C).

Previous Page 1734

CONTENTS

Next Page


i n not mistaking the deformity for the effects of compression by a l ocal tumour. Porencephaly Porencephaly has been classified as congenital or acquired. The congenital form is due to localised agenesis of the cortical mantle resulting in the formation of a cavity or a lateral slit through which the lateral ventricle communicates with the convexity of the brain. The cavity is lined by ependyma and laterally by a thin pia-ependynial layer which may rupture into the subarachnoid space. In less severe forms the cavity may be reduced to a lateral slit lined by ependyma and partially fused. The acquired type is secondary to any type of cerebral destructive process, ranging from trauma to infarction. Sometimes called false porencephaly, such cases are better labelled by their aetiological cause if this is known, e.g. post-traumatic or postinfaretion cerebrat cavities. Porcncephalic cavities or clefts can be identified or suspected by ultrasound in the neonate or infant. In children or adults they are demonstrated by CT or MRI (Fig. 57.19). Schizencephaly This term is used for a form of porencephaly in which clefts, commonly bilateral, extend from the ventricles to the convexity in the opcrculoinsular regions. Injury to the germinal matrix in the early stage of gestation may result in a loss of the full thickness of cerebral substance, with a cleft extending from the ventricle to convexity subarachnoid space (Fig. 57.20). The lesion is usually symmetrical and may be associated with a single ventricle and other malformations, especially heterotopias with -topic grey matter lining the cleft margins. CT or MRI in the axial plane will demonstrate the lesions well. Holoprosencephaly This comprises a complex craniocerebral and facial anomaly classified in grades of severity from alobar through semilobar to lobar, which is the least severe. It results from a failure of normal development of the forebrain (prosencephalon). The severe abhor form is incompatible with survival beyond infancy. The child has multiple craniofacial anomalies (cleft lip and palate, hypotelorism, anophthalmia or cyclopia). Intracranially the midline structures (falx, corpus callosum, septum pellucidum and olfactory bulbs) are absent and the two hemispheres are replaced by a single large ventricle with a thin ri m of cortical mantle. The thalami are fused, unlike hydranencephaly, where they remain normal. Semilobar holoprosencephaly, the intermediate form, shows less severe facial deformity and the brain has a thicker mantle with recognisable occipital horns (Fig. 57.21). In the mildest (lobar) type there are well-formed lateral ventricles and a recognisable third ventricle, though the septum and Sylvian fissures are usually absent. The grosser forms may be suspected from the clinical appearance of the infant, but imaging will be required in all varieties to delineate the intracranial anomalies.

Fig. 57.21

(A,B) Semilobar holoprosencephaly.

Fig. 57.22

(A -C) Hydranencephaly

The diagnosis in the neonate can he suggested by ultrasound, and the anatomical feature can be delineated by CT or MRI (Fig. 57.22).


Hydranencephaly This condition is generally the result of a major destructive process, often in the prenatal or perinatal period, and results in massive intracerebral cavitation which can resemble gross ventricular dilatation. It tends to affect the anterior parts of the hemisphere, suggesting major involvement of the areas supplied by the internal carotids. The falx is present (a distinguishing feature from severe holoprosencephaly), and the posterior fossa and basal ganglia appear normal. The ventricular dilatation resembles gross hydrocephalus, but the infant's head is not enlarged.

Cortical neurones are formed in the embryonal germinal matrix which lies adjacent to the lateral ventricles. They then migrate centrifugally out to the surface of the brain where they form the cerebral cortex, the deeper layers being constituted first. Interruption of this process, which takes place at the end of the second month of fetal life, may give rise to various anomalies including lissencephaly, pachygyria, polymicrogyria and heterotopic grey matter. Lissencephaly (agyria) This, the most severe form of migrational defect, is characterised by the absence of sulci and convolutions in the cortex, and is the normal appearance before the seventh month of fetal life. The Greek word can be translated as `smooth brain'. It may be spontaneous or inherited as an autosomal recessive. Its persistence to term may be total or involve only part of the hemispheres, and some degree of pachygyria may be associated. Clinically there is severe mental retardation, fits and decerebrate rigidity. There is often a characteristic nodule of calcification in the septum pellucidum near the foramen of Monro, which can be seen at simple X-ray or CT examination. The latter, like MRI (Fig. 57.23), will identify the agyria and pachygyria. Pachygyria In this condition the convolutions and cortex are abnormally wide and thick (pachygyria = thick gyri). The condition results from disturbance of migration at a later stage than

Previous Page

CONTENTS


173 5

ital anomalies, but isolated nodular lesions have been mistaken for neoplastic areas at CT and they occasionally contain calcified foci. At MRI these lesions may give lower signal than does normal grey matter (Fig. 57.25).

Phakoma (from the Greek phakos = l entil) was the term used for a l entil-shaped object such as a spot on the body or in the retina. The term 'phakomatoses' is used to refer to a group of diseases of different aetiologies but having in common lesions of the skin, retina and nervous system and the fact that they are developmental. They are also referred to as `neuroectodermal dysplasias', and include tuberous sclerosis, neurofibromatosis, Sturge-Weber syndrome and von Hippel-Lindau disease.

Fig. 57.23 sulci.

Cerebral dysplasia with gross pachygyria and poorly formed

agyria and can occur together with it and other anomalies such as heterotopic grey matter (Figs 57.23, 57.24). Polymicrogyria This condition is the result of a disturbance at the latest stage of neuronal migration, causing multiple small gyri to be formed. It may be focal or diffuse and accompanied by epilepsy, focal neurological deficits and mental retardation. Heterotopias These may be observed as an isolated phenomenon due to arrest of migrating neurones in the form of small masses of grey matter. They may form subependymal nodules protruding i nto the lateral ventricles similar to those seen in tuberous sclerosis, or lie in the centrum ovale or subcortically, near the caudate nuclei, in the cerebellar white matter and in the brainstem. These may be nodular or band like. There are usually associated congen-

Tuberous sclerosis (synonyms: epiloia, Bourneville's disease) The l esions of this condition can affect the skin and nervous system but many other systems, including the bones (see Ch. 37), respiratory system (Ch. 4), kidneys (Ch. 32), heart and skeletal muscle. In the brain there are multiple areas of heterotopia containing abnormal and giant glial cells and lying in the subcortical and paraventricular regions as well as the white matter and cerebellum. These often contain calcification which can be recognised on a simple X-ray film. Typically the calcifications are widely scattered small rounded discrete lesions, though they can be irregular, sinuous or densely nodular. Their multiple scattered nature should suggest the diagnosis (Fig. 53.45). Rarely a paraventricular nodule near the foramen of Monro becomes large enough to produce obstructive hydrocephalus by developing into a low-grade giant cell astrocytoma (Fig. 57.26A). I maging At CT the characteristic and diagnostic feature is the presence of small nodular protrusions into the ventricles, usually with calcified foci. The latter are of course much easier to see at CT (Fig. 57.26B,C) than at simple X-ray. MRI shows the areas of dysplasia better than CT but has the disadvantage of not imaging the calcification. On T,-weighted images


Fig. 57.24 3D surface rendering MR images in two different patients showing a malformation of cortical development in fronto-parietal region (A) and also in both parasagittal regions (B).

Previous Page 1736

CONTENTS

Next Page



Fig. 57.25 (A) Multiple anomalies including agenesis of corpus callosum, subependymal heterotopia, subcortical heterotopia and polymicrogyria. (B) Diffuse subependymal heterotopia. (C) Unilateral nodular and subependymal heterotopia (arrow). (D) Band-like heterotopic grey matter (arrows) in a patient with epilepsy. Similar but less obvious lesions on right side. Positron density weighted MRI.

Previous Page

CONTENTS


Fig. 57.27

MRI studies of a patient with tuberous sclerosis shows cortical tubers as: (A) low signal (T,-weighted); (B) high signal (T i - weighted). Calcification is not specifically identified but appears as low signal in the i ntraventricular tubers.

Fig. 57.28

1737

In the CNS NF1 is associated with glial tumours, including optic nerve gliomas (Fig. 57.28B) and cerebral and brainstem astrocytoma, and also with anomalies of migration resulting in various types of dysplasia and heterotopia in the white matter. Hamartomas with no mass effect are frequently present, leading to areas of high signal on T, MR studies. Generalised or focal cerebral atrophy may also be seen, associated with mental retardation. Vascular dysplasias involving the terminal internal carotid or anterior or middle cerebral origins have also been described in NF I, as has the resulting moyamoya disease. Neurofibromatosis 2 (NF2 or central neurofibromatosis), is much l ess commonly seen, affecting only I in 50 000 of the population. It is also inherited as an autosomal dominant but is due to a defect of chromosome 22. The CNS lesions are quite different from those of


(A) Neurofibromatosis showing absence of sphenoid wing on the right. Reversal negative at wider window for bone detail. (B) Axial MRI (T,-weighted) shows right optic nerve glioma as low-signal rounded mass i n a patient with neurofibromatosis.

tubers appear as irregular areas of increased signal cortically, in the white matter, or paraventricular, though calcification will appear as l ow-signal areas (Fig. 57.27). T,-weighted images usually show the lesions as giving low signal and occasionally they resemble cysts.

Neurofibromatosis This manifests in several different ways, the commonest being peripheral neurofibromatosis (von Recklinghausen's disease) now classified as NFI and affecting 1 in 4000 of the population. It is an autosomal dominant familial disorder associated with a defect in the long arm of chromosome 17, and the CNS is affected in some 15% of cases. The skull and spine are also frequently involved. The skull lesions have been described in Chapter 53; absence of the greater wing of the sphenoid is one of the causes of the `bare orbit' causing unilateral exophthalmos (Fig. 57.28A). Spinal scoliosis or kyphoscoliosis is present in onethird of the cases, and dural ectasia and lateral thoracic meningocele are other spinal manifestations. Paraspinal neurofibromas may also occur and can occasionally become malignant (10%).

Fig. 57.29 (A) Sturge-Weber syndrome. Predominantly occipital cortical calcification with some atrophy. Axial T 2- weighted image (B) i n a patient with Sturge-Weber syndrome shows atrophy of the left cerebral hemisphere. Cortical T 2 l ow signal in the left occipital region is consistent with calcification. There is also some thickening of the calvarium in the left frontal region. On the postcontrast TI axial image (C) there is marked superficial enhancement in the occipital and temporal regions due to the presence of a pial angioma. There is atrophy of the right cerebellar hemisphere (D)-crossed cerebellar diaschisis.

Previous Page 1738

CONTENTS

Next Page


NF1 and consist of bilateral acoustic neuromas which may be associated with schwannomas of other cranial nerves and with meningiomas. In the spine bony abnormalities are absent but large bilateral neurofibromas may be seen, sometimes at every level (see Fig. 2.51). Spinal meningiomas also occur as does ependymoma of the cord or filum. The intracranial lesions can be characterised by CT or MRI and are discussed below under cerebral tumours. Other features of neurofibromatosis are described elsewhere and i nclude skeletal deformities (Ch. 40) and a rare association with renal artery stenosis or aneurysm. There is also a very rare association with phacochromocytoma. Sturge-Weber disease (encephalotrigeminal angiomatosis)

This condition involves a flat and often extensive unilateral angioma of the `port wine' stain type in the facial and scalp distribution area of the trigeminal nerve. It is accompanied by a leptomeningeal angiomatosis in the parieto-occipital area on the same side, which is associated with underlying cortical gliosis and calcification. Clinically there is epilepsy and a deficit corresponding to the cortical lesion. The cortical calcification gives rise to the characteristic tramline calcification seen on simple radiographs (Fig. 3.46). This represents the margins of sulci where the calcification is seen end-on. The occipital cortical calcification is also well shown by CT, but the classic tramline appearance is obscured at normal window since CT shows all the extensive calcification and not just that seen end-on at the sulcal margins (Fig. 57.29A). The gyriform nature of the calcification is more apparent when viewed at wide window. Occasionally CT will demonstrate bilateral changes, an important finding if hemispherectomy for intractable epilepsy is being considered. MRI may show the calcification in the cortex but mainly by susceptibility effects, especially on gradient-recalled echo sequences. It also shows cortical atrophy, and the extent of the angiomatosis may he delineated by intravenous gadolinium (Fig. 57.29B-D). The cortex often seems to enhance as well and the choroid plexus on the side of the lesion is conspicuously larger.

Von Hippel-Lindau disease (retinocerebellar angiomatosis) This rare familial disease is characterised the occurrence of multiple haemangioblastomas in the retina and cerebellum. They may also be found in the spinal cord and rarely in the cerebrum. There is also an association with visceral tumours and cysts, particularly renal and pancreatic tumours. Patients usually present in their late teens or early twenties.

I maging The intracranial lesions can be demonstrated by CT or MRI but angiography is more specific (Fig. 57.30). It is important to realise that haemangioblastomas occur much more commonly as isolated lesions, single or multiple in the cerebellum, and without the retinal lesions of von Hippel-Lindau disease. They can also occur as isolated lesions in the spine. Their diagnosis is discussed i n the sections on cerebral tumours and spinal tumours (see also Fig. 57.27). Ataxia telangieetasia This condition presents in infants, who

show cerebellar ataxia when starting to walk followed by rapid deterioration with choreoathetosis. The mucocutaneous telangiectasias appear at 3-6 years and there is a high incidence (10-15%) of malignancy (lymphoma, and leukaemia in children and epithelial in adults). The disease is an autosomal recessive, with an incidence of 1 in 40 000 live births. Imaging by CT or MRI shows isolated gross atrophy of the cerebellum with dilated sulci and a dilated fourth ventricle. Haemorrhage as a result of cerebral telangiectasias has also been described. Such an appearance in a young child should always raise the diagnosis of ataxia-telangiectasia. L'hermitte-Duclos disease This is a rare slow-growing cerebellar

malformation which presents as a mass lesion in young adults (20-40 years). It is usually associated with Cowden's syndrome (benign skin tumours, GI polyps, goitre and breast cancer) but can occur alone. CT shows a low-density cerebellar area with thickened folia and no enhancement after contrast. Calcification may be present. MRI shows low signal on T and high signal on T, weighting, and some enhancement after L~,idolinium.


Fig. 57.30 A vascular mass is shown in the left cerebellar hemisphere in a patient with Von Hippel-Lindau syndrome. The mass shows marked enhancement on the coronal postcontrast image (A) and has a dense vascular blush on the lateral vertebral angiogram (B) with evidence of early venous shunting. Two small enhancing superficial spinal haemangioblastomas are also present in the cervical and upper thoracic region on the T, postcontrast sagittal image (C).

CONTENTS

Previous Page


1739

These affect mainly the middle-aged and elderly, with the exception of secondary neuroblastoma which occurs mainly in children. Secondary tumours

Cerebral neoplasms can be classified in several ways. Topographically they may be supratentorial, posterior fossa or intermediate (straddling the tentorial notch or foramen magnum). They can also be intra-axial or extra-axial. In this section they will be discussed mainly on a histopathological basis as listed in Table 57.3.

Frequency

In the past many neurosurgical series have been analysed to show the relative frequency of the different cerebral tumours. However, surgical series arc selected, do not reflect the natural incidence and grossly underrate the incidence of metastases. This is because most patients with cerebral metastases are not operated on by neurosurgeons but are treated by oncologists and radiotherapists. A truer incidence is provided in the pathological series of 5199 cases analysed by Zimmerman (1971), which lists the incidence as in Table 57.1. Age incidence Cerebral tumours are predominantly tumours of adult life with a peak incidence of 13 cases per 100 000 population at age 55-65. They are relatively uncommon in infants and children at 2 cases per 100 000. Primary tumours

While examples of most primary intracranial t umours can occur at any age, it is helpful in differential diagnosis to know that certain tumours occur mainly in certain age groups. Thus the four main tumours of childhood occur in the posterior fossa and are cerebellar glioma, pontine glioma, medulloblastoma and choroid plexus papilloma. Table 57.2 lists tumours by their usual age group.

I n adults, supratentorial tumours outnumber posterior fossa tumours by a ratio of 7 to 3, but in children this ratio is reversed, and posterior fossa tumours are the most common. Location

These, the commonest intracranial tumours, vary greatly in malignancy. This has led to specific names being used based on the histol ogical appearances, e.g. astrocytoma and oligodendroglioma for well-differentiated slow-growing tumours, and glioblastoma and spongioblastoma multiforme for highly malignant ones. Most workers however favour grading by degree of malignancy judged by the proportion of mitotic figures seen at microscopy. The grades range from I, which is relatively benign, through grade II to grades III and IV which are highly malignant (Kernohan). Some 50% of gliomas arc of grades Ill and IV, and about 25% of grade I. Gliomas can occur at any site in both children and adults, but gliomas of the optic chiasm and of the brainstem are particularly common in children, where they have a peak incidence at 1-6 years and 1-4 years respectively. Cerebellar astrocytoma is also commoner in children, with a peak incidence at 5-10 years. I n the last decade the World Health Organisation (WHO) has introduced a more logical system based on histopathology (Table 57.3), which is now generally accepted.

This is the largest group of primary brain neoplasms; they vary greatly in malignancy and can exhibit either a circumscribed or diffuse growth pattern-50% of these tumours are supratentorial, 35% in the cerebellum and 15% in the brainstem. It is the most common primary brain tumour of childhood, accounting for between 30 and 40% of all CNS tumours; diffuse astrocytomas are more common in adults whereas in children the majority of astrocytomas arc of the pilocytic type, only the cerebral hemispheres and brainstem being more commonly affected by the diffuse types. Tumours with a circumscribed growth pattern will initially be described followed by the more diffuse, infiltrative type. Modern high-resolution CT will localise and characterise the vast majority of cerebral astrocytomas, and the same is true of MRI. Skull radiographs are of limited value showing calcification in only a small proportion of astrocytomas. General signs of raised intracranial pressure may be present, as described in Chapter 53, as may lateral displacement of the calcified pineal.


Gliomas Metastases Meningiomas Angiomas Pituitary adenomas Acoustic tumours Congenital tumours Granulomas Miscellaneous

31.4

20.3

15.4 5.9

4.4 1.5

2.0 0.4

12.3

Brainstem glioma, optic nerve glioma Medulloblastoma, cerebellar astrocytoma, papilloma choroid plexus, pinealoma, craniopharyngioma Ependymoma Glioma, meningioma, acoustic neuroma, pituitary tumour, hemangioblastoma Meningioma, acoustic tumour, glioblastoma

0-5

5-15

1 5-30 30-65

65+

Pilocytic astrocytomas These low-grade (WHO grade 1) tumours account for 2-6%0 of all primary intracranial neoplasms and usually present in childhood between the ages of 5 and 15 years. They characteristically occur in the cerebellum, hypothalamus or optic nerves and involvement of the latter is a feature of neurofibromatosis type 1. Most series report an equal s.'x incidence though some series do report a slight female preponderance. Cerebellar tumours occur equally within the vermis and cerebellar hemispheres (Figs 57.31-57.33). The tumours are well circumscribed and encapsulated growing mainly by expansion. The cerebellar tumours are usually large and can be cystic, solid or both. On CT 40% of the tumours are solid, 10-20% have calcification

Previous Page 1740

CONTENTS

Next Page


1.

A strocytic tumours

11.

Tumours of the sellar region

1.1

Astrocytoma-WHO grade II

11.1

Pituitary adenoma

1.2

Anaplastic (malignant) astrocytoma-WHO grade III

11.2

Pituitary carcinoma

1.3

Glioblastoma-WHO grade IV

11.3

Craniopharyngioma

1.4

Pilocytic astrocytoma-WHO grade I

12.

Haematopoietic tumours

1.5

Subependymal giant cell astrocytoma-WHO grade I

12.1

Primary malignant lymphoma

1.6

Pleomorphic xanthoastrocytoma

13.

Germ cell tumours

2.

Oligodendroglial tumours

13.1

Germinoma

2.1

Oligodendroglioma-WHO grade II

13.2

Embryonal carcinoma

2.2

Anaplastic oligodendroglioma-WHO grade III

13.3

Yolk sac tumour

3.

Ependymal tumours

13.4

Choriocarcinoma

3.1

Ependymoma-WHO grade I

1 3.5

Teratoma

3.2

Anaplastic (malignant) ependymoma-WHO grade III

1 3.6

Mixed germ cell tumours

3.3

Myxopapillary ependymoma

1 4.

Tumours of the meninges

3.4

Subependymoma-WHO grade I

1 4.1

Meningioma

Mixed gliomas

1 5.

Non-meningothelial tumours of the meninges

4.1

Mixed olig-astrocytoma-WHO grade I

1 5.1

Benign mesenchymal

4.2

Anaplastic (malignant) oligo-astrocytoma-WHO grade III

1 5.2

Malignant mesenchymal

4.3

Others (e.g. ependymal astrocytomas)

1 5.2.1 Chondrosarcoma

5.

Neuroepithelial tumours of uncertain origin

1 5.2.2 Haemangiopericytoma

5.1

Polar spongioblastoma-WHO grade IV

1 5.3

5.2

Astroblastoma-WHO grade IV

1 5.4

Haematopoietic neoplasms

5.3

Gliomatosis cerebri-WHO grade IV

1 5.5

Tumours of uncertain histogenesis

6.

Choroid plexus tumours

1 5.5.1 Haemangioblastoma

6.1

Choroid plexus papilloma

1 6.

Tumours of cranial and spinal nerves

6.2

Choroid plexus carcinoma

1 6.1

Schwannoma

7.

Neuronal and mixed neuronal-glial tumours

1 6.2

Neurofibroma

7.1

Gangliocytoma

1 6.3

Malignant Schwannoma

7.2

Dysplastic cerebellar gangliocytoma (L'hermitte-Duclos)

1 7.

Local extension from regional tumours

Ganglioglioma

1 7.1

Paraganglioma

Anaplastic (malignant) ganglioglioma

1 7.2

Chordoma

4.

7.3 7.4 7.5 7.6 7.7 7.8

8. 8.1 8.2 8.3

9. 9.1 9.2 9.2.1 9.3 9.4 9.5

Primary melanocytic lesions

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ Desmoplastic infantile ganglioglioma

1 7.3

Chondroma

Central neurocytoma

1 7.4

Chondrosarcoma

Dysembryoplastic neuroepithelial tumour

1 8.

Metastatic tumours

Olfactory neuroblastoma (esthesioneuroblastoma)

1 9.

Unclassified tumours

Pineal tumours

20.

Cysts and tumour-like lesions

Pineocytoma

20.1

Rathke cleft cyst

Pinealoblastoma

20.2

Epidermoid

Mixed pineocytoma/pinealoblastoma

20.3

Dermoid

Embryonal tumours

20.4

Colloid cyst of the third ventricle

Medulloepithelioma

20.5

Enterogenous cyst

Primitive neuroectodermal tumours

20.6

Neuroglial cyst

Medulloblastoma

20.7

Granular cell tumour (choristoma)

Neuroblastoma

20.8

Hypothalamic neuronal hamartoma

Retinoblastoma

Ependymoblastoma

and up to 70% are cystic. Forty-seven per cent of the cystic slightly hyperdense on CT. Enhancement of the cyst wall indicates tumours have an associated mural tumour nodule which is iso- or it is tumour lined. On MRI the tumours are hypointense to isoinhyperdense and enhances; the cyst wall is composed of surroundtense on T,-weighted images and hyperintense on T,-weighted i ng compressed or gliotic parenchyma and thus may appear images and enhancement patterns are similar to those seen on CT.

Previous Page

CONTENTS


Fig. 57.31 Axial T2 -weighted image (A) in a patient with a juvenile pilocytic astrocytoma of the right cerebellar hemisphere showing a large cystic component, a little peritumoural vasogenic oedema and hydrocephalus due to obstruction of the fourth ventricle. The T,-weighted coronal postcontrast image (B) shows a superior enhancing mural nodule and minor peripheral enhancement of the cyst wall.

1741

Fig. 57.33 MRI studies of cystic juvenile astrocytoma. (A) T,-weighted: tumour and cyst are of different low signals. (B) T 2 -weighted: tumour and cyst given uniform high signal.

Fig. 57.32 Cerebellar astrocytomas. (A) Large nodular calcification in a right hemisphere tumour. (B) Large cystic tumour with only fair marginal enhancement. (C) Near midline tumour with irregular marginal enhancement.


Fig. 57.34 Axial T,-weighted MRI (A) showing enlargement of the cranial portions of both optic nerves and the chiasm in keeping with an optic glioma. On the postcontrast T,-weighted coronal images (B,C) there is some enhancement of the tumour.

Enhancement of the solid tumours is variable ranging from complete to absent enhancement. Optic pathway gliomas can arise anywhere along the optic pathway, from orbit to the visual cortex ( Figs 57.34-57.36). Between 50 and 85% involve the chiasm or

hypothalamus, and 60% of these will be pilocytic astrocytomas histologically. On imaging the optic nerve is typically enlarged and buckled and enhances: although cystic changes may he present calcification is not a feature in the absence of irradiation, which

Previous Page 1742

CONTENTS

Next Page


Fig. 57.35 These sagittal T,-postcontrast images show a large enhancing sellar/suprasellar chiasmal and hypothalamic mass (A,B), which extends posteriorly along the optic tract, consistent with an optic glioma.

Fig. 57.36 (A) Coronal MRI section through suprasellar cistern and sella shows chiasm swollen mainly on left side (arrow). The infundibulum is also seen. (B) Postgadolinium the tumour enhances, mainly on the left. (C) Left parasagittal section through the swollen chiasm (arrow).


Fig. 57.37 Pontine glioma. (A) Axial T 2 i mage through the posterior fossa shows an expanded pons with increased signal, with flattening of the fourth ventricle and pontine cistern. (B) This case shows similar changes to A, with slight marginal enhancement after contrast medium. (C) Sagittal T, view shows an expanded pons and medulla compressing the pontine cistern and displacing aqueduct and fourth ventricle backwards.

may be useful in differentiating this lesion from an optic nerve sheath meningioma. The tumours involving the chiasm or hypothalamus are often large, lobulated and well circumscribed at prescaitation and can be partially cystic. Although haemorrhage and necrosis is uncommon, they can invade local structures and be associated with perilesional oedema. On CT they are hypodense and on MRI hypointense on T,-weighted sequences and hyperin-

tense on T2-weighted sequences with enhancement of the solid components (Fig. 57.37). These are also WHO grade I tumours and in 90% of cases are associated with tuberous sclerosis (TS) (Fig. 57.26); conversely 15% of patients with TS will develop this well-demarcated tumour usually in their teens Subependymal giant cell astrocytomas

Previous Page

CONTENTS

or early twenties, which is located in the vicinity of the foramen of Monro, attached to the caudate head nucleus and protrudes into the ventricle. As this tumour is covered by an intact layer of ependyma, seeding via the CSF is unlikely, but calcification is a frequent imaging feature. On CT the tumours are hyopodense to isodense and on MRI they are usually hypointense to isointense on T,- and hyperintense on T,-weighted images, with strong enhancement. Hydrocephalus is a commonly associated feature. Pleomorphic xanthoastroeytoma This tumour category is a new i nclusion in the WHO 1993 classification with behaviour intermediate between that of a low grade II circumscribed lesion and a more infiltrative one (Fig. 57.38); it presents in children and young adults usually with seizures and has a predilection for the temporal lobe, and is rarely seen in the cerebellum or spinal cord; there is no sex predilection. It arises from subpial astrocytes which may explain its tendency to arise or involve the surface of the brain. On imaging the typical appearance is one of a superficial cortical temporal lobe mass which is partially cystic but associated with a tumoral mural nodule which is often adherent to t he surface of the brain and thus there may be associated reactive meningeal thickening. On CT the solid tumour components are heterogeneous and can be either hypodense or hyperdense; calcification is variable but as the tumour is relatively slow growing and superficial there may be scalloping of the overlying calvarial inner table. On MRI the solid tumour components are hypoi ntense or isointense on T,- and hyperintense on T,-weighted irnages. The tumoral nodule enhances, but enhancement of the cyst wall is variable.

Next Page

most frequently arises in the cerebral hemispheres though the cerebellum and brainstem may also be affected. It tends to infiltrate along white matter tracts, but the blood-brain barrier remains intact. Necrosis and haemorrhage are usually absent. On imaging there is asymmetrical enlargement within the affected brain; lesions are poorly demarcated and of low density on CT, but are well circumscribed on MRI and hypointense on T,- and hyperintense on T2weighted sequences (Fig. 57.37). Contrast enhancement is notably absent in WHO grade II tumours, but they can transform into higher grade lesions. Anaplastie astrocytoma This WHO grade III tumour is less frequent than either the aggressive glioblastora multiforme or the l ower grade 11 astrocytoma, and has imaging appearances intermediate between the two. It occurs in a slightly older population than the lower grade astrocytomas. The imaging appearances can be similar to those of WHO grade II tumours, but are often more heterogeneous on T,-weighted images with more frequent enhancement. Vasogenic oedema and mass effect may be more pronounced but necrosis or heterogeneous ring enhancement should not be a feature of this tumour grade.

Diffuse astrocytoma These are poorly marginated and infiltrate the surrounding brain; the histological grade varies between WHO grade 11 and grade IV, the latter being the most aggressive form, the glioblastoma multiforme (GBM), The WHO grade II tumour accounts for 10% of all intracranial and one-third of all primary CNS tumours. It is uncommon in the first decade, but more common in older children and young adults up to the age of 40-45 years, being relatively uncommon after the age of 65 years. Slightly more males than females are affected and it

Glioblastoma multiforme (Figs 57.39-57.42) This WHO grade IV tumour is the most aggressive and poorly differentiated tumour accounting for 15-20% of all intacranial twnours. It usually occurs after the age of 40 and is the most common astrocytic tumour after the age of 70. The majority are of astrocytic origin, though occasionally poorly differentiated ependymomas or oligodendrogliomas are described as glioblastomas multiforme (GBM). As these tumours may arise from pre-existing lower grade lesions, appearances can be complex and approximately 5% are multicentric and 5% will seed via the CSF. On CT and MR these tumours are of complex density and T, and T, signal respectively with areas of solid and necrotic tumour, resulting in heterogeneous, solid and ring enhancement. Haemorrhage may also be a feature of these tumours and lesions are associated with a large amount of vasogenic oedema and can extend along the corpus callosum. Unfortunately no i maging technique can determine the true extent of the lesion, although will identify the main bulk of tumour.

Fig. 57.38 A large complex mass in the left cerebral hemisphere shown on this postcontrast CT (A) and coronal T,-weighted MR image (B) was proven to be a pleomorphic xanthoastrocytoma. It has a large cystic component and the peripheral lobulated solid component shows marked enhancement. Some focal resorption of the overlying skull vault is shown on the coronal MR image.

Fig. 57.39 Axial MRI study (T,-weighted) shows tumour enhancement after IV gadolinium, resembling enhanced CT study. Mass effect and herniation under the falx are well shown. Glioblastoma multiforme.


Previous Page 1744

CONTENTS

Next Page


Fig. 57.40 Axial unenhanced CT (A) showing a large complex bifrontal partially cystic haemorrhagic tumour proven to be a glioblastoma. The T 2 signal characteristics are complex (B) and the coronal T, unenhanced image (C) shows large T, hyperintense components due to the presence of methaemoglobin. There is no significant vasogenic oedema despite tumour size but there is resultant obstructive hydrocephalus with early transependymal oedema around the occipital horns.


Fig. 57.41 Axial CTs pre- (A) and post- (B) contrast showing a slightly hyperdense mass precontrast involving the splenium of the corpus callosum and which shows marked enhancement. On the T 2- weighted axial image (C) the large amount of vasigenic oedema is clearly shown and there is some mass effect on the posterior aspects of the lateral ventricles. Although this mass was a glioblastoma the radiological differential would include primary central nervous system lymphoma.

Oligodendroglioma (Fig. 57.43) These glial tumours arise from

tumours and cyst formation or haemorrhage in 20%, necrosis is not commonly seen. On CT the calcification may be central, peripheral or even gyriform or ribbon-like and the tumour is hypo- or isodense. Seventeen per cent of tumours scallop and become adherent to the overlying skull vault due to their super icial location and slow growth. On MRI tumours are usually hypointense on T,- and hyperintense on T 2 -weighted images to grey matter and nearly 50% show variable enhancement. Vasogenic oedema is not a striking feature.

oligodendrocytes which make the myelin sheaths for the central portions of the cranial and spinal nerves but often tumours comprise a mixture of oligodendrocytes and astrocytes; for a tumour thus to be classified as an oligodendroglioma, most neuropathologists require a preponderance of neoplastic oligodendrocytes in the order of 75-90%. Oligodendrogliomas account for 5-10% of all intracranial neoplasms and 18% of all gliomas. These tumours occur slightly more frequently in males, usually between the age of 35-40 years, and 85% are located in the frontal lobes, often deep within the white matter but growing towards the surface. They are slowly growing, unencapsulated and less infiltrative than Ependymoma (Figs 57.44-57.47) These are low-grade tumours diffuse astrocytomas but may infiltrate the overlying cortex. arising from the ependyma forming the epithelial lining of Although dystrophic tumoral calcification occurs in 90% of the ventricular system, cerebral hemispheres, brainstem and

Previous Page

CONTENTS


Fig. 57.42 A heterogeneous mass proven to be a glioblastoma is shown i n the left cerebral hemisphere on the T 2 - weighted axial image (A) associated with mass effect and vasogenic oedema. The axial T, postcontrast i mage (B) shows marked enhancement of the tumour but also evidence of subependymal enhancement, thus spread.

1745

Fig. 57.44 A large avidly enhancing ependymoma is shown on the sagittal T, postcontrast image (A) occupying the lower part of the fourth ventricle, compressing the medulla and extending through the foramen of Magendie into the upper cervical canal. The axial T 2 i mage (B) shows that the mass is hyperintense and also extends out through the lateral recesses i nto the cerebellopontine angles, particularly on the right.


Fig. 57.45 T2 -weighted MRI study of tumour protruding into left lateral ventricle. The tumour is of mixed but mainly low signal, suggesting multilocular cysts. Ependymoma.

Fig. 57.43 Axial T2 (A) and T, postcontrast (B) images showing a large l eft frontal non-enhancing tumour involving cortex and white matter which was proven to be an oligodendroglioma. There is only minimal associated mass effect.

cerebellum, central canal of the spinal cord and tip of the filum terminale, and they account for 3-5% of all intracranial tumours and 9% of all glial tumours. Their incidence is higher in the paediatric population (approximately 17%) and they are slightly more common in males. Sixty-five per cent occur infratentorially, 25% supratentorially and 10% in the spinal cord. Overall the most common location is within the fourth ventricle, arising from the ependyma lining the floor, and as they are firmly fixed tumour resection is often incomplete. On imaging they typically present as intraventricular masses which take the shape of the ventricle, and those in the fourth ventricle may extend out through the midline foramen of Magendie or laterally through the foramina of Luschka into the cerebello-pontine angles and thus seed through subarachnoid space. There may be associated obstructive hydrocephalus and they need to be differentiated from medullo-

Fig. 57.46 CT demonstrates moder- Fig. 57.47 Strongly enhancing ate contrast enhancement in tumour tumour extending from lateral i nvading and obstructing fourth ven- recess of fourth ventricle into tricle. cerebellopontine angle. Ependy-

moma.

blastomas, which tend to calcify less frequently. The cerebral hemisphere ependymomas tend to arise adjacent to the ventricular system and are more frequently calcified or cystic than infratentorial tumours. CT shows a mass of slightly higher density than normal brain, which enhances uniformly or patchily. MRI shows ependymomas to have similar characteristics to gliomas, but to be more often associated with small cysts which may be multil ocular. Their characteristic site projecting into and from the

Previous Page 1746

CONTENTS

Next Page


fourth ventricle, or paraventricular and extending into a ventricle, should suggest the diagnosis (Fig. 57.45). Subependymoma This variant contains both ependymal and astrocyte cells and occurs mainly in elderly males. It also presents as an intraventricular mass in the lateral or fourth ventricle but is relatively benign and does not disseminate.

Neuroepithelial tumours of uncertain origin Gliomatosis cerebri This is relatively uncommon occurring mainly in adults in their twenties to forties, and with an equal sex incidence. Prognosis is poor as the tumour invasion is often extensive, multilobar and bihemispheric, but despite its extent patients usually present with few clinical symptoms at the time of diagnosis. On CT gliomatosis cerebri is typically hypodense and enhancement infrequent, occurring in about 12% of cases. On MR1 the affected areas are fairly ill defined and are iso- or hypointense on T,- and hyperintense on T,-weighted sequences. Enhancement on MRI is seen in approximately 50% of tumours and is solitary or multifocal. Although the affected brain appears enlarged, surprisingly significant mass effect is uncommon.

Tumours of the choroid plexus Choroid plexus neoplasms arise from the secretory choroid epithelium of the ventricular system and account for I-2% of all i ntracranial tumours. They are most frequently seen in childhood, 85% occurring below the age 5 and 40% in the first year of life. There are two main types-the well-differentiated choroid plexus papilloma and the malignant choroid plexus carcinoma. They are i nvariably intraventricular tumours except those that arise in the region of the foramen of Luschka. Choroid plexus papillomas occur most commonly in the first year of life and are classified as a WHO grade I tumour. In children they occur most commonly in the lateral ventricles centred on the trigone, whereas in adults in the fourth ventricle. Approximately 30% of choroid plexus tumours are carcinomas, WHO grade Ill-IV, and they present at a slightly later age (2 years) than the papillomas.

Overall the tumours are lobulatcd and friable and highly vascular and thus may present with an intraventricular haemorrhage. Approximately 65% of patients present with hydrocephalus which may be localised to the affected ventricle or more widespread due to affects on CSF production and absorption or ventricular obstruction. On CT the tumours are iso- or hypcrdensc intraventricular masses, with punctate calcification but avid homogeneous enhancement. If in the fourth ventricle it will be difficult to differentiate from a mcdull oblastoma or ependymoma (Fig. 57.48). In the lateral ventricle (more commonly the left) it lies in the region of the atrium, which is dilated. The ventricles may also be generally dilated from overproduction of CSF. On MRI the tumours are usually hyo- to isointense on T,weighted sequences with lower signal foci, either due to calcification or vascular signal flow voids. The resultant T z signal of the lesions is rather heterogeneous but on T,-weighted images oedema is frequently visible, extending into the brain substance near the site of attachment; enhancement is avid. Parenchymal invasion or metastatic nodules within the ventricles suggests choroid plexus carcinoma.

Neuronal and mixed neuronal-glial tumours This is a group of slow-growing benign tumours consisting of a mixture of differentiated neoplastic glial and neuronal elements or more rarely just neuronal elements (ganglioneuroma, WHO grade 1), and tumours occasionally undergo malignant transformation. They can occur anywhere in the central nervous system. Gangliogliomas These are WHO grade I or grade II tumours, which have both neuronal and glial elements; they can occur at any site but have a predilection for the temporal lobes. A cystic form with a predilection for the cerebellum has however been described. Only 10% of these tumours behave in an aggressive fashion. They are more common in children who often present with a long history of seizures. The tumours are well circumscribed, peripheral in location but may be cystic and calcification may be prominent. On CT they are usually isodense to hypodense and up to 57% are cystic and 40% calcified. Up to 50% exhibit either homogeneous or heterogeneous enhancement and occasionally the peripheral tumours result in re-modelling of the overlying skull vault inner table. On MRI the lesions are either solid or cystic masses which are usually hypointense to grey matter on T, and hyperintense on T,. Occasionally the cystic lesions present as a cyst with an enhanci ng mural nodule. Vasogenic oedema is usually mild.


Fig. 57.48 Axial T2 - weighted image (A,B) showing a lobulated vascular mass within the third ventricle which shows avid uniform enhancement on the coronal postcontrast T,-weighted image and is consistent with a choroid plexus papilloma. The lateral ventricles are markedly dilated due to obstructive hydrocephalus. (C) CT shows a rounded mass of increased attenuation with a low-density centre occupying the fourth ventricle. (D) After contrast medium, a slightly higher section shows strong uniform enhancement.

Previous Page

CONTENTS


1747

Fig. 57.49

Neuronal and mixed neuronal-glial tumours. A poorly defined hyperintense lesion consistent with a DNT is shown in the right temporal lobe on this T2 -weighted image (A). It involves cortex and white matter and the coronal T, unenhanced image (B) shows it to have a medial cystic component. The lateral patchy signal hyperintensity is likely to be calcification. There is no enhancement of this lesion (C). L'hermitte-Duclos disease This is a rare slow-growing cerebellar malformation due to the presence of a dysplastic gangliocytoma, which is of WHO grade 1. Pathologically it is thought to represent a hamartoma. A familial occurrence has been reported and associated developmental abnormalities include megalencephaly, heterotopia, microgyria, hydromelia and polydactyly. It is also associated with Cowden's syndrome (benign skin tumours, GI polyps, goitre and breast cancer) or neurofibromatosis but can occur alone. It typically presents as a mass lesion in young adults (20-40 years); CT shows a low-density cerebellar area with thickened folia and no enhancement after contrast. Calcification may be present. MRI shows low signal on T, and high signal on T. weighting with a typical striated appearance and some enhancement after gadolinium. These are WHO grade II tumours arising Central neurocytoma from small differentiated neurones typically seen in young adult male patients and less frequently in children or older adults. It presents as an intraventricular mass near the foramen of Monro, centred on the septum pellucidum and sometimes involving the third ventricle. Calcification and cystic change are common. They are usually well-circumscribed tumours, but are often locally recurrent after surgical excision, and may also seed through the CSF.

CT shows an isodense mass which enhances strongly after contrast and which may contain calcification. MRI reveals a heterogeneous tumour, hypodense on T, weighting and hyperdense on T z weighting, which enhances with gadolinium. Dysembryoplastic neuroepithelial tumour These mixed WHO grade I glioneuronal tumours most often presents with seizures and usually affect children, males more frequently than females, but are also encountered in the adult population. More than 80% of lesions occur in the temporal lobe (Fig. 57.49). Although they are cortically based lesions, only 38% exclusively involve the cortex and usually there is some involvement of underlying subcortical white matter. As they are slow growing and often superficial in location, scalloping of the calvarial outer table may be a feature. On CT they are often of low density; calcification is seen in 12.5% of lesions and cystic change is not uncommon. On MRI they are hypointense on T, and hyperintense on T 2 to grey matter with a variable enhancement pattern seen in approximately one third of tumours. There is usually very little associated perilesional vasogenic oedema or mass effect.

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ pineal parenchymal tumours

( Figs 57.50,57.51)

There is some variation in size of the normal pineal gland. Glands which are I cm or larger in diameter are usually cystic, and may be

Pineal tumours. Axial postcontrast CT (A) showing an enhancing mass centred on the pineal region invaginating into the third ventricle and resulting in obstructive hydrocephalus with transependymal oedema. This was a proven pineoblastoma. On the axial T 2 - weighted MR i mage (B) the mass is of slightly heterogeneous T2 signal; enhancement is also slightly heterogeneous on the sagittal T, postcontrast image (C). Fig. 57.50

Previous Page 1748

CONTENTS

Next Page


Fig. 57.51

Pineal tumour. (A) Isodense mass with calcification indenting back of third ventricle. (B) Enhancement after contrast medium (L30, W80). a conspicuous incidental finding on T,-weighted images as rounded l esions of high signal. However, unlike neoplasms, they do not i ndent the back of the third ventricle or obstruct the mouth of the aqueduct and produce hydrocephalus. Pineal region tumours constitute approximately 1 % of adult and 1 0% of all paediatric brain tumours. The pineal parenchymal tumours account for up to 30% of pineal region tumours and range from the moderately differentiated pineocytoma to the primitive pineoblastoma; in fact pineal germ cell tumours are much more common in this region, accounting for 50-70% of tumours. Pincoblastomas occur in the paediatric population, with a peak i ncidence in the first decade and rarely over the age of 30, whereas pincocytomas in the fourth decade, but in both cases there is no sex predilection. On CT these tumours are either hypodense or isodense and 50% are calcified. On MRI they are either hypointense or isoi ntense on T,; pineoblastomas tend to be more isointense on T, and pineocytomas more hyperintense. Enhancement is prominent and hydrocephalus may be seen in up to 90% of cases due to indentation of the tectal plate and obstruction of the cerebral aqueduct. Pincocytomas are better defined than pincoblastomas; the latter tend to be rather heterogeneous in appearance, with areas of cystic change, haemorrhage and necrosis, and CSF seeding occurs in 1 0%. There is a rare association between bilateral retinoblastomas and pineoblastomas (trilateral retinoblastoma) which in most cases presents before the age of 2 and is hereditary.

Germ cell tumours. A large heterogeneous T 2 signal mass is shown in the pineal region on the T 2 -weighted axial image (A) proven to be a pineal region teratoma. The mass comprises both solid and cystic components and there is heterogeneous enhancement after contrast on the sagittal T,-weighted image (B) and the mass results in obstructive hydrocephalus. Fig. 57.53

Table 57.4

Pineal tumours: different types found in 54 cases

1. Pineal-cell tumours

a. Pineoblastoma b. Pineocytoma 2. Glial tumour

8 5

a. Astrocytoma b. Glioblastoma 3. Germ cell tumours a. Germinoma

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ b. Teratoma

c. Embryonal cell carcinoma d. Choriocarcinoma

18 5 1 1

Other tumours which occur in this pinial situation and may sometimes enter into differential diagnosis include germ cell tumours (Figs 57.52, 57.53), arachnoid cysts in children and pineal teratoma. These are described elsewhere in the chapter. Finally glial tumours may also be found at this site (Table 57.4).

Tumours with neuroblastic or glioblastic elements (embryonal tumours) Primitive neuroectodermal tumours (PNETs)/medu11oblastomas They are predominantly tumours of childhood and com-

Fig. 57.52 (A,B) Pineal tumour. Large enhancing and partly calcified

mass. Heavy calcification is commonest in germ cell tumours.

prise about 3% of large pathological series of brain tumours, but comprise 15-20c/ of the malignant brain tumours of childhood, and 30-40% of childhood posterior fossa tumours. The cell of origin is the external granular cell, an embryological precursor of the internal granular cell layer of the cerebellar folia. They typically occur in the posterior fossa (85%) and may present as a congenital lesion but in the majority of cases present between the ages of 5 and 15 years. Although grossly they appear quite well

Previous Page

CONTENTS


circumscribed they are often invasive at their margins; 80% of midline tumours, arise in the roof of the fourth ventricle in the region of the inferior cerebellar vermis and inferior medullary velum. They usually grow to fill much of the fourth ventricle, and may extend out through the midline foramen of Magendie, but l ateral extension through the foramina of Luschka is uncommon, compared with ependymomas; 90% are associated with obstructive hydrocephalus at presentation. Forty per cent of patients have metastases at presentation, either as a result of seeding via the

1749

Prolactinomas manifest with gonadal dysfunction in both sexes and with galactorrhoea in a proportion. Symptoms can be associated with large tumours, but many are microadenomas when they first present. Since hyperprolactinaemia can also be produced by a variety of different drugs. it i s important to exclude this cause before imaging investigation is begun. Although MRI is the preferred imaging modality for assessing

pituitary adenomas as it avoids beam hardening artefacts and can evaluate local tumour invasion and compression of critical strucCSF or direct CNS metastases or both. This is one of the few tures such as the chiasm more accurately, CT will better demonmalignant brain tumours known to metastasise systemically, and strate destruction of the sellar floor. For large adenomas producing this is usually to bone, occurring in about 5% of cases. chiasmal compression CT will show the extent and relationships of Occasionally the tumour may primarily involve the cerebellar the suprasellar component as well as the enlarged sella, but sagittal hemisphere, and this pattern is more frequent in adult patients. and coronal reformats after intravenous contrast medium may he This tumour may occasionally arise within the brainstem, espenecessary to assess the full extent of a large adenoma. Macrocially the pons, and these tumours are associated with a worse adenomas first extend into the normal suprasellar cistern and then prognosis than those originating in the cerebellum. On CT the into the anterior end of the third ventricle. Eventually they can posterior fossa tumours are isodense or hypodense, with clumpextend as high as the foramen of Monro, and are often quite asymlike calcification present in 20%. On MRI they are isointense or metrical. The tumour can also extend laterally into the cavernous hypointense on T,- and isointense or hyperintense on T,-weighted sinuses and temporal lobe or downward into the sphenoid sinus. i mages. T, signal intensity may be heterogeneous due to calcificaLarge tumours may also extend subfrontally or above and behind tion, cystic change, haemorrhage or necrosis. Tumour enhancethe sella (Figs 57.54-57.63). ment is seen in 90% of cases. On MRI the normal pituitary yields a homogeneous brain-like Supratentorial PNETs represent 30%: of adenomas and are indicated

by high signal on T i - weighted images. Tumours isointense with brain on a variety of sequences are more likely to be hard whereas

Fig. 57.60 Pituitary adenomas. (A,B) Two different cases shown by MRI. (A) Sagittal section (T,weighted). Suprasellar extension as well as downward extension through the sphenoid sinus and backward extension through the dorsum and clivus. (B) Coronal section (T,-weighted). The tumour is displacing the cavernous sinuses laterally. The internal carotids appear black because of flow defects. (C,D) Sagittal sections of large adenoma before and after gadolinium enhancement. The anterior third ventricle is indented and the chiasm is stretched over the tumour.

Previous Page 175 2

CONTENTS

Next Page


Fig. 57.63 Mixed T 2 l ow signal is shown in the sellar region on the T 2 - weighted axial image (A) in a patient with a known pituitary macroadenoma and who presented acutely with symptoms and signs of pituitary apoplexy. On the coronal T -weighted unenhanced image (B) the sell and suprasellar mass is largely isointense but with strands of hyperintensity. The T 2 and T, signal characteristics suggest the presence of deoxyhaemoglobin and methaemoglobin in a subacute haemorrhage within a pituitary macroadenoma.

Fig. 57.65 Craniopharyngioma. Cystic tumour with solid enhancing component. Before (A) and after (B) enhancement.

Fig. 57.66 Cystic suprasellar mass with enhancing capsule. Craniopharyngioma. Some Rathke's cleft cysts can produce a similar appearance. Before (A) and after (B) enhancement.

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ Fig. 57.64 Axial CT (A) and T2-weighted MRI (B) showing a l argely cystic lobulated suprasellar mass. There is a small focus of calcification posteriorly within the l esion and appearances are those of a craniopharyngioma. The coronal T1 postcontrast image (C) is of another patient with a craniopharyngioma which has both cystic and solid enhancing components.

t umours which differ from normal brain arc generally softer and easier to remove.

Fig. 57.67 Craniopharyngioma. Sagittal midline MRI (T,-weighted) shows high-signal suprasellar mass due to cystic tumour. (Courtesy of Dr Gordon Thomson and Bristol MRI Centre.)

Craniopharyngioma These tumours of epithelial origin account

i s present in over 80% of the childhood cases, but is often absent

for up to 3% of primary intracranial tumours. They have a bimodal

i n the less common adult cases. They usually grow above the sella

age distribution, with most occurring in children and adolescents

and adhere tightly to the floor of the third ventricle and infundibu-

usually between the ages of 6 and 10, but a second smaller peak

lum, making surgical extirpation difficult or impossible. A small

occurs in patients in their fifties. In fact half of childhood supra-

proportion (15%) grow into the pituitary fossa and can produce

sellar tumours are craniopharyngiomas (Fig. 57.64). Calcification

enlargement and deformity of the sella. Some of these tumours

CONTENTS

Previous Page

A. Common

Pituitary adenoma Craniopharyngioma Aneurysm Suprasellar meningioma

Next Page

B. Rare

Optic chiasm glioma Hypothalamic glioma Germinoma Ganglioglioma Metastasis Arachnoid cyst Epidermoid Rathke's cleft cyst Inflammatory masses Histiocytosis X

may appear slightly ectopic and to lie in the third ventricle or behind the sella. Histologically they are of two main types-adamantinomatous and squamous-papillary; the former are the more common variety and are often lobulated masses with cystic components which can be hyperintense on T,- and T,-weighted sequences if the cyst contents are proteinaceous. The solid components enhance and calcification is a feature, but this is better demonstrated with CT. The squamous-papillary type are less often calcified or cystic and tend to have a more uniform rounded configuration. The solid components of the tumour are isointense on T, and hyperintense on T, to grey matter, and the solid components enhance. This variety of craniopharyngioma is more common in the adult population (Figs 57.65-57.67). Differential diagnosis A wide variety of mass lesions can occur i n the suprasellar region. Many of these can be indistinguishable clinically from pituitary adenomas, and some can resemble them at imaging (Box 57.3). Many of the lesions listed have features that readily differentiate t hem from pituitary adenomas, and these are described under the i ndividual lesions. Calcification, if present, suggests a craniopharyn*ioma; if the shape is an arc or a ring, aneurysm or craniopharyn gioma is possible. Suprasellar meningiomas can also calcify but present in an older age group, and calcification, if present, is homogeneous or nodular and never in an arc. Arachnoid cysts arc easily i dentified by their CSF density and thin capsule at CT or MRI.

Lymphoma of brain. (A) CT before contrast. (B) After contrast. Isodense tumour with little mass effect enhances strongly with contrast. (C) MRI, T,-weighted, in another patient. Low-density paraventricular mass involves thalamus and extends into third and left lateral ventricles. Fig. 57.69


Secondary lymphoma of the brain is exceedingly rare, though it can occur. More usually, however, lymphoma of the brain is found without any systemic involvement. This was previously an uncommon tumour accounting for I I/ of primary brain neoplasms. but with the recent AIDS epidemic this tumour has been encountered with increasing frequency particularly in younger adults, where there is a relationship with Epstein-Barr virus exposure (Fig. 57.68);

Fig. 57.68

Bilateral complex intrinsic masses are shown in the cerebral hemispheres in a patient with known AIDS. On biopsy these were proven to be primary central nervous system lymphoma. There is a large amount of vasogenic oedema on the T 2 - weighted image (A) associated with the right frontal mass and there is extension across the corpus callosum. The coronal T, unenhanced image (B) shows faint T, high signal in the right frontal mass consistent with methaemoglobin. Postcontrast (C) the right frontal lesion shows thick irregular ring enhancement. Subependymal enhancement around the anterior aspects of the frontal horns is also evident, a characteristic feature of cerebral lymphoma.

Previous Page 1754

CONTENTS

Next Page


i nterestingly there has also been an increase in non-HIV-related sporadic primary CNS lymphoma (PCNSL), where the peak incidence occurs in patients in their fifties (Fig. 57.69). Most of the tumours tend to be high-grade B-cell tumours and the AIDS-related tumours tend to have a poor prognosis, but that of the sporadic tumours has i mproved with a median survival of 5 years. In the HIV-infected population the tumour tends to be multifocal with periventricular spread, and necrosis and haemorrhage are common, whereas the sporadic form may be limited to one or two sites within the brain. Lesions are often located deep within the cerebral hemispheres, but can occur in the posterior fossa. On CT they are often hyperdense pre-contrast and show either solid or ring enhancement, On MRI signal intensity of the masses is variable and can be isointense to grey matter on T, or isointense or hyperintense on T, and signal abnormality and enhancement often extends along an ependymal surface. Vasogenic oedema and mass effect are variable and, like gliomas, PCNSL can extend along the corpus callosum. In the HIVi nfected population the main radiological differential of PCNSL is toxoplasma abscesses.

(A) There i s a high-density parasagittal mass with irregular calcification. (B) The tumour enhances strongly with contrast agent. Parasagittal meningioma. Fig. 57.70

Germ cell tumours account for eneraliscd changes are seen with viruses, in particular with paraventricular and cortical calcifications. cytomegalovirus, herpes simplex and the JC papovavirus responsPML is due ible for PML. The HIV virus itself has been found in the brain of Progressive multifoeal leueoeneephalopathy (PML) AIDS patients at autopsy, and is considered to be responsible for a t o a papovavirus and occurs only in special pathological cirdirect cytopathic effect leading to the encephalopathy, dementia and cumstances. It particularly affects patients with compromised atrophy frequently encountered. i mmunity, c.g. Sufferers from chronic lymphatic leukaemia or Hodgkin's disease, or those receiving immunosuppressive I maging CT and MR/ will identify focal lesions such as therapy. abscesses due to toxoplasmosis (Fig. 58.44) or cryptococcus

Previous Page

1794

CONTENTS

Next Page

A TEXTBOOK OF RADIOLOGY AND IMAGING l eptomeningitis or as intracerebral granulomas, although both may be present. Clinically patients with chronic leptomeningitis present with hydrocephalus and those with granulomas with focal signs, depending on the site of the lesion (Fig. 58.53). CT In the meningitic form CT will demonstrate communicati ng hydrocephalus associated with enhancing hyperdensities in the basal cisterns, an appearance which is non-specific. Granulomas are also hyperdense and enhance with contrast medium. They may be multiple and small, resembling metastases, although there is no surrounding oedema. A single large lesion is occasionally seen and simulates a neoplasm.

(A) Axial T2 -weighted image in a patient with AIDS showing widespread signal abnormality in the white matter of HIV encephalitis. There i s also background atrophy with ventricular enlargement and mild sulcal widening. (B) AIDS encephalitis. Fig. 58.52

(Fig. 58.43). The appearances seen in PML and involvement by l ymphoma have been described above. In some cases brain biopsy may be necessary to distinguish between abscess and neoplasm. However, toxoplasrnosis is the commonest cause of focal lesions in AIDS patients, and anti-infective treatment with pyrimethamine and sulfadiazine may be given a trial before resorting to brain biopsy. Non-specific changes such as brain atrophy with cortical shrinkage and ventricular dilatation are readily identified by CT or MRI. The latter can also demonstrate non-specific white-natter signal changes on Tfl-weighted images in 50% of patients who are HIV positive and in 70% of cases with established AIDS (Fig. 58.52). They are thought in most cases to he due to a direct HIV infection. Other techniques that have been used to investigate HIV encephalopathy include MR spectroscopy and radioisotopes (PET and SPECT). In both cases sensitivity is higher than MRI but specificity is low and neither method is as yet widely accepted.

MRI A variety of changes have been described. Many of these have not been detected by CT, but they are non-specific. The commonest abnormalities are areas of hyperintensity on T,-weighted i mages in the cerebral white matter, usually located more superficially than the lesions of multiple sclerosis, although pcriventricular lesions do occur. Surface granulomas may be shown, especially adjacent to the major cisterns (Fig. 58.50), and also deeper lesions involving the basal ganglia typically with marginal enhancement after contrast. Behfet's syndrome MRI commonly reveals non-specific changes, characterised by hyperintensity on T-weighted images. The optic chiasm, hypothalamus and brainstem are frequently involved.

Cerebral atrophy may be focal or generalised. Focal atrophy may be vascular, infective or traumatic in origin and the causes have been discussed above. It may also be degenerative, as in olivopontocerebellar atrophy, which is discussed


Sareoidosis I nvolvement of the CNS is not uncommon, occurring in around I5° of affected patients. It manifests either as a basal

below.

Genera/iced atrophy is more commonly degenerative or idiopathic, although vascular, inflammatory, toxic and traumatic types also occur or may he associated. Toxic atrophy may follow alcohol or drug abuse. Generalised atrophy of the brain is a routine concomitant of the ageing process and is a normal finding in the elderly, increasing with age. The loss of neural tissue in the senile brain occurs in a cranium of unchanged size and is therefore compensated by an

Fig. 58.53 (A) Axial MR section with gadolinium enhancement (T,-weighted) shows superficial peduncular lesions (arrows). Sarcoidosis. (B) Intracanalicular optic neuritis secondary to sarcoid. Left optic nerve shows high signal on T,-weighted MR. (C) Sarcoidosis with marked meningeal thickening.

Previous Page

CONTENTS


1795

none are reliable and hence are rarely used in most centres. Radionuclide cisle rnographv or CT cisie rnographv have been advocated. With both methods the ventricles of hydrocephalus patients may be outlined at 12-24 hours and remain so for 48-72 hours: normal patients or those with atrophy show little radionuclide or contrast within the ventricles but much over the cortex. However these methods are rarely used. Pseudoatrophy Fluid shift from the brain may result from high doses of steroids and from anti diuretics. It is also described in anorexia, protein starvation, and severe dehydration. The resulting pseudoatrophy may simulate the CT appearances of true atrophy, Generalised atrophy. (A) Dilated lateral ventricle. (B) Enlarged sulci and interhemispheric fissure (L36, W80). Fig. 58.54

but is reversible with appropriate treatment.

i ncreased volume of CSF, which occupies the resulting enlarged ventricles, sulci and subarachnoid space. I maging CT or MR/ of patients with cerebral atrophy shows i n varying degrees: I. Ventricular dilatation with rounding of the ventricular angles 2. Enlargement of the cerebral and cerebellar sulci 3. Increased width of the subarachnoid space and basal cisterns, with increased prominence of the intcrhemispheric and sylvian fissures. Diagnosis is usually obvious (Fig. 58.54) but can occasionally he more difficult in patients with communicating hydrocephalus, particularly in the elderly patient with `normal-pressure' or 'lowpressure' hydrocephalus. Communicating hydrocephalus ( Fig. 58.55) The ventricular bodies and frontal horns are often more rounded: the temporal horns are also characteristically more dilated and prominent. They are usually normal or relatively small with atrophy. In addition the sulci are less prominent with hydrocephalus, rather than dilated as in atrophy. Periventricular lucencies at CT, usually most

In this section, the following will he considered: I. The dementias Alzheimer's disease (AD) Frontotemporal dementia (FTD) Lewy body dementia Vascular dementia (including CADASIL) 2. Extrapyrarnidal and other movement disorders with or without dementia Parkinson's disease Multiple system atrophy (MSA) Progressive supranuclcar palsy (PSP) Motor neurone disease (MND) Corticobasal degeneration (CB D) Huntingdon's disease 3. Prion diseases Sporadic and iatrogenic Creutzfeldt-Jacob disease (CJD) Familial CJD New variant CJD (nvCJD) Gcrstmann-Strfussler-Scheinker disease (GSSD) Fatal familial insomnia (FFI) 4. Cerebellar atrophy.


marked around the frontal horns, are also a feature of hydrocephalus, and are due to seepage of CSF through the ependyma. Nevertheless, cases do arise where imaging remains equivocal and in these cases further investigation may be required. The problem is usually to decide whether patients with dementia are suffering from normal pressure hydrocephalus (NPH) or merely atrophy, as the former have been said to benefit from a shunting operation. Many investigations have been advocated to distinguish central atrophy (mainly ventricular enlargement) from NPH, and

These are usually diseases of the ageing population. Most patients progress to death within 5-10 years. In the early stages, the diseases can he difficult to differentiate from normal ageing, or between one another at a clinical level. Interest has been greatly enhanced in recent years by advances in molecular pathology and

Fig. 58.55 (A-D) Communicating hydrocephalus complicating meningitis. All four ventricles are dilated and the sulci are effaced.

Previous Page 1796

CONTENTS

Next Page


Fig. 58.56 (A) Alzheimer's disease. Coronal MRI showing bilateral symmetrical atrophy of each hippocampus and parahippocampal gyrus. (B,C) Frontotemporal dementia (non-Pick). Coronal MRI showing severe atrophy of the anterior part of the left temporal lobe and mild enlargement of the frontal horn of the left lateral ventricle. (D) True Pick's disease, showing more mild left temporal lobe atrophy.

cortices at end-stages. On MR1 or CT, accentuated atrophy in medial temporal regions and relative sparing of the cerebellum c.rn have predictive value (Fig. 58.56A). This category includes a variety of diseases such as Pick's disease, semantic dementia, the primary progressive aphasias and other mainly frontal or temporal lobar degenerations. Some are characterised by very specific isoforms of tau protein, such as Pick, and others are non-tau, such as MND-like inclusion body frontal lobe degeneration. Clinically early stages are characterised by behavioural or speech disturbances. On MRI and CT, striking accentuated atrophy in temporal and/or frontal lobes (usually mainly temporal), with an equally striking anteroposterior gradient of decreasing severity, can be seen, which will then have considerable positive predictive value (Fig. 58.56B). Frontotemporal dementia (FTD)

genetics. As well as neurone loss and gliosis, most of these conditions are characterised pathologically by intracytoplasmic inclusions in neurones or glia, or mainly extracellular amyloid deposits. Many of the proteins in these materials have been characterised and the genes which encode them identified. Familial trends are recognisable in an increasing proportion. At this moment, however, these diseases arc still classified syndromically, but the molecular pathologies suggest that there is enough phenotypic variation for syndromes to overlap, adding to diagnostic difficulties. This is a rapidly evolving area. The role of imaging traditionally has been to exclude symptomatic causes such as neoplasms, chronic infections, arteritis, diffuse vascular malformations, or to diagnose intercurrent disease. The development of newer forms of treatment, both symptomatic and directed at molecular mechanisms, has increased the need for precise early diagnosis, and the importance of the positive predictive value of imaging has become considerably enhanced. Because structural imaging has been perceived as wanting in this area, clinical roles for functional imaging using radionuclides (PET and SPECT) or functional or perfusion MRI have been advocated. Despite a large and encouraging literature developed over many decades, functional imaging has proved disappointing as a discriminator, especially in difficult cases, or in early stages of disease, and is little used by many centres except in ongoing research.


This is characterised pathologically by neurofibrillary tangles within neurones, and neuritic plaques. Progressive neurone loss, manifesting on CT and MRI mainly as volume loss (atrophy), occurs in the hippocampus in the preclinical and early stages and progresses to involve temporal and medial parietal lobes, only affecting primary motor and sensory Alzheimer's disease

The molecular structure and genes for Lewy bodies have not been fully characterised as yet. Difficult to distinguish from AD clinically, fluctuation in cognition and visual hallucinations are suggestive, together with overlaps with Parkinson's disease. There arc no suggestive MR1 or CT features. Lewy body dementia

Vascular dementia The almost iniquitous finding of some ischaemic damage in cerebral white matter and basal ganglia in MRI in the ageing population, its poor correlation with cognitive i mpairment, and the relative rarity of extensive cortical infarction manifesting mainly as dementia, have cast serious doubt on the existence of this as an independent entity. A familial form affecti ng younger patients is currently very topical: CADASIL (cerebral autosomal dominant arteriopathy, subcortical infarcts and l eucoencephalopathy); stroke and migraine as well as dementia are common clinical features. The gene is known, but sporadic mutations occur. On MR1 there is usually striking confluent i schaemic damage throughout the cerebral white matter (includi ng temporal lobes, often spared in arteriosclerotic disease) and basal ganglia (Fig. 58.57).

Parkinson's disease Lewy body intracellular inclusions and neurone loss, maximal in the pars compacta of the substantia nigra, are the pathological hallmarks. Loss or reduction of the

Previous Page

CONTENTS

Next Page

Corticobasal degeneration This is another tau-opathy. MRI is usually unhelpful, but some cases show selective asymmetric, mainly parietal lobe atrophy, not seen in other conditions. Huntington's disease A direct gene test is now available for a gene which encodes for the protein Huntington. On CT and MRI the caudate nuclei are selectively atrophic in established cases, giving rise to a characteristic appearance of the frontal horns of the lateral ventricles.

Fig. 58.57 T2- weighted axial images in a patient with CADASIL (cerebral autosomal dominant arteriopathy, subcortical infarcts and leucoencephalopathy) showing predominantly confluent signal hyperintensity in the cerebral white matter, fairly symmetrical in distribution. The changes are particularly marked in the temporal lobes (A). The basal ganglia also exhibit a lacunar state (B).

normally higher signal of the substantia nigra between the cerebrat peduncle and red nucleus in the midbrain on T 2 - weighted MRI has been emphasised by some but is highly unreliable. Multiple system atrophy This is characterised by mainly glial inclusions consisting of alpha-synuclein. There are two main forms. (I) The pontocerebellar type: cerebellum and pons show atrophy on MRI, pontine involvement being characterised by loss of pontine nuclei and shrinkage of the middle cerebellar peduncles (Fig. 58.58). (2) Striatonigral type: the putamen may show reduced signal on MRI due to increased iron deposition and a thin l ongitudinal cleft may appear in its lateral portion; neither are specific.

Prion diseases These are the spongiform encephalopathies affecting mainly grey matter. Some arc hereditary, others are sporadic and transmissible. MRI and CT usually show only progressive cerebral cortical atrophy in CJD, but diffusion-weighted i mages often show altered diffusion in large areas of cerebral cortex. In a minority of cases (10%), diffuse, symmetrical mildly increased signal in the caudate nucleus and putamen may be seen. In new variant CJD, symmetrically increased signal in the pulvina of each thalamus is a consistent and specific diagnostic feature on MR1, sometimes also involving the medial thalamus and upper midbrain. Progressive mainly cerebellar atrophy may be seen on MRI i n GSSD. cerebellar atrophy Apart from those already mentioned, the main causes are paraneoplastic, toxic (e.g. phenytoin, alcohol), genetic (spinocerebellar atrophies, Friedreich's ataxia, hereditary telangiectasia), and of course, unknown. Involvement of the pons on MRI usually indicates multiple system atrophy; atrophy of the l ower brainstem and upper spinal cord is suggestive of Friedreich's ataxia (Fig. 58.59). W ilson's disease (hepatolenticular degeneration) Hepatolenticular degeneration is a familial metabolic disease due to an abnormality of copper metabolism in which cirrhosis of the liver is associated with degeneration of the corpus striatum. It usually commences in the second decade. Clinically there is progressive rigidity and tremor, dysarthria and dysphagia, and in the terminal stages some degree of dementia. A characteristic feature is a ring of brown pigpi--


Deposition of tau protein and neurone loss accentuated in the midbrain tegmentum may be indicatedon MRI by shrinkage and slight signal changes in the midbrain, sparing the cerebral peduncles. Progressive supranuclear palsy

Motor neurone disease Non-(au inclusions and degeneration in spinal motor neurones and pyramidal tracts usually leave no indications on MRI. Occasionally signal change in the tracts can be seen in the brain.

Fig. 58.58 Axial T2 - weighted i mage in a patient with multisystem atrophy showing marked atrophy of the pons and middle cerebellar peduncles. Abnormal T2 high signal is present in the pons in the form of a cross (hot cross bun sign) with more confluent signal hyperintensity in the middle cerebellar peduncles. The cerebellum is also atrophic and the 4th ventricle is consequently enlarged.

CT The brain may show no significant changes, but lowdensity lesions may be seen in the basal ganglia of more advanced cases. MRI MRI may show increased signal in the basal ganglia on T,-weighted images and low signal on T,-weighted images. The

Fig. 58.59 Cerebellar atrophy showing widened sulci and fissures. The brainstem is also atrophic.

CONTENTS

Previous Page

1798

Next Page


Major I diopathic Familial Hypoparathyroidism Pseudohypoparathyroidism Minor or atypical Fahr's syndrome

Fig. 58.60

T2- weighted axial images demonstrating basal ganglia, thalamic and midbrain lesions in Wilson's disease.

l entiform nucleus is most commonly affected. In a few cases low signal has been observed in Tz - weighted sequences, associated with either copper or iron deposited in the basal ganglia.

Bilateral symmetric calcification of the basal ganglia is not uncommon. Pathologically, the heaviest calcifications arc found in the globus pallidus, caudate nucleus and putamen. The dentate nuclei of the cerebellum and the internal capsule may also be involved. Smaller collections, which are only rarely visible at imaging, occur at the junction of cortex and white matter and in the cerebellar cortex. Histopathologically the calcifications are deposits of tiny pericapillary calcospherites plus calcification in the medial walls of tiny arteries and veins (Fig. 58.61).

Cockayne's syndrome Carbon monoxide poisoning Lead poisoning Toxoplasmosis Mineralising microangiopathy Secondary hyperparathyroidism Mitochondrial cytopathy

Familial calcification is very rare but when it occurs affects young as well as elderly members of the same family. Hvpoparathyroidisrn, either idiopathic or following thyroidcctomy, is the most important aetiologic cause because it is potentially treatable. Pseudohvpoparathvroidism (Albright's syndrome) may also he associated with similar calcification. This rare condition is described in the bone section (see Ch. 42). Fahr's syndrome (idiopathic familial cerebrovascular ferrocalcinosis) is a rare familial condition commencing in childhood and characterised by the deposition of iron and calcium in the basal ganglia, dentate nuclei and subcortical regions. It presents with spasticity and choreoathetoid movements, proceeding to progressive mental deterioration (Fig. 58.62). Cockavne's syndrome, which is inherited as an autosomal recessive, is characterized by dwarfism and progcria. There is microcephaly and a thick cranial vault as well as the intracranial calcification (Fig. 58.63). The calcification in toxoplasmosis is linear or nodular rather than amorphous (Fig. 58.43) and in the other conditions the diagnosis is usually clear from the history and clinical findings. Carbon monoxide poisoning is usually due to attempted suicide. Classic cases show necrosis of the globus pallidus, giving rise to bilaterally symmetric small lucent areas in this characteristic site. Calcification is unusual but has been described. Mineralising inicroangiopathv is also referred to as disseminated necrotising leucocncephalopathy. It is a rare condition seen mainly i n patients with leukaemia treated with cerebral radiotherapy foll owed by intrathecal mcthotrexate. In the acute phase there is bilat-


The major causes, and some other conditions in which calcification of a less typical type may be found in the basal ganglia, are listed in Box 58.3. Idiopathic calcification is by far the commonest type and is seen mainly in the elderly, where it may be regarded as a normal variant of little pathological significance. The condition is usually encountered as a chance finding either at simple X-ray of the skull (see Ch. 53) or more frequently at CT. The latter is far more sensitive than simple X-ray and has shown the condition to be more common than previously realised (Fig. 58.54), being present in 0.6 1-/(- of one l arge series of 7000 consecutive CT examinations.

Fig. 58.62

Calcification of the basal ganglia. Fig. 58.61

Axial CT scans demonstrating extensive calcification in the basal ganglia, thalami, subcortical white matter, brainstem and dentate nuclei of the cerebellum in Fahr's syndrome.

Previous Page

CONTENTS

Next Page

Changes in the white matter are well shown by CT, but there is no doubt that MRI is even more sensitive in demonstrating whitematter lesions. The changes in specific pathological conditions arc described below, but it should be realised that there are also many physiological causes of changes in white-matter density at CT or in i ts signal at MRI. Neonates, particularly premature infants, normally show low-attennation white matter at CT and poorer differentiation between white and grey matter at MRI. This is due to incomplete myclination and i mproves as myclination progresses. Low-attenuation white matter at CT and abnormal contrast at MRI is also seen in the elderly, usually in association with generalised senile atrophic changes. Other causes of generalised low attenuation of the white matter at CT and altered signal at MRI include uraenric and hepatic coma and hspertensive crises: maligncuii disease and muscular dystrophy are other rare causes. Vasogenic oedema associated with tumour or trauma can also produce focal or more generalised low density, as can radiation damage. Fig. 58.63 Axial T 2- and T,-weighted images in Cockayne's syndrome demonstrating gross supratentorial atrophy and diffuse signal abnormality and reduction in bulk of the white matter. T, high and Tz low signal in the basal ganglia is indicative of calcification.

Multiple sclerosis (MS) This is one of the commonest neurological disorders and is characterised by disseminated plaques of demyelination and gliosis throughout the neuraxis. The sites of election are:

oral leucoencephalopathy, and CT may show bifrontal low density i n the white matter, but atrophy follows later. Mineralising microangiopathy shows low attenuation in the white matter, mainly near the corticomedullary junction and with relative sparing of the deeper white matter. Thin reticular or serrated linear calcification is also present at the corticomedullary junction as well as in the basal ganglia and posterior fossa. giving rise to a very characteristic appearance. Secondary hsperpai-uth~roidism is often associated with extensive calcification in the falx and tentorium, and this is evident at simple X-ray of the skull. Rarely, intracerebral calcification is identified at CT and Occurs in the Nasal ganglia and also in the white matter.

I. 2. 3. 4. 5.

Pcriventricular Optic pathways Brainstem Cerebellar white matter and peduncles Spinal cord.

Most authorities consider the aetiology to be immunopathological, and there is an increased incidence of certain tissue type antigens i n the affected patients. There is also raised gamma globulin in the CSF. Young adults are primarily affected, with an increased i ncidence in the colder wet temperate zones of the northern hemisphere.


Under this heading we shall discuss conditions of differing aetiol ogies which have in common predominant involvement of the white matter. White-matter diseases have been classified as dvsnwelinating, where there is abnormal formation or maintenance of myelin, and denivelinating, where there is destruction of normally formed myelin (Box 58.4). To these may he added miscellaneous causes of what may he termed 'secondary' demyelination inflammatory and postinllanmnatory lesions and neoplastic, vascular, toxic and other disorders. Some of these have already been discussed, e.g. disseminated encephalomyelitis, progressive multifocal encephalopathy (PML) and Reye's syndrome. Other conditions which fall into this category are described below.

Dysmyelinating Leucodystrophies Metabolic disorders Demyelinating (myelinoclastic) Multiple sclerosis

Schilder's disease Central pontine myelinolysis Marchiafava-Bignami disease Secondary to other disorders

CT CT may show no abnormality even in the acute stages. In about one-third of patients, small low-density areas are seen in the white matter, particularly adjacent to the atria or in other periventricular sites (Fig. 58.64). In the acute stage these may show marked enhancement (Fig. 58.64B.C), but chronic lesions fail to enhance. Occasionally enhancing small lesions are seen which were isodense before adding contrast medium. Classic cases show no mass effect or surrounding oedema, but very rarely mass effect and oedema are encountered which can simulate tumour, particularly with a single large lesion. MRI This is far more sensitive than CT in the demonstration of MS plaques. and it has been claimed that an accuracy of nearly 100% can be obtained with modern machines. The most characteristic appearance is that of periventricular nodular hyperi ntense lesions on T,-weighted images, most numerous posteriorly (Fig. 58.65), and plaques are also well shown at the grey-white matter interfaces. MRI can even image lesions in the spinal cord, as well as the brainstem and cerebellum. Multiple cerebral lesions are demonstrated in over half the patients presenti ng with a single episode of optic neuritis, or with clinically i solated brainstem or cord lesions. As already noted, sonic lesions may have enough mass effect to simulate a tumour, although this is rare. Others may show a central area of greater signal intensity. resembling a target.

Previous Page 1800

CONTENTS

Next Page


Fig. 58.64 Multiple sclerosis. (A) Paraventricular and white-matter low-density areas. (B,C) Enhancing paraventricular and white-matter low-density areas. (D) Axial MRI section J, -weighted). Multiple sclerosis plaques are seen as areas of high signal.

Contrast enhancement after giving gadolinium occurs in the acute phase, presumably indicating activity, as it later subsides. It should be remembered that the MRI lesions are non-specific and must be considered in the clinical context. MRI is so sensitive that it shows similar periventricular lesions in around 25% of elderly patients. These may represent ischaemic lesions, as they are particularly noted in patients with multi-infarct dementia. Similar l esions have also been noted following radiotherapy and in encephalomyelitis. Schilder's disease (diffuse sclerosis) This condition affects children as well as young adults. Histopathologically, the lesions are

si milar to those of MS, but are more extensive, and clinically the course is continuous and progressive, even fulminant, rather than i ntermittent and relapsing. The main lesions extend through the corpus callosum into both parieto-occipital regions.

MRI This shows the white-matter lesions better than CT, and their true extent is more clearly reflected. Central pontine myelinolysis This rare condition, primarily affecting the pontine white matter, was originally thought to be due to chronic alcoholism and malnutrition, but was later shown to be, associated with many other chronic illnesses, including neoplasia and liver and kidney diseases. The underlying cause may be an abnormality of plasma sodium, and in most cases hyponatremia has been observed, although the relationship is not a si mple one. It is often fatal, but recovery can occur, with varying degrees of neurological deficit.

I maging CT may demonstrate low density in the affected pons, usually only when the disease is well advanced. MRl i s more sensitive and reveals more extensive change. There is increase in T, and T, within the pons, and changes can be more widespread, extending into the midbrain and thalami and even into the subcortical white matter.


CT Bilateral low-attenuation areas are shown in both parietooccipital regions which enlarge progressively but not symmetrically. Contrast enhancement is uncommon.

Marchiafava-Bignami disease This is another rare condition,

with a peculiar white-matter localisation. The corpus callosum is mainly affected, although other central pathways may also be i nvolved. There is an association with chronic alcoholism, but this i s not invariable, and chronic cyanide intoxication can produce si milar lesions. Radiation leucoencephalopathy Radiation used in the treatment

The T 2 axial image (A) demonstrates multiple hyperintense l esions in the cerebral white matter, many of which are periventricular in distribution and typical of multiple sclerosis. A couple of the lesions are associated with perilesional oedema, giving a target appearance, and on the postcontrast Tl coronal image (B) show patchy and ring enhancement; these lesions are plaques of active demyelination. Fig. 58.65

of malignant tumours can itself produce harmful effects. This is mainly due to occlusion of small vessels and resulting infarction. In the acute stage there are changes in the white matter indistinguishable from tumour at both CT and MRI. Since the radiation damage is sometimes delayed, and can occur months later this can give rise to a difficult differential diagnosis from tumour recurrence or extension both clinically and on imaging. In this situation PET may help by showing hypermetabolism in tumour recurrence and hypometabolism in radiation necrosis (see below). Radiation changes will eventually progress to local atrophy, although white-matter low density on CT and signal change on MRI will persist (Fig. 58.66). The association between radiotherapy, methotrexate and PML has been discussed above.

Previous Page

CONTENTS

Next Page

Fig. 58.66 (A-C) Axial CT scans demonstrating confluent and symmetrical low-density change in the deep and peripheral supratentorial white matter in metachromatic leucodystrophy. (D) Axial MR Tz W shows widespread low density in white matter.

Leucodystrophy is a term commonly used to refer to dysmyelination disorders. They arc mainly due to genetic defects in the formation and maintenance of myelin, usually present in infants and children, with increasing developmental delay, progressive dementia and neurological deficits. They include the following conditions, which, like most disorders of lipid metabolism, are inherited as autosomal recessives and are associated with lysosomal enzyme deficiencies:

There is mental retardation or regression and other neurological which are progressively fatal. Globoid leucodystrophy (Krabbe's disease) This is characterised

by a lack of the enzyme P-galactocerebrosidase. It presents in i nfancy with retardation and spasticity, and is usually fatal by the second year. Rarely, it presents in later childhood with a more chronic course and early visual failure (Fig. 58.67). Spongiform degeneration The condition is so-called because of

• Metachromatic leueodystrophy • Globoid leucodystrophy (Krabbe's disease) • Spongiform degeneration (Canavan's disease).

the characteristic spongy degeneration produced in the white matter of the affected infant. It is one of the few leucodystrophies that produces enlargement of the head, others being Alexander's disease and GM, gangliosidosis (Tay-Sachs disease).

Metachromatic leucodystrophy This is one of the commoner

X-linked leucodystrophies These are less frequent, but two

hereditary leucodystrophies; clinical symptoms usually commence in infancy, although they can be delayed into adolescence or later. It is due to a deficiency of the enzyme arylsulphatase A.

i mportant ones are adrenoleucodystrophy and PelizaeusMerzbacher disease.


Adrenoleucodystrophy This is an X-linked genetic disease that was at one time confused with Schilder's disease. It affects

(A, B) Axial T z signal abnormality in the periventricular occipitoparietal white matter is fairly symmetrical and extends across the corpus callosum; there are zones of differential T z signal hyperintensity representing the different phases of demyelination and appearances are typical of X-linked adrenoleucodystrophy. On the T, coronal image a band of enhancement is present at the active edge of the demyelination. Fig. 58.68

Fig. 58.67 (A,B) Axial CT and T 2 - weighted MRI i n Krabbes's disease

showing focal bilateral thalamic and posterior internal capsular lesions which are calcified on CT. T z signal abnormality is also present in the deep periventricular white matter with some cavitatory change adjacent to the frontal horns.

Previous Page

1802

CONTENTS

Next Page


boys aged 4-6 years and occasionally a little older. The victims develop adrenal insufficiency, resembling that in Addison's disease, together with dementia, cortical blindness, ataxia and spasticity (Fig. 58.68). Pelizaeus-Merzbacher disease This is also X-linked. It presents mainly in infancy, but can begin later. There is progressive dementia, ataxia and nystagmus, often with extrapyramidal feaCures such as dystonic postures and torsion spasms. The metabolic basis remains obscure. Alexander's disease i s one of the few leucodystrophies that

occurs sporadically and is characterised histopathologically by the abundant presence of Rosenthal fibres in the affected brain. Definitive diagnosis thus depends on brain biopsy (Fig. 58.69). Batten's disease (neuronal ceroid lipofuscinosis) This was also

referred to as cerebrnmacrrlur degeneration, and in contrast to the above the brain is atrophic, often markedly so. CT CT in the leucodystrophies shows low-density areas in the white matter which are bilateral but not necessarily symmetric and become more extensive as the disease process progresses ( Fig. 58.56). In adrenoleucodvsnopliv the parieto-occipital regions are i nvolved and in thin curvilinear or serrated rim of contrast enhancement may be seen, representing perivascular inflammation at the margins of the extending process. Such enhancement is not seen in the other leucodystrophies described above. Another unusual finding described in adrenoleucodystrophy is dystrophic calcification in the affected white matter. Krahbe's disease has been described as showing increased density in the basal ganglia as well as hypodensity of the white matter, while A lexander's disease has been reported as showing unusual enhancement of the caudate nuclei and periventricular areas in one reported case. Considerable atrophy has been reported 1 7 to accompany the white-matter changes in P( /j aeus-Merzbacher disease.

( Figs 58.68, 58.69). In adrenoleucodvshophv maternal carriers have in some cases shown white-matter changes with mild neurol ogical symptoms. Similar findings have been reported in carriers of Pelizaeus-Merzbacher disease.

This is a wide group of metabolic disorders caused by enzyme defects resulting in accumulation of toxic metabolites in the body and in the brain, where they interfere with normal myelin synthesis or cause destruction of myelin sheaths. They present in infants and with improved methods of biochemical assay many different types have now been defined, some of which are listed in Box 58.5. Diagnosis is usually made biochemically but imaging may help by showing the degree of myelin damage or the effect of treatment on progression. I maging CT may show small low-density foci in the white matter but MRI shows the white-matter lesions best as small areas of high signal on T,-weighted studies, and lying subcortically or i n the centrum ovate. Sagittal T sections may demonstrate small nonl ow-density lesions in the corpus callosum. The findings are nonspecific. Phenylketonuria This is a congenital disorder of amino acid

metabolism due to absence of phenylalanine hydroxylase which hydrolyses phenylalaninc to tyrosine. Severe mental impairment tray ensue if the condition is unrecognised and untreated, and

Phenylketonuria Hyperphenylalaninaemia Tyrosinaemia Maple syrup urine disease Argininosuccinicaciduria Citrullinaemia Ornithine transcarbamylase deficiency Hyperargininaemia

Homocystinuria Cystathioninuria Hypermethioninaemia Propionic acidaemia Methylmalonic acidaemia Non-ketotic hyperglycinaemia Lactic acidaemia


MRI MRI is more sensitive than CT in the demonstration of white-matter disease. Changes can be diagnosed earlier, are shown more clearly, and appear more extensive. T 2 -weighted sequences produce a high signal from affected white matter

Fig. 58.69 (A) Axial T 2 and (B) T, coronal postcontrast images in a patient with Alexander's disease showing signal abnormality in the white matter, which has a predilection for the frontal lobes. There is some basal ganglia involvement and enhancement postcontrast.

(A) Axial T 2 -weighted MRI showing bilateral parietal periventricular signal abnormality in a patient with classical phenylketonuria. (B) Axial CT showing extensive very low density change in the cerebral white matter and thalami in maple syrup urine disease. Fig. 58.70

Previous Page

CONTENTS


1803

Binswanger's disease (subcortical arteriosclerotic encephalopathy) There is a frequent association between hypertension,

arteriosclerosis and dementia or other neuropsychiatric manifestations. The deep penetrating arteries of the brain are affected, l eading to ischaemic changes in the deep white matter, associated with generalised atrophy. Some authorities consider this to be a form of multi-infarct dementia, but others regard the changes as non-specific senile changes. CT There is generalised atrophy with multifocal or generalised l ow attenuation of the white matter, particularly around the ventricles (Fig. 58.72). There is no enhancement after contrast medium. Other findings reported in some cases include lacunar infarcts in the basal ganglia. However, similar low-attenuation changes may be seen in the brains of many elderly people who are normotensivc and without dementia or other neurological signs. Lacunar infarcts and atrophy are also non-specific. MRI MRI will demonstrate the white-matter changes more sensitively than CT and will also show the generalised atrophy. Periventricular lesions are well shown but it should be realised that these are not specific and similar lesions may be seen in some symptomless elderly patients. Fig. 58.71 (A) Multiple dilated perivascular spaces are present on the T 2 - weighted axial image in this patient with mucopolysaccharidoses. (B) Tz sagittal image through the craniocervical junction in Morquios' syn-

drome showing a small foramen magnum and some cervical cord impingement. There is odontoid hypoplasia and ligamentous and dural thickening contributing to the small foramen magnum. there are alterations in the hemisphere myelin resembling leucodystrophy. The white-matter changes have been characterised by MRI (Fig. 58.70). Mueopolysaccharidoses The skeletal abnormalities seen in these

l ysosomal storage disorders have been described in Chapter 35. Diagnosis is made by the characteristic urinary mucopolysaccharides combined with the clinical picture (Fig. 58.71).

Mitochondrial cytopathy This is a group of diseases that have in common a defect in mitochondria, with consequent deficiency of enzymes controlling oxidative phosphorylation and/or the respiratory chain. They frequently i nvolve muscle and may cause a proximal myopathy. Involvement of the CNS can result in very varied clinical presentations associated with elevation of the serum and CSF lactate (Figs 58.73, 58.74). The typical presentations include: I . Myoclonic epilepsy with ataxia 2. Ophthalmoplegia with retinal degeneration, which is often associated with heart block and elevated CSF protein 3. Stroke-like episodes, often associated with episodic vomiting and cortical blindness.


I maging The brain shows atrophy and hydrocephalus and in some cases megalencephaly. White-matter changes appear as diffuse low attenuation at CT or as high signal on T,-weighted MRI images.

Fig. 58.72 (A,B) CT of elderly demented patient. Cortical atrophy i s mild but there are extensive areas of patchy l ow density in the deep cerebral

white matter. Binswanger's disease.

The disease is transmitted through the maternal cytoplasm and is non-mendelian. The affected persons are often short in stature.

Fig. 58.73 (A) Axial T 2- and (B) coronal T 1 - weighted images showing extensive signal abnormality in the cortex and white matter of the right cerebral hemisphere in MELAS. There is some mass effect and note that the signal change involves all three vascular territories, although is largely posterior in distribution. There is also cerebellar atrophy.

Previous Page 1804

CONTENTS

Next Page


Partial seizures (localisation related, focal) Simple

• • • •

(consciousness unimpaired)

Motor Somatosensory Autonomic Psychic

Complex (impaired consciousness) • Beginning as simple • Complex at onset • Secondary generalised

Fig. 58.74 Bilateral mature cystic cleft-like lesions have occurred in the l entiform nuclei in this patient with a mitochondrial cytopathy. (A) Axial MR T2 W. (B) Coronal MR T 2 W.

Generalised seizures • Absence seizures • Myoclonic seizures • Clonic seizures • Tonic seizures • Tonic-clonic seizures • Atonic seizures

I nfantile and childhood epilepsies • Infantile spasm • Lennox-Gastaut syndrome • Cryptogenic myodonic epilepsy Fig. 58.75 Mitochondrial cytopathy. Four-year-old boy with ataxia, myoclonus and drowsiness. MRI (T 1 -weighted) reveals l oci of high signal in the lentiform nuclei and parietal white matter extending into posterior limbs of internal capsule.

Adolescent epilepsies • Early morning myoclonus with tonic-clonic seizures • Myoclonus simple absence Progressive myoclonic epilepsies • Lipidoses • Ceroid lipofuscinosis (Kufs' disease) • Lafora body disease • Sialidosis • Mitochondrial cytopathy • Baltic myoclonus


The histopathology includes subacute necrotising encephalomyelopathy associated with proliferation of capillaries and glia and with spongiform degeneration. The distribution, reflected in the MRI appearances, includes symmetric necrosis in the basal ganglia, spongiform change with increase in T, and T 2 i n the thalamus, brainstem and/or hemispheric white matter (Fig. 58.75). In the more chronic presentations such as Kearn-Sayer syndrorne there may be basal ganglia calcification associated with atrophy, which may be marked in the cerebellum and brainstem. The stroke-like episodes arc reflected in focal abnormalities closely resembling infarcts.

In the generalised epilepsics imaging is usually negative or shows only evidence of trauma due to the seizures; but if a causative focal lesion is found, the diagnosis will change from generalised to secondary generalised epilepsy. In patients with seizures of recent onset, imaging (either CT or MRI) is recommended in most cases: in developed countries, causative lesions will be found in only 5-10% (usually neoplasms); in some less developed countries it can be as high as 50% due to tuberculosis or cysticerosis. Malformations of cortical development, whether focal, lobar or generalised, are much more common in focal childhood epilepsies, and are rarely found with late-onset

seizures. Box 58.6 gives the classification of seizures by the Commission In patients with partial epilepsy of habitual type, the causative of Classification of the International League Against Epilepsy; Box 58.7 gives the classification of myoclonus and epilepsy lesions are shown by MRI in about 20% of cases overall, and in up to 60% in some types. The lesions found, in approximate order of syndromes. prevalence, are: Partial seizures are further subdivided into the suspected lobe or region of origin based on seizure semiology. Thus we have tempo• Hippocampal sclerosis (over 50% of the total) (Figs 58.76, ral lobe seizures (mesial or lateral temporal neocortex), frontal 58.77) (jacksonian, supplementary motor, premotor, cingulate, orbito• Benign neoplasms frontal) and parieto-occipital. There is also an important distinction • Cortical scars • Vascular lesions (Fig. 58.78) between habitual and recent onset seizures, and symptomatic epilepsy where seizures occur as part of other diseases. • Malformation of cortical development.

CONTENTS

Previous Page

Next Page

Fig. 58.76 (A) Diagram of hippocampus and its relationships. A. Lateral aspect of left temporal lobe. B-E Coronal sections from before backward at levels shown in A. Ch F = choroidal fissure; F = fimbria; MD = margo denticulatus (dentate gyrus); HFi = hippocampal fissure; TH = temporal horn; CAI = four regions within hippocampus itself; PHG = parahippocampal gyrus; UN = uncus; EC = entorhinal cortex; UG = uncinate gyrus; SLG = semilunar gyrus (part of amygdala); GA = gyrus ambiens; HF = hippocampal formation; OF = uncal fissure; CS = collateral sulcus; UN = uncal notch; SUB = subiculum. (B) Coronal MRI section through left medial temporal lobe at level of C above. T,-weighted.

appropriate for investigating habitual epilepsy in general. HS usually seems unilateral on imaging, so diagnosis is affected by comparing the two sides. Measurements of volume and relaxometry are not usually necessary but may aid detection of bilateral abnormalities. CT does not detect HS at all.


Neoplasms These are benign and are usually centred on cerebral

cortex of usually temporal or frontal lobes. They are often circumscribed. The commonest in this context is dysplastic neuroepithelial tumour (DNT), described elsewhere.

Cortical scars These may be small or extensive. They usually n hippocampal subfields CAI and CA4 manifests on MRI as (I) involve frontal lobes. They result from trauma, infarction or infeci tions, in descending order of frequency. reduction in volume, and 2 in creased signal on T 2 - weighted i mages. These features are confined to the hippocampus. Vascular lesions These are usually cavernous and sometimes Although many extrahippocampal abnormalities have been arteriovenous malformations involving cortex of usually frontal described on MR1, none are reliable in isolation and the signifior temporal lobes. Their appearances on MRI and CT are cance and even the existence of some are doubtful. The optimal described elsewhere. MRI protocol for detection of HS is as follows. Dual-echo fast spin-echo acquisition in the coronal plane, slice Malformations of cortical development (MCD) These can take thickness 5 mm, and a T,-weighted volumetric acquisition also in many forms, and appearances on MRI and CT are described elsethe corona) plane, with slice thickness 1.5 min or less. The dual where. An important one in the context of partial epilepsy is focal echo acquisition is designed to maximise contrast resolution for cortical dysplasia, which manifests on MRI as focal cortical detection of T,-dependent signal change, and the volumetric acquithickening, blurring of the underlying grey-white matter interface sition for volume loss. We usually also add a coronal FLAIR acquiand variable signal change in the underlying white matter. Most sition: it sometimes aids recognition of snmatl cortical abnormalities i nvolve frontal or temporal lobes; they usually are 1-2 cam in size, that might otherwise be overlooked. This protocol is therefore but can be more extensive. They are indistinguishable from cortiHippocampal sclerosis (HS) Neurone loss and gliosis maximal

O

Previous Page

1806

CONTENTS

Next Page



Fig. 58.77 (A) Corona) MRI sections, T2 - weighted, through hippocampal bodies, corresponding to level E of Fig. 58.76A. The left hippocampus is smaller than the right and shows higher signal. Hippocampal sclerosis. (B) Coronal MRI (FLAIR) image at level of hippocampal body accentuating high signal in left hippocampus. (C) T,-weighted image at similar level from a 3D accumulation, showing the left hippocampus to be much smaller than the right.

cal tubers in tuberous sclerosis. A special case is the hypothalamic hamartoma, found very frequently in patients with gelastic seizures.

refractory to drug treatment may benefit from surgery designed to cure or ameliorate the epilepsy itself. Seizure outcomes are best after resection of hippocampal sclerosis, about 80% of carefully selected patients can expect to become seizure-free and off drugs.

Neurosurgery and epilepsy

Focal resections of neocortical lesions usually result in only 50% or l ess being seizure-free, and resection of histologically normal tissue, less than 20%. Non-resective surgery is used sometimes to

Symptomatic epilepsy may require neurosurgery to treat progressive pathology for its own sake. Patients with habitual epilepsy

Previous Page

CONTENTS

Next Page

In each examination a series of closely spaced scans is made in both the coronal and sagittal planes by altering the angulation of the transducer to the fontanclle. The aim is to include as much as possible of the brain on each scan. A more limited access may also be possible through the posterior fontanelle and through the sutures. Occasionally it may be of value to image directly through the skull vault. The larger the fontanclle and therefore the freer the access, the better will he the images. As the fontanelle begins to decrease in size the access becomes more limited and the visualisation of the brain less complete. Ultrasound studies of the brain are most valuable in the first 6 months of life. Indications The two most common abnormalities of the neonatal

Fig. 58.78 Coronal MRI (T 2 -weighted). Abnormal vessels are shown in the right medial temporal lobe involving the hippocampus. Angioma.

ameliorate devastating epilepsy: corpus callosectomy, functional hemipherectomy and multiple subpial cortical transection. Diagnostic neurosurgical procedures include depth electrode and surface grids placement, usually designed to test a hypothesis concerning site of seizure onset.

Functional imaging in epilepsy PET and especially SPECT have a long history of clinical application in epilepsy. However, in recent years their role has diminished because of the much greater specificity of MRI. The reliability of l ocalisations achieved by PET and SPECT in MRI-negative cases i s questionable; depth electrodes or surface grids are necessary and they often fail to confirm localisations in such cases. Some units still practise what has been termed multimodality convergence in presurgical work-up: the results from MRI, SPECT (and PET), neuropsychology and videotefemetry are compared to clinical features, and surgery is considered in the context of how well they converge. I n most units, however, less than 1 0% of MRI-negative cases come to surgery. MRI therefore is paramount.

brain requiring confirmation or exclusion by imaging techniques are hydrocephalus and intracranial haemorrhage. These are the major indications for an ultrasound scan in the neonatal period. It i s now widely accepted that ultrasound reliably and accurately demonstrates the size of the lateral ventricles and also sensitively detects the presence of intracerebral haemorrhage. Hydrocephalus may result from congenital malformation. Haemorrhage into the ventricles or subarachnoid space may also cause hydrocephalus. It may also develop following intrauterine infection such as toxoplasmosis. The definitive surgical management of hydrocephalus i s to perform a shunt procedure of some type. The success of the procedure and the early complications such as shunt failure may also be followed by ultrasound. The incidence of intracerebral haemorrhage is considerably i ncreased in premature infants. With the growth of high-risk nurseries, allowing the survival of very premature infants, this complication is increasingly frequently seen and needs evaluation. The diagnosis of intracerebral and intraventricular haemorrhage may be made with ultrasound. Meningitis is a comparatively common infection in some parts of the world. In young infants the progress of the disease and the results of medical and surgical treatment may be monitored effectively and simply with ultrasound. The common complications are


Technique The ideal equipment for the examination of the infant brain is a high-resolution sector real-time scanner, fitted with a 5 MHz (or higher frequency) transducer. The scan head should i deally be as small and manoeuvreable as possible. The sector field of view is particularly well suited for showing the maximum area of information through a small acoustic window. This type of equipment is usually mobile. It is taken to the infant and the transducer applied to the head via a port-hole in the incubator. In this way the baby is almost undisturbed by the examination. If this type of equipment is not available it is possible and reasonable to examine the brain with either conventional static or linear-array real-time equipment, provided that a suitable high-frequency transducer is available (5 MHz). The linear-array transducer has a long rectangular scan head and field of view. Provided the acoustic window is suitable, it is increasingly used to optimise detail.

ventricular enlargement, subdural c%%usions and cerebral oedema,

all of which may be well shown with trans fontaneIlar scanning.

Ultrasound is able to demonstrate the normal brain structure in considerable detail. The coronal sections are particularly valuable for showing focal abnormalities as advantage may be taken of the brain's symmetry. A series of landmarks in both coronal and sagittal planes has been described which allows recognition of the position at which the scan has been taken, and subsequently the acquisition of standard reproducible sections. The position of these sections is illustrated diagrammatically in Figure 58.79. Within the brain, structures containing CSF, such as the ventricles and cisterns, appear anechoic; normal brain tissue generates low to mid-level echoes; higher-level echoes are generated from the cerebellum, the sulci and vascular structures. It is likely that a great deal of the internal echo reflection in the brain is due to the collagen content of the pia, ependyma and the vascular sheaths. One of the most echogenic intracerebral structures seen with ultrasound is choroid plexus. This stretches around the floor of the lateral ventricles from the foramen of Monro into the temporal horns, enlarging at the level of the glomus in the trigone before sweeping down into the temporal horns. The bones of the vault also produce highly echogenic reflections and so may he useful as landmarks.

Previous Page 1808

CONTENTS

Next Page


Fig. 58.79 Diagram showing three standard sagittal sections (A) and six coronal sections (B).

In the first coronal plane the transducer is angled anteriorly in the fetus. It is also present in about half of all full-term infants. It is fontanelle, producing a section through the frontal lobes, seen sepaseen more commonly in premature infants. Its incidence decreases rated centrally by the interhemispheric fissures and the falx. The sharply with age, and by 6 months compares with the low percentl ateral ventricles will not be seen in this section. The far field landage reported in adults. The roof of the lateral ventricle is formed by mark is the high-amplitude echoes from the orbital roofs and the the corpus callosum, which is echogenic, and the floor by the head central echoes from the ethmoid complex extending downward to a of the caudate nucleus, with lower-level echoes. l ower level. With the transducer angled vertically the fourth coronal plane As the transducer is angled slightly back to the second position passes through the third ventricle which is not usually resolved in the far field landmark changes so that the lesser wings of the sphethis plane when of normal size. A very prominent far field landnoid, with the greater wings of the sphenoid behind, are seen mark is formed by the paired C-shaped echoes from the paraforming the anterior floor of the temporal fossa. The anterior horns hippocampal gyri and medial surface of the temporal lobes of the lateral ventricles may be seen in this section as small slit-like (Fig. 58.81). Part of the sylvian fissure will again be seen in this spaces either side of the central midline echo. section. The bodies of the lateral ventricles now lie more horizonAngling the transducer slightly further back so that it is almost tally and reflections from the choroid plexus may be seen in the vertical produces the third coronal section where the most promifloor of the ventricles. nent landmarks are the sylvian fissures. These are seen just behind Angling the transducer posteriorly on the fontanelle leads to the the lesser wings of the sphenoid and run laterally and outward, fifth coronal section where the prominent landmark is the highly becoming Y-shaped between the temporal and frontal lobes. Within echogenic tentorium and cerebellum shaped like an inverted V. this echo complex, pulsation from branches of the middle cerebral Slight further angulation will bring into view the echogenic diverartery will be readily seen. This section lies just anterior to the gent bands characteristic of the glomus of the choroid plexus. These foramen of Monro, and between the lateral ventricles the echoare the landmarks for the sixth section (Fig. 58.82). In the normal free box-like structure of the cavun septi pellucidi may be seen infant the choroid plexus will always be seen in this section, (Fig. 58.80). This is present during fetal life where it is used as a although the fluid-filled ventricles will not always be separately disl andmark to optimise a measurement of biparietal diameter in the tinguished around the choroid.


Fig. 58.80 A coronal section in the third position showing the bodies of the lateral ventricles (V) and the sylvian fissures (SF). Lying between the l ateral ventricles is the cavum septi pellucidi (CSP).

Fig. 58.81 A coronal section in the fourth position showing the prominent landmark of the parahippocampal gyri (HG). V = lateral ventricles.

Previous Page

CONTENTS

Next Page

Rotating the transducer through 90° into the sagittal plane, the first sagittal section is taken in the midline. The sulcal detail is usually prominent and the cingulate sulcus containing the branches of the anterior cerebral artery can be identified. The echogenic cerebellum is seen posteriorly, and anterior to this the fourth ventricle can be recognised. Above this the third ventricle and the massa intermedia are demonstrated, and often the aqueduct of Sylvius is outlined. The clivus forms the far field landmark (Fig. 58.83). The second parasagittal section is angled about 15' away from the midline and shows almost the full sweep of the lateral ventricle containing the echogenic choroid plexus. Within the sweep of the ventricles the rounded mass is formed by the thalamus and caudate nucleus (Fig. 58.84). The far field landmark is the floor of the middle fossa. The third parasagittal section is angled outward about 30°, lateral to the ventricle and through the sylvian fissure. Fig. 58.82 A coronal section in the sixth position showing the characteristic echogenic choroid plexus (CP).

The lateral ventricles in the normal neonate may be small and difficult to define accurately. The mean width of the lateral ventricle in the full-term infant is 12 mm and in the 30-week premature infant 9 min, measured at the level of the body of the lateral ventricle (coronal section 3). In the past the measurement of the width of the l ateral ventricle has usually been expressed as a ratio of the distance to the skull vault. In axial CT scanning, deviation of the medial walls of the lateral ventricle is an early sign of dilatation but the normal ratio of up to 30% appears too large when measured by ultrasound in the coronal plane and the ratio may not be valid in this circumstance. The trigone and occipital poles of the ventricles are generally the l argest part of the entire ventricular system, but no convenient landmark exists to monitor changes in size accurately and in any event there may be considerable variation in the shape of these structures. However, early dilatation is often most prominently featured in the trigone, occipital and temporal horns, possibly reflecting the lesser compliance of the supporting structures in this region. Ultrasound has proved to be accurate and reliable in detecting and grading the severity of hydrocephalus and is ideally suited to following progress. Figure 58.85 is a coronal section at the level of the bodies of the lateral ventricles showing the appearances of established hydrocephalus. Further evaluation of hydrocephalus to establish a precise aetiology may require diagnostic procedures in addition to ultrasound. However, in certain instances, ultrasound may provide most or all of the answers. In addition to the diagnosis and follow-up of hydrocephalus, ultrasound may also be helpful in following shunt procedures, particularly in the evaluation of complications. Besides hydrocephalus other cystic lesions within the brain may show well. In hvdrcniencephaly no brain parenchyma is present above the level of the midbrain and no cortical mantle is seen. The Dandy-W alker syndrome has a characteristic appearance where cystic dilatation of the fourth ventricle can be recognised ( Fig. 58.86). This condition must, however, be distinguished from a retrocerebellar arachnoid cyst which may produce a similar appearance. In the latter instance, however, a normal fourth ventricle is i dentified compressed anteriorly by the cyst (Fig. 58.87). A wide variety of other cystic lesions has been described, such as holoprosencephaly (Fig. 58.88).


Fig. 58.83 A sagittal section in the midline showing the echogenic cerebellum and the fourth ventricle (arrow) posteriorly.

Fig. 58.84 An angled sagittal section showing the full sweep of one l ateral ventricle around the caudate nucleus and thalamus.

Previous Page 181 0

CONTENTS

Next Page


Fig. 58.85 (A) A coronal section showing the typical appearance of hydrocephalus involving the lateral and third ventricles. The section is through the foramen of Monro. Echogenic haemorrhage is within the ventricle. (B) A ventricular shunt in position more posteriorly.

Fig. 58.86 (A) A coronal section. (B) A sagittal section showing the characteristic findings of the DandyWalker syndrome. There is cystic dilatation of the fourth ventricle filling the posterior fossa. A shunt is in place.


Fig. 58.87 (A) A coronal section. (B) A sagittal section showing the findings of a retrocerebellar arachnoid cyst for comparison. Note that in the sagittal section the cerebellum (Cb) and the fourth ventricle (iv) can be seen compressed forward by the cyst.

The germinal matrix i s a neural vascular tissue in the fetus which is normally involuted by term. It is situated subependymally in the ventricles and is prominent in the groove between caudate nucleus and thalamus. This is a frequent site for haemorrhage in premature infants (Fig. 58.89).

The vascular choroid is also an important site for cranial haemorrhage in infants. The shape of the choroid plexus as it surrounds the caudate nucleus and thalamus is fairly constant so that any i rregular increase in size is suspicious of haemorrhage (Fig. 58.90). The symmetry of the two sides may also be of value in detecting abnormality.

Previous Page

CONTENTS

Coronal section showing the single fused ventricle typical of holoprosencephaly.

Next Page

Fig. 58.88

A sagittal section showing the characteristic appearances of advanced periventricular leucomalacia (PML) as periventricular cystic spaces. Fig. 58.91

58.92 Coronal section showing extensive reflective intraparenchymal haemorrhage. Fig.


A coronal section in a premature infant showing a typical reflective haemorrhage (H) from the germinal matrix. A mass effect from the haemorrhage is distorting and elevating the lateral ventricle on this side. Fig. 58.89

Fig. 58.90 A posterior coronal section showing asymmetric enlargement of one choroid plexus (H) typical of haemorrhage.

White-matter haemorrhage occurs in both preterm and full-term i nfants. In the former it is associated with perii'entricular leucomalacia (Fig. 58.91). This lesion is attributed to underperfusion of the

boundary zones between different arterial territories within the

periventricular white matter, which may develop during episodes of hypotension. Ultrasound may show a periventricular increase in echo amplitudes, with follow-up studies showing formation of cystic spaces. In term neonates, intraparenchymal haemorrhage may be diffuse and petechial in the cortical areas. The diagnosis of intracerebral, periventricular and intraventricul ar haemorrhages may be made reliably using ultrasound where the haemorrhage shows as a brightly reflective area (Fig. 58.92). This is mainly due to the fibrin mesh formed, which produces multiple reflecting interfaces for the ultrasound beam. Subarachnoid haemorrhage is not reliably demonstrated using ultrasound, but subdural haemorrhage is usually shown as a crescentic echo-poor region separating brain from the skull vault (Fig. 58.93). Haemorrhage, particularly in premature infants, may rupture into the lateral ventricles. This may enlarge the ventricles, either due to brain tissue loss and atrophy or to hydrocephalus due to CSF flow obstruction. The combination of both may occur, and monitoring to show progress or stabilisation is important in these circumstances. Trauma to the infant brain due to birth or accidental injury produces a combination of the types of haemorrhagic or ischaemic change already described. Non-accidental injury to the brain may produce pathognomonic shearing injuries at the grey-white matter i nterface. This appears typically as small linear slit-like cavities, optimally demonstrated with the linear-array transducers.

Previous Page 1812

CONTENTS

Next Page


Fig. 58.93 A transcranial section showing a large crescentic subdural haematoma.

Numerous different approaches have been employed for brain i maging using radionuclides. Anatomical studies using pertechnetate and other extracellular fluid tracers have been superseded by CT and MRI. Techniques for mapping regional cerebral blood flow have i mportant clinical applications, which are described below. Specific agents targeted at neurotransmitter sites and other receptors are used extensively in research and are beginning to enter the clinical arena, particularly in relation to Parkinson's disease. Tracer studies of CSF flow are now rarely required, but may occasionally be helpful.

neuronal activity. Inert gases arc effective markers of' regional blood flow, but have technical drawbacks. Krypton-81m with its l ow solubility in blood and its short half-life of 14 s, allows continuous measurement of regional blood flow, but the same properties require its delivery by intracarotid infusion. A rebreathi ng method using xenon-133 (half-life 5.3 days) offers a noni nvasive method for assessment of regional blood flow, but the greater solubility and long half-life of xenon allow recirculated activity in the scalp, which causes image noise. A more practical approach emerged with the development of tracers based on small lipophilic molecules which readily cross the blood-brain barrier and arc retained within brain cells. The first of these to be widely used was iodoamphetamine labelled with iodine-123 (]"I-IMP), but for most clinical applications agents labelled with "°'Tc and commercially available in kit form have major practical advantages. Two of these agents are currently widely availableexametazime (`°I"Tc-HMPAO) and ethylcysteinedimer (`fI"TcECD). HMPAO has a high extraction efficiency by brain tissue. About 5% of the injected dose is taken up by brain tissue and only a small proportion of this is washed out in the first few hours after i njection, giving an ample time window for imaging. ECD has some practical advantages in that it is more stable in vitro and its extraction efficiency is slightly better than HMPAO, but the kinetics of ECD in abnormal brain are as yet less well established than those of HMPAO.

` 91 "Tc-HMPAO (exametazime) requires no patient preparation. The typical activity of 500 MBq is injected intravenously with the patient in a quiet stable environment. Images may be obtained from 20 min to several hours after injection because the tracer distribution in the brain is stable during this time. Dedicated multidetector i maging devices will produce the best resolution, but diagnostic results may also be obtained using a rotating gamma camera, taking care to position the camera head as close as possible to the patient during rotation. Volumetric data are displayed in standardised axial. coronal and sagittal planes, and colour displays may be used to i ncrease the conspicuity of low contrast features in the reconstrutted images.


Dedicated PET imagers provide the best spatial resolution for scintigraphic studies of the brain. Proximity to a cyclotron is essential for receptor studies using short-lived tracers labelled with carbon-11, or flow studies using oxygen-I5 labelled water as a marker. Although much of the research and development work in

brain metabolism has involved the use of PET tracers for blood flow, glucose metabolism using '"FDG, neurotransmitter and other receptor activity, these have not yet been widely translated into routine clinical applications. A brief review of the current and potential clinical applications of PET is given in Chapter 59. SPECT using a conventional single or multiheaded gamma camera mounted on a rotating gantry offers tomographic imaging of useful diagnostic quality. Improved resolution is obtained by using a dedicated single-photon brain imaging array, with the disadvantage that the device is unsuitable for other scintigraphic applications.

Because brain tissue has no mechanism for storage of oxygen or glucose, local blood flow in the normal brain correlates well with

Absolute uptake of exametazime is affected by numerous factors which vary between individual subjects, so interpretation generally relies upon the use of specific reference areas in the image to give an indication of normal flow. For most purposes, the cerebell ar hemispheres may be taken as normal flow indicators. Cortical uptake is normally fairly homogeneous and white-matter structures are identified as areas of low uptake in the subcortical. periventricular and capsular areas. Increased separation between the heads of the caudate nuclei indicate dilatation of the lateral ventricles. Sagittal slices are helpful in comparing regional uptake i n frontal, parietal and occipital lobes with the cerebellum, although cortical atrophy may lead to partial volume effects causing spurious hypoperfusion on sagittal images, particularly in the frontal region. Coronal images overcome this problem and should also demonstrate the mesial and lateral components of the temporal cortex.

Previous Page

CONTENTS

I n acute stroke, symptoms and signs of brain injury are associated with changes in regional cerebral blood flow; however, the consequent structural abnormalities may take hours or even days to become visible on CT or MRI. For example, 8 h after the clinical onset of acute stroke, only about 20% of patients will show CT abnormalities in the affected area, whereas 90% will show local blood flow abnormalities on SPECT imaging. Although the dingnosis of stroke is essentially clinical, early SPECT imaging may be used to establish the presence and extent of the blood flow deficit. The magnitude of the deficit, when SPECT is performed in the first few hours after onset of a stroke, is a fairly good indicator of the eventual outcome-the more extensive the initial deficit, the worse the prognosis for recovery. A further guide to prognosis may be obtained by repeating the SPECT study about 24 h after onset, an early recovery in the perfusion deficit being a good prognostic i ndicator. After the initial ischaemic insult, cell death in the infarcted area effectively reduces the oxygen and glucose requirement so that a continuing deficit in blood flow may not result in further neuronal damage. However, a common finding in the subacute phase of stroke is that the blood flow deficit shown on SPECT is more extensive than the structural changes of infarction which appear on CT. This discrepant area, described as the SPECT 'penumbra', may i ndicate an area of ischaemic but viable tissue which is potentially recoverable. The appearance of 'luxury perfusion' commonly occurs 2-28 days after the onset of a stroke. This is a variety of reactive hyperaemia with increased flow through the infarcted area. Experimental work suggests that much of this flow is non-nutritive, so that the normal consistent relationship between How and the uptake of oxygen and glucose in the local tissues is lost. However, the uptake of HMPAO is unchanged or in some cases more avid than normal, so SPECT studies carried out during this period arc likely to underestimate the true ischaemic deficit. The correlation of l ong-term abnormalities on SPECT with chronic infarcts shown on CT is fairly good, except that the resolution of SPECT is i nadequate for small cortical and lacunar infarcts. Crossed cerebellar diaschisis describes the phenomenon in which cerebral infarction leads to diminished perfusion in the contralateral cerebellar hemisphere, due to loss of linkages which normally stimulate neuronal activity in the cerebellum. This deficit is often seen when there is extensive cortical infarction, and it persists throughout the period of luxury perfusion in the subacute phase. A similar reactive fall in perfusion may occur in the cortical areas which overlie deep-seated infarcts. Since cerebellar blood flow is usually used as a reference point for assessing regional cortical flow, it is i mportant to bear in mind the possibility of unilateral perfusion loss when cerebellar activity is asymmetric. As a prelude to surgery or angiographic intervention, SPECT i maging may be used to assess the degree and extent of functional deficit which is associated with arterial stenosis or occlusion. An alternative approach uses acetazolamide stress testing to measure perfusion reserve. Acetazolamide (Diamox) is a carbonic anhydrase i nhibitor which mimics the effect on the cerebral circulation of i ncreasing the concentration of circulating carbon dioxide. An intravenous injection of I g of acetazolamide normally causes an i ncrease in cerebral blood flow that reaches a peak about 20-30 min

Next Page

later and then declines gradually over the next 1-2 h. The degree of augmentation of flow after Diamox, described as `cerebral perfusion reserve', may be as much as 40% increase over resting flow in normal subjects. The degree of reserve in patients is often unrelated to the severity of vascular stenoses, particularly in the extracranial arteries. Loss of perfusion reserve carries a poor prognosis, with a high risk of subsequent stroke, and this assessment may be used as a selection criterion for patients who arc candidates for carotid artery surgery or stenting. Pre- and post-Diamox studies can he carried out either on the same day or on consecutive days, and subtraction images can be used to highlight the areas of abnormal reserve. Diamox should not be given to patients who suffer from migraine, since it may induce severe headache. In summary, the value of rCBF SPECT imaging in suspected stroke is: 1. For triage-to distinguish stroke from TIA in the acute stage. 2. To show the extent and distribution of abnormality, which will help to identify the likely site of arterial occlusion (Fig. 58.94). 3. To demonstrate areas at risk of further damage, and those areas with possibility of recovery (Fig. 58.95). 4. To monitor thrombolytic therapy. Dementia A gradual and global reduction in cortical blood flow is a feature of normal ageing in asymptomatic subjects. Specific variations in rCBF are associated with different types of dementia. Alzheimer's disease accounts for about 50% of patients with progressive dementia. Pathology shows loss of neurones, fibrillary tangles, amyloid deposition. granulovacuolar degeneration and neuritic plaques. Clinical criteria for Alzheimer's disease (AD) are helpful, but are relatively non-specific, so that autopsy studies show the typical pathological changes of AD in only 50% of patients with

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ 99

"Tc-exametazime brain SPECT: axial (A), Coronal (B), right parasagittal (C) and left parasagittal (D) sections in a patient with massive i nfarction of the right middle cerebral artery territory. Note severe ischaemia o5 the Srontai, tempura% and panetai cortex and also oI the basaS ganglia on the right.

Fig. 58.94

Previous Page

1814

CONTENTS

Next Page


99 ,

" Tc-exametazime brain SPECT: axial, coronal and right parasagittal images showing very extensive perfusion deficits during the acute ischaemic phase (top row) and substantial improvement several months later after clinical recovery (bottom row). Fig. 58.95

clinically `probable' AD. The addition of SPECT rCBF studies i mproves the accuracy of diagnosis to the extent that 80% of patients with clinically probable AD and positive brain SPECT are found to have typical pathological changes at autopsy. The typical defect of AD is a reduction of flow in the posterior parietal and temporal areas (Figs 58.96, 58.97). This pattern of abnormality is highly specific in discriminating between AD and normals, and a little less specific in distinguishing between AD and other demential, but still highly characteristic. The blood flow changes may precede the onset of clinical symptoms. Brain atrophy is also a feature of AD, but is non-specific. The blood flow response to Diamox is relatively normal in AD. Flow to the basal ganglia is normal in most patients with AD, while in Parkinson's disease cortical flow is well maintained. However, there is a recognised subgroup of patients who present with a combination of Parkinson's disease and AD, and these show the typical cortical abnormalities of AD. An important differential diagnosis to be made by SPECT is i n distinguishing AD from similar clinical presentations in occa-

19 mTc-exametazime brain SPECT: axial (top row) and left and right parasagittal (bottom row) sections in a patient with suspected Alzheimer's disease. The distribution is primarily posterior but quite asymmetric, unlike Fig. 58.96.

Fig. 58.97

sional patients with multiple sclerosis, and in those with multiple drug and alcohol abuse. Lewy body disease (LBD) is the cause of about 20% of dementias in patients over 70. Lewy bodies are intracellular inclusions found in the basal ganglia and substantia nigra in Parkinson's disease, but when they are also found in the cerebral cortex they are associated with progressive dementia. The perfusion abnormality is less specific than that of AD, but broadly similar in distribution, although the calcarine and occipital cortices are often also affected.

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ 99 mTc-exametazime brain SPECT: axial (top row) and left and right parasagittal (bottom row) sections in a patient with suspected Alzheimer's disease. Note the posterior distribution of the major ischaemic areas.

Fig. 58.96

91 "Tc-exametazime brain SPECT: axial (top row) and left and right parasagittal (bottom row) sections in a patient with dementia of frontal lobe type. Note the ischaemic lesions are predominantly in the frontal lobes on both sides.

Fig. 58.98

Previous Page

CONTENTS

99

-Tc-exametazime brain SPECT: axial (top row) and left and right parasagittal (bottom row) sections in a patient with dementia of frontal lobe type. Note the marked frontal and temporal ischaemia with central atrophy as well.

Fig. 58.99

Progression is typically faster than with AD. In addition, the basal ganglia are typically abnormal in LBD, but normal in AD. Patients with LBD, who show the typical pattern of bilateral frontal, temporal, parietal and occipital abnormalities of flow, may also have a parkinsonian presentation, and in other cases LBD and AD may coexist in the same patient. Frontal lobe and frontotemporal dementias, i ncluding Pick's disease, account for about 15% of demented patients. SPECT studies show symmetrical reduction in rCBF in both frontal lobes i nitially, and in later stages both temporal lobes are affected as well (Figs 58.98, 58.99). In multiinfarct dementia ( MID) flow abnormalities are related to multiple episodes of cortical infarction. The distribution of abnormalities is sporadic but tends to be related to the major arterial territories, rather than the lobar distributions seen i n AD and the other primary neuronal dementias. Abnormalities are often asymmetrical, and may involve frontal, temporal and parietal l obes, usually with multiple areas of diminished flow. The uptake of CBF agents into white matter is normally about one-third of that in grey matter. Focal white-matter infarcts (Binswanger's disease) may show no abnormality on SPECT studies in the early stages, but more advanced disease can be detected as diffuse or focal loss of white-matter uptake, often associated with cortical infarcts and atrophy. In summary, the aims of rCBF imaging in dementia are:

Next Page

the mesial cortex-is successful in eliminating seizures in over 90% of cases where the abnormal focus can be identified preoperatively. Results are much worse if the focus is not identified. The accuracy of external EEG alone is about 25 (/( and, although intracranial EEG gives a much better accuracy of 85-90%, it is an i nvasive and costly procedure. Mesial temporal sclerosis shown on MRI is often associated with the abnormal focus, but the sensitivity of about 60%- is still less than desirable for selecting patients for surgery. In focal epilepsy, the affected area of the brain shows abnormal blood flow. During focal fits, the flow to the abnormal cortex is sharply increased, and it then falls in the immediate postictal period. In about half of patients, blood flow to the abnormal focus i s lower than normal in the resting (interictal) phase. A recent metaanalysis of SPECT imaging in focal epilepsy suggested that the sensitivity for detecting the abnormal focus was 97%In when the SPECT was obtained during seizure, 75% for examinations clone in the postictal stage, and 44%- for interictal studies. Those patients in whom flow is normal during the ictal or immediate postictal phases will show a marked hypoperfusion in the interictal phase, so although ictal studies are the most sensitive, interictal studies are also worthwhile. Ictal studies require patients to he under continuous observation, with a tracer dose ready prepared and available for i mmediate injection when a seizure occurs. Since the tracer is fixed within the first few minutes after injection, imaging can he delayed until the patient has been stabilised. The pattern of abnormality seen in CBF during seizures may vary from day to day even in the same patient, so multiple studies are worthwhile. Patients with bilateral foci are less likely to benefit from surgery. Chronic motor seizures are usually associated with a frontal cortex focus, again best identified in ictal studies. About one-third of patients with refractory temporal lobe seizures over a long period will develop diminished flow in the cerebellum. Although the main value of radionuclide studies is to identify a temporal lobe focus as a prelude to surgery, SPECT may also be used to exclude from surgery those psychiatric patients with nonepileptic seizures in whom perfusion remains normal throughout.


1. To distinguish the different aetiologies of dementia, as treatment and prognosis arc both influenced by the cause of the disease. 2. To detect disease as early as possible. as treatment is more likely to be effective than in the later stages of progression. 3. To show the extent and severity of the lesion and to indicate prognosis. 4. In follow-up, to monitor the progress of disease and the effect of treatment.

Epilepsy Surgery for intractable focal epilepsy-usually excision of the anterior two-thirds of the temporal lobe and in some cases the whole of

Delayed neurological sequelac of trauma after head injury are characteristically associated with blood flow abnormalities in the acute stage. SPECT studies in the early period after head injury may give a useful prognostic indicator, as those with normal SPECT are unlikely to develop late scquelae.

Several tracers have been developed for investigating the pathophysiology of movement disorders and their response to drug therapy. The preferred agent for demonstrating the presynaptic dopamine transporter is 1 23 1-FP-CIT (ioflupane). '- 3 1-beta-CIT has si milar properties but its uptake into the target sites is slower, so i maging has to he done 18-24 h after injection, while with '- 3 1-FPCIT images can be obtained on the same day as injection. For i nvestigation of the postsynaptic dopamine receptors, the favoured agent is 1 '3 1-iodobenzamide (IBZM).

Previous Page 1816

CONTENTS

Next Page


normal controls, i.e. 1.5-2 times the uptake in the rest of the brain. Initial changes in IPD occur in the putamen and later progress to the caudate. The abnormality is often bilateral but also typically asymmetrical, being worse on the clinically more affected side. Similar findings to those in IPD are seen in patients with progressive supranuclear palsy and multiple system atrophy. Asymmetry is more likely in IPD but there is too much overlap to make a distinction i n individual cases. Uptake of dopamine transporter tracers in the striatum is normal in patients with drug-induced parkinsonism and in Dopamine transport imaging: interpretation Normal subjects those with essential tremor. Patients with `vascular Parkinson's show a crescent-shaped area of specific uptake in the corpus disease' show an overlap between normal and IPD appearances, but in striatum, a lesser degree of uptake in the substantia nigra and these cases deep infarcts may be visible on CT. little or none in white matter. The specific uptake in the striaDopamine receptor levels normally decline with age, so in poputum is normally about four times that of the rest of the lation controls the reduction in uptake of tracers in the corpus brain. declines by about 2-3% per year. Consecutive studies in IPD show

Dopamine transport imaging: technique Uptake of these agents is related to the presynaptic dopamine transporter activity. Patient preparation includes 5 days of thyroid blockade with potassium iodide. Dopaminergic drugs need not be stopped but tricyclic antidepressants may interfere with the technique and should be withheld--prior to the test. High-resolution SPELT images are obtained 1 3 3-6 h after intravenous injection of 150-185 mBq of '- I-FP-C1T, or 1 8-24 h after a similar activity of 1 23 1-beta-LIT.

Application: the investigation of movement disorders

The

diagnosis of idiopathic Parkinson's disease is confirmed in only about one-half of patients with the initial clinical diagnosis of Parkinson's. About one-quarter of presenting patients have parkinsonian syndromes, which may be drug-induced or associated with other degenerative disorders including Huntington's chorea, Wilson's disease, progressive supranuclear palsy and multiple system atrophy. In the remainder of patients presenting with Parkinson-like symptoms, the true diagnosis may be essential tremor, early dementia or `gait apraxia', all conditions that do not benefit from anti-Parkinson medication. SPELT imaging of the dopamine transporter system helps to distinguish between the various neuropathologies that may be associated with parkinsonian clinical presentation (Fig. 58.100). Idiopathic Parkinson's disease (IPD) is characterised by loss of dopamine-containing cells in the substantia nigra. SPELT i maging shows reduced uptake in the corpus striatum, the degree of which is proportional to the severity of clinical manifestations. A qualitative grading system may be used to indicate the degree of abnormality:

a much more rapid decline of 6 10% per year. Consecutive studies in patients with essential tremor and in parkinsonian syndromes associated with cerebrovascular disease, drug therapy, previous trauma and psychogenic origins typically show a normal rate of decline of dopamine receptors. In summary, the objectives of clinical SPELT imaging for movement disorders are: 1. To differentiate IPD from essential tremor and other mimics of IPD. 2. To establish the diagnosis of IPD as early as possible (images may be abnormal 3 years or more before the onset of clinical symptoms). 3. To monitor disease evolution, particularly the effect of neurorescue' drugs in slowing or arresting the progression of the disease. 1 23

The postsynaptic D2-receptor ligand iodobenzarnide ( IBZM) may be useful in some clinical situations but its role has not yet been firmly established. Initial studies suggest that uptake of this agent is normal or increased in IPD, particularly in the early stages


Grade I-reduced uptake in putamen, caudate normal. Grade 2 -reduced uptake in both putamen and caudate. Grade 3 -reduced uptake in caudate, zero uptake in putamen.

Studies on a mixed group of patients with Parkinson's disease showed uptake in the striatum to be approximately half that of

Fig. 58.100 1 23 1 -beta CIT SPECT imaging showing normal distribution of activity in a patient with essential tremor (A) and diminished putamen and caudate uptake i n a patient with Parkinson's disease (B). (Courtesy of Dr M. Buxton-Thomas.)

of the disease, while uptake is reduced in progressive supranuclear palsy, multiple system atrophy, Huntington's chorea and Wilson's disease. A further application of 231-FP-CIT may be in the differentiation 1 of Alzheimer's disease (normal striatal uptake) from Lewy body dementia (similar pattern to IPD). Radionuclide cisternography CSF dynamics can be investigated by injecting a suitable tracer into the subarachnoid space, either by lumbar puncture or preferably by lateral cervical or direct eistemal puncture. One such agent is indiumI Il-DTPA, which is not absorbed by the meninges, and has a sufficiently long half-life (68 h) to allow useful images to be obtained at 24 or 48 h after injection. The normal pattern of flow of CSF is from the lateral ventricles, where it is formed, out through third and fourth ventricles to the basal cisterns, where it mixes with spinal CSF, and then over the surface of the cerebral hemispheres, where absorption takes place. Obstruction to flow may prevent the appearance of the tracer over the cerebral hemispheres on delayed images and in severe cases CSF may apparently reflux into the lateral ventricles. The main application for this technique has been in patients with normal pressure hydrocephalus, in whom the demonstration of obstruction to flow of CSF was used to assist in the selection of patients for bypass surgery. The same technique has also been used to search for CSF leaks following trauma, or in patients with sponta-

Previous Page

CONTENTS


neous intracranial hypotcnsion in whom leaks of CSF arc often found in the cervical or dorsal spine. However, in both of these applications, cisternography is now considered to add little to assessment based on high-resolution CT and clinical criteria.

Textbooks Atlas, S. W. (1991) MR/ of the Brain and Spine. New York: Raven Press. Barkovich, A. J. (1990) Pediatric Neuroimaging. New York: Raven Press. Brain Tumours (1999) Neuroimaging Clinics of Noiih America, 9(4). Graham,d P. L. (eds) (1997) Greerrf ell's Neuropathology, I., on 6th n. London: Arnold. Intracranial Infectious Diseases (2000) Neuroimmging Clinics of'Norlh America, 10(2). Kleihues, P., Cavenee, W. K. (2000) World Health Organization Classification of Tumour. s. Pathology and Genetics of Tumours of the Nervous System. Lyon: [ARC Press. Orison, W. W.. Lewine, J. D., Sandes, J. A., Hartshorne, M. A. (1995) Functional Brain Imaging. St Louis: Moshy. Osborn, A. G. (1994) Diagnostic Neuroradiologv. St Louis: Mosby. Russet, , t ecl n L. T. (1988) Pathology of Tmrour -s of the Nervous System 5th edn. London: London: Arnold_ Scott, W. (1995) MR of the Brain and Spine, 2nd edn. New York: Raven Press. CT Altman, N. R.. Alhnan, D. H., Sheldon H., et ad (1984) Holoprosencephaly classified by CT. American Journal of Neuroradiologv, 5, 433_437. Ambrose, J. (1973) Computerised transverse axial scanning (tomography). 2. Clinical applications. British Journal of Radiology, 46, 1 023-1047. Barber. C..1., Rowlands, P. C., McCarty, M., et al (1990) Clinical utility of' cranial CT in HIV positive and AIDS patients with neurological disease. Clinical Radiology, 42, 1 64-165. Bentsen, J. R., Wilson, G. H., Iletmer, E., Winter, J. (1977) CT in i ntracranial cysticercosis. Journal of Computer Assisted Tomography, 1,464-471. Brismar, J., Ageel, A., Gascon, G., et al (1990) Malignant hyperphenylalaninemia: CT and MRI of the brain. American Journal of Neuroradiologv, 11, 1 35-138. Brismar, J., Brismar G., Coates, R.. et al (1990) Increased density of the thalamus in Ct of patients with GM2 gangliosidoses. American Journal of Neuroradiologv, 11, 1 25-130. Cone, L., Spinavasan, H., Romanui, F. C. A. (1990) Granular cell tumour (ehoristoma) of the neurohypophysis. American Journal of Nerroradiology,11.403- 406. Davis, P. C., Hoffman, J. C. Jr., Tindall, G. T., Braun. L F. (1985) Prolactin secreting pituitary microadenomas-inaccuracy of high resolution CT i maging. American Journal of Roentgenology, 144, 151-156. Deibler, C., Dusser, A., Dulac, O. (1985) Congenital toxoplasmosis. Clinical and radiological evaluation of the cerebral lesions. Neurorcrdiology, 27. 1 25-130. Dome, H. L.. O'Gorman, A. M., Melanson, D. (1986) CT of intracranial an liomas. American Journal o Neuroradiolo 7, 281-285. Fitz, C R. (1983) Holoprosencephaly and related entities. Neurorcrdiology, 25,225-285. Ganti, S. R., Hilal, S. K., Stein, B. M., Silver, A. J.. Mawad, M., Sane, P. (1986) CT of pineal region tumours. American Journal of Neuroradiologv, 7,97-104. Holtas, S., Nyman. U., Cronquist, S. (1984) CT of malignant lymphoma of the brain. Neuroradiology, 26, 33-38. Hounsfield, G. N. (1973) Computerised transverse axial scanning (tomography). I . Description of system. British Journal of Radiology, 46. 1 016-1022. Illurn, F., Dupont, E. (1985) Prevalences of CT detected calcification in the basal ganglia. Neuroradiologv, 27,32-37. Jack, C. R., Reese, D. F., Scheithauer, B. W. (1985) Radiographic findings in 32 cases of primary CNS lymphoma. American Journal of Neuroradiologv, 6,899-904.

1817

Klein, H. M., Bcrtalantfi, 1.. Mayfrank, A., et al (1994) 3D spinal CT for neurosurgical planning. Netroradiology, 36, 435-439Levy, M., Rosenbloom. S.. Perkett, R. V. (1986) Neuroradiological findings in AIDS: review of 200 cases. American Journal of Neuooradiology, 7. 833-839 . Liang, E. Y., Shan. M., Hsiang, J. H. K. (1995) Detection and assessment of i ntracranial aneurysms; value of CT angiography with shadow surface display. American Journal of Roentgenologv, 165, 1497-1502. Lott, P. R., Bellinger. W. E., Quisling, R. C. (1986) Suhcortical arteriosclerotic encephalopathy: the CT spectrum. Arerican Journal of Neurorodiolo e, 7. 817-822. Naidich, T. P.. Fulling, K. H. (1983) Congenital malformations of the brain. Neuroradiologv, 5, 1 77-304. Post, M. J. D., Kursonoglu, S. J., Hensky, G. T., ('ban, ,1. C., Moskowitz, L. B., Hoffman, T. A. (1985) Cranial CT in AIDS. American Journal of Neuroradiologv, 6. 743-754. Ronald, A. A., Patel, M., Casey, S., et al (1995) Evaluation of circle of Willis with 3D CT angiography in patients with suspected intracranial aneurysms. American Journal of'Neuroradiologv, 15, 1571-1579. Savoiardi, M., Strada, L., Paserini, N. (1983) Intracranial cavernous hacmangiomas review of 36 operated cases. American.lournal of Neuroradiologv, 4, 945-950. Suss, P. A., Maravillc, K. R., Tompson, J. (1986) MR of intracranial cysticercosis: comparison with CT and anatomopathologicad features. American Journal of Neuroradiologv, 7, 235 242. Sutton, D., Claveria, L. E. (1977) Meningiomas diagnosed by scanning. A review of 100 intracranial cases. In: Du Boulay, G. H., Moseley, T. F. (eds) Computerised Aria l Tomography in Clinical Practice, pp. 1 02-110. Berlin: Springer. Wclchman, J. M. (1979) CT of intracranial tuberculomata. Clinical Radiology. 30, 567-573. Zeumcr, H.. Schonsky, B., Sturm, K. W. (1980) Predominant white matter i nvolvement in subcortical arteriosclerotic encephalopathy (Binswangcr's disease). Journal of Computer Assisted Tomography, 4, 1 4-19. Zimmerman, H. M. (1971) The ten most common types of brain tumour. Seminars in Roentgenology. 6, 48-54. MRI Boltshauser, E., Schinzcl, A., Wichmann, W., Haller, D., Valavanis, A. (1988) Pelizaeus Merzhacher disease: identification of heterozygotes with magnetic resonance imaging. Hunan Genetics, 80, 363-394. Braffman, B. H., Bilanick, L. T., Zimmerman, R. A. (1988) Central nervous system manifestation of the phakomatoses on MR. Radiologic Clinics of North America, 26, 773-782. Bryant, R. N., Wells, S. W., Hiller, T..1., et al (1997) Infarct like lesions in the brain at MR imaging of the elderly. Radiology, 202, 47-54. Coates, T. L.. Hinshaw, D. B.. Peckman, N., et al (1989) Paediatric choroid plexus neoplasms: MR. CT and pathologic correlation. Radiology, 173, 81-90. Earnest, F. I. V., Kelly, P. J., Scheithauer, B. W., ct al (1988) Cerebral astrocytomas: histopathofogic correlation of MR and CT contrast enhancemeat with stereotactic biopsy. Radiology, 166, 823-830. Ellis, P. K. (1996) Case report; I, hermitte-Duclos disease; enhancement following gadolinium DTPA. Clinical Radiology, 51, 222-224. Elstcr, A. D.. Challa, V. R., Gilbert. T. H., et al (1989) Haemangiomas: MR and histopathologic features. Radiology, 170, 857-863. Feinberg, D. A., Mark, A. S. (1987) Human brain motion and cerebrospinal fluid circulation demonstrated with MR velocity imaging. Radiology, 163, 793-798. Freeborough, P. A., Woods, R. P., Fox, N. C. (1996) Accurate registration of serial 3D MR brain imaging in neurodegenerative disorders. Journal of Computer Assisted Tomography, 20. 1 012-1022. Freeman, M. P., Kessler, R. M., Allen J. H., et al (1987) Craniopharyngioma: cases-Journal CT and MR Imaging in nine of Computer Assisted Tomography, 11. 810-8 I8. Hayman, L. A., Taber, K. H., Ford. T. J.. Bryan, R. N. (1991) Mechanisms of MR signal alteration by acute intracianial blood clot: old concepts and new theories. American ./ournal of Neuroradiologv, 12, 897-907. Hesselink, J. R.. Dowd, C. F.. Healy, M. L., et al (1988) MR Imaging of brain contusions: a comparative study with CT. American .loitrrul of Neuroradiologv, 9. 269-275.


Previous Page 1818

CONTENTS

Next Page


I makita, S., Nishimura, T., Yamada, N., et al (1988) Magnetic resonance i maging of cerebral infarction: time course of Gd-DTPA enhancement and CT comparison. Neuroradiology, 30, 372-381. Jack, Jr, C. K. (1993) Epilepsy; surgery and imaging. Radiology, 89, 635-646. Joubert, M., Eisenring, J. J., Robb, J. P., Andermann, F. (1969) Familial agenesis of the cerebellar vermis. Neurology, 19, 813-825. Kucharczyk, J., Kucharczyk, W., Bern, I., et al (1988) Histochemical characterisation and functional significance of the hyperintense signal on MR images of the posterior pituitary. American Journal of Neuroradiology, 9,1079-1083. Kulkarni, M. V., Lee, K. F., McArdle, C. B., et al (1988) 1.5 T MR Imaging of pituitary microadenomas: technical considerations and CT correlation. American Journal of Neuroradiology, 9, 5-15. Lee, S. R., Sanches, J., Mark, A. S., et al (1989) Posterior fossa haemangioblastomas MR Imaging. Radiology. 171, 463-470. Li, L. M., Fish, D. R., Sisedya, S. M., et al (1995) High resolution MRI in adults with partial or secondary generalised epilepsy attending a tertiary referral unit. Journal of Neurology, Neurosurgery and Psychiatry, 59, 384-387. McAndrew, P. T., Land, N., Sellar, R. J. (1995) A case of intracranial septated arachnoid cyst. Clinical Radiology, 50, 502-504. McArdle, C. B., Richardson, C. J., Nicholas, D. A., et al (1987) Developmental features of the neonatal brain: MR Imaging. I. Grey-white matter differentiation and myelination. II. Ventricular size and extracerebral space. Radiology, 162, 223-245. McMurdo, S. K., Brand-Zawadski, M., Bradley, W. G., et al (1986) Dural sinus thrombosis: a study using intermediate field strength MR Imaging. Radiology, 161, 83-86. Marshall, V. G., Bradley. W. G., Marshall, C. E., et al (1988) Deep white matter infarction: correlation of MR imaging and histopathological findings. Radiology, 167, 517-522. Miller, D. A., Albert, P. I., Berkhof, F., et al (1996) Guide lines for the use of MR techniques in monitoring the treatment of MS. Annals of Neurology, 39,6-16. Moseley, I. (1995) Imaging the adult brain. Journal of Neurology, Neuro.surge rv and Psychiatry, 58, 7-21. Muller-Forell, W., Schroth, G., Egan, P. J. (1988) MR Imaging in tumours of the pineal region. Neuroradiology, 30, 224-229. Pickworth, S., Leutloth. U. (1996) CT and MRI findings in primitive neuroectodermal tumours in adults. British Journal of Radiology, 69, I -5. Pirker, W., Asenbaum, S., Bencsits, G., et al (2000) SPECT in multiple system atrophy, supranuclear palsy, and cortico basal degeneration. Movement Disorders, 15, 1158-1167. Post, M. J. D., Tate, L. G., Quencer, R. M., et al (1988) CT, MR and pathology in HIV encephalitis and meningitis. American Journal of Neuroradiology, 9, 469-473. Roosen, N., Sahlen, D., Stork, W., et al (1987) Magnetic resonance imaging of colloid cysts of the third ventricle. Neuroradiology, 29, 1 0-2 I. Savader, S. J., Murtagh, F. R., Savader, B. L., et al (1989) Magnetic resonance imaging of intracranial epidermoid tumours. Clinical Radiology, 40, 282-290. Schwaighofer, B. W., Hesselink, J. R., Press, G. A., et al (1989) Primary i ntracranial CNS lymphoma: MR manifestations. American Journal of Neuroradiology, 10, 725-732. Sherman, J. L., Citrin, C. M., Sangarossa, R. E., et al (1986) The MR appearances of CSF How in patients with ventriculomegaly. American Journal of Neuroradiology, 7, 1 025-1031 . Smirniotopoulos, J. G., Murphy, F. N. (1992) The phakomatoses. American Journal of Neuroradiology, 13, 725-746. Stack, J. P., Ramsden, R. T., Antowa, N. M., et al (1988) Magnetic resonance i maging of acoustic neuromas: the role of gadolinium-DTPA. British Journal of Radiology, 61, 800-806.

Stevens, J. M. (1998) Neuroradiology in epilepsy. In: Scaravilli, F. (ed.) Neuropathology of Epilep.s_v, pp. 77-139. London, World Scientific. Stevens, J. M.. Fox, N. (2001) Structural imaging. In: Hodges, J. R. (ed.) Early Onset Dernentia. A Multidisciplinary Approach, pp. 124-142. Oxford: Oxford University Press. Tolly, T. L., Wells, R. G., Sty. J. R. (1989) MR features of the fleeting lesions associated with Epstein-Barr virus infection. Journal of Computer Assisted Tomography, 13, 665-669Van Paerschen, W., Sisodiya, S., Connelly, A.. et al (1995) Quantitative hippocampal MRI and intractable temporal lobe epilepsy. Neurology, 45, 2233-2240. Weishman, U. C., Free, S. L., Everitt, A. D., et al (1996) MRI in epilepsy with a fast FLAIR sequence. Journal of Neurology, Neurosurgerv and Psychiatry, 61, 357-361. Yeakley, J. W., Kulkarmi, M. V., McArdle. C. B., et al (1988) High resolution MR imaging ofjuxtascllar meningiomas with CT and angiographic correlation. American Journal of Neuroradiology, 9, 279-283. Yoshiura, T., Wu, 0., Sorensen. A. (1994) Diffusion MRI, perfusion MRI and spectroscopy. Newnimaging Clinics of North America, 9, 439-453. Zee, C. S., Segall, H. D., Boswell, W., et al (1988) MR Imaging of neurocysticercosis. Journal of Computer Assisted Tomography, 12, 927-936. Ultrasound Babcock, D. S., Hann, B. K., Lequesne, G. W. (1980) B mode grey scale ultrasound in the head of the newborn and young infant. American.lournal of'Roentgenologv, 134, 457-468. Brown, J. K., Minns, R. A. (1993) Non accidental head injury with particular reference to whiplash shaking injury. Developmental Medicine and Child Neurology, 35, 849-869. Cremin, B. N., Chilton, S. J., Peacock, W. J. (1983) Anatomical landmarks in anterior fontanelle ultrasonography. British Journal of Radiology, 56, 511-526. Dewbury, K. C.. Bates, R. I. (1983) Neonatal intracranial haemorrhage: the cause of the ultrasound appearances. British Journal of Radiology, 56, 783-789. Fisher, R. M., Lipinski, J. K., Cremin, B. J. (1984) Ultrasonic assessment of i nfectious meningitis. Clinical Radiology. 35, 267-273. Gaissie, G., Roberts, M., Bouldin, T. W., Scatlifi, J. H. (1990) The echogenic wall in intracranial haemorrhage; sonographic pathological correlation. Paediatric Radiology, 20. 297-300. Jaspan. T. (1994) Assessment of infant intracranial trauma by high resolution ultrasound. BMUS Bulletin, 2, 1 4-22. Jaspan, T., Narborough, G., Punt, J. A. G., Lowe, 1. (1992) Cerebral contusion tears as markers of child abuse - detection by cranial sonography. Pediatric Radiology, 22, 237-245. Khaw, K-T. (1997) Does carotid duplex imaging render angiography redundant before carotid endarterectomy? British Journal of Radiology, 70,235-238. Levene, M. 1. (1981) Measurement of the growth of the lateral ventricles in preterm infants with real time ultrasound. Archives of Disease in Childhood, 56, 900-904.


Radionuclide Imaging Asenbaum, S., Baumgartner, C. (2001) Nuclear medicine in the preoperative evaluation of epilepsy. Nuclear Medicine Communications, 22, 835-840. Benamer, T. S., Patterson, J., Grosset, D. G., et al (2000) Accurate differentiation of parkinsonism and essential tremor using visual aseesment 1 1 231 of[ 1 23 11-FP-CIT SPECT imaging: the 1 -FP-CIT study group. Movement Disorders, 15, 503-510. lacoboni, M., Baron. J. C., Frackowiak, R. S., et al (1999) Emission tomography contribution to clinical neurology. Clinical Neurophvsiology,

Previous Page

CONTENTS

Next Page

59

The applii ttion of MR techniques to medicine appears to advance each year with a relentless pace and it would almost seem foolish for someone to suggest what its limitations will be and when these will be reached. We constantly see refinements and improvements in existing techniques, providing enhanced diagnostic information. This is hand

in hand with the emergence of new applications, which have been enabled by hardware advances linked to new sequence and acquisition protocols, which in turn reflect the development of new contrast agents. Increasingly, MR is providing new, non invasive alternatives to existing imaging techniques, often with improved diagnostic mlormatron. There is also a move towards more quantitative imaging techniques, providing data quicker than using other modalities and without the requirement for multiple imaging investigations. There is always the temptation to suggest that, as MR becomes more advanced and faster, the possibility of reduced examination times may result in higher patient throughputs. However, it is often more realistic to accept that these advances, while giving some benefits in shorter examination times, will in the main result in the ability to make additional measurements, producing more extensive and improved quality of diagnostic information. I n this section, covering the recent advances of MRI, we hope to provide an insight into a few examples of where MR may he positioned relative to the `routine' in 2-3 years time. Hopefully, if the development of MR continues unabated at its current pace, further new and exciting advances will rapidly supersede the information prcsented here.

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ Fig. 59.1 Siemens Symphony 1.5 tesla MR system typical of the patient friendly features, incorporating short magnet bore length and flared patient aperture. (Courtesy of Siemens Medical Solutions.)

Standard clinical imaging systems have generally embraced patient comfort aspects in terms of their design. Superconducting magnets have become shorter in bore length and the internal patient apertures are enhanced by use of strongly flared entrances (Fig. 59.1). These changes in bore size have been accomplished without compromising the maximum image field of view or main magnetic field homogeneity. The development of MR systems designed for specialist niche applications has continued. Smaller, low-field, open resistive systems still find widespread applications in, for example, extremity

1819

Previous Page

1820

CONTENTS

Next Page

A TEXTBOOK OF RADIOLOGY AND IMAGING For general clinical imaging, most users now choose 1.5 T superconducting systems, with field strengths of I T and below primarily being applied in niche situations. There is an increasing emergence of very high field (VHF) systems, with a growing number of 3 T systems being used for routine and research applications and a small number of 7-9 T whole-body magnets installed worldwide. For an optimised 3 T magnet, there are significant advantages for many neurological applications, in particular spectroscopy, tMRI and spin-tagging perfusion imaging. In terms of general imaging, there arc significant hurdles to overcome at 3 T. The main issues relate to increased T, relaxation times, RF power problems in relation to specific absorption rate (SAR) and receiver coil design issues for efficient transfer of a uniform B, (radiofrequency excitation) field. However, for neurological applications the improved signal-to-noise can give significant benefits (Fig. 59.3).

Fig. 59.2 An interventional MR suite incorporating a high-field superconducting magnet and a digital X-ray fluoroscopy facility. Patient transfer between the MR system and the X-ray fluoroscopy unit is possible while still within the general environment of an operating theatre. Patient coordinates relative to the two imaging modalities can be maintained via a patient couch transport system. (Courtesy of Philips Medical Systems.) musculoskelctal imaging. In addition, higher field (0.5-1.0 T) open architecture systems are emerging, essentially aimed at interventional MR installations. These new interventional systems, although significantly heavier than standard superconducting magnets (typically around 10 tons and above), provide the ability to perform surgical intervention during imaging, along with therapeutic applications such as focused radiofrequency and ultrasound. Uniquely, MR techniques can provide measurements of tissue temperature during therapy, or in the future, monitor drug delivery using new MR-visible agents. Furthermore, recent trends in the development of interventional installations have seen the linking of MR systems with digital X-ray fluoroscopic units in a single examination/operating room or in contiguous rooms, thus allowing patient movement between modalities (Fig. 59.2).

Magnetic field gradient coils continue to develop, with increasingly greater maximum amplitudes and faster rise times. Maximum amplitudes are now typically 30-40 mT/m or even 50-60 mT/m on some of the dedicated cardiac and neurological systems operating at reduced field of view with shorter gradient coils. At optimum performance, many systems now accomplish these maximum amplitudes with rise times of 200 ps or below, at a 100% duty cycle. As a result of reduced duration of the RF pulse, improvements in gradient performance result in reduced minimum echo and repetition ti mes (TE and TR), and improved spatial resolution via increased matrices and reduced image fields of view (Fig. 59.4). In addition, high-performance gradients provide the ability to perform echoplanar imaging (EPT), and enhancements in many other sequence applications. Improvements in diffusion-weighted sequences are enabled by higher b-values (sensitivity to diffusion) and increased speed of acquisition. Contrast-enhanced angiography (CE-MRA) benefits from shorter TEs. Bolus injection-based dynamic perfusion i maging achieves greater temporal resolution via shorter TRs, and


Fig. 59.3 (A) T2 - weighted FSE and (B) time-of-flight angiography of a cerebral arteriovenous malformation acquired at 3 Tesla. (Courtesy of Malcolm Randall, VA Medical Center, University of Florida.)

Previous Page

CONTENTS

The effects of improved gradient performance on duration of gradient pulses. Increased gradient rise times result in a reduction in the total pulse length within a sequence for the same useful pulse length. Fig. 59.4

cardiac imaging has improved significantly, with faster acquisitions now permitting real-time imaging and navigator-echo corrections. Ultimately, the limit to gradient performance development is constrained by considerations for patient safety. Peripheral nerve stimulation is now a real safety issue when running many routine rapid sequences. Unfortunately, the geometry of the patient within the magnet and individual patient sensitivity to the phenomena are factors dictating the probability of induced nerve stimulation. Therefore, it is for this reason that current estimations on the likelihood of any possible effects are empirical. Some manufacturers are now producing double gradient systems and gradient coil inserts which reduce the high dB/dt (rate of change of magnetic field) values produced at the ends of standard gradient designs, while facilitating the possibility of even greater maximum amplitudes and rise times.

Next Page

Manufacturers continue to produce an ever-expanding series of receiver coils, many of which are aimed at specific applications and parts of the body. The majority arc now quadrature/circularly polarised in operation, with resulting improved signal-to-noise ratio. Increasingly, new coil designs are enhancements on previous coils, based on developments in techniques, for example combined neck and head coils for neurological applications and peripheral leg coils for contrast-enhanced angiography. Phased array coils are now considered standard for spine, body and pelvic examinations. The principle of phased array coils is to combine the signal-to-noise properties of small surface coils with large image fields of view (Fig. 59.5). Wherever possible, phase array technology is being applied to other body areas such as the shoulder, to give the benefits of greater signal-to-noise. The increasing use of several coils at the same time, and multiple-element phased array coils, has led to the provision of multiple data receive channels and the option of duplexing coils into a single channel. Some systems have eight independent data channels, allowing the possibility of up to 16 separate coils or coil elements to be used simultaneously. Alongside these developments have been significant enhancements in data handling and processing power.

Significant developments in computing power have had to be made in order to benefit from the many hardware innovations already discussed. Rapid imaging strategies can easily produce hundreds of i mages per examination. Dedicated array processors can now perform thousands of Fourier transforms per second, providing the ability to achieve real-time imaging. The use of multiple-element phased array coils and multiple data acquisition channels also require enhanced computing power, if data reconstruction times are not to be detrimental to new imaging strategies. Innovative techniques, for manipulating k-space data, have produced significant image quality benefits and new methods for reducing image acquisition times. Partial k-space acquisition strategies, such as half-Fourier and partial echo acquisitions, are now widely used techniques. Developments in the flexibility of k-space filling routines have enabled improvements for triggered acquisition routines in cardiac imaging and non-triggered methods in freebreathing sequences for abdominal imaging. Reordering of k-space filling can now produce optimum contrast rendition in contrastenhanced angiography. New partial/parallel acquisition techniques, such as SENSE and SMASH, have provided the ability to produce


Principle of phased array coil technology. Multiple coil arrays provide the benefits of large fields of view along with the signal-to-noise properties of small single-coil elements when signals are combined in an array combination. Fig. 59.5

Example of the application of SENSE to abdominal breath-hold imaging for an elderly patient unable to sustain a long breath-hold. (A) 23 s without SENSE; (B) 12 s with SENSE. (Courtesy of Philips Medical Systems.) Fig. 59.6

Previous Page 1822

CONTENTS

Next Page


Fig. 59.7 (A,B) Rapid T 2 -TSE i maging using SENSE i ll ustrating the

i maging of a fetus in utero. The fast imaging in less than 20 s minimises any possible artefacts from fetal movement with high signal-to-noise and high spatial resolution. (Courtesy of Philips Medical Systems.)

The refinement of fast/turbo spin-echo (FSE/TSE) sequences to permit single shot acquisitions has made breath-hold Tn- weighted sequences a practical technique. Long echo train lengths of 128 spin echos, often linked with a half Fourier k-space filling ( HASTE), produce heavily T,-weighted images from a single excitation pulse, as shown schematically in Figure 59.8. The visualisation of the static, long T, pancreaticobiliary fluids resulting from long effective echo times gives perfect contrast for MRCP techniques (Fig. 59.9). Single thick-slab acquisitions, which mimic maximum intensity projection (MIP) images, can be made in 2-3 s. Multiple 2D slices can be acquired in 10-15 s or even 3D variations in a single breath-hold. Data can then be viewed as an MIP, with the ability to remove overlying high signal from stomach or small bowel.

significant reductions in acquisition times for sequences setup with a single signal excitation/average. These new techniques require multiple coil elements in specific directions within the desired i maging volume. Existing coils permit the use of these new techniques; however, new designs consisting of coil arrays along multiple axes will extend their application. The resulting shorter acquisition times will benefit CE-MRA, for example in breath-hold abdominal imaging and cardiac studies, where temporal resolution is paramount. Figures 59.6 and 59.7 show examples of the benefits of the shorter acquisition times in minimising the deleterious effects of patient movement.

The hardware developments already described have enabled a series of new sequences to be applied to routine clinical practice. These sequences, in conjunction with phase array technology, have permitted refinements to existing examination types as well as the opportunity for MR to be used in innovative and exciting areas of diagnostic imaging and spectroscopy. The following examples of clinical applications serve to reflect the breadth of the advances that MR has achieved.


Fig. 59.9 Single-shot breath-hold MRCP examination. Heavily T 2- weighted FSE/TSE i mages demonstrate high contrast fluid signal from

pancreatic and biliary systems.

Fig. 59.8 Schematic diagram for a single shot FSE/TSE sequence. An echo train length

of 128 180 pulses, following a single 90 excitation pulse, fills half of k-space. The effective echo time TE ett corresponds to those data lines written to the centre of k-space. The significant number of very long TE ethos give a heavily T 2 -weighted contrast.

CONTENTS

Previous Page

High-performance gradients linked to automated table movements enables rapid imaging, covering the whole of the body. Fast T,- and T,-weighted FSE/TSE sequences can be planned with automatic, sequential movements of the patient through the MR scanner. The whole body can then be scanned in coronal or sagittal planes, giving full body views (Fig. 59.10). Additional transverse imaging can be carried out as required at each relevant table position incrementation, while still maintaining short, practical examination ti mes. The technique offers exciting alternative imaging methods for some aspects of oncological screening. Whole-body STIR i maging can provide a sensitive and fast screening process for hone t umours and lymph nodes.

-

Rapid diffusion-weighted imaging and T,' - weighted, gadoliniumbased perfusion imaging require high specification gradients. Fast acquisitions minimise problems with patient motion for diffusion i mages and provide good temporal resolution for dynamic, firstpass perfusion studies. Diffusion- and perfusion-weighted protocols can provide key, unique diagnosis for the assessment of acute stroke patients. Early diagnostic imaging can give essential information,

Next Page

which can dictate strategies for patient management using new antithrombolitic drugs. New high-specification systems can now perform a comprehensive brain study in a practical examination ti me. Diffusion and perfusion sequences can be added to routine neurological sequences such as FSE/TSE or GRASE T, FLAIR and time of flight angiography. Subsequent analysis of the diffu sion data can give an apparent diffusion coefficient map (ADC) to complement the diffusion weighted images. Similarly, analysis of the dynamic perfusion images produces time to peak (TTP) and mean transit time (MTT) maps, along with integral data implying relative cerebral blood volume (rCBV). An example of a typical stroke examination protocol, which i ncludes routine imaging along with diffusion and perfusion sequences, is shown in Figure 59.11 . The value of imaging acute patients in the first few hours following a stroke is illustrated by the diffusion and perfusion data, as conventional imaging shows no change until approximately 24 h after the stroke.

CE-MRA has recently developed into a standard vascular imaging routine for systems with high-performance gradient systems. The

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ Fig. 59.10 (A) Rapid whole-body screening using a moving table facility. Four sets of coronal images with both T, and T2 weighting are acquired to cover the entire length of the body. (B) Additional transverse images are also acquired at chosen anatomical positions. In total, 750 images are produced in 5 min. Note the high resolution with demonstration of a contained prostatic carcinoma (low signal, arrow) in the left peripheral zone. (Courtesy of Philips Medical Systems.)

Previous Page 1824

CONTENTS

Next Page


Fig. 59.11 Comprehensive stroke examination performed after (a) 6 h and (b) 24 h from an acute stroke. The protocol includes routine brain sequences T2` i n addition to diffusion weighted and dynamic perfusion protocols. The whole examination can be carried out in less than 10 min. The early perfusion and diffusion imaging clearly shows areas of infarct not visible in the standard T 2 and FLAIR images until after 24 hours. (Courtesy of Philips Medical Systems.) technique does not rely on either the inflow signal enhance-

enhancement following the administration of a gadolinium-based

ment associated with time-of-flight, or the signal phase shift for

contrast agent.

moving spins utilised in phase-contrast techniques. Essentially, it is

The high equipment performance requirements for CE-MRA relate to the ability to perform a 3D T I - weighted acquisition while a

based on the simple, longstanding principle of T 1 - weighted signal


Fig. 59.12 (A,B) Contrast-enhanced abdominal angiography. A maximum intensity projection image from the CE-MRA 3D acquisition performed during a 15 s breath-hold and triggered using real-time image monitoring of the contrast bolus arrival. The centre of k-space is acquired at the start of the sequence, using an elliptical filling strategy to provide optimal signal as the bolus of contrast passes through the vasculature. (Courtesy of Siemens Medical Solutions.)

Previous Page

CONTENTS

bolus of contrast passes through the particular vasculature of i nterest. Therefore, short repetition times, using a flip angle optimised for the shortened T, of contrast-enhanced blood, are needed to ensure the 3D acquisition takes place while the bolus is within the area of interest. For abdominal and thoracic applications, the acquisition additionally needs to be made within a breath-hold period (Fig. 59.12). Furthermore, exceptionally short echo times (typically less than 1.5 ms) are also required to minimise intravoxel signal phase losses due to flow. Real-time imaging routines are used to monitor the arrival of the bolus of contrast in order to ensure correct timing of the acquisition. The CE-MRA sequences are also linked to dedicated k-space filling strategies as alternatives to the standard linear filling routines. These specialised strategies enable the centre of k-space (which controls low spatial frequency, contrast information), to be captured at the peak of the signal enhancement from the bolus of contrast agent (Fig. 59.12). Filling strategies i nclude elliptical or centric k-space trajectories, with the central portion of the k-space volume being filled at the start of the 3D sequence (see Fig. 15.134). The use of short echo times in CE-MRA minimises the possi bility of dephasing signal losses arising from complex or turbulent flow. The well-recognised signal voids associated with complex and turbulent flow in time-of-flight angiography and exaggeration of vessel stenosis do not normally occur in CE-MRA. Therefore CEMRA offers significant improvements over standard MR angiographic techniques, providing diagnostic information to rival conventional angiography in a non-invasive way. Adaptation of bolus contrast injection strategies is used to perform multiple acquisitions over different areas of the body for peripheral angiography. Typically, a steady-state injection of contrast is linked with automatic incrementation of the patient couch position. Acquisitions are performed over three to four positions. The series of images can then be linked together for display purposes and maximum intensity projection images generated (see Fig. 15.149). Furthermore, multi-

Next Page

planar reconstructions and surface rendering routines can be used to produce virtual endoscopy images. Further important contributions to vascular imaging have been made by the use of high-resolution imaging in conjunction with dedicated small field-of-view, vascular coils. The main application has been in the visualisation of thrombus and plaque formation for carotid artery imaging (Fig. 59.13). The dedicated coils can be used to obtain high signal-to-noise, high-resolution, black-blood images demonstrating areas of thrombus.

The ability to perform rapid sequences using cardiac triggering with the latest highest performing gradient systems has catapulted MR as an important tool for cardiac imaging. One of the main strengths of MR is the versatility it brings to cardiac imaging, permitting investigations previously performed using several different imaging modalities to be carried out in a single, one-stop cardiac examination. Established techniques of black-blood prepared FSE/TSE and bright-blood, cine gradient-echo acquisitions (TrueFISP/ BalancedFFE/FIESTA) can provide several slices in a single breathhold. Higher performing gradients, linked with improved cardiac triggering regimens, permit multiple cardiac phases to be measured with excellent R-to-R wave coverage. Figure 59.14 demonstrates examples of the different types of image data that can contribute to a generalised cardiac examination. Motion of the myocardium can be imaged and assessed quantitatively using tagging routines, which use a narrow grid of saturation lines applied repeatedly across the image field of view during the cardiac cycle. Perfusion can be assessed using a rapid series of gated acquisitions following a bolus of contrast. Measurements can be made in terms of signal changes in the myocardium. The technique can be used in conjunction with stress induction, using, for example, dobutamine administration, to give diagnostic information that matches both ultrasound- and radioisotope-based techniques. Reductions in image acquisition times and the use of the latest navigator-echo correction methods have enabled coronary artery i maging to be a realistic technique for MR. Issues do remain, however, for patients who are unable to hold their breath for even short periods. Interestingly, it has been found that a very useful byproduct of contrast administration during the cardiac examination has been the development of myocardial viability measurements. Images acquired approximately 10 min postcontrast, using an inversion recovery sequence to null-out normal myocardium, demonstrate residual signal from trapped contrast within any infarcted myocardial tissue.

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ Fig. 59.13 High-resolution carotid imaging for the visualisation of thrombus. Transverse images through the area of thrombus show excellent correlation with histology.

Clinical spectroscopy continues to develop and remains a key research tool for the investigation of many neurological diseases, i ncluding stroke, psychiatric disorders, multiple sclerosis, HIV and oncology. The use of shorter echo times to provide data on important, short T 2 metabolites and multiple voxel acquisitions (chemical shift imaging, CSI) has extended the clinical application of spectroscopy. A huge reduction in examination times for spectroscopy measurements and automated setup and adjustment routines for shimming and water suppression have helped ease 'H-spectroscopy

Previous Page 1826

CONTENTS

Next Page


Fig. 59.14 A comprehensive MR cardiac examination. The one-stop i nvestigation provides diagnostic information on: (A) myocardial motion using saturation tagging techniques to visualise wall movement through the cardiac cycle; (B) myocardium viability using an inversion recovery technique, 10 min postcontrast administration; (C) black blood-prepared FSE anatomical images showing atheromatous plaques (arrows); (D) coronary artery angiography; (E) myocardial perfusion-weighted images acquired immediately after a bolus injection of contrast; (F) tine gradient echo, bright-blood images, for ventricular volume analysis. (Courtesy of GE Medical Systems.)


Fig. 59.15 Chemical shift imaging in the brain. Multiple voxels are acquired over slices through the brain. Analysis of the spectra provides quantification of the measured metabolites using an overlaid map on a reference anatomical image. Individual spectra can be viewed and interrogated from within the metabolite maps. The spatial distributions of the various metabolites seen can give important clinical information. (Courtesy of GE Medical Systems.)

i nto the clinical routine. Single voxel acquisitions, including system setup, can now be made in about 10 min, with only a further 10 min being required for multiple slice CSI. This reduction in examination time and the increased numbers of clinical MR systems with the capability to perform MR spec-

troscopy have resulted for the first time in spectroscopy being used as a routine clinical tool. It has been in the invcstieation of brain tumours that the most significant diagnostic benefits have been seen. MR spectroscopy can provide important information to aid the identification and grading of brain tumours, discriminating

Previous Page

CONTENTS

l esions from other pathologies and differentiating tumour from postradiotherapy signal changes. FSE/TSE approaches have permitted the extension of the single voxel techniques to CSI methods where several slices are acquired, each consisting of a 32 x 32 voxel array. Importantly, the CSI data allow metabolite maps to be created and overlaid upon conventional MR images as an alternative to standard MR spectra (Fig. 59.15). Individual spectra from each voxel can be viewed as required. The metabolite maps provide the clinician with spectroscopic data in an image-like format, which l ends itself more readily to a routine clinical interpretation. An important, new application of 'H-MR spectroscopy has been in the study of prostate cancer. Single voxel or CSI can be used to identify tumour and monitor disease spread. Within the prostate, MR signals from citrate only occur within malignant tissue, therefore the visualisation of citrate is indicative of tumour. Using CSI techniques, the individual voxel resolution is sufficient to monitor disease spread from within the gland into, for example, the seminal vehicles, and also to monitor response to treatment therapies (Fig. 59.16). The analysis of citrate signal in each spectrum provides information on disease spread. In a similar manner to the brain metabolite maps, citrate maps can be overlaid on to reference T2 anatomical images of the prostate. Evidence appears to suggest that the best quality data are currently obtained using endorectal coils.

Over a decade, extracellular space (ECS) gadolinium chelate contrast agents (the first being gadolinium DTPA (Magnovist, Schering, AG)) have been the only agents available for MRI. Newer contrast agents have been, and are currently being, developed, with some of these agents due for general release shortly, further extending the role of MRI into new areas, as well as consolidating existing practices.

Next Page

(iv) blood-pool agents. New developments in each of these areas are becoming available, extending the clinical utility of MRI. Extracellular space agents are routinely used in MR in CNS and body imaging, and have a high safety profile. Recent developments i n the assessment of brain function described above, CE-MRA and the improved detection of intracranial disease with double or triple dosing (0.2-0.3 mmol/kg body weight) require the use of a rapid bolus of increased volumes of contrast. Newer agents with lower viscosity and increased concentration of gadolinium (e.g. I molar formulation gadobutrol (Gadovist, Schering, AG), as opposed to the original 0.5 molar agents, can reduce the volume load injected by a half. Extracellular space agents, using a dynamic acquisition, can be used for improved detection of liver lesions. There are, however, several other agents (e.g. Mn DPDP (Teslascan, Nycomed): Gd EOB-DTPA (Eovist, Schering); Gd BOPTA/dimeglumine ( MultiHance, Bracco)) that are actively taken up by hepatocytes and lead to prolonged T, shortening of the liver. These hepatobili ary agents remain in the liver for a long period of time, allowing a wide imaging window for delayed scanning, providing increased flexibility in sequence selection and imaging planes compared with the ECS agents. The hepatobiliary agents provide an increased liver-to-lesion contrast and improved lesion conspicuity. Reticuloendothelial system agents are iron oxide-based particles with a covering of polysaccharides that specifically target the RES or blood pool. They can be formulated in different sizes: those of >50 nm.are termed superparamagnetic iron oxides (SPIO), and those of >-1.0), the microbubbles are disrupted and burst. This behaviour of echo-enhancing agents results in new varieties of imaging techniques for these agents. At the most basic level they can be used to improve the signal-to-noise ratio in spectral, colour and power Doppler examinations to rescue inadequate examinations. Transit times can also be measured and arteriovenous shunts can be identified. Using low-MI imaging allows detection of the agent in the blood vessels and tissues with minimal destruction of the microbubbles; this allows imaging of the agent in blood vessels on real-time imaging. Pulse inversion imaging detects the nonlinear echoes from microbubbles and enables areas containing agent to be distinguished from areas without agent (Fig. 59.30B). High-MI imaging destroys the agent but this produces high-amplitude signals from areas where the agent is present (stimulated acoustic emission (Fig. 59.30C)), which contrast with the absent signals from areas without the agent. This is particularly useful in the liver, where it has been discovered that several agents have a delayed hepatosplenic-specific phase (Fig. 59.31). Sequential highMI scans at intervals can be used to provide an estimate of blood flow into the area of interest, calculated from the rate of reaccumulation between scan sweeps. Research is now developing in the area of tissue-specific agents containing a variety of therapeutic substances. These could be targeted at specific areas of disease, such as tumours or blood clots; when the agent has accumulated in the target area, the microbubbles are then destroyed using ultrasound at the appropriate fre-


These were originally developed to improve the signal-to-noise ratio in Doppler examinations but, by the time the first ones became

Fig. 59.30 (A) shows a liver tumour on real-time imaging; (B) shows low MI imaging of the vascular phase; (C) shows the same lesion with high MI i maging producing bubble destruction and stimulated acoustic emission. (Courtesy of Dr P. Sidhu, Kings College Hospital.)

Previous Page

CONTENTS

i maging and visual inspection with injection of patent blue dye has produced the best results. The radiotracer technique gives better detection of nodes in unexpected sites, and with dynamic imaging is more likely to indicate the sequence of drainage when several nodes are shown. The ideal tracer would be easy to prepare, stable and securely labelled, and would migrate rapidly from the injection site to the sentinel nodes. Early studies have used various technetium-labelled colloids, but more recently technetium-dextran compounds have shown promise. The technique requires injection of the tracer directly into or closely adjacent to the primary tumour, or into the overlying skin if the tumour is superficial. Imaging over the next 1-4 h typically i dentifies one or more sentinel nodes, the position of which can then be marked on the patient's skin. Surgical excision, which may follow immediately after the i maging or the next day, is assisted by using a handheld scintillation detector (gamma probe) to find the sites of tracer concentration. Injection of patent blue dye immediately prior to surgery improves the detection rate of sentinel nodes, and offers the additional advantage of identifying the lymphatic vessels that lead from the tumour towards the sentinel nodes. The successful application of a sentinel node technique requires effective expertise not only in imaging but particularly also in surgical and pathological techniques. The routine use of immunohistochemistry has improved the sensitivity of conventional H & E staining for nodal staging, and the introduction of reverse transcriptase-polymerase chain reaction (RT-PCR) offers a further improvement in accuracy. However, the sensitivity of pathological examination for detecting microscopic tumour deposits in otherwise normal lymph nodes remains uncertain, as does the ultimate significance of such micrometastases.

Most of the early experience with sentinel node techniques is in breast cancer and melanoma.

Next Page

lymph nodes is a powerful indicator of prognosis from the time of presentation. As with breast cancer, the majority of new patients show negative sentinel nodes. Patients spared extensive nodal dissection on the basis of an initial negative sentinel node biopsy have shown on follow-up a small but worrying incidence of subsequent relapse, so the adoption of this technique as best surgical practice is still in doubt. Other tumours Promising results for sentinel node imaging and biopsy have been obtained in various gastrointestinal tumours, particularly colorectal and oesophageal cancer. The original description of the techniques arose from observations in patients with penile cancer; more recent applications include patients with cancer of the cervix, vulva, squamous tumours of the head and neck, and papillary carcinoma of the thyroid.

In common with other clinical imaging techniques, ultrasound has benefited significantly from the massive improvements in computing power and data storage that have taken place over the last few years. These have allowed major advances in signal generation and processing techniques to be achieved. In parallel with these developments in data processing, the advent of echo-enhancing agents has opened up many new opportunities for developing the role of ultrasound as a diagnostic and therapeutic technique. The rapid advance of computing power, together with the trend to miniaturisation, has enabled manufacturers to provide either more processing power in the same size of system, or the same functionality in a smaller system. Several handheld systems are available with excellent frame rates and resolution comparable to high-end systems of 4-5 years ago, and which increase the availability of ultrasound in many areas of medical practice (Fig. 59.23).


Breast cancer About one-third of patients with breast cancer

have axillary nodal disease at the time of presentation. Routine dissection of the axilla creates morbidity and is of no benefit to that large group of patients whose nodes are negative. Identification of sentinel nodes in breast cancer is achievable in about 95% of cases, with greater than 95% success in correctly i dentifying whether the axilla is involved. Follow-up studies have shown that nodal recurrence in patients whose axilla is negative at the time of sentinel node biopsy is very rare. Sentinel node biopsy is inappropriate for patients with clinically i nvolved axillary nodes, and is also unreliable in patients who have had just radiotherapy. The application of sentinel node biopsy in patients with multifocal tumours and in those with previous local excision of the primary breast lesion is still not clear. Sentinel node techniques can be carried out on patients with impalpable tumours by using ultrasound-guided i njection. Melanoma The patterns of lymphatic drainage from melanoma are less consistent than with breast cancer, and simultaneous drainage to multiple sites is not uncommon. For this reason, i maging after local radiotracer injection is of critical importance and the distinction of primary from second tier nodes may be helpful in planning subsequent surgery. The status of sentinel

Fig. 59.23 Two modern ultrasound systems, the larger one on the left is a fully configured system; the smaller one on the right on the couch is nevertheless capable of abdominal and small parts examinations, including Doppler studies.

Previous Page

CONTENTS

All aspects of Doppler signal generation, reception and subsequent processing are undergoing intense development. Harmonic i maging, using harmonic frequencies that are generated at two, three or four times the fundamental frequency, reduces the amount of noise in the image and enhances the relative signal-to-noise ratio. Initially used with echo-enhancing agents, it is now also used for i mproved real-time imaging, especially in technically `difficult' patients. Harmonics are produced only by pulses with high amplitude (a violin string stroked gently produces a single, pure tone; stroked more forcefully, it produces a much coarser sound due to the harmonics that are generated), so pulses in the focused beam tend to produce harmonics, whereas those in side lobes and neverberations are too weak, thus reducing noise and speckle in the i mage (Fig. 59.24A,B). Pulse inversion harmonic techniques produce pulses of alternating phase, with the returning echoes being summated; these might be expected to cancel themselves but, because of the non-linear effects relating to the propagation of ultrasound waves through tissues, summation of pulse pairs produces a detectable signal containing diminished amounts of noise and clutter. As with òrdinary' harmonic imaging, pulse inversion techniques can be used with both echo-enhancing agents and i maging techniques. Further variations in pulse generation can be used, including pulse coding techniques, initially developed for radar transmissions, and tailoring of the pulses to the characteristics of specific echo-enhancing agents. Compound imaging can he achieved by interrogating a volume of tissue with scan lines steered electronically at different angles; this results in noise and speckle reduction, together with improved visualisation of margins and small structures, such as microcalcifications. Initially used for

Next Page

superficial scans, it has now been further developed for abdominal use (Fig. 59.24). Echoes from flowing blood can sometimes be seen on real-time i maging, typically in the common femoral veins. With modern systems and appropriate processing techniques, these can be displayed in real time to show flowing blood without the use of Doppler techniques (Fig. 59.25). The stiffness or compliance of different tissues can be assessed using ultrasound, the technique is known as elastographY . A set of digital radiofrequency (RF) echo data from the region of interest is obtained. The tissue is then compressed, usually by the transducer, and a further set of RF echo data is obtained. This is then crosscorrelated with the initial data and any changes in arrival time of pulses from equivalent volumes of tissue can be identified. An estimate of the local tissue strain can he made and from this the degree of tissue stiffness can be assessed. This technique is still being developed but shows some promise for identifying abnormal areas

A 'B flow' image of an internal carotid artery stenosis with the echoes from the moving blood clearly shown in the lumen of the vessel. (Courtesy of GE Ultrasound.) Fig. 59.25

̶̰ϴϧϭήΘ̰ϟ΍ ήθϧ ̶Η΍έΎθΘϧ΍ ϪδγϮϣ ζϧ΍ΩέϮϧ Fig. 59.24 (A) shows a kidney at fundamental tissue imaging frequency; (B) shows the reduction in noise obtained by imaging at harmonic frequencies;

and (C) shows further improvement of the image by using dynamic compound i maging.

Previous Page 1834

CONTENTS

Next Page


Fig. 59.26 3D ultrasound. (A) shows a reconstruction of a fetal face and arm; (B) shows planar reconstructions from the data volume shown at the bottom right. (Courtesy of Acuson.)

in tissues, which may not have sufficiently different echo properties to register on an amplitude image, such as a scirrhous breast carcinoma compared with a fibroadenoma. The increased computing power now available allows the storage and manipulation of the large amounts of data required for 3D i maging. The two-dimensional ultrasound beam is swept through the region of interest and the data processed to produce a threedimensional representation of the volume scanned, which can then he viewed and manipulated in a variety of ways (Fig. 59.26). Increased capabilities for data storage also allow extended field of view imaging, which allows the display of larger regions of interest, such as muscles and tendons, or large masses (Fig. 59.27).

5 MHz; today the equivalent frequencies arc 5-7 MHz and 1 2-15 MHz. Some specialist applications, such as dermatology and i ntravascular ultrasound, use transducers with frequencies up to 30-40 MHz. Broadband technology allows the transducer's operati ng frequency to be adjusted to the scanning conditions and harmonic and non-linear imaging techniques to be undertaken. Multielement array transducers have made it possible to focus and steer the ultrasound beam electronically; different groups of elements can be used for different tasks, such as pulsed Doppler and imaging. The elements in these transducers have been arranged i n a single row along the length of the transducer, allowing beam manipulation in the lateral (across the slice length) direction only.


Developments in transducer design and construction have been responsible for many of the advances in ultrasound techniques and applications over the years. New materials are being developed to i mprove transducer performance. Multiple elements, broadband transmission and reception, and improved beam forming techniques have all contributed to improved performance in terms of both spatial and contrast resolution. In the 1980s abdominal scans were typically performed at 3.5 MHz, and `small parts' scanning at

Fig. 59.27

Extended field of view image of an iliopsoas bursa.

Fig. 59.28 Diagram of a 2D transducer showing the improvement in beam formation and the associated improvement in spatial and contrast resolution. (Courtesy of GE Ultrasound.)

Previous Page 183 6

CONTENTS

Next Page


Fig. 59.31 (A) shows two focal areas of relatively reduced echoicity in a fatty liver; (B) shows filling in of these areas after the administration of an echo-enhancing agent with a li ver-specific phase, confirming the presence of normal liver tissue in these areas consistent with focal fatty sparing in a fatty liver. (Courtesy of Professor D. 0. Cosgrove, Hammersmith Hospital.) quency and thereby release chemotherapeutic or thrombolytic agents within the tumour or blood clot. Despite developments in other imaging techniques, such as CT, MRI and computed radiography, the role of ultrasound continues to expand. Its unrivalled ability to show events within the body in real ti me, the continuing improvements in contrast sensitivity and resolution, technological advances in equipment, together with the emergence of new techniques such as elastography, will mean that the prominent role of ultrasound in clinical imaging will continue to expand over the years to come.

Over the last decade, computed tomography has developed rapidly, from conventional, single-slice machines through helical (spiral) CT to the current multidetector (multislice) scanners. A number of advances in hardware and software have enabled this: in particular, the use of slip rings to allow the scanner gantry to rotate continuously in one direction; the development of X-ray tubes with great heat capacity for long continuous X-ray exposure (up to 60 s); cheaper, more stable solid-state detectors and multidetector arrays; i mproved computer power and hardware to handle and store the l arge amounts of data produced; and improvements in software to allow effective image postprocessing.

which is reconstructed into axial slices by interpolating between adjacent points in the helix.

Helical CT calls for continuous X-ray exposure over the entire period of the scan (up to 60 s) with reasonable levels of tube current. The tube heating associated with this requires a tube with a very high heat capacity, and these have been developed especially for helical CT. Delays due to overheating of the tube can cause problems in high throughput departments, and where high tube load examinations are carried out (for example, back-to-hack helical scans for triphasic liver CT). This is not such a problem with multislice CT because the scan times are so much quicker. reducing tube l oading compared with single-detector helical CT.

For many years CT scanners have used xenon gas detectors which have a conversion efficiency (X-rays to signal strength) of around 60%. This diminishes further if the detectors are not maintained, but xenon detectors are relatively stable. Solid-state crystal detectors may have conversion efficiencies of nearly 100%, with a consequent potential for a 40% reduction in patient radiation dose for the equivalent scan appearances. Technical difficulties of stability of output, continuing emission of light after the X-rays have terminated (afterglow), and other problems with respect to the size of the front face of the individual detectors and the intcrspace material between adjacent detectors, have been largely overcome or cornpensated for in the data processing. Developments in solid-state detectors and the ease with which they can be stacked in parallel adjacent channels have facilitated the development of multislice scanners. Although different approaches to the multidetector array have been adopted by the various manufacturers,


Prior to the development of slip ring technology, the high tension supply to the gantry-mounted X-ray tube was made via cables. The gantry therefore had to return to the starting position after each rotation to unwind the cables. During this period the table would i ncrement to the next slice position, and the patient would begin another cycle of suspended respiration. Slip rings (large circumference electrically conducting rings) allow power to be conducted to the tube on the CT gantry via electrical brushes instead of cables, The gantry is therefore able to rotate continuously in one direction. The data from the detectors may also be transferred back to the reconstruction computer by slip rings. This has several advantages: with single-slice CT, the cycle time per slice was reduced, allowing a number of slices to be acquired within a single breath-hold; a faster rotation speed can be achieved giving shorter scan acquisition ti mes; continuous exposure is possible at a single table position, allowing functional studies, better temporal resolution, and `CT fluoroscopy' for interventional procedures. The combination of continuous exposure with continuous table movement through the gantry is the basis of helical CT. A helical volume dataset is acquired,

all current multislice machines are able to acquire four slices per rotation. With rotation times as low as 0.5 s, the slice acquisition rate of a multislice scanner is thus eight times that of a typical single-slice helical unit. Scanners capable of acquiring 16 slices per rotation (32 slices/s) are presently under evaluation. The width of the detector array produces problems of image

quality due to the cone shape of the X-ray beam. Special reconthe b ion algorithms are needed to compensate for the divergence of the beam along the long axis of the patient.

With helical and particularly multislice CT, a vast amount of data per unit time is produced from the gantry detectors. The rapid trans-

Previous Page

CONTENTS

Next Page

fer of the data to the reconstructing computer is crucial and may be carried out via specialised cables or radiofrequency/infrared links. With multislice CT datasets typically containing 250-300 images, rapid reconstruction is vital. Powerful computers are needed to apply the interpolation and reconstruction algorithms to give adequate reconstruction speed. Current scanners reconstruct each slice i n less than 1 s, but further improvements are likely to be needed as multislice scanners evolve and produce even larger numbers of thinner slices. Rapid links between the scanner console and other elements in the scanner network, such as workstations, PACS systems and teleradiology units, are important, and each of these elements must have sufficient computing power to handle the large datasets ergonomically. Archiving of data should be efficient and economical.

Continuous scanning in the same table location results in a repeatedly updated image at the same slice location. Using rapid reconstruction algorithms (often omitting some of the usual data filtration steps), images can be updated several times a second with an effective delay between acquisition and display of less than half a second. This gives a `near real-time' image, which can be used to guide interventional procedures. Radiation dose to the patient is significantly higher than from conventional fluoroscopy, so minimum tube current settings and minimum exposure times should be used. The potential for increased dose to the operator is also high with this technique. Interventional devices designed to keep the operator's hands out of the primary beam, audible timers, fluoroscopy exposure time limits and routine logging and review of fluoroscopy times arc advocated.


Although similar in many respects to helical CT, multidetector CT represents a considerable advance. The essential difference lies in the detector array, which contains multiple detector elements. Most current (circa 2002) machines allow four slices to be acquired si multaneously, scanners acquiring 16 simultaneous slices are on the market and detector arrays in some machines already contain 32 rows of detectors. Large volumes can be scanned rapidly with relatively narrow slice width (typically 0.5-3.0 mm). The advantages of multidetector CT can be summarised as improvement in speed, quality and scope. The manufacturer's eventual aim must be to acquire all the sections required in a single rotation of the gantry, although currently the corrections to the acquired data for the conebeam geometry of the volume swept by the X-ray beam are still under development. Consequently the increasing number of detecfor rows in current machines are being used to reduce the individual slice thicknesses more than increasing the total beam width. As multislice CT is up to eight times as fast as single-detector helical CT, there is even less likelihood of movement or respiration artefacts in routine scanning. Examination of the chest, for example, is feasible in less than IO s. It is particularly suited to imaging of children, trauma patients and the seriously ill, monitored patient where speed is important (Fig. 59.32). Further refinements arc possible in contrast examinations, such as precise arterial and portal phase liver scans, and reduction in contrast medium dose is seen in vascular studies because of the shorter acquisition times.

Fig. 59.32 (A) Sagittal view of entire spine in trauma patient. Single acquisition with 2.5 mm slices throughout; (B) local axial view through cervical fracture.

As there is no longer any compromise between volume coverage and slice width, thin slices can be used in all cases. This leads to better spatial resolution inplane and along the z-axis. Multiplanar reconstruction (MPR) and other image processing can he carried out routinely, thus improving diagnostic precision (Fig. 59.33). For high-definition imaging of small volumes, very thin slices can give virtually isotropic imaging (where spatial resolution in any reconstructed plane approaches that of the inplane resolution) (Fig. 59.34). The thinnest slice collimation is currently achieved by collimating the X-ray beam to half the width of the two central detector rows, to give two slices each half the width of the individual detector row. Thus, however many simultaneous sections the scanner is capable of as a maximum, only two are available at the thinnest slice width. Such accurate beam collimation is technically difficult, and may be aided by postpatient collimation, with consequent potential for significantly increased patient radiation dose. Helical CT led to many new CT applications but many of these have only become clinical reality with the advent of multislice CT. The large-volume, thin-slice datasets obtained are well suited to all forms of image processing. Possibilities include CT angiography,

Previous Page

183 8

CONTENTS

Next Page



Fig. 59.33 (A) Routine multislice abdominal study with intravenous and oral contrast in patient with cystic renal abnormality; (B) showing quality of coronal reconstruction through kidneys; (C) renal arteries (three) clearly visible.

CT of the colon with `virtual' endoluminal views (Fig. 59.35), virtual bronchoscopy, routine MPR from axial acquisition in sinus/ENT and orthopaedic imaging, cardiac CT with prospective and retrospective ECG-triggered gating and functional and perfusion imaging.

As with helical CT, pitch can be increased to increase coverage

or reduce dose or acquisition time. However, the relationship between pitch and image degradation is complex and non-linear with multislice CT. Optimal pitch values are given by the scanner manufacturer.

Fig. 59.34 (A) Axial 0.5 mm acquisition through petrous temporal bone with overlapping reconstructions; (B) coronal reconstructions from axial dataset have similar resolution to source images (isotropic resolution).

In 1989 the National Radiological Protection Board estimated that CT scanning accounted for 2% of radiological investigations but 20% of the population collective radiation dose. By 2000, further

Previous Page

CONTENTS

Next Page

Fig. 59.35 (A) 3D image of air-insufflated colon; (B) endol uminal view of colon ('virtual colonoscopy') demonstrates a rounded submucosal lesion seen at conventional colonoscopy; i nset image, oblique axial reformat profiles the lesion and demonstrates fat density within it, confirming a colonic lipoma.

studies suggested the latter figure had risen to 40% in the UK and 67% in one USA department. The routine use of thin slices does confer a slight increase in radiation dose, partly because increasing rows of thinner detectors i n multislice machines will require more rows of interspace material, resulting in `dead space'-irradiated but not contributing to i mage data. This had been partly offset by solid-state detectors, used in all multislice machines, which have X-ray conversion efficiencies approaching 100%. However, this leaves negligible scope for further improvements in detector efficiency. Multidetector scanners may use slightly overlapping pitch values in routine practice, as accurate prepatient collimation of the X-ray beam to the exact li mits of the detector rows requires very high engineering precision. This may also increase radiation dose if scanning parameters are not adjusted. It has been estimated that absorbed dose may be up to 40% higher for a multislice helical CT examination than for the equivalent conventional CT scan. Otherwise, dosimetry is similar to single-slice helical scanners, with the caveat that the speed and versatility of these scanners may encourage scanning of larger volumes or multiphasic examinations, with an associated increase in radiation exposure. Protocols for scanning the liver, for example, may now include arterial, early venous and late venous scans (triphasic). Such extended scanning protocols will inevitably suffer from a `law of diminishing returns'. For every individual patient, a clear indication for each phase of the scan is required, performing only those phases relevant to the investigation, This clearly necessitates the direct supervision of every examination by an experienced radiologist. The development of new indications for multislice CT and the increasing demand for established CT indications will also result in increased population radiation exposure from CT. The role of the experienced radiologist in vetting requests for CT to minimise unnecessary examinations remains of paramount importance. Reduction in scanner mAs to the lowest values consistent with acceptable noise in the resultant i mage is a further responsibility of the supervising radiologist. This is of the greatest importance in the scanning of children, where the consequences of radiation exposure are potentially greater and yet the small size of the patient makes reduction in mAs easier to

achieve. Modulation of the X-ray beam mAs during the scan can be used to reduce patient radiation dose with negligible effect on i mage quality. The chest, abdomen and pelvis are usually thinner from front to back than side to side, consequently a lower mAs suffices in those portions of the gantry rotation where the beam traverses the patient in these directions. Dose reductions of 25-40% can be achieved by this method. Reconstruction algorithms designed to reduce streak artefact (particularly useful for scans in the pelvis or across the shoulders) can also be used to reduce radiation dose by reducing the selected mAs, and can be combined with the modulated mAs described above.


The CT computer workstation is now, more than ever, a pivotal part of the scanning unit. Radiologist interaction with the workstation is now routine, and many radiologists will report entirely from the workstation. A fast and ergonomic system is therefore essential. The computer must be powerful enough to handle and store the large datasets with ease. A typical workstation will contain software to allow viewing in the axial and reconstructed planes (including curved planes), surface shading, maximum and minimum intensity projection, endoluminal imaging (virtual endoscopy) and volume rendering. Volume rendering has become available on CT scanners relatively recently. The software assigns variable degrees of opacity and brightness to all voxels within the dataset within a certain HU range. Parameters can be changed interactively to give varying degrees of translucency and conspicuity to different structures. The result is a more comprehensive image which can show soft tissue, bone and vascular features (Figs 59.36, 59.37). As the algorithm uses the whole dataset, it requires great computing power to work effectively. An advantage of volume rendering is that very little editing or segmentation of the dataset is needed compared with maximum intensity projection. Cardiac imaging using CT is becoming increasingly important. Gating techniques can be used to freeze cardiac motion. Coronary artery calcification assessment is now well established with CT

Previous Page

1840

CONTENTS

Next Page


Fig. 59.38 Coronary calcification scoring. Calcification is highlighted in each vessel (here in left anterior descending) and a cumulative score obtained.

Because of the large number of images produced in a typical multislice dataset, printing of all images to hard copy is no longer feasible. Thin slice images can be fused together into a thicker i mage for presentation to the hard-copy printer. As Much CT reporti ng is currently performed from the workstation, images from abnormal studies only may be selected from the workstation for printing by the radiologist. As PACS systems become more preval ent, the need for hard-copy imaging will decline further.

Fig.

59.36

Volume-rendered image of thoracoabdominal aortic

aneurysm.

The ability to acquire CT data continuously while moving the patient through the scanner aperture confers a number of significant advantages. The speed of acquisition allows large volumes, for example the entire chest, to be scanned within a single hreath-hold. This guarantees contiguity of the reconstructed slices, avoiding the misregistration often seen with conventional single-slice CT. caused by variable depth of respiration. Rapid scanning also minimises the risk of movement artefact. The helical dataset can he thought of as a volume of data within which scans can he reconstructed at any level. This means that overlapping slices can be reconstructed without the need for further scanning or X-ray exposure. The above features give a marked


Volume-rendered image of abdomen combined with cutsurface reformations to demonstrate viscera. Fig. 59.37

(Fig. 59.38), and angiographic techniques to visualise the coronary arteries directly are being evaluated. Tumour permeability imaging and functional and perfusion studies in various sites are possible. Modality matching software allows, for example, CT and PET i mages to be superimposed, giving anatomical and functional information in one image.

Fig. 59.39 3D segmented CT angiogram image of renal transplant artery with arterial stent.

Previous Page

CONTENTS

i mprovement in spatial resolution in the z-axis (long axis of the patient), allowing good quality MPR and 3D reconstruction from t he dataset. A further advantage of rapid scanning is that contrast medium can be deployed with greater precision, whether used for imaging of solid organs or vascular structures. Data can be acquired entirely within the arterial phase of a peripheral intravenous injection, giving the possibility of performing angiographic studies (Fig. 59.39). By increasing the speed of table movement during data acquisition, coverage in a given time can be increased and radiation dose reduced. The relationship between slice width and table travel per gantry rotation is known as pitch. The disadvantage of increasing pitch is that the image quality decreases as pitch increases, primarily due to an increase in the effective slice width. Pitches greater than 1.5-2.0 are not recommended. By using multiple banks of detectors, multislice scanners acquire several interlaced datasets si multaneously. This can be used to both decrease the individual slice collimation (slice thickness) and increase the table increment per tube rotation, giving rise to faster scans with even better z-axis resolution. The relationship between slice collimation and table i ncrement per rotation is then modified by the number of simultaneous slices in order to retain an effective pitch ratio of less than 2. I mage quality and radiation dose do not inherently differ in any significant way from conventional single-slice CT. However, the rapidity and case of scanning may mean that larger volumes than before are routinely imaged. Moreover the versatility of helical CT i s such that more CT is being used in clinical practice. The radiation dose to the population from CT is therefore likely to be i ncreasing. The massive increases in computer power, coupled with t he more rapidly acquired, thinner slice data from multislice helical CT, has resulted in a plethora of image display techniques such as surface-shaded 3D. maximum and minimum intensity projections, and endoluminal surface reconstructions such as virtual colonoscopy. The clinical value of these techniques is currently being actively investigated.

Next Page

an explanation of the action taken being provided in the latter two cases. 3. The examination appointment is made, with a date, time and room booking. 4. Any patient instructions or pharmacological preparation are dispensed to the patient. 5. On the day of the examination a DICOM (digital imaging and communication in medicine) modality worklist is generated, includi ng that patient's examination scheduled to be acquired on a specific piece of imaging apparatus. 6. The completed examination is verified as being technically adequate by the person acquiring the examination (usually a radiographer/technologist). 7. The examination is sent to the PACS and PACS confirms receiving the images it contains using DICOM storage commitment. 8. Postprocessed images or later image acquisitions relating to the same examination (imaging study) may be added to the examination on PACS using DICOM performed procedure step (PPS). 9. The examination is viewed on a diagnostic PACS workstation, i n conjunction with the clinical details relating to the examination (as stated when the examination was requested) and any pertinent past imaging examinations. 1 0. The examination is reported (dictated) using a digital speech recognition system and the resultant text report generated is visible on PACS and the radiological information system (RIS), as well as being accessible on the hospital information system (HIS) and via the electronic patient record (EPR). 1 1. The report is verified (electronically signed) by one of the reporting radiologists. 1 2. Possible addenda are added to the report when the examination is later reviewed or audited, and these addenda are electronically signed by the radiologists responsible for them. Integration It is clear that a number of information technology systems are involved in the digital workflow scenario outlined above, and all these must be integrated seamlessly with one another for the digital workflow to function smoothly. In most health care environments there will be a number of legacy computerised systems already in place, which function well in isolation but are not integrated with one another. It is essential that demographic (textural) data, such as the patient's name and personal details, be entered only once on the hospital network of computer systems, to avoid typographical errors generating spurious patients with different hospital numbers for the same individual. It should also be possible for the person accessing the system to navigate from one dataset to another on the same patient, without having to log-in again, or to re-enter the patient's demographic details. Thus one virtual database must exist linking all these disparate computer systems. Even if a completely new health care institution is being created, different medical specialists may wish to install electronic systems from different manufacturers, choosing the `best of breed' information technology (IT) solution best suited to fulfilling their specialised requirements. rather than making do with the solution that happens to be provided by a single vendor chosen to equip the entire institution. This important issue of streamlining workflow and integrating the various hospital 1T systems is being addressed by the ME (Integrating the Healthcare Enterprise) initiative launched at the Radiological Society of North America (RSNA) international


A number of hospitals throughout the world have recently become filmless using a picture archiving and communication system ( PALS), but the major challenge now is to re-engineer the whole clinical worktlow process to become digital, so that hospitals may become paperless as well as filmless, not just with respect to radiol ogical examinations but for the whole of clinical workflow. Seamless integration of all the various electronic systems throughout the hospital (and beyond, if telemedicine is being considered) is t he critical issue that now needs to be solved.

I . The examination is electronically requested, with the relevant clinical hi~tocy Wecails) provided. 2. The request is justified electronically or manually, as appropriate to the investigation, and accepted, modified or rejected, with

Previous Page

CONTENTS

meeting 1999. This initiative is bringing together for the first time vendors of clinical information systems with manufacturers of PACS and imaging equipment, and there is thus broad industry support for the work of IHE. IHE is not introducing new standards, but merely facilitating those which already exist by defining profiles that specify an unambiguous way to communicate integration requirements. W eb-based integration One of the simplest solutions to this i ntegration dilemma is to use a web-based protocol so that all the different computer systems, such as the RIS and the HIS, are merely web-based clients, allowing the user to browse different datasets on the same patient in much the same way as the worldwide web allows users to access different computer servers around the world over the internet, and to pass easily (by pointing and clicking) from one internet site to another. PACS manufacturers are using web browsers to allow review of PACS images and reports from ordinary desktop PCs. Indeed, a growing number of PACS vendors are using web technology for their entire PACS, not just for review workstations. Web technology used in this manner has much to commend it: upgrades can be carried out on central servers without the need for software packages to be upgraded on many specialised workstations; training requirements are minimal, as the majority of modern health care professionals are familiar with the use of web browsers; and navigation between different modules is facilitated by pointing and clicking on hyperlinks. DICOM Digital imaging and communication in medicine is a standard, now accepted worldwide by vendors and users of imaging equipment, which precisely describes the various attributes of pieces of apparatus one might wish to connect together on a network, and thereby specifies whether they will be able to integrate with one another and to what extent they will interact. HL7 Health level 7 is a standard which is less well defined than DICOM and is used by vendors of patient information systems such as HIS and RIS. Originally a standard used predominantly by North American manufacturers, it has now superseding other, less well accepted European standards. HL7 and DICOM equipment can be relatively easily integrated together using standard brokers (protocol converters).

Next Page

data such as patient demographics (name, address, hospital number, family doctor name, and other personal information on the patient), the numerical results from laboratory tests (biochemistry, haematology, bacteriology, virology, immunology), and reports from clinical investigations (such as endoscopic and bronchoscopic studies, pathological examinations of tissue): as well as scheduling information such as outpatient clinic appointments, hospital admissions, the bed state and occupancy, and patient discharge information. The HIS is currently the most important computer system in the majority of extant hospitals. It forms a major part of the EPR, but also needs to integrate with the RIS and other specialised reporting systems (such as in histopathology) so that all the radiological and other reports are visible from within the HIS. HIS needs to integrate with PACS for prefetching old examinations on patients attending outpatient clinics but not necessarily having a new radiological examination. (These patients will not be scheduled on RIS if they are not undergoing a new radiological procedure and therefore their old images will not be prefetched by the RIS.) Electronic remote requesting system (RRS) Such systems are

also referred to as remote order entry and order communications (order comms). They provide an electronic means of requesting and booking investigative examinations from a remote location (wards, outpatient clinics and general practitioner surgeries). It is suggested that the entire health care facility be served by a common electronic remote requesting system (RRS), but that the requests to different departments will have customised front-end software. Thus there will be menus tailored to suit an imaging request, a haematology request, a biochemistry request, a histopathology request, and so forth. In this way the remote requesting system can be kept as generic as possible. The RRS is frequently regarded as being a module of the HIS, but must be integrated with the RIS and with the PACS. There needs to be a very close integration of the RRS with the RIS because a list of the previous imaging examinations that a patient has had, with their reports (known to the RIS), must automatically be presented to, or easily accessed by, the requesting doctor. Also, procedures booked on the RRS will need to be closely linked with the RIS procedure code, and the RRS needs access, via the RIS, to the room availability for booking the various imaging procedures within the imaging department. The clinical details pertaining to the requested imaging examination arc entered onto the system (typed or entered by a speech recognition system) and are thus legible, and already in digital form, thereby rendering the examination-requesting process paperless. An RRS is very advantageous in the presence of a PACS since it ensures the typed clinical details relating to the examination will appear on the diagnostic PACS workstations in conjunction with the associated images at the time of reporting. There is thus no need for these details to be typed from a paper request form, nor for paper request forms to he digitised.


The electronic patient record (EPR) The EPR (or electronic medical record, EMR) can be thought of as an almost virtual entity linking the other health care computer systems together (Fig. 50.40); however, it must also ultimately accommodate the electronic patient case notes, discharge summaries, clinic letters and prescribing data. The databases of all the various computer systems in the hospital need to be linked (or be the same database) so that, when any relevant items of that patient's demographic details are entered on any part of the system (patient name, hospital number, NHS number, date of birth, etc.), that patient will he accessed. Input of demographic data only once minimises error. Any update to a patient's demographic data on any part of an integrated EPR system (e.g. the HIS, RIS or PACS) means that that updated information is propagated to all systems automatically, ensuring that all systems contain accurate information. Hospital information system (HIS) Other names for the HIS

i nclude the CIS (clinical information system) and PAS (patient administration system). The HIS is a computer system that holds

Radiological information system (RIS) The RIS is the computer system responsible for the scheduling of the various rooms within the imaging department and booking of examinations into those rooms, and it serves as the repository of the radiological reports once they have been typed. The workflow of the imaging department and its radiologists, radiographers and typists can be audited from the RIS data. Every imaging procedure is allocated a

Previous Page

CONTENTS

Next Page

specific RIS code. Many patients are known to the RIS but not to caesium iodide) produce light photons locally, which arc then conthe HIS, as they have only attended the hospital for an imaging vertcd into electrical signals. The earlier photostimulable phosphor procedure (referred directly from their family doctors) but have plate technology, generally known as computed radiography (CR), not been seen as an inpatient or outpatient at the hospital and is the most widespread means of acquiring plain digital radiotherefore will not have been given an HIS-based hospital number. graphic images, although in fact this technology produces an anaThis illustrates one reason why the various IT computer systems logue signal in photons, which is converted to a digital signal as a i n the hospital must share a common database, or have their datafinal step by cathode ray tube. bases seamlessly integrated. When a patient is scheduled for an PACS Regardless of the size of the PACS project, it should i maging procedure, the RIS must first check to see whether a always be founded on a formal contract between the institution and patient with the same (or very similar) name already exists in the the PACS vendor, and it should define in detail the responsibilities HIS/EPR database, and only if no match is found will a new id of each. Different vendors arc contributing to the project. it is (identification) number be allocated to the patient. highly desirable to have one who contractually assumes ultimate Modern RIS and imaging acquisition modalities support the responsibility for the integration of each part of the network so DICOM service of `modality worklists', whereby a list of patients there can be no argument when an interface problem arises. whose examinations are to be acquired on a particular modality is Clearly a PACS must replace the traditional roles of conventional passed automatically to that piece of image acquisition apparatus film, namely: the acquisition, display, transport, storage and (e.g. a CT scanner, a computed radiography phosphor plate imaging retrieval of images. The major added value provided by a PACS is device), thus generating a worklist. This is an extremely important the institution-wide efficiency with which those roles are achieved. aspect of departmental workflow facilitated by DICOM, as it obviThis operational efficiency is only truly realised when film has been ates the need for the operator to retype the patient name and other withdrawn and the entire institution has an electronic digital i dentification details onto the acquisition apparatus and thus avoids workflow. The two critical characteristics of a PACS are its perforthe generation of typographical errors and the mismatch of data that mance and the access it provides to radiological examinations. would otherwise result. This is particularly important when The physical architectures of most PACS installations today are working in a PACS environment because images on the acquisition similar, consisting of fast networks connecting multiple servers, modalities will only be accurately sent to their destination on PACS modalities and workstations. The physically distributed, but logii f the demographic details match exactly; if they do not, and the cally centralised, architecture provides the most powerful capabilii mages are still sent, they end up as òrphan' examinations, usually ties to PACS users, especially in large PACS installations. l abelled ùnspecified', and require manual reallocation or merger The network is the critical link for all the individual C0111with the correct PACS entry by a PACS systems administrator. Such ponents in a PACS. Its performance is a major factor in determining housekeeping database `clean-up' operations are time consuming, the ability of a PACS to support the examination volume of the i nefficient, wasteful of resources and interfere with departmental hospital. workflow. Most PACS products now use industry-standard networks, usually consisting of a high-speed network (backbone) that spans Acquisition devices When purchasing new apparatus (or PACS the entire institution, with branches that connect individual coincomponents), it is crucial for radiologists to specify the role that ponents or groups of components to the backbone. Networks can be they expect that apparatus to play in the setting of departmental differentiated in several ways, including cable type, speed, commuworkflow so that it can be established which DICOM attributes nication protocol and topology. that piece of apparatus will need to support. The DICOM attribEthernet is the most common network in use today: 100 megabit utes of any piece of modern imaging equipment are listed in its per second (100 Mbps) is the minimum Ethernet speed required for DICOM conformance statement, and a technical expert will thus efficiently transmitting the large amount of data contained in radiobe able to determine whether or not two pieces of DICOM con- logical imaging examinations (8-10 megabytes of data in a chest formant apparatus will be able to interface with each other and radiograph) in a PACS network. carry out the desired functions, as this will depend upon cornparFor a PACS used for high-volume primary diagnosis (at least i ng their respective DICOM conformance statements and deter50 000 procedures per year), the network traffic is very large. Such mining which common DICOM attributes they possess. DICOM a system should not share a network with other functions in the i s an international standard composed of numerous subparts. It is enterprise if it is to guarantee consistently high performance. This continually being expanded and improved in such a way that none usually means that a network will have to be installed specifically of the additions negates pre-existing parts of the standard, but for the PACS. builds upon them. All PACS products have a database to store information identifyI n a digital workflow environment, images clearly need to be ing patients and cataloguing their imaging studies. Most PACS acquired digitally rather than in analogue format. This is already i nstallations use a single relational database (like Sybase, Oracle or the case for most of the specialised examinations: computed tomogMicrosoft Access) to store identifying information for patients raphy, magnetic resonance, digital fluoroscopy and angiography, and studies. The user interacts directly with the database to find radioisotope scanning, the newer ultrasound scanners and positron the imaging study of current interest, and thus the database peremission tomography. The bulk (>-70%) of the imaging workload of formance is critical to the performance of the whole PACS. any imaging department, however specvaXise(X, is composed Of plain The radiologists' participation in the digital woY )'N process is radiographic image acquisition. The newer direct digital acquisition highly dependent upon having a user-friendly seamless reporting systems for plain radiography capture remain very expensive, but process in place. The PACS software plays a crucial role in facilitatproduce a signal in direct digital format. Some are more directly i ng rapid reporting from PACS, which must be at least as rapid as digitally acquired than others: those using a scintillator (generally reporting from hard-copy film in a conventional film-based radio]-


Previous Page

CONTENTS

ogy department. It is also essential, as discussed above, to have satisfactory integration of other computerised health care systems closely related to reporting from PACS: in particular, the electronic RRS (which supplies the typed clinical details associated with each i maging study); the speech recognition system (which immediately converts the radiologists' dictated reports into text on the system); and the RIS (which stores these typed reports). The PACS software concepts required for efficient soft copy reporting are: availability of relevant historical images for comparison (often known as prefetching), compilation of reporting worklists, ordered arrangements of the current and previous images on the PACS workstation monitors by default (often called default display protocols (DDPs), or hanging protocols), and reporting or ' dictation macro' software to facilitate the mechanics of the softcopy reporting process. The compilation of worklists of unreported i maging studies (particularly for plain radiography reporting) should he automatic but flexible and institution configurable. The triggers for building such worklists are generally within the RIS. Easy intranet e-mail access from all modules forming part of the EPR, but especially the electronic RRS and the PACS, is desirable in allowing rapid correspondence with requesters of i maging procedures. This can be useful, for example, in alerting a requesting clinician to a rejected or modified request for an i maging procedure on the RRS, or drawing his or her urgent attention to an abnormal imaging result. Restricting the e-mail access to the intranet rather than the internet reduces the stringency of requirements for protective firewalls and the need for data encryption. E-mail

On-line journals The reporting radiologist can benefit from accessing on-line journals and textbooks as sources of reference during reporting. Ideally, these should be accessible from the same workstation the radiologist is using for viewing PACS i mages and dictating reports on the speech recognition systems, as it is cumbersome and time consuming to have to log into, and activate, a separate computer system for every task and this interferes with streamlined workflow.

Next Page

It is clear that digital workflow (filmless and paperless) throughout a hospital, or other health care enterprise, is utterly dependent upon seamless integration of the various computer systems contributing to that workflow. It requires efficient appropriate networks and conscicntious round-the-clock maintenance and management.

MRI Castillo, M. (ed.) (1998) Proton spectroscopy of the brain. Neuroinur,ging Clinics ofNorth Ameri(a, 8(4). Drayer, B. P. (ed.) (2001) Selected topics in neuroradiology. Neurointaging Clinic,, of North America. 11(3). Elster, A. D.. de Roos. A. ( 1999) Multiple perspectives in MR contrast. Journal of Computer Assisted Tonngraphv. 23 (suppl. I ), S I-S217. Heisernian. J. E. (ed.) ( 1999) Fast scan and echoplanar MR imaging. Neurounaging Clinics ofNortlr America, 9(2). Melhem, E. R. (ed.) (2002) Diffusion imaging. Neuroinraging Clinics of North America, 12(1). Naidich, T. P.. et al (cds) (2001) Anatomic basis of functional imagine. Neuroirnaging Clinics of North America. 1 1 (2). Nesbit, G. M. (ed.) (1999) Acute ischemic stroke. Neuroiniagilig Clinics of' North America, 9(3). Padhani, A. R.. Nccma n. M. (2001) Challenges for imaging angiogenesis. British Journal of Radiology, 74. 886-890. Semelka. R. C. (ed.) (2001) MR imaging of the liver 1: techniques and contrast agents. Magnetic Resonance haaging Clinics of North America, 9(4). Truwit, C. L. (ed.) (2001) MR-guided therapy in neurosurgery. Neuroinutging Clinics of Nortli America, 11(4). Yucel, E. K. (ed.) (1999) New techniques in body MR imaging. Magnetic Resonance Imaging Clinics of North America, 7(2). Magnetic Resonance in Medicine and , lournal of Magnetic Resonance hanging ( published by John Wiley & Sons. Chichester) arc both official journals of the International Society for Magnetic Resonance in Medicine: good sources lot- original articles and works in progress.


Image library/museum and teaching modules The transfer of digital images to an on-line image library and/or personal teachi ng or academic folders should be considered part of the digital workflow. This area has been rather neglected until recently. being seen as a non-essential part of digital workflow of interest only to academic centres (even though it is actually of concern to the majority of radiologists and other clinicians) and possibly also because it was not considered by PACS vendors to be a financially rewarding area for investment. Speech recognition dictation system Speech recognition (voice activation) systems are now beginning to achieve sufficient accuracy (>_95%) to be acceptable for clinical use and use in radiol ogical reporting (dictation). The accuracy and speed of learning of such systems is an important issue affecting departmental workflow: radiologists require an extensive customised word database for their work (which is specific to imaging and to the appropriate form of English used to generate reports), and time spent correcting the text must be low enough not to impact significantly on the time spent generating the report. It is crucial that a speech recognition system be integrated into the departmental RIS and into the hospital PACS, rather than acting as a stand alone system, which is still often the case.

Radionuclide imaging Barrington, S. F. (2000) Clinical uses of PET in neurology. Nuclear Medicine Conununicutions, 21, 237-240. Bar-Shalom. R., Valdivia. A. Y., Blaufox. M. D. (2000) PET imaging in oncology. Seminars in Nuclear Medicine. 30. 1 50-185. Beyer, T., Townsend, D., Brun, T., et al (2000) A combined PET/CT scanner for clinical oncology. Journal of Nuclear Medicine. 41. 1 369-1379. Haigh. P. 1 . . Hansen, N. M., Giuliano. A. E., et al (2000) Factors affecting sentinel node localisation during pre-operative breast lymphoscintigraphy. .Journal of Nuclear Medicine, 41, 1 682-1688. Kcshtgar, M. R. S., Waddington, W. A., Lakhani, S. R., Ell, P. J. (1999) The Sentinel Node in Surgical Oncology. Berlin: Springer. Maisey. M. N. (2000) Clinical PET in cardiology and cardiac surgery. Nuclear Medicine Connnunications, 21, 234-236. Nunan. T. O., Haiti, S. F. (2000) PET in oncology II-other wmours. Nuclear Medicine Cononunications, 21. 2 2 9-233. O'Doherty, M. J. (2000) PET in oncology I-lung, breast, soft tissue 2 2 sarcoma. Nuclear Medicin' Connnunications. 21. 2 4-2 9. Ultrasound Harvey, C. J.. Blomley. M. J. K., Eckersley. R. J., Cosgrove. D. O. (2001) Review: Developments in ultrasound contrast media. European Journal of Radiology, 11. 675-689. Laing. F. C. (2001) Update on ultrasonography. Radiologic Clinics of North America, 39, 385-590. (1999) Ultrasound today and tomorrow: new deselopments and contrast media. European Radiology. 9(suppl. 3), 5298 S414. WFUM13 Symposium on the Future of Biomedical Ultrasound (2000) A celebration of medical and biological ultrasound at the dawn of the third millennium. tiltrasound in Medicine and Biology, 26. S I-S 172.

Previous Page

CONTENTS

CT Broderick, L. S., Shemesh, J., Wilensky, R. L. (1996) Measurement of coronary artery calcium with dual-slice helical CT compared with coronary angiography: evaluation of CT scoring methods. interobserver variations, and reproducibility. American Journal of Roentgenologv. 167, 439-444. Calhoun. P. S., Kuszyk, B. S., Heath, D. G., Carley, J. C., Fishman. E. K. (1999) Three-dimensional volume rendering of spiral CT data: theory and method. Radiographics, 19, 745-764. Daly, B., Templeton, P. A.. Krebs, T. L., Carroll, K., Wong You Cheong, J. J. (1998) Evaluation of biopsy needles and prototypic needle guide devices for percutaneous biopsy with CT fluoroscopic guidance in simulated organ tissue. Radiology, 209, 850-855. Dawson, P., Lees, W. R. (2001) Multi-slice technology in computed tomography. Clinical Radiology, 56, 302-309.

Next Page

De Mey, J., Op de Beeck, B., Meysman, M., et al (2000) Real time CT-fluoroscopy: diagnostic and therapeutic applications. European Journal of Radiology, 34, 32-40. Golding, S. J., Shrimpton, P. C. (2002) Radiation done in CT: are we meeting the challenge? British Journal of Radiology, 75, 1 -4. Johnson, C. D., Dachman, A. H. (2000) CT colonography: the next colon screening examination? Radiology, 216, 331-341. Johnson, P. T., Halpern, E. J., Kuszyk, B. S., et al (1999) Renal artery stenosis: CT angiography-comparison of real-time volume-rendering and maximum intensity projection algorithms. Radiology. 211, 337-343. Kalender. W. A. (1995) Thin-section three-dimensional spiral CT: is isotropic i maging possible? Radiology, 197, 578-580. Remy-Jardin, M., Remy, J. (1999) Spiral CT angiography of the pulmonary circulation. Radiology, 212, 615-636.


Previous Page

CONTENTS

Next Page

.



.

Previous Page

CONTENTS

The following illustrations show diagrammatically the dates of appearance of the primary and secondary ossification centres in the various regions of the body. Primary ossification centres are identified by capital letters; secondary ossification centres are identified by small letters.

Next Page

These three epiphyses fuse with one another in the 6th year and with the shaft in the 20th year.

Fig. A2 Ossification of bones of the elbow. A, 8th week of fetal life. B, 8th week of fetal life.


Fig. Al Ossification of bones of the shoulder. A, 8th week of fetal life. B, 5th and 6th weeks of fetal life, fuse about the 45th day of life. C, 8th week of fetal life. a. Head of humerus 1 year b. Greater tuberosity 3 years Lesser tuberosity 5 years c. Two epiphyses for acromion 15-18 years, fuse by 25th year d. Middle of coracoid process 1 year, fuses 15th year e. Root of coracoid process 17 years, fuses 25th year f. Inferior angle of scapula 14-20 years, fuses 22-25 years g. Medial border of scapula 14-20 years, fuses 22-25 years h. Sternal end of clavicle 18-20 years, fuses 25th year.

a. b. c. d. e. f.

External epicondyle 10-12 years, fuses 17-18 years Internal epicondyle 5-8 years, fuses 17-18 years Capitellum 1-3 years, fuses 17-18 years Head of radius 5-6 years, fuses 16-19 years Trochlea 1 1 th year, fuses 18th year Olecranon 1013 years, fuses I6-20 years.

Previous Page 1848

CONTENTS

Next Page


Fig. AS

Ossification of bones of the

knee. A, 7th fetal week. B, 8th fetal

week. C, 5 years. a.

At birth, fuses 20th year

b. Tibia) tubercle 5-10 years, fuses 20th year c. 4th year. fuses 25th year d. At birth, fuses 20th year. Fig. A3 Ossification of the wrist, carpus, and hand. A, Capitate 4 months. B, Hamate 4 months. C, Triquetral 3 years. D, Lunate 4-5 years. E, Trapezium 6 years. F, Trapezoid 6 years. G, Scaphoid 6 years. H, Pisiform 11 years. I, Metacarpals 10th fetal week. J, Proximal phalanges 11th fetal week. K, Middle phalanges 12th fetal week. L, Distal phalanges 9th fetal week. M, Middle phalanx of 5th digit 14th fetal week. a.

Lower end of radius 1-2 years, fuses 20th year

b.

Lower end of ulna 5-8 years, fuses 20th year

c.

Metacarpal heads 2; years, fuse 20th year

d.

Base of proximal phalanges 2; years, fuse 20th year

e.

Base of middle phalanges 3 years, fuse 18-20 years

f. g.

Base of distal phalanges 3 years, fuse 18-20 years Base of I st metacarpal 2z years, fuses 20th year.

Fig. A6

Ossification of bones


of the foot. A, Calcaneus 6th

fetal month. B, Talus 6th fetal month. C, Navicular 3-4 years. D, Cuboid at birth. E, Lateral

cuneiform 1 year. F, Middle cuneiform 3 years. G, Medial cuneiform 3 years. H, Metatarsal shafts 8-9th fetal week. I, J, K, Phalangeal shafts 10th week.

a. Metatarsal epiphyses 3 years, fuse 17-20 years

Fig. A4

Ossification of bones of the hip. A, 7th fetal week.

a.

1 st year, fuses 18-20 years

b.

3-5 years, fuses 18-20 years

c.

8-14 years, fuses 18-20 years.

b. Proximal phalangeal base 3 years, fuse 17-20 years c. Middle phalangeal base 3 years, fuse 17-20 years d. Distal phalangeal base 5 years, fuse 17-20 years e. Posterior epiphysis of calcaneus 5th year, fuses at puberty.

CONTENTS

Previous Page

Next Page

Note: Atlas: 3 centres; one for each lateral mass and one for the anterior arch-unite 6-8 years Axis: 5 primary centres; one for the body, two for the arch, two for the odontoid process 2 secondary centres; one for the tip of the odontoid, one for the i nferior rim of the body 7th Cervical vertebra: extra primary centre for each costal process Lumbar vertebrae: have two extra secondary centres for the mamillary processes.

Fig. A7 Ossification of bones of the pelvis. A, Ilium 3rd fetal month. B, Ischium 4th fetal month. C, Pubis 5th fetal month. Note: i schiopubic ramus fuses at 7th year. a. b. c. d. e.

Detailed assessment of bone age is discussed in Greulich, W. W., Pyle, S. I. (1959) Radiographic Atlas of Skeletal o,/the Development Hand and Wrist. Stanford University Press.

cartilage-2

or more centres Y-shaped Iliac crest Anterior inferior iliac spine Pubic symphysis Ischial tuherosity.

Appear about puberty-fuse 20-25 years.


Fig. A8 Ossification of vertebrae. A, Body 8th fetal week. B, Neural arch-one for each side-8th fetal week.

a. h. c. d.

Transverse process Spinous process Upper surface of the body Lower surface of the body

All appear at about 16 years and fuse by the 25th year.

Previous Page

CONTENTS

Next Page

APPENDIX B

The terms included in this glossary are some of those commonly used for CT scanning and subsequent image manipulation. Some of the terms are common to the manipulation of other digital images, being generic to image manipulation rather than to CT. Other terms have accepted clinical, radiological or anatomical meanings but are included here as they carry specific connotations when applied to CT scanning. Some terms vary between manufacturers. Algorithm Computer program which modifies the raw data to alter the characteristics of the final image; e.g. bone or high spatial frequency algorithm gives high spatial resolution to the image but increases image noise, soft-tissue or low spatial frequency algorithm reduces image resolution but reduces image noise. Artefact A fault in the image not representing an object in the scanned volume. Artefacts may be generated by the imaging equipment, by patient motion during scanning, or by many other confounding factors. Back projection The computer technique used to convert the raw CT data i nto all image. Filtered back projection is performed, the digital filtration alters the data prior to back projection in order to optimise the image spatial resolution and noise (see Algorithm) and to reduce artefacts. Collimation The use of lead shutters to control the effective thickness of the X-ray beam and thus the slice thickness. Prepatient collimation is next to the X-ray tube. Positioning of the shutters requires high precision and accuracy as the beam will diverge, thus the shutter collimation has to be less than the nominal slice thickness. Prepatient collimation markedly affects patient radiation dose. Postpatient collimation is next to the detectors. This will reduce scatter and also controls the nominal slice thickness. Thinsection CT can be more easily acquired with postpatient collimation, but accurate alignment between pre- and postpatient collimation is required to optimise radiation (lose. Cone beam Descriptive term for the shape of the 'swept volume' passed through by the CT beam for a detector row away from the centre of the fanbeam on multislice CT. Because this beam is not perpendicular to the detectors, it generates image artefacts. Corrections for the 'cone beam geometry' of the data acquisition from multislice scanners are under development to enable wider banks of detectors to further increase the volume scanned per rotation. Contiguous Sections which are nominally immediately adjacent to each other. Contrast resolution Ability to resolve structures according to density difference. Influenced on CT by the noise in the image, improved by higher mAs (radiation dose). Convolution Computer manipulation of raw data to generate an image (usually by back projection). CTDI CT dose index. A measurement of radiation dose per 100 mAs on tile surface of. 1 cm into or at the centre of a plexiglass cylinder 16 or 32 cm in diameter. Used to make comparisons of dose between scanners as simple mAs exposure levels do not give comparable doses between different scanners. CT fluoroscopy Rapid serial display of images from a single-slice location to provide an image updated between one and several times a second, used to guide interventional procedures by CT. Fast image reconstruction algorithms, often without the usual correction steps and utilising some of the

already back-projected data from the previous image, can result in a moving i mage in `near real-time', less than 250 ms later. Detector Hardware used to convert incident X-rays into electrical impulses. Xenon gas detectors are stable and relatively inexpensive hut only approximately 60% efficient. Solid-state crystal detectors arc expensive and some are less stable but are over 99% efficient in converting X-rays to electrical i mpulses. Higher conversion efficiency can allow radiation dose reduction. Direct coronal Patient positioning to allow scans to occur in the coronal plane. Direct sagittal Patient positioning to allow scans to occur in the sagittal plane. Dose-radiation dose Measurement of radiation dose on CT is more complicated than for other radiological techniques as the patient is irradiated from all directions around the circumference of the scan (unless a partial scan is performed). Calculation of the effective dose can he performed from a combination of measurements and computer simulations. Radiation doses for CT tend to be higher than for most other diagnostic imaging techniques. While only 2% of radiological examinations are by CT, they account for 20% of the population diagnostic radiology dose in the UK. Dynamic scan Rapid scan acquisitions, which may all be at the sank site over a period of time (dynamic sequential scan) or with table increments between each scan (dynamic incremental scan). Dynamic scanning on conventional equipment is being superseded by spiral scanning. Flying focal spot Electromagnetic oscillation of the X-ray tube focal-spot position during scanning, performed at every data acquisition point during gantry rotation, give two sets of data from slightly different positions for each gantry position. consequently improving spatial resolution to a level that would otherwise require twice as many detectors, each of half the front face area. Gantry The scanner gantry contains the X-ray tube and detectors mounted around a central tunnel through which the patient and table pass. The gantry can be angled away from the vertical over a limited range. Some spiral scanners also have the generator mounted on the gantry. Helical scanning Correct terminology for what is usually referred to as spiral scanning, where the use of slip rings allows the X-ray tube and detector assembly to rotate continuously in one direction at 0.5-I revolution per second. Continuous smooth table motion results in a helical path of the X-ray beam through the patient with no pauses between adjacent loops of the helix. Images can be reconstructed centred on any point in the helix, allowing overlapping sections to be generated. Hounsfield unit (HU) This is the corrected or accurate value of the CT number, measured from the scanner image, although CT number and Hounsfield unit are often used interchangeably. The Hounsfield unit is defined by the relationship between the linear attenuation value of the material being scanned and that of water. A vacuum has a Hounsfield unit value of -1000, while water has a value of zero. The upper end of the scale is open ended but image data storage limitations usually result in a scale from -1024 to +3096 (12 hit). HRCT High-resolution computed tomography usually applied to pulmonary i maging. See CT Scanning of the Lungs, Chapter I. Interpolation A mathematical process used to estimate the value to ascribe to a point lying between known measurements. Interpolation is used in spiral


1850

Previous Page

CONTENTS

scanning using data Itom the adjacent loops of a helix to produce a correction for the helical scan path. Matrix A grid of picture elements (pixels) from which the CT image is coot posed. Most scanners use 5 12 x 512 pixel matrices: 256 x 256 and 320 x 320 are becoming less commonly used as scanner resolutions improve and data storage media get cheaper. Many scanners now have display consoles with 1024 x 1024 matrices. Maximum intensity projection Computed image derived from a stack of CT i mages which displays the slack from any direction, with the image density related to the peak pixel value seen from that direction throughout the stack volume. This technique is used to demonstrate vascular enhancement in an ' angiogram-like' way. Minimum i ntensity projection As for maximum intensity projection but with i mage density related to the minimum pixel value. Multislice CT CT scanner in which the detector bank is split into several adjacent rows so that the data acquisition from a single gantry rotation can he used to produce multiple adjacent thin slices from a beam collimated to one thicker section. Pitch (Pitch prime) The usual definition of pitch (see above) can Cause contusion when applied to nmItislice CT. A scanner with a 2 cut wide beam hitting four 5 mm detector rows to give four simultaneous sections with a t able increment of 2 cm per rotation has a pitch of 4 (table increment over i ndividual detector width) but a pitch prime of I (table increment over nominal beam width). Pitch factor In spiral scanning, this is the ratio of the distance travelled by the t able per rotation of the gantry to the X-ray beam collimation. A pitch factor of I is standard: increasing the pitch actor to 1.25 or 1.5 results in thicker effective slices, more image noise, greater image artefacts, and lower patient radiation dose. It requires greater interpolation of data but allows a shorter scan time or an extended scan volume. Pixel Picture clement in the inutge display (cf. Voxel). Postprocessing Manipulation of the image data. Preprocessing Manipulation of the 'raw data' to enhance certain features it) subsequent image reconstructions. QCT Quantitative computed lomography-the use of a calibration procedure to identify the values of the CT numbers in an image in terms of a reference material. QCT is most commonly used to measure the mineral density of trabecular hone in the spine. Techniques have been described using two different scan energies (kVp) to allow for correction of the error introduced by the presence of marrow fat in trahecular hone (dual energy QCTDEQCT). QCT can also he calibrated to measure the iron content of the liver. Real-time CT Rapid image reconstruction such that the image is available almost as soon as the data acquisition for that slice is finished. Sec CT fluoroscopy.

Next Page

Reconstruction The production of an image floor the 'raw data'. Reformation (reformatted image) The generation of images in an alternative plane from the original CT images. Reformats can now he in any plane or even curved, as computer speeds have improved. Multiple overlapping sections such as those obtained froin spiral acquisitions produce markedly i mproved reformat image quality. Region of interest (ROI) Any Part of a CT image on which measurements are being made-particularly measurements of the mean CT number. Scout view CT scanner digital radiograph acquired by keeping the X-ray tube and detectors stationary whilc moving the table through the gantry. The resultant image can he used to identify the start and stop points for the subsequent CT sections. Segmentation Removal of part of each image in a stack in order to allow subsequent manipulations to be applied to the area of interest, e.g. removing the spine from an angiographic CT to allow MIP projections without the bone densities obscuring the vascular contrast densities. Slice thickness The nominal thickness of the slice which the image represents. As the X-ray beam is divergent, the sampled volume in the patient is t hicker around the edges than centrally. The alignment of pre- and postpatient collimation may also affect the true slice thickness. Image noise is reduced with thicker slices but partial volume averaging effects become apparent in larger structures. Slip ring A component of the gantry transferring data or electrical energy to or from the stationary (fixed) part of the scanner to the rotating part of the gantry, allowing continuous rotation of the gantry. Low-voltage slip rings can supply current to an onboard generator. high-voltage slip rings allow the generator to he separate from the gantry and supply high kV to the tube. Spatial resolution Ability to resolve structures according to size. Formally assessed by calculation of the modulation transfer function. Influenced on CT by choice of field of view (targeting), mAs (radiation dose) and reconstruction algorithm (bone or high spatial frequency algorithm). Spiral Sec Helical. Surface rendered A computer-generated image of the surface defined by a specified minimum pixel intensity in a stack of images. Used to generate a 'three-dimensional' image. Surface shaded Computer enhancement of a three-dimensional CT image to give a 'realistic' shaded effect to the surface of the object. Voxel The volume within the scan plane that is represented by a pixel in the t riage. Window The window width is the range of CT numbers that will occupy the entire greyscale of the display monitor: the window level is the CT number at the midpoint of that range. Altering the width and level therefore alters the contrast and brightness of the image, allowing the entire image density range to he interrogated on the monitor.


Previous Page

CONTENTS

Next Page

APPENDIX C

Aliasing An artefact that occurs, as a result of the nature of the image encodi ng process, when the field of view (FOV) is smaller than the area being i maged. There is folding-over of the anatomy outside the FOV hack into the i mage. Also called 'fold-over' or 'wrap-around' artefact. Angular frequency Frequency of oscillation or rotation (measured in radians/second) represented by the Greek symbol omega (w). w = 27rv, where v is the frequency in Hertz. Artefact An error that occurs in the reconstructed image which does not correspond to any anatomical region or pathological lesion within the patient. There are many sources of artefact in MRI. Averaging A method for improving the signal-to-noise ratio (SNR). The data lines in k-space are acquired multiple times by repeat acquisitions of the sane phase encoding steps. The same MR signal is added up and the sum divided by the number of signals (i.e. number of signal excitations (NEW)/ signal averages (NSA)). B„ The symbol used for the static main magnetic field in the MR system measured in Tesla. The orientation of this field is along the z axis. Il l The symbol used to denote the iadiofrequency (RF) magnetic field in an MR system measured in TesIA Bandwidth A measure of the range of frequencies within which the MR system is tuned to receive the signal (receiver bandwidth) or the range of frequencies within an RF pulse delivered by the transmitter (transmitter bandwidth). Alteration in the receiver bandwidth affects the SNR, with narrowing of the bandwidth increasing SNR and vice versa. RF pulses of different bandwidths arc used for various applications within pulse sequences. Coil Single or multiple loops of wire (or other electrical conductor such as tubing) designed to either produce (transnit) a magnetic field from current flowing through the wire or measure (receive) an induced voltage in the loop arising from a changing magnetic field. Diamagnetic A substance that will slightly reduce the strength of a magnetic field in which it is placed. The polarity of the magnetisation induced in a diamagnetic substance (most organic material) is opposite to that of the surrounding magnetic fields, and has negative magnetic susceplibi/itv. These substances have no unpaired orbital electrons. Diffusion A process whereby water and other sniall molecules in a tissue undergo random microscopic translational movement from Brownian motion (thermal processes). MR is a sensitive technique for measuring diffusion effects, e.g. in strokes, differentiating cysts from solid tumour and in CSF dynamic studies. Echo rephasing The re-establishment of MR signal coherence which is achieved with either a 180° RF pulse or by gradient switching. Echo train length (ETL) This is the number of echoes individually phase encoded for a fast (turbo) spin-echo sequence. The ETL corresponds to the number of lines of k-space measured per repetition tune (TR) interval. ETLs can range typically from 3 to 128 depending on the pulse sequence type. Eddy currents Small electrical currents induced by the changing magnetic fields within the gradients coils or the structure of the magnet. These i nduced electrical currents degrade image quality unless either properly compensated for or cliniinated.

Even echo rephasing Re-establishment of spin-echo coherence of moving spins on symmetric even echoes (e.g. 2, 4, 6, etc.) in multiecho sequences as a result of sequential integration of signal phase shifts summing to zero. Fat/water suppression A method that suppresses the signal within the i maging volume from cither fat or water protons by application of a frequency-selective saturation RF Pulse. Ferromagnetic A substance, such as iron, cobalt or nickel, that is attracted by a magnetic field and retains its magnetism once removed from the effects of the magnetic field. Flip angle The angle through which the magnetisation vector moves, relative to the longitudinal axis of the static magnetic field, as a result of the application of an RF pulse. This variation in flip angle is used in gradient-echo i maging to obtain the various tissue-weighted images (a low flip angle (typi cally 10-30°) produces a T,-weighted image and a 90° flip angle pulse provides T i - weighting). Flow-related enhancement A process by which the signal intensity of moving fluids can be increased as compared with signal from stationary tissue; it occurs when in-flowing, unsaturated fully magnetised spins replace saturated spins within the imaging slice between successive RF pulses. Fourier transform A mathematical process by which the frequency components of a signal are separated from its amplitudes as a function of time, and vice versa. Free induction decay (FID) A transient signal that occurs as the transverse magnetisation decays towards zero following application of an RF pulse. Gauss (G) An old unit used for measuring magnetic field strength. The internationally accepted unit is the Tesla ( I Tesla = 10 000 Gauss). The earth's magnetic field is approximately 0.5 Gauss. Gradient coils These are magnetic coils designed to alter the main magnetic field (in order that the magnetic field is stronger in some areas compared with other parts) by a few per cent. The magnetic field gradient generated by a coil is controlled by the electrical current passed through the coil. They are used to localise a slice and spatially encode slice information. Gradient echo (GE/GRE) A basic pulse sequence which only uses magnetic field gradient reversal to rcphase the transverse magnetisation and produce echoes of the MR signal. This allows shorter repetition tines thus faster scanning, and flip angles less than 90° (see flip angle). I mage acquisition time This is the scanning time to produce a set of images from a measurement sequence. For a 21) sequence it is the product of the repetition time (TR), number of signal excitations/averages ( NEX /NSA ). and the number of phase encoding steps (N,). For a 3 1) volume sequence the acquisition tittle is as for the 2D sequence together with the product of the number of partitions (z phase encoding steps, N,). For fast (turbo) 2D sequences the acquisition time is the product of the 2D sequence divided by t he echo train length (ETL). Inhomogeneity The slight variation in uniformity of the static magnetic field expressed in parts per million (ppm) as a fractional deviation from the average value of the field. Inversion recovery (IR) A basic pulse sequence which inverts the magnctisation and measures the recovery rate as the nuclei return to equilibrium. This rate of recovery is dependent on the T i relaxation rate.


1852

Previous Page

CONTENTS

Inversion time (TI) The time front the centre of the first invention pulse (180°) to the beginning of the second (90") excitation pulse in an IR sequence. K-space The .space which is filled with information Ilet - nicel r ate duty matri.C) and undergoes a Fourier transform to form the MR image. The points al the centre of this data matrix represent low spatial frequencies (which give gross signal levels, i.e. contrast in the MR image). Increasing the offset from the centre corresponds to higher spatial frequencies (edge detection, i . e. image sharpness). By manipulating K-space faster sequences can he i mplemented as in fast ( or 1111-ho) i maging. Larmor equation w = yB„ '['he expression of the proportional relationship between the precessional angular frequency of a nuclear magnetic moment ( w. expressed in Hertz) and the main magnetic field (B,,, in Tesla). The gyrolttagnetic ratio (y) is a proportionality constant fixed for the nucleus. being 42.6 MHz/Tcsla for hydrogen. For B„ of 1.5 T w=42.6 x 1.5 = 64 MHz. Longitudinal magnetisation (M,) The part of the macroscopic magnetisation vector parallel to the main static magnetic field (B„). Following RF excitation the M, returns to its equilibrium value (M„) as a function of the characteristic tittle constant T, of the tissues excited. Longitudinal relaxation time See T 1 . Magic angle artefact I n joint imaging if a tendon is orientated at a certain angle (55" ) to the static main magnetic field, the tendon appears brighter on T 1 - and proton density (intermediate)-weighted images but normal low signal on T,-weighted images. This artefactnal increase in signal within the magic tendon is termed the angle c//sect and can he potentially confused with pathology. This effect is due to the highly ordered structure of tendons, which arc composed of type I Collagen. This ordered structure restricts the motion and orientation of water protons, with consequent reduction in the dipolar interaction and thus an increase in signal. Magnetic moment A measure of the magnitude and direction (vector quantity) of the magnetic properties of an object or particle that cause it to align with the static main field and create its own local magnetic field. Magnetic resonance ( MR) The absorption of emission of electromagnetic energy by nuclei (hydro('en protons in imaging) in a static magnetic field following excitation by a RF pulse. The resonant frequency of the RF pulse and entitled signal is proportional to tile magnetic field (the relationship given by the Larntor equation). Magnetic resonance signal The electromagnetic signal produced by the precession of the transverse magnetisation of the spins. The rotation of the transverse magnetisation induces a voltage in the receiver coil. It is this voltage that is amplified by the receiver and forms the signal. Magnetic susceptibility The ability Of a substance to become magnetised or to distort a magnetic field, denoted as the Greek symbol X. This is described in terms of diarrutgaene, paramagnetic, xuperparamugnetk and ferromagnetic materials and refers to their respective electronic magnetic moments. Magnetisation transfer contrast (MTC) This is a relatively new technique in which the image contrast is Manipulated by selectively saturating a pool of protein-hound water. It should be noted that protons in the protein-bound water have a resonant frequency, approximately 500-2500 Hz away front t he bulk water protons. By applying an off-resonance pulse (i.e. the centre frequency of the pulse being 1000-2000 Hz removed front the Larntor frequency for protons together with a wide bandwidth) these protons are suppressed. Because the protein-bound water and the bulk water protons are in rapid exchange, the saturation is transferred to the bulk phase of the water protons. This leads to a reduction in the signal from hulk water. MTC i s used in the following situations: (i) in MRA of the brain as a means of enhancing visualisation of small peripheral vessels and aneurysms by Suppressing the background brain tissue; (ii) in conjunction with gadolini um chclate enhancement, staking the enhanced areas store conspicuous: (iii) combined with T,-weighted images in the detection of early dentyclination or protein destruction. Maximum intensity projection (MI P) An algorithm for producing multiple projections from a 2D or 31) volume data set. This volume data set is processed along selected angles in which the highest signal intensity pixel i s projected as a 2D image. This technique is used ill MRA , whereby flowing blood has a high signal intensity, and in MR cholangiopancreat ography (MRCP) in which bile is of high signal. MRA Magnetic resonance angiography. NEX Number of signal excitations repeated in a given acquisition. This is a way of increasing SNR at the expense of time. Nuclear spin An intrinsic property of certain nuclei (i.e. those with an odd number of neutrons and/or protons, e.g. IH, "F 2 ' Na 13 C, 110) that gives than angular ntontentunt and magnetic moment. The spins of nuclei have

Next Page

characteristic fixed values. Nuclei of paired neutrons and protons align to cancel out their spins and thus do not resonate. Off-centre field of view A field of view (FOV) that is not centred at the i soccntre of the magnet. Paramagnetic This describes a substance (e.g. gadolinium, ntethacmoglobin) with positive magnetic susceptibility (opposite of diamagnetic). The substances align themselves with the static magnetic field and allect the rcla.xation times of tissues containing them. Gadolinium-chelate contrast agents are used for this effect, reducing the T, relaxation times (with subsequent i ncrease in signal) of tissues containing them. Phase-contrast (PC) angiography A 21) or 3D imaging technique relying on velocity induced phase shifts to distinguish [lowing blood front stationary tissues. This technique relies on the fact that the phase gain of flowing blood is proportional to its velocity, assuming constant velocity. The mugnil ode i mage shows the vasculature with the direction of flow given by the phase i mage (also called phase map). Phase encoding This allows the localisation of an MR signal by applying a series of varying phase encode gradient pulses in order to alter the phase of spins prior to signal readout. The spins themselves retain memory of the effect of the separate phase encode pulses. Pixel An acronym for picture element describing the smallest component of a digital inmage display. Proton density (p rho) weighted image These intages show contrast related to the number of mobile protons in the structure and require scanning parameters that minimise the effects of T, (long TR) and T, (short TE) to obtain the appropriate weighting. Images usually have T,, T, and hulk flow contributions, however, and are thus better termed r ims rruediatr' ( between T, and T,) weighted. Repetition time (TR) The time between successful excitations Of a slice (i.e. t he time front the beginning of one RF pulse sequence to the stmt of the next). In cardiac Bated studies this can vary depending on the heart rate of t he patient and as such an e//e(tive repetition time is obtained. Saturation A non-equilibrium state in which equal numbers of spins are aligned with and against the magnetic field. This occurs immediately foll owing a 90" RF pulse with the longitudinal magnetisation aligned in the x-y (transverse) plane and thus the system is said to he saturated. A few moments later. with some T, recovery, the system is partially saturated. Saturation effects are used in the slice direction to minimise signal from following blood. Signal-to-noise ratio (SNR) The ratio of signal amplitude to the noise. Methods of improving the SNR arc: increasing the another of signal excitations (N EX) with subsequent time penalty; increasing the field of view ( FOV) with corresponding reduction in spatial resolution of the resultant i mage, and increasing the strength of the main magnetic field used. Spin echo (SE) The reappearance of an MR signal, after the initial signal ( FID) has disappeared following a 90" RF pulse, by the application of a refocusing 180° RF pulse. This refocusing Pulse results in an effective reversal of the dephasing of the spins by eliminating the elects of inhoniogencity in the main magnetic field. Superparamagnetic This describes a substance (c.g. haeniosiderin. Superparamagnetic iron oxide liver. lymph node. and oral bowel contrast agents) with magnetic susceptibilities 100-1000 times stronger than paramagnetic substances. These substances produce marked shortening of the T_, relaxation time, producing areas of signal void. T r relaxation ti me Also called the /anginulinu/ or spin-/office relaxation t i me, referring to the time taken for the spins to g i ve the energy obtained from the initial RF pulse hack to the surrounding lattice (environment( and return to equilibrium. T, relaxation time represents the tints required for t he longitudinal magnetisation (M,) to increment from 0 to 63ki of its final maximum value. T, relaxation time This is termed tran.sverse or spin-spin relaxation time, and is the characteristic time constant for loss of phase coherence among spins orientated al an angle to the static main magnetic field. This time constant arises from interaction between the spins (hence the term 'spin-spin'). This can he mathematically represented by the time required for file transverse magnetisation (M, or M v ) to decay to about 37'/ of its maximum value. TE This refers to the echo time, i.e. the time between the centre of the excitation pulse and the peak of the echo. TI This refers to the inversion time and represents the tine after the middle of t he 180" RF inverting pulse in the IR sequence to the middle of the 90" read pulse. This time monitors the amount of longitudinal magnetisation ( M,) and cam be varied in the I R sequence. e.g. as i n STIR (short TI (or tau)


Previous Page 1854

CONTENTS

Next Page


I nversion Recovery) where the TI interval is short, being at the null point (zero signal) for fat; and FLAIR (Fluid Attenuation Inversion Recovery) where the TI interval is very long, being at the null point for fluids (e.g. CSF). Time-of-flight (TOF) angiography Also referred to as in-flow angiography. It is used in 2D or 3D mode based on the principle of flow-related enhancement, i.e. blood flowing into a slice appearing brighter than surrounding tissue because it has not become saturated by previous RF pulses. Voxel An acronym for volume element, being a 3D region of an imaged object represented in 2D by a pixel.

Adamis, M. K. (1996) Glossary of MR terms. In Edelman, R. E.. Hesselink, .1. R., Zlatkin. M. B. (eds) Clinical Magnetic Resonance Imaging, 2nd edn, pp. 2177-2196. Philadelphia: WB Saunders. am/A nswers in Magnetic Elster, A. D., Burdetle J. H. (2001 ) Questions Resonance Imaging. St Louis: Mosby. Floyd. L. J., Williams, R. F., Stark, D. D. (1999) Glossary. In Magnetic Resonance Imaging, 3rd edn. St Louis: Mosby. Hashemi, R. H., Bradley, G. W. (1997) MRI: The Basics. Bahimore Williams & Wilkins.


CONTENTS

Previous Page

Next Page

S1

Cr

EDTA

GFR measurement

IV

3

0.006

0.008

No

67

Ga

Gallium citrate

I nfection /Inflammation and tumour imaging

IV

150

1 7

12

Stop

75

Se

23-Seleno-25homo-taurocholate (SeHCAT)

Bile salt absorption

Oral

0.4

0.3

03

No

Krypton gas

Lung ventilation

Inhalation

6000

0.2

Textbook of Radiology and Imaging (Vol. 2)

Textbook of Radiology and Imaging (Vol. 1)

Emergency Radiology: Imaging and Intervention

Head and Neck Imaging (2 Vol set )

Learning Vascular and Interventional Radiology (Learning Imaging)

Radiologia-Radiology-Advanced Imaging Of The Abdomen

Radiology Fundamentals: Introduction to Imaging and Technology

Percutaneous Tumor Ablation in Medical Radiology (Medical Radiology Diagnostic Imaging)

New Practical Chinese Reader, Textbook Vol. 2

Textbook of in vivo Imaging in Vertebrates

Musculoskeletal Imaging Cases (McGraw-Hill Radiology)

Paediatric Imaging Manual (Notes in Radiology)

Urogenital Imaging (Direct Diagnosis in Radiology)

The ESC Textbook of Cardiovascular Imaging

Combat Radiology: Diagnostic Imaging of Blast and Ballistic Injuries

Diagnostic Nuclear Medicine (Medical Radiology Diagnostic Imaging)

Cardiac Imaging (Direct Diagnosis in Radiology)

Pediatric Radiology Review (Contemporary Medical Imaging)

Essential Radiology: Clinical Presentation Pathophysiology Imaging

Kelley's Textbook of Rheumatology, 7th Edition (2 Vol. Set)

Textbook of Dr. Vodder's Manual Lymph Drainage: Vol 2 Therapy

Radiology

Case Studies in Medical Imaging: Radiology for Students and Trainees

Vascular and Interventional Radiology; 2 edition

Duke Radiology Case Review: Imaging, Differential Diagnosis, and Discussion

Radiology

Head and Neck Cancer Imaging, Second Edition (Medical Radiology Diagnostic Imaging)

Textbook of Radiology and Imaging (Vol. 2)

Textbook of Radiology and Imaging (Vol. 1)

Emergency Radiology: Imaging and Intervention

Head and Neck Imaging (2 Vol set )

Learning Vascular and Interventional Radiology (Learning Imaging)

Radiologia-Radiology-Advanced Imaging Of The Abdomen

Radiology Fundamentals: Introduction to Imaging and Technology

Percutaneous Tumor Ablation in Medical Radiology (Medical Radiology Diagnostic Imaging)

New Practical Chinese Reader, Textbook Vol. 2

Textbook of in vivo Imaging in Vertebrates

Musculoskeletal Imaging Cases (McGraw-Hill Radiology)

Paediatric Imaging Manual (Notes in Radiology)

Urogenital Imaging (Direct Diagnosis in Radiology)

The ESC Textbook of Cardiovascular Imaging

Combat Radiology: Diagnostic Imaging of Blast and Ballistic Injuries

Diagnostic Nuclear Medicine (Medical Radiology Diagnostic Imaging)

Cardiac Imaging (Direct Diagnosis in Radiology)

Pediatric Radiology Review (Contemporary Medical Imaging)

Essential Radiology: Clinical Presentation Pathophysiology Imaging

Kelley's Textbook of Rheumatology, 7th Edition (2 Vol. Set)

Textbook of Dr. Vodder's Manual Lymph Drainage: Vol 2 Therapy

Radiology

Case Studies in Medical Imaging: Radiology for Students and Trainees

Vascular and Interventional Radiology; 2 edition

Duke Radiology Case Review: Imaging, Differential Diagnosis, and Discussion

Radiology

Head and Neck Cancer Imaging, Second Edition (Medical Radiology Diagnostic Imaging)

Recommend Documents