Echocardiography inPediatric and CongenitalHeart Disease

Dedications

To my parents, CheToo and SoFa Lai; my wonderful wife, Lydia; and my children, Justin and Amanda.Wyman W. Lai

To my wife Benedikte, my daughter Virginie and my son Francis. For all the time I could not spend with them.Luc L. Mertens

To my husband Bruce Randazzo and my children Jake, Isabel and Ethan for supporting me always.Meryl S. Cohen

To my wife Judith and sons Omri and Alon with Love.Tal Geva

Echocardiography inPediatric and CongenitalHeart DiseaseFrom Fetus to Adult

Second Edition

Edited by

Wyman W. Lai MD, MPHProfessor of Pediatrics at CUMCColumbia University Medical Center;Director, Noninvasive Cardiac ImagingMorgan Stanley Children’s Hospital of NewYork-PresbyterianNew York, NY, USA

Luc L. Mertens MD, PhDSection Head, EchocardiographyThe Hospital for Sick Children;Professor of PediatricsUniversity of TorontoToronto, ON, Canada

Meryl S. Cohen MDProfessor of PediatricsPerelman School of Medicine, University of Pennsylvania;Medical Director, EchocardiographyProgram Director, Cardiology FellowshipThe Cardiac CenterThe Children’s Hospital of PhiladelphiaPhiladelphia, PA, USA

Tal Geva MDProfessor of PediatricsHarvard Medical School;Chief, Division of Noninvasive Cardiac ImagingDepartment of CardiologyBoston Children’s HospitalBoston, MA, USA

This edition first published 2016 © 2016 by John Wiley & Sons Ltd.First edition published 2009 © 2009 by John Wiley & Sons Ltd.

Registered office: John Wiley & Sons, Ltd, The Atrium, Southern Gate, Chichester, West Sussex,PO19 8SQ, UK

Editorial offices: 9600 Garsington Road, Oxford, OX4 2DQ, UKThe Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK111 River Street, Hoboken, NJ 07030-5774, USA

For details of our global editorial offices, for customer services and for information about how to apply for permission to reuse thecopyright material in this book please see our website at www.wiley.com/wiley-blackwell

The right of the author to be identified as the author of this work has been asserted in accordance with the UK Copyright, Designsand Patents Act 1988.

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by anymeans, electronic, mechanical, photocopying, recording or otherwise, except as permitted by the UK Copyright, Designs and PatentsAct 1988, without the prior permission of the publisher.

Designations used by companies to distinguish their products are often claimed as trademarks. All brand names and product namesused in this book are trade names, service marks, trademarks or registered trademarks of their respective owners. The publisher is notassociated with any product or vendor mentioned in this book. It is sold on the understanding that the publisher is not engaged inrendering professional services. If professional advice or other expert assistance is required, the services of a competent professionalshould be sought.

The contents of this work are intended to further general scientific research, understanding, and discussion only and are not intendedand should not be relied upon as recommending or promoting a specific method, diagnosis, or treatment by health sciencepractitioners for any particular patient. The publisher and the author make no representations or warranties with respect to theaccuracy or completeness of the contents of this work and specifically disclaim all warranties, including without limitation anyimplied warranties of fitness for a particular purpose. In view of ongoing research, equipment modifications, changes in governmentalregulations, and the constant flow of information relating to the use of medicines, equipment, and devices, the reader is urged toreview and evaluate the information provided in the package insert or instructions for each medicine, equipment, or device for,among other things, any changes in the instructions or indication of usage and for added warnings and precautions. Readers shouldconsult with a specialist where appropriate. The fact that an organization or Website is referred to in this work as a citation and/or apotential source of further information does not mean that the author or the publisher endorses the information the organization orWebsite may provide or recommendations it may make. Further, readers should be aware that Internet Websites listed in this workmay have changed or disappeared between when this work was written and when it is read. No warranty may be created or extendedby any promotional statements for this work. Neither the publisher nor the author shall be liable for any damages arising herefrom.

Library of Congress Cataloging-in-Publication Data

Echocardiography in pediatric and congenital heart disease : from fetus to adult / edited by Wyman W. Lai, Luc L. Mertens,Meryl S. Cohen, Tal Geva. – Second edition.

p. ; cm.Includes bibliographical references and index.ISBN 978-0-470-67464-2 (cloth)I. Lai, Wyman W., editor. II. Mertens, Luc, editor. III. Cohen, Meryl, editor. IV. Geva, Tal, editor.[DNLM: 1. Heart Defects, Congenital–ultrasonography. 2. Echocardiography–methods. 3. Heart Defects,

Congenital–diagnosis. WG 141.5.E2]RJ423.5.U46618.92′1207543–dc23

2015033801

A catalogue record for this book is available from the British Library.

Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronicbooks.

Cover images: Reproduced from images within the book.

Set in 9.5/12pt MinionPro by Aptara Inc., New Delhi, India

1 2016

Contents

Contributors, vii

Preface, xi

About the Companion Website, xii

Part I Introduction to Cardiac Ultrasound Imaging

1 Ultrasound Physics, 3Jan D’hooge and Luc L. Mertens

2 Instrumentation, Patient Preparation, and Patient Safety, 19Stacey Drant and Vivekanand Allada

3 Segmental Approach to Congenital Heart Disease, 31Tal Geva

4 The Normal Pediatric Echocardiogram, 44Wyman W. Lai and H. Helen Ko

Part II Quantitative Methods

5 Structural Measurements and Adjustments for Growth, 63Thierry Sluysmans and Steven D. Colan

6 Hemodynamic Measurements, 73Mark K. Friedberg

7 Systolic Ventricular Function, 96Luc Mertens and Mark K. Friedberg

8 Diastolic Ventricular Function Assessment, 132Peter C. Frommelt

Part III Anomalies of the Systemic and PulmonaryVeins, Septa, and Atrioventricular Junction

9 Pulmonary Venous Anomalies, 157David W. Brown

10 Systemic Venous Anomalies, 180Leo Lopez and Sarah Chambers

11 Anomalies of the Atrial Septum, 197Tal Geva

12 Ventricular Septal Defects, 215Shobha Natarajan and Meryl S. Cohen

13 Ebstein Anomaly, Tricuspid Valve Dysplasia, and RightAtrial Anomalies, 231Frank Cetta and Benjamin W. Eidem

14 Mitral Valve and Left Atrial Anomalies, 243James C. Nielsen and Laurie E. Panesar

15 Common Atrioventricular Canal Defects, 259Meryl S. Cohen

Part IV Anomalies of the Ventriculo-arterial Junctionand Great Arteries

16 Anomalies of the Right Ventricular Outflow Tract andPulmonary Valve, 281Matthew S. Lemler and Claudio Ramaciotti

17 Pulmonary Atresia with Intact Ventricular Septum, 297Jami C. Levine

18 Abnormalities of the Ductus Arteriosus and PulmonaryArteries, 317Bhawna Arya and Craig A. Sable

19 Anomalies of the Left Ventricular Outflow Tractand Aortic Valve, 336John M. Simpson and Owen I. Miller

20 Hypoplastic Left Heart Syndrome, 357David J. Goldberg and Jack Rychik

21 Aortic Arch Anomalies: Coarctation of the Aortaand Interrupted Aortic Arch, 382Jan Marek, Matthew Fenton, and SachinKhambadkone

22 Tetralogy of Fallot, 407Shubhika Srivastava, Wyman W. Lai, and Ira A.Parness

23 Truncus Arteriosus and Aortopulmonary Window, 433Timothy C. Slesnick, Ritu Sachdeva, Joe R. Kreeger,and William L. Border

24 Transposition of the Great Arteries, 446Luc Mertens, Manfred Vogt, Jan Marek, and Meryl S.Cohen

25 Double-Outlet Ventricle, 466Leo Lopez and Tal Geva

26 Physiologically “Corrected” Transposition of the GreatArteries, 489Erwin Oechslin

v

vi Contents

Part V Miscellaneous Cardiovascular Lesions

27 Hearts with Functionally One Ventricle, 511Stephen P. Sanders

28 Echocardiographic Assessment of Functionally SingleVentricles after the Fontan Operation, 541Marc Gewillig and Luc Mertens

29 Cardiac Malpositions and Heterotaxy Syndrome, 558Irene D. Lytrivi and Wyman W. Lai

30 Congenital Anomalies of the Coronary Arteries, 584J. Rene Herlong and Piers C. A. Barker

31 Vascular Rings and Slings, 609Andrew J. Powell

32 Connective Tissue Disorders, 624Julie De Backer

33 Cardiac Tumors, 641Michele A. Frommelt

Part VI Anomalies of Ventricular Myocardium

34 Dilated Cardiomyopathy, Myocarditis, andHeart Transplantation, 653Renee Margossian

35 Hypertrophic Cardiomyopathy, 677Colin J. McMahon and Javiar Ganame

36 Restrictive Cardiomyopathy and Pericardial Disease, 694Cecile Tissot and Adel K. Younoszai

37 Other Anomalies of the Ventricular Myocardium, 719Rebecca S. Beroukhim and Mary Etta E. King

Part VII Acquired Pediatric Heart Disease

38 Kawasaki Disease, 739Erik C. Michelfelder and Allison Divanovic

39 Rheumatic Fever and Rheumatic Heart Disease, 750Bo Remenyi, Andrew Steer, and Michael Cheung

40 Infective Endocarditis, 763Manfred Otto Vogt and Andreas Kuhn

Part VIII Special Techniques and Topics

41 Transesophageal and IntraoperativeEchocardiography, 777Owen I. Miller, Aaron J. Bell, and John M. Simpson

42 3D Echocardiography, 791Folkert Jan Meijboom, Heleen van der Zwaan, andJackie McGhie

43 Pregnancy and Heart Disease, 815Anne Marie Valente

44 Fetal Echocardiography, 834Darren P. Hutchinson and Lisa K. Hornberger

45 The Echocardiographic Assessment of PulmonaryArterial Hypertension, 872Lindsay M. Ryerson and Jeffrey F. Smallhorn

APPENDIX Normal Echocardiographic Values forCardiovascular Structures, 883

Index, 903

Contributors

Vivekanand Allada MDProfessor of PediatricsUniversity of Pittsburgh School of MedicineDirector of Clinical Services, Pediatric CardiologyChildren’s Hospital of Pittsburgh of UPMCHeart InstitutePittsburgh, PA, USA

Bhawna Arya MDAssistant ProfessorDepartment of PediatricsUniversity of Washington School of MedicineSeattle Children’s HospitalSeattle, WA, USA

Piers C.A. Barker MDAssociate Professor of Pediatrics and Obstetrics/GynecologyDuke University Medical CenterDurham, NC, USA

Aaron J. Bell MDDepartment of Congenital Heart DiseaseEvelina London Children’s Hospital;Guy’s & St Thomas’ NHS Foundation TrustLondon UK

Rebecca S. Beroukhim MDDirector, Fetal EchocardiographyInstructor in PediatricsMassachusetts General HospitalBoston, MA, USA

William L. Border MBChB, MPH, FASEDirector of Noninvasive Cardiac ImagingMedical Director, Cardiovascular Imaging Research Core (CIRC)Children’s Healthcare of Atlanta Sibley Heart Center;Associate ProfessorEmory University School of MedicineAtlanta, GA, USA

David W. Brown MDAssistant Professor of PediatricsDepartment of CardiologyBoston Children’s Hospital;Department of PediatricsHarvard Medical SchoolBoston, MA, USA

Frank Cetta MDProfessor of Medicine and PediatricsMayo ClinicRochester, MN, USA

Sarah Chambers MDThe Pediatric Heart Center at the Children’s Hospital at MontefioreNew York, NY, USA

Michael Cheung BSc, MB ChB, MRCP(UK), MDAssociate ProfessorHead of CardiologyRoyal Children’s Hospital;Department of PaediatricsUniversity of Melbourne;Heart Research GroupMurdoch Childrens Research InstituteMelbourne, VIC, Australia

Meryl S. Cohen MDProfessor of PediatricsPerelman School of Medicine, University of Pennsylvania;Medical Director, EchocardiographyProgram Director, Cardiology FellowshipThe Cardiac CenterThe Children’s Hospital of PhiladelphiaPhiladelphia, PA, USA

Steven D. Colan MDProfessor of PediatricsHarvard Medical School;Boston Children’s HospitalBoston, MA, USA

Julie De Backer MD, PhDSenior LecturerDepartment of Cardiology and Medical GeneticsUniversity Hospital GhentGhent, Belgium

Jan D’hooge MDProfessorDepartment of Cardiovascular SciencesCatholic University of Leuven;Medical Imaging Research CenterUniversity Hospital GasthuisbergLeuven, Belgium

Allison Divanovic MDAssistant Professor of PediatricsUniversity of Cincinnati College of MedicineThe Heart InstituteCincinnati Children’s Hospital Medical CenterCincinnati, OH, USA

vii

viii Contributors

Stacey Drant MDAssociate Clinical ProfessorDirector, Pediatric Echocardiography LaboratoryChildren’s Hospital of PittsburghPittsburgh, PA, USA

Benjamin W. Eidem MD, FACC, FASEProfessor of Pediatrics and MedicineMayo ClinicRochester, MN, USA

Matthew Fenton MB, BS, BScConsultant Paediatric CardiologistCardiothoracic UnitGreat Ormond Street Hospital for Children and Institute ofCardiovascular SciencesLondon, UK

Mark K. Friedberg MDAssociate Professor of PediatricsThe Labatt Family Heart CenterThe Hospital for Sick ChildrenUniversity of TorontoToronto, ON, Canada

Peter C. Frommelt MDProfessor of PediatricsDirector of EchocardiographyDivision of Pediatric CardiologyMedical College of WisconsinChildren’s Hospital of WisconsinMilwaukee, WI, USA

Michele A. Frommelt MDAssociate Professor of PediatricsChildren’s Hospital of WisconsinMilwaukee, WI, USA

Javiar Ganame MD, PhDDepartment of MedicineDivision of CardiologyMcMaster UniversityHamilton, ON, Canada

Tal Geva MDProfessor of PediatricsHarvard Medical School;Chief, Division of Noninvasive Cardiac ImagingDepartment of CardiologyBoston Children’s HospitalBoston, MA, USA

Marc Gewillig MD, PhD, FESC, FACC, FSCAIProfessor Paediatric & Congenital CardiologyUniversity Hospitals LeuvenLeuven, Belgium

David J. Goldberg MDAssistant ProfessorDivision of Pediatric CardiologyThe Children’s Hospital of Philadelphia;Perelman School of Medicine, University of PennsylvaniaPhiladelphia, PA USA

J. Rene Herlong MDAssociate Clinical Professor of Pediatric CardiologySanger Heart and Vascular InstituteLevine Children’s HospitalCharlotte, NC, USA

Lisa K. Hornberger MDProfessor of Pediatrics and Obstetrics & GynecologyDirector, Fetal and Neonatal Cardiology ProgramSection Head, Pediatric EchocardiographyStollery Children’s HospitalEdmonton, AB, Canada

Darren P. Hutchinson MBBS, FRACPFetal & Pediatric CardiologistDepartment of Pediatric CardiologyThe Royal Children’s HospitalMelbourne, VIC, Australia

Sachin Khambadkone MDPaediatric and Adolescent Cardiology, Interventional CardiologistGreat Ormond Street Hospital for Children and Institute of CardiovascularSciencesLondon, UK

Mary Etta E. King MDAssociate Professor of PediatricsMassachusetts General HospitalBoston, MA, USA

H. Helen Ko BS, RDMS, RDCSMount Sinai Medical CenterNew York, NY, USA

Joe R. Kreeger RCCS, RDCSTechnical Director, EchocardiographyChildren’s Healthcare of Atlanta Sibley Heart CenterAtlanta, GA, USA

Andreas Kuhn MDDepartment of Pediatric Cardiology and Congenital Heart DiseaseDeutsches Herzzentrum MunchenTechnische Universitat MunchenMunich, Germany

Wyman W. Lai MD, MPHProfessor of Pediatrics at CUMCColumbia University Medical Center;Director, Noninvasive Cardiac ImagingMorgan Stanley Children’s Hospital of NewYork-PresbyterianNew York, NY, USA

Matthew S. Lemler MDProfessor of PediatricsDivision of CardiologyUniversity of Texas Southwestern;Director, Echocardiography LaboratoryChildren’s Medical Center of DallasDallas, TX, USA

Contributors ix

Jami C. Levine MS, MDAssociate in CardiologyAssistant Professor of PediatricsBoston Children’s HospitalBoston, MA, USA

Leo Lopez MDAssociate Professor of Clinical PediatricsThe Pediatric Heart CenterChildren’s Hospital at MontefioreNew York, NY, USA

Irene D. Lytrivi MDAssistant Professor of PediatricsIcahn School of Medicine at Mount SinaiDivision of Pediatric CardiologyMount Sinai Medical CenterNew York, NY, USA

Jan Marek MD, PhDAssociate Professor of CardiologyInstitute of Child HealthUniversity College London;Director of EchocardiographyConsultant Pediatric and Fetal CardiologistGreat Ormond Street Hospital for ChildrenLondon, UK

Renee Margossian MDAssistant Professor of PediatricsHarvard Medical SchoolDepartment of CardiologyBoston Children’s HospitalBoston, MA, USA

Jackie McGhieCongenital Cardiac Ultrasound SpecialistDepartment of CardiologyErasmus University RotterdamRotterdam, The Netherlands

Colin J. McMahon MB, BCh, FRCPI, FAAPConsultant Paediatric CardiologistDepartment of Paediatric CardiologyOur Lady’s Hospital for Sick ChildrenCrumlin, Ireland

Folkert Jan Meijboom MD, PhDStaff PhysicianDepartment of Pediatrics and CardiologyAcademic Medical Centre UtrechtUtrecht, The Netherlands

Luc L. Mertens MD, PhDSection Head, EchocardiographyThe Hospital for Sick Children;Professor of PediatricsUniversity of TorontoToronto, ON, Canada

Erik C. Michelfelder MDAssociate Professor of PediatricsUniversity of Cincinnati College of MedicineThe Heart InstituteCincinnati Children’s Hospital Medical CenterCincinnati, OH, USA

Owen I. Miller FRACP, FCSANZ, FRCPCHHead of Service/Clinical Lead, Congenital Heart DiseaseConsultant in Pediatric and Fetal CardiologyEvelina London Children’s Hospital;Guy’s & St Thomas’ NHS Foundation Trust;Honorary Senior Lecturer, Kings College LondonLondon, UK

Shobha Natarajan MDAssistant Professor of Clinical PediatricsThe University of Pennsylvania School of MedicineDivision of CardiologyThe Children’s Hospital of PhiladelphiaPhiladelphia, PA, USA

James C. Nielsen MDAssociate Professor of Pediatrics and RadiologyChief, Division of Pediatric CardiologyStony Brook Children’sStony Brook, NY, USA

Erwin Oechslin MD, FRCPC, FESCDirector, Toronto Congenital Cardiac Centre for AdultsThe Bitove Family Professor of Adult Congenital Heart Disease;Professor of Medicine, University of TorontoPeter Munk Cardiac CentreUniversity Health Network/Toronto General HospitalToronto, ON, Canada

Laurie E. Panesar MDAssistant Professor of PediatricsDirector, Fetal and Pediatric EchocardiographyStony Brook Children’sStony Brook, NY, USA

Ira A. Parness MDProfessor of PediatricsDivision of Pediatric CardiologyMount Sinai Medical CenterNew York, NY, USA

Andrew J. Powell MDDepartment of Cardiology, Boston Children’s Hospital;Associate Professor of PediatricsDepartment of Pediatrics, Harvard Medical SchoolBoston, MA, USA

Claudio Ramaciotti MDProfessor of PediatricsDivision of CardiologyUniversity of Texas Southwestern;Children’s Medical Center of DallasDallas, TX, USA

x Contributors

Bo Remenyi MDMenzies School of Health ResearchCharles Darwin UniversityDarwin, NT, Australia

Jack Rychik MDProfessor of PediatricsDivision of Pediatric CardiologyThe Children’s Hospital of Philadelphia;Perelman School of Medicine, University of PennsylvaniaPhiladelphia, PA USA

Lindsay M. Ryerson MD, FRCPCAssistant Clinical ProfessorUniversity of Alberta;Division of Pediatric CardiologyStollery Children’s HospitalEdmonton, AB, Canada

Craig A. Sable MDDirector, Echocardiography and TelemedicineChildren’s National Medical CenterWashington, DC, USA

Ritu Sachdeva MDMedical Director, Cardiovascular Imaging Research CoreChildren’s Healthcare of Atlanta Sibley Heart Center;Associate Professor of PediatricsEmory University School of MedicineAtlanta, GA, USA

Stephen P. Sanders MDProfessor of Pediatrics (Cardiology)Harvard Medical School;Director, Cardiac RegistryDepartments of Cardiology, Pathology, and Cardiac SurgeryBoston Children’s HospitalBoston, MA, USA

John M. Simpson MD, FRCP, FESCConsultant in Fetal and Paediatric CardiologyDepartment of Congenital Heart DiseaseEvelina Children’s Hospital;Guy’s and St Thomas’ NHS Foundation TrustLondon, UK

Timothy C. Slesnick MDDirector of Cardiac MRIChildren’s Healthcare of Atlanta Sibley Heart Center;Assistant Professor of PediatricsEmory University School of MedicineAtlanta, GA, USA

Thierry Sluysmans MD, PhDProfessor of PediatricsUniversite Catholique de Louvain;Director, Cliniques Universitaires Saint LucBrussels, Belgium

Jeffrey F. Smallhorn MBBS, FRACP, FRCPCProfessor of PediatricsUniversity of Alberta;Division of Pediatric CardiologyStollery Children’s HospitalEdmonton, AB, Canada

Shubhika Srivastava MBBSAssociate Professor of PediatricsDivision of Pediatric CardiologyMount Sinai Medical CenterNew York, NY, USA

Andrew Steer MDDepartment of General Medicine and Centre for International Child HealthDepartment of PaediatricsRoyal Children’s HospitalMelbourne, VIC;Group A Streptococcal Research Group, Murdoch Childrens Research InstituteMelbourne, VIC, Australia

Cecile Tissot MDAttending PhysicianPediatric Cardiology UnitThe University Children’s Hospital of GenevaGeneva, Switzerland;The Heart InstituteChildren’s Hospital ColoradoAurora, CO, USA

Anne Marie Valente MDAssociate Professor of Pediatrics and Internal MedicineDepartment of Cardiology, Department of PediatricsBoston Children’s Hospital;Division of Cardiology, Department of MedicineBrigham and Women’s Hospital;Harvard Medical SchoolBoston, MA, USA

Heleen van der Zwaan MD, PhDDepartment of CardiologyErasmus University RotterdamRotterdam, The Netherlands

Manfred Otto Vogt MD, PhDProfessor of Pediatric CardiologyDepartment of Pediatric Cardiology and Congenital Heart DiseaseDeutsches Herzzentrum MunchenTechnische Universitat MunchenMunich, Germany

Adel K. Younoszai MDAssociate Professor, University of Colorado DenverDirector of Cardiac Imaging and Fetal CardiologyThe Heart InstituteChildren’s Hospital ColoradoAurora, CO, USA

Preface

For the first edition of this textbook, we had set out to fill a voidfor an updated comprehensive resource on all aspects related toechocardiography in patients with congenital heart disease, fromthe fetus to the adult. We felt it was important to include detailedinformation about the anatomy and pathophysiology of eachlesion and to describe the goals and techniques of the echocar-diographic examinations for diagnosis, guidance of treatment,and monitoring after intervention. In addition to diagrams andstill images, hundreds of videos were provided to illustrate keyanatomic and functional issues mentioned in the text. We werepleased by the overwhelmingly positive response that the firstedition of this book received.

In this second edition, we made improvements in multipleareas. The discussion of fundamental concepts of echocardio-graphy and the sections on imaging techniques were updated toinclude advances in knowledge and improvements in ultrasoundtechnology. Our coverage of congenital lesions was expandedto include separate chapters on the post-Fontan patient and onpregnancy and heart disease. Each of the lesion chapters nowhas a section highlighting the Key elements of the echocardio-gram(s). Finally, all of the figures and videos were reviewed withthe goal of upgrading and standardizing image quality and dis-play technique.

The field of echocardiography remains dynamic and con-stantly evolving. Nevertheless, the mainstay for education in

clinical echocardiography continues to be the written text illus-trated with images. As we try to expand the resources availableto trainees, practitioners, and educators, we are constrained bythe publishing format currently available for textbooks. There-fore, our second edition remains mostly accessible in print orPDF format. The videos have moved from being primarily avail-able on DVD to a companion website. Future efforts will nodoubt benefit from greater electronic access to the text andimages.

As with the first edition, this book is the product of manyexcellent contributions from the best physician and sonographerexperts in the field. We are indebted to their dedication to thefield and their commitment to education. We remain grateful toour mentors and colleagues, and we continue to be inspired byour trainees. The staff at Wiley-Blackwell have been truly sup-portive, and the professional appearance of this book is due totheir many contributions. Finally, this and all projects in whichthe editors are involved remain possible only with the unwaver-ing support of our families and friends.

Wyman W. Lai, MD, MPHLuc L. Mertens, MD

Meryl S. Cohen, MDTal Geva, MD

xi

About the Companion Website

This book is accompanied by a companion website:

www.lai-echo.com

The companion website includes over 580 video clips, referenced at the end of the chapters throughout the book.

xii

PART I

Introduction to CardiacUltrasound Imaging

CHAPTER 1

Ultrasound Physics

Jan D’hooge1 and Luc L. Mertens2

1Department of Cardiovascular Sciences, Catholic University of Leuven; Medical Imaging Research Center, University Hospital Gasthuisberg, Leuven, Belgium2Cardiology, The Hospital for Sick Children; Department of Pediatrics, University of Toronto, Toronto, ON, Canada

Physics and technology of echocardiography

Echocardiography is the discipline of medicine in which imagesof the heart are created by using ultrasound waves. Knowledge ofthe physics of ultrasound helps us to understand how the differ-ent ultrasound imaging modalities operate and also is importantwhen operating an ultrasound machine to optimize the imageacquisition.

This section describes the essential concepts of how ultra-sound waves can be used to generate an image of the heart. Cer-tain technological developments will also be discussed as well ashow systems settings influence image characteristics. For moredetailed treatment of ultrasound physics and imaging there isdedicated literature, to which readers should refer, for example[1,2].

How the ultrasound image is createdThe pulse–echo experimentTo illustrate how ultrasound imaging works, the acoustic “pulse–echo” experiment can be used:1 A short electric pulse is applied to a piezoelectric crystal. This

electric field will induce a shape change of the crystal throughreorientation of its polar molecules. In other words, due toapplication of an electric field the crystal will momentarilydeform.

2 The deformation of the piezoelectric crystal induces a localcompression of the tissue with which the crystal is in contact:that is, the superficial tissue layer is briefly compressed result-ing in an increase in local pressure; this is the so-called acous-tic pressure (Figure 1.1).

3 Due to an interplay between tissue elasticity and inertia, thislocal tissue compression (with subsequent decompression,i.e., rarefaction) will propagate away from the piezoelectriccrystal at a speed of approximately 1530 m/s in soft tissue(Figure 1.2). This is called the acoustic wave. The rate ofcompression/decompression determines the frequency of the

wave and is typically 2.5–10 MHz (i.e., 2.5–10 million cyclesper second) for diagnostic ultrasonic imaging. As these fre-quencies cannot be perceived by the human ear, these wavesare said to be “ultrasonic.” The spatial distance between sub-sequent compressions is called the wavelength (λ) and relatesto the frequency (f) and sound velocity (c) as: λf = c. Dur-ing propagation, acoustic energy is lost mostly as a result ofviscosity (i.e., friction) resulting in a reduction in amplitudeof the wave with propagation distance. The shorter the wave-length (i.e., the higher the frequency), the faster the particlemotion and the larger the viscous effects. Higher frequencywaves will thus attenuate more and penetrate less deep intothe tissue.

4 Spatial changes in tissue density or tissue elasticity will resultin a disturbance of the propagating compression (i.e., acous-tic) wave and will cause part of the energy in the wave tobe reflected. These so-called “specular reflections” occur, forexample, at the interface between different types of tissue (e.g.,blood and myocardium) and behave in a similar way to opticwaves in that the direction of the reflected wave is determinedby the angle between the reflecting surface and the incidentwave (cf. reflection of optic waves on a water surface). Whenthe spatial dimensions of the changes in density or compress-ibility become small relative to the wavelength (i.e., below∼100 μm), these inhomogeneities will cause part of the energyin the wave to be scattered, that is, retransmitted in all possi-ble directions. The part of the scattered energy that is retrans-mitted back into the direction of origin of the wave is calledbackscatter. Both the specular and backscattered reflectionspropagate back towards the piezoelectric crystal.

5 When the reflected (compression) waves impinge upon thepiezoelectric crystal, the crystal deforms. This results in rela-tive motion of its (polar) molecules and generation of an elec-tric field, which can be detected and measured. The ampli-tude of this electric signal is directly proportional to theamount of compression of the crystal, that is, the amplitude ofthe reflected/backscattered wave. This electric signal is called

Echocardiography in Pediatric and Congenital Heart Disease: From Fetus to Adult, Second Edition. Edited by Wyman W. Lai, Luc L. Mertens, Meryl S. Cohen and Tal Geva.© 2016 John Wiley & Sons, Ltd. Published 2016 by John Wiley & Sons, Ltd.Companion website: www.lai-echo.com

3

4 Part I Introduction to Cardiac Ultrasound Imaging

Acoustic pressure

Density/

Pre

ssure

C

C

−

+

+

−

Depth

Figure 1.1 Local tissue compression due todeformation of the piezoelectric crystal whenapplying an electric field.

the radio-frequency (RF) signal and represents the ampli-tude of the reflected ultrasound wave as a function of time(Figure 1.3). Because reflections occurring further away fromthe transducer need to propagate further, they will be receivedlater. As such, the time axis in Figure 1.3 can be replaced bythe propagation distance of the wave (i.e., depth). The signaldetected by the transducer is typically electronically ampli-fied. The amount of amplification has a preset value but canbe modified on an ultrasound system by using the “gain”button (typically the largest button on the operating panel).

Importantly, the overall gain will amplify both the signaland potential measurement noise and will thus not affect thesignal-to-noise ratio.In the example shown in Figure 1.3, taken from a water tank

experiment, two strong specular reflections can be observed(around 50 and 82 μs, respectively) while the lower ampli-tude reflections in between are scatter reflections. In clinicalechocardiography, the most obvious specular reflection is thestrong reflection coming from the pericardium observed in theparasternal views as a consequence of its increased stiffness with

Depthλ

Depth

Depth

Density/

Pre

ssure

Density/

Pre

ssure

Density/

Pre

ssure

Figure 1.2 The local tissue compression/decompression propagates away from itssource at a speed of approximately 1530 m/sin soft tissue.

Chapter 1 Ultrasound Physics 5

Time (μs)

20 30 40 50 60 70 80 90 100 110

Am

plit

ude (

V)

20

15

10

5

0

–5

–10

–15

–20

Figure 1.3 The reflected amplitude of the reflected ultrasound waves as afunction of time after transmission of the ultrasound pulse is called theradio-frequency (RF) signal.

respect to the surrounding tissues. The direction of propagationof the specular reflection is determined by the angle between theincident wave and the reflecting surface. Thus, the strength ofthe observed reflection will depend strongly on the exact trans-ducer position and orientation with respect to the pericardium.Indeed, for given transducer positions/orientations, the strongspecular reflection might propagate in a direction not detectableby the transducer. For this reason, the pericardium typically doesnot show as bright in the images taken from an apical transducerposition. In contrast, scatter reflections are not angle dependentand will always be visible for a given structure independent ofthe exact transducer position.

The total duration of the above described “pulse–echo” exper-iment is about 100 μs when imaging at 5 MHz. The reflectedsignal in Figure 1.3 is referred to as an A-mode image (“A”referring to “Amplitude”) and is the most fundamental formof imaging given that it tells us something about the acousticcharacteristics of the materials in front of the transducer. Forexample, Figure 1.3 clearly shows that at distance of ∼3.7 cmin front of the transducer the propagation medium changesdensity and/or compressibility with a similar change occur-ring at a distance of ∼6.3 cm (these distances correspond to50/82 μs × 1530 m/s – which is the total propagation distanceof the wave – divided by two as the wave has to travel back andforth). The 2.6 cm of material in between these strong reflectionsis acoustically inhomogeneous (i.e., shows scatter reflections)and thus contains local (very small) fluctuations in mass densityand/or compressibility while the regions closer and further awayfrom the transducer do not cause significant scatter and wouldthus be acoustically homogeneous. Indeed, this A-mode imagewas taken from a 2.6 cm thick tissue-mimicking material (i.e.,gelatin in which small graphite particles were dissolved) put in awater tank.

Grayscale encodingSince the A-mode image presented above is not visually attrac-tive, the RF signal resulting from a pulse–echo experiment isprocessed in the following manner:1 Envelope detection: The high-frequency information of the

RF signal is removed by detecting the envelope of the signal(Figure 1.4). This process is also referred to as “demodulation.”

2 Grayscale encoding: The signal is subdivided as a function oftime in small intervals (i.e., pixels). Each pixel is attributed anumber, defined by the local amplitude of the signal, rangingbetween 0 and 255 (28 or 8-bit image). “0” represents “black;”“255” represents “white;” a value in between is represented bya grayscale. By definition, bright pixels correspond to high-amplitude reflections. This process is illustrated in Figure 1.4.Please note that a different kind of color encoding is also pos-sible simply by attributing different colors to the range of val-ues between 0 and 255. For example, shades of blue or bronzeare also popularly used. The choice of the color map usedto display an image is a matter of preference and can eas-ily be changed on all ultrasound systems. Nowadays, mostultrasound systems have 12- or 16-bit resolution images (i.e.,encoding 4096 or 65536 gray/color levels).

3 Attenuation correction: As wave amplitude decreases withpropagation distance due to attenuation (mostly due to con-version of acoustic energy to heat), reflections from deeperstructures are intrinsically smaller in amplitude and there-fore show less bright. In order to give identical structureslocated at different distances from the transducer a simi-lar gray value (i.e., reflected amplitude), compensation forthis attenuation must occur. Thus, an attenuation profile asa function of distance from the transducer is assumed, whichallows for automatic amplification of the signals from deeperregions – the so-called automated time-gain compensation(TGC), also referred to as depth-gain compensation. As thepre-assumed attenuation profile might be incorrect, sliders onthe ultrasound scanner (TGC toggles) allow for manual cor-rection of the automatic compensation and will result in moreor less local amplification of the received signal as requiredto obtain a more homogenous brightness of the image. Inthis way, the operator can optimize local image brightness.It is recommended to start scanning using a neutral settingof these sliders, as attenuation characteristics will be patientand view specific. For each view the TGC can be optimizedmanually.

4 Log-compression: In order to increase the image contrast inthe darker (i.e., less bright) regions of the image, gray val-ues in the image may be redistributed according to a loga-rithmic curve (Figure 1.5). The characteristics of this com-pression (i.e., local contrast enhancement) can be changedthrough settings on the ultrasound scanner. They will notchange the ultrasound acquisition as such (and thus impactimage quality) but will merely influence the visual representa-tion of the resulting image. This is similar to what is nowadaysvery common in digital photography, where ambient lighting

6 Part I Introduction to Cardiac Ultrasound Imaging

Time (μs)20 30 40 50 60 70 80 90 100 110

Am

plit

ude (

V)

20

15

10

5

0

–5

–10

–15

–20

180120 50 65 70 75 100 60 53 65 72 167 73 68 101 68 85 48

Figure 1.4 The RF signal is demodulated in order to detect its envelope. This envelope signal (bold) is color encoded based on the local signal amplitude.

0 50 100 150

Original gray value

200 250

Pro

cessed g

ray v

alu

e

250

200

150

100

50

0

Figure 1.5 The logarithmic compression curve.

Chapter 1 Ultrasound Physics 7

Transducer

Transducer

Object Object

(a) (b)

Transducer

Transducer

Figure 1.6 Translation of the ultrasoundsource results in a linear image format (a)whereas pivoting results in sector images (b).

and/or contrast can be retrospectively enhanced. The settingon the system impacting the visual aspect of the image is theso-called “dynamic range” which will change the number ofgray values used and will therefore result in “hard” (i.e., almostblack-and-white images without much gray) or “soft” imagesfor low and high dynamic range, respectively.

Image constructionIn order to obtain an ultrasound image, the above procedures ofsignal acquisition and post-processing are repeated.

For conventional B-mode imaging (“B” referring to “bright-ness” mode), the transducer can either be translated (Fig-ure 1.6a) or tilted (Figure 1.6b) within a plane between two sub-sequent “pulse–echo” experiments. In this way, a conventional2D cross-sectional image is obtained. The same principle holdsfor 3D imaging by moving the ultrasound beam in 3D spacebetween subsequent acquisitions.

Alternatively, the ultrasound beam is transmitted into thesame direction for each transmitted pulse. In that case, an imageline is obtained as a function of time, which is particularly use-ful to look at motion. This modality is therefore referred to asM-mode (motion-mode) imaging.

Image artifactsSide lobe artifactsIn the construction of an ultrasound image, the assumption ismade that all reflections originate from a region directly in frontof the transducer. Although most of the ultrasound energy isindeed centered on an axis in front of the transducer, in prac-tice part of the energy is also directed sideways (i.e., directed off-axis). The former part of the ultrasound beam is called the mainlobe whereas the latter is referred to as the side lobes (Figure 1.7).

Because the reflections originating from these side lobes aremuch smaller in amplitude than the ones coming from the mainlobe, they can typically be neglected. However, image artifactscan arise when the main lobe is in an anechoic region (i.e., acyst or inside the left ventricular cavity) causing the relative con-tribution of the side lobes to become significant. In this way, asmall cyst or lesion may be more difficult to detect, as it appearsbrighter than it should due to spillover of (side-lobe) energyfrom neighboring regions. Similarly, when using contrast agents(see later) the reflections resulting from small side lobes maystill become significant as these agents reflect energy strongly.As such, increased brightness may appear in regions adjacent toregions filled with contrast without contrast being present in theregion itself.

Tra

nsd

uce

r

Re

fle

cte

d

am

plit

ud

e

Time

Figure 1.7 Reflections caused by side lobes(red) will induce image artifacts because allreflections are assumed to arrive from themain ultrasound lobe (green).

8 Part I Introduction to Cardiac Ultrasound Imaging

Transducer

Reflected

pulse at transducer

surface

Transmitted

pulse

Reflecte

d

am

plit

ude

Time

Figure 1.8 A transmitted wave (green) will reflectand result in an echo signal (green). The reflectedwave will, however, partially reflect at thetransducer surface (red) and generate secondarysignals (red).

Reverberation artifactsWhen the reflected wave arrives at the transducer, part of theenergy is converted to electrical energy as described in theprevious section. However, another part of the wave is sim-ply reflected on the transducer surface and will start propagat-ing away from the transducer as if it were another ultrasoundtransmission. This secondary “transmission” will propagate in away similar to that of the original pulse, which means that it isreflected by the tissue and detected again (Figure 1.8).

These higher-order reflections are called reverberations andgive rise to ghost (mirror image) structures in the image. Theseghost images typically occur when strongly reflecting structuressuch as ribs or the pericardium are present in the image. Sim-ilarly, as the reflected wave coming from the pericardium isvery strong, its backscatter (i.e., propagating again towards thepericardium) will be sufficiently strong as well. This wave willreflect on the pericardium and can be detected by the trans-ducer after the actual pericardial reflection arrives. In clinicalpractice, this causes a ghost image to be created behind thepericardial reflection that typically appears as a mirror imageof the left ventricle around the pericardium in a parasternallong-axis view.

Shadowing and dropout artifactsWhen perfect reflections occur, no acoustic energy is transmit-ted to more distal structures and – as a consequence – no reflec-tions from these distal structures can be obtained. As a result, avery bright structure will appear in the image followed by a sig-nal void, that is, an acoustic shadow. For example, when a metal-lic artificial valve has been implanted, the metal (being verydense and extremely stiff) can cause close to perfect ultrasound

reflections resulting in an apparently anechoic region distal tothe valve. This occurs because no ultrasound energy reachesthese deeper regions. Similarly, some regions in the image mayreceive little ultrasound energy due to superficial structuresblocking ultrasound penetration. Commonly, ribs (being moredense and stiff than soft tissue) are fairly strong reflectors at car-diac diagnostic frequencies and can impair proper visualizationof some regions of the image. These artifacts are most commonlyreferred to as “dropout” and can only be avoided – if at all – bychanging the transducer position/orientation.

When signal dropout occurs at deeper regions only, theacoustic power transmitted can be increased. This will obvi-ously result in more energy penetrating to deeper regions andwill increase the overall signal-to-noise ratio of the image (incontrast to increasing the overall gain of the received signalsas explained earlier). However, the maximal transmit powerallowed is limited in order to avoid potential adverse biolog-ical effects. Indeed, at higher energy levels, ultrasound wavescan cause tissue damage either due to cavitation (i.e., the for-mation of vapor cavities that subsequently implode and gener-ate very high local pressures and temperatures) or tissue heat-ing. The former risk is quantified in the “mechanical index”(MI) and should not pass a value of 1.9 while the latter is esti-mated through a “thermal index” (TI). Both parameters aredisplayed on the monitor during scanning and will increasewith increasing power output. The operator thus has to finda compromise between image quality (including penetrationdepth) and the risk of adverse biological effects. In case pene-tration is not appropriate at maximal transmit power, the oper-ator has to choose a transducer with lower transmit frequency(see earlier).

Chapter 1 Ultrasound Physics 9

1 2 3 4 5 6 7 1 2 3 4 5 6 7

Figure 1.9 An array of crystals can be used tosteer the ultrasound beam electronically byintroducing time delays between the activation ofindividual elements in the array.

Ultrasound technology and imagecharacteristics

Ultrasound technologyPhased array transducersRather than mechanically moving or tilting the transducers,as in early generation ultrasound machines, modern ultra-sound devices use electronic beam steering. To do this an arrayof piezoelectric crystals is used. By introducing time delaysbetween the excitation of different crystals in the array, the ultra-sound wave can be sent in a specific direction without mechan-ical motion of the transducer (Figure 1.9). The RF signal for atransmission in a particular direction is then simply the sum ofthe signals received by the individual elements. These individualcontributions can be filtered, scaled, and time-delayed separatelybefore summing. This process is referred to as beam-formingand is a crucial element for obtaining high-quality images. Thescaling of the individual contributions is typically referred to asapodization and is critical in suppressing side lobes and thusavoiding the associated artifacts.

This concept can be generalized by creating a 2D matrix ofelements that enables steering of the ultrasound beam in threedimensions. This type of transducer is referred to as a matrixarray or 2D array transducer. Because each of the individualelements of such an array needs electrical wiring, manufactur-ing such a 2D array remained technically challenging for manyyears, in part because of the limitation of the thickness of thetransducer cable. Generally, these obstacles have been overcomeand 2D arrays are now readily available.

Second harmonic imagingWave propagation as illustrated in Figure 1.2 is only valid whenthe amplitude of the ultrasound wave is relatively small (i.e., theacoustic pressures involved are small). Indeed, when the ampli-tude of the transmitted wave becomes significant, the shape ofthe ultrasound wave will change during propagation, as illus-trated in Figure 1.10. This phenomenon of wave distortion dur-ing propagation is referred to as nonlinear wave propagation. Itcan be shown that this wave distortion causes harmonic frequen-cies (i.e., integer multiples of the transmitted frequency) to be

Depth

Depth

Depth

De

nsity/

Pre

ssu

re

Density/

Pre

ssure

Density/

Pre

ssure

Figure 1.10 Nonlinear wave behavior results inchanges in shape of the waveform duringpropagation.

10 Part I Introduction to Cardiac Ultrasound Imaging

generated. Transmitting a 1.7-MHz ultrasound pulse will thusresult in the spontaneous generation of frequency componentsof 3.4, 5.1, 6.8, 8.5 MHz, and so on. These harmonic compo-nents will grow stronger with propagation distance. The rate atwhich the waveform distorts for a specific wave amplitude is tis-sue dependent and characterized by a nonlinearity parameter, β(or the so-called “B/A” parameter).

The ultrasound scanner can be set up to receive only the sec-ond harmonic component through filtering of the received RFsignal. If further post-processing of the RF signal is done inexactly the same way as described earlier, a second harmonicimage is obtained. Such an image typically has a better signal-to-noise ratio by avoiding clutter noise due to (rib) reverberationartifacts. This harmonic image is commonly used in patientswith poor acoustic windows and poor penetration. Althoughharmonic imaging increases signal-to-noise, it has intrinsicallypoorer axial resolution as elucidated later. Please note that higherharmonics (i.e., third, fourth, etc.) are present but typically falloutside the bandwidth of the transducer and thus remain unde-tected. Harmonic imaging has become the default cardiac imag-ing mode for adult scanning on many systems. It is typicallyunnecessary to use harmonic imaging in young infants but isoften required to enhance the image in the older patient. Switch-ing between conventional and harmonic imaging is done bychanging the transmit frequency of the system. For lower fre-quency transmits, it will automatically enter a harmonic imagingmode, which is indicated on the display by showing both trans-mit and receive frequencies (i.e., 1.7/3.4 MHz). When a singlefrequency is displayed, the scanner is in a conventional (i.e., fun-damental) imaging mode. For pediatric scanning, especially insmaller infants, fundamental imaging is the preferred mode dueto its better spatial resolution.

Contrast imagingAs blood reflects little ultrasound energy it shows dark in theimage. For some applications (such as myocardial perfusionassessment) it can be useful to artificially increase the bloodreflectivity. This can be achieved using an ultrasound contrastagent. As air is an almost perfect reflector of ultrasound energygiven it is very compressible and has a low density compared tosoft tissue, it often is used as a contrast agent. As such, the injec-tion of small air bubbles (of diameter similar to that of red bloodcells) will increase blood reflectivity. This can be done using agi-tated saline or by using ultrasound contrast agents that encapsu-late air bubbles in order to limit diffusion of the air (or a heaviergas) in blood. Contrast imaging can be used to help in visualizingthe endocardial border by enhancing the difference in gray valuebetween the myocardium and the blood pool (i.e., left ventricu-lar opacification) or to detect shunts. Similarly, contrast agentscan be used to increase the brightness of perfused (myocar-dial) tissue although artifacts become more prominent and maymake interpretation less obvious. At present, different contrastagents are commercially available for clinical use but none of

them have been approved for pediatric use. Especially in thepresence of right-to-left shunting, contrast agents should be usedcarefully.

Image resolutionResolution is defined as the shortest distance at which two adja-cent objects can be distinguished as separate. The spatial reso-lution of an ultrasound image varies depending on the positionof the object relative to the transducer. Also the resolution in thedirection of the image line (range or axial resolution) is differ-ent from the one perpendicular to the image line within the 2Dimage plane (azimuth or lateral resolution), which is differentagain from the resolution in the direction perpendicular to theimage plane (elevation resolution).

Axial resolutionIn order to obtain an optimal axial resolution, a short ultra-sound pulse needs to be transmitted. The length of the trans-mitted pulse is mainly determined by the characteristics of thetransducer. Current transducers can generate multiple frequen-cies influencing axial resolution by selecting different frequen-cies. The bandwidth is most commonly expressed relative to thecenter frequency of the transducer. A typical value would be80% implying that for a 5-MHz transducer the absolute band-width is about 4 MHz. This type of transducer can thus gen-erate/receive frequencies in the range of 3–7 MHz. The abso-lute transducer bandwidth is typically proportional to the meantransmission frequency. A higher frequency transducer will thusproduce shorter ultrasound pulses and thus, better axial resolu-tion. Unfortunately, higher frequencies are attenuated more bysoft tissue and are impacted by depth (see earlier). As such, acompromise needs to be made between image resolution andpenetration depth (i.e., field of view). In pediatric and neona-tal cardiology, higher frequency transducers can be used thatincrease image spatial resolution. Generally, for infants 10–12-MHz transducers are used resulting in a typical axial resolutionof the order of 250 μm.

Most systems allow changing the transmit frequency of theultrasound pulse within the bandwidth of the transducer. Assuch, a 5-MHz transducer can be used to transmit a 3.5-MHzpulse which can be practical when penetration is not sufficientat 5 MHz. The lower frequency will result in a longer transmitpulse with a negative impact on axial resolution. Similarly, forsecond harmonic imaging a narrower band pulse needs to betransmitted, as part of the bandwidth of the transducer needs tobe used to be able to receive the second harmonic. As such, inharmonic imaging mode, a longer ultrasound pulse is transmit-ted (i.e., less broadband) resulting in a worse axial resolution ofthe second harmonic image despite improvement of the signal-to-noise ratio. Therefore, some of the cardiac structures appearthicker, especially valve leaflets. This should be considered wheninterpreting the images.

Chapter 1 Ultrasound Physics 11

Figure 1.11 Introducing time delays during transmission of individualarray elements (left) allows for all wavelets to arrive at a particular point(focus) simultaneously. Similarly, received echo signals can be time delayedso that they constructively interfere (receive focus).

Lateral resolutionLateral resolution is determined by the width of the ultra-sound beam (i.e., the width of the main lobe). The narrower theultrasound beam, the better the lateral resolution. In order tonarrow the ultrasound beam, several methods can be used butthe most obvious one is focusing. This is achieved by introduc-ing time delays between the firing of individual array elements(similar to what is done for beam steering) in order to assure thatthe transmitted wavelets of all individual array elements arriveat the same position at the same time and will thus construc-tively interfere (Figure 1.11). Similarly, time delaying the reflec-tions of the individual crystals in the array will make sure thatreflections coming from a particular point in front of the trans-ducer will sum in phase and therefore create a strong echo signal(Figure 1.11). Because the sound velocity in soft tissue is known,the position from which reflections can be expected is known ateach time instance after transmission of the ultrasound pulse. Assuch, the time delays applied in receive can be changed dynam-ically in order to move the focus point to the appropriate posi-tion. This process is referred to as dynamic (receive) focusing .In practice, dynamic receive focusing is always used and doesnot need adjustments by the operator in contrast to the transmitfocus point whose position should be set manually. Obviously,to resolve most morphologic detail, the transmit focus shouldalways be positioned close to the structure/region of interest.Most ultrasound systems allow selecting multiple transmit focalpoints. In this setting, each image line will be created multipletimes with a transmit pulse at each of the set focus positions andthe resulting echo signals will be combined in order to gener-ate a single line in the image. Although this results in a morehomogeneous distribution of the lateral resolution with depth,this obviously implies that it takes more time to generate a sin-gle image and thus will result in lowering the frame rate (i.e.,temporal resolution).

The easiest way to improve the focus performance of a trans-ducer is by increasing its size (i.e., aperture). Unfortunately, thefootprint needs to fit between the patient’s ribs, thereby limitingthe size of the transducer and thus limiting lateral resolution of

the imaging system. For an 8-MHz pediatric transducer, realisticnumbers for the lateral resolution of the system are depth depen-dent and are approximately 0.3 mm at 2 cm going up to 1.2 mmat 7 cm depth.

Elevation resolutionFor elevation resolution, the same principles hold as for lateralresolution in the sense that the dimension of the ultrasoundbeam in the elevation direction will be determinant. However,most ultrasound devices are still equipped with 1D array trans-ducers. As such, focusing in the elevation direction needs to bedone by the use of acoustic lenses (similar to optic lenses acous-tic lenses concentrate energy in a given spatial position), whichimplies that the focus point is fixed in both transmit and receive(i.e., dynamic focusing is not possible in the elevation direction).This results in a resolution in the elevation direction that is worsethan the lateral resolution. The homogeneity of the resolution isalso worse with depth. Moreover, transducer aperture in the ele-vation direction is typically somewhat smaller (in order to fit inbetween the ribs of the patient) resulting in a further decrease ofelevation resolution compared to the lateral component. Newersystems with 2D array transducer technology have more similarlateral and elevation image resolution. Matrix array transducersnot only create 3D images but also allow generating 2D imagesof higher/more homogeneous spatial resolution.

Temporal resolutionTypically, a 2D pediatric cardiac image consists of 300 lines. Theconstruction of a single image thus takes about 300 × 100 μs(the time required to acquire one line as explained earlier) or30 ms. About 33 images can be produced per second, which issufficient to look at motion (e.g., standard television displaysonly 25 frames per second). With more advanced imaging tech-niques such as parallel beam forming, higher frame rates can beobtained (70–80 Hz). In order to increase frame rate, either thefield of view can be reduced (i.e., a smaller sector will requirefewer image lines to be formed and will thus speed up the acqui-sition of a single frame) or the number of lines per frame (i.e.,the line density) can be reduced. The latter comes at the costof spatial resolution, as image lines will be further apart. Thereis thus an intrinsic trade-off between the image field-of-view,spatial resolution and temporal resolution. Most systems have a“frame rate” button nowadays that allow changing the frame ratealthough this always comes at the expense of spatial resolution.Higher frame rates are important when the heart rate is higheras is often the case in pediatric patients and when you want tostudy quickly moving structures like valve leaflets or myocardialwall motion.

Image optimization in pediatric echocardiographyAll the aforementioned principles can be used to optimizeimage acquisition in the pediatric population. As children aresmaller, less penetration is required. The heart rates are higher

12 Part I Introduction to Cardiac Ultrasound Imaging

requiring higher temporal resolution and, as structural heartdisease is more common in the pediatric age group spatialresolution needs to be optimized to obtain the best possiblediagnostic images. Image optimization will always be a com-promise between spatial and temporal resolution. A few gen-eral recommendations can be made which can help in imageoptimization:1 Always use the highest possible transducer frequency to opti-

mize spatial resolution. For infants high-frequency probes(8–12 MHz) must be available and used. Often differ-ent transducers have to be used for different parts of theexam. So, for instance, for subcostal imaging in a new-born, a 5- or 8-MHz probe can be used while for theapical and parasternal windows a 10–12-MHz probe oftenprovides better spatial resolution. For larger children andyoung adults 5-MHz and rarely 2.5–3.5-MHz probes canbe used, although for the parasternal windows the higher-frequency probes generate good quality images also in thispopulation.

2 Especially in smaller children harmonic imaging does notnecessarily result in better image quality due to its intrinsi-cally lower axial resolution. In general, fundamental frequen-cies provide good-quality images. Harmonic imaging is gen-erally more useful in larger children and adults.

3 Gain and dynamic range settings are adjusted to optimizeimage contrast so that the structures of interest can be seenwith the highest possible definition. TGCs are used to makethe images as homogeneous as possible at different depths.Image depth and focus are always optimized to image thestructures of interest.

4 For optimizing temporal resolution the narrowest sector pos-sible is to be used.

5 Depth settings are minimized to include the region of interest.

Doppler imaging

Continuous-wave DopplerWhen an acoustic source moves relative to an observer, the fre-quencies of the transmitted and the observed waves are different.This phenomenon is known as the Doppler effect. A well-knownexample is that of a whistling train passing a static observer: theobserved pitch of the whistle is higher when the train approachesthan when it moves away.

The Doppler phenomenon can be used to measure tis-sue and blood velocities by comparing the transmitted to thereceived ultrasound frequency. Indeed, when ultrasound scatter-ing occurs at stationary tissues, the transmitted and reflected fre-quencies are identical. This statement is only true when attenua-tion effects are negligible. In soft tissue there will be an intrinsicfrequency shift due to frequency-dependent attenuation. Whenscattering occurs at tissue in motion (Figure 1.12), a (additional)frequency shift – the Doppler shift (fD) – will be induced that is

T

T

Figure 1.12 The Doppler effect will induce a frequency shift of thetransmitted ultrasound wave when the reflecting object is in motion.

directly proportional to the velocity (v) by which the tissue ismoving:

fD = −2v cosθfT∕c

where fT is the transmit frequency, θ is the angle between thedirection of wave propagation and the tissue motion, and c isthe velocity of sound in soft tissue (i.e., 1530 m/s). Note that formotion orthogonal to the image line, θ = 90◦ and the Dopplershift is zero regardless of the amplitude of the tissue velocityv. The Doppler phenomenon thus only allows measuring themagnitude of the velocity along (parallel to) the image line (i.e.,motion towards or away from the transducer) while motionorthogonal (perpendicular) to the line is not detected. In prac-tice, an angle up to 20◦ is considered acceptable in order toobtain clinically relevant Doppler measurements. Ideally, theangle should always be minimized by selecting the proper imageline for the Doppler recording or by repositioning the ultrasoundtransducer. Notably, the velocities cannot be overestimated dueto angle dependency; thus, the highest value recorded is closestto the true one. Moreover, in case the angle between the flowand the ultrasound line is known (i.e., θ in the above equation),the velocity estimate can be corrected for this angle. Althoughthis is possible in laminar flow conditions (such as flow in anonstenosed vessel), the flow direction is typically not known incardiac applications. Angle correction is therefore typically notmade in cardiac ultrasound and should be used with care whenapplied.

In order to make a Doppler measurement, one piezoelectriccrystal is used for transmitting a continuous wave of a fixed fre-quency and a second crystal is used to continuously record thereflected signals. Both crystals are embedded in the same trans-ducer, and the frequency difference, that is, the Doppler shift, iscontinuously determined. The instantaneous shift in frequencyis converted to a velocity according to the Doppler equationand is displayed as a function of time in a so-called spectro-gram. However, because different velocities are present within

Chapter 1 Ultrasound Physics 13

Figure 1.13 Example of a continuous-waveDoppler spectrogram of the left ventricularoutflow tract.

the ultrasound beam for any given time instance during the car-diac cycle, a range of Doppler frequencies is typically detected.Thus, there is a spectrum of Doppler shifts measured and dis-played in the spectrogram (Figure 1.13) – hence the name “spec-tral Doppler” often encountered in literature. Depending onthe clinical application, the speed at which the spectrogram isupdated can be modified (i.e., the sweep speed). For example,to look at beat-to-beat variations during the respiratory cycle, alow sweep speed can be used while a high sweep speed is neededwhen looking at flow characteristics within a single cardiac cycle.

Finally, as typical blood velocities cause a Doppler shift in thesonic range (20 Hz to 20 kHz), the Doppler shift itself can bemade audible to the user. A high pitch (large Doppler shift) cor-responds to a high velocity whereas a low pitch (small Dopplershift) corresponds to a low velocity. As such, the user gets bothvisual (spectrogram) and aural information on the velocitiesinstantaneously present in the ultrasound beam.

The system described earlier is referred to as the continuous-wave (CW) Doppler system. As an ultrasound wave is trans-mitted continuously, no spatial information is obtained. Indeed,all velocities occurring anywhere within the ultrasound beam(i.e., on the selected ultrasound line of interrogation) will con-tribute to the reflected signal and appear in the spectrogram. Infact, as the ultrasound signal weakens with depth due to atten-uation, velocities close to the transducer will intrinsically con-tribute more than the ones occurring further away. For mostapplications CW Doppler is combined with 2D imaging whichallows the operator to align the Doppler signal based on theanatomical information. Sometimes the relatively large footprintof the probe does not allow a good Doppler alignment and theuse of a small blind probe can help with obtaining better Doppleralignment. This can be used, for instance, for measurement of

a peak gradient across the aortic valve in case of aortic valvestenosis.

Optimization of CW Doppler1 Alignment with the direction of the measured velocity should

be optimized.2 The gain control affects the ratio of the output signal strength

to the input signal strength. The gain controls should bemanipulated to produce a clean uniform profile without any“blooming.” The gain controls should be turned up to overem-phasize the image and then adjusted down. This will preventany loss of information due to too little gain.

3 The compress control assigns the varying amplitudes a certainshade of gray. If the compress control is very low or high thequality of the spectral analysis graph will be affected and thismay lead to erroneous interpretation.

4 The reject button eliminates the smaller amplitude signals thatare below a certain threshold level. This will help to provide acleaner image and may make measurements more obvious.

5 The filter is used to reduce the noise that occurs from reflec-tors that are produced from walls and other structures that arewithin the range of the ultrasound beam.

Pulsed-wave DopplerThe “pulse–echo” measurement described previously can berepeated along a particular line in the ultrasound image at agiven repetition rate (referred to as the “pulse repetition fre-quency,” or PRF). Rather than acquiring the complete RF sig-nal as a function of time, in pulsed-wave Doppler mode, a singlesample of each reflected pulse is taken at a fixed time after thetransmission of the pulse (the so-called range gate or sample vol-ume). Assuming the position of the scattering sites relative to the

14 Part I Introduction to Cardiac Ultrasound Imaging

T

Echo 1

Echo 2

Echo 3

Echo 4

Echo 5

Time

Time

Am

plitu

de

Echo 1

Echo 2

Echo 3

Echo 4

Echo 5

Figure 1.14 Schematic illustration of theprinciple of the pulsed-wave Doppler system.

transducer remains constant over time, all reflections (thereforeall samples taken at the range gate) will be identical. However,when the tissue is moving relative to the transducer, the ultra-sound wave will have to travel further (motion away from thetransducer) or less far (motion toward the transducer) betweensubsequent acquisitions. As a result, the reflected signal willshift in time and the sample taken at the range gate will change(Figure 1.14).

It can be demonstrated that the frequency of the signal con-structed in this way is directly proportional to the velocity of thereflecting object following the same mathematical relationshipthat is given in the Doppler equation. For this reason, this imag-ing mode is referred to as pulsed-wave (PW) Doppler imagingdespite the fact that the Doppler phenomenon as such is notexploited.

Note that motion orthogonal to the direction of wave propa-gation will not induce a significant change in propagation dis-tance for the ultrasound wave; it will not result in a significant

time shift of the reflected signal and will thus not be detected bythe system.

For a PW Doppler system, velocities are displayed as a func-tion of time in a spectrogram similar to what is done for the CWDoppler system (Figure 1.15). However, the velocities displayedin a PW spectrogram are occurring within a specific region (thesample volume) within the 2D image.

In practice, more than one sample is taken at the range gate.Moreover, it can be demonstrated that the accuracy of the veloc-ity estimate is better for narrow band pulses (i.e., ultrasoundpulses containing relatively few frequencies). By changing thesize of the range gate, the bandwidth of the transmitted pulsecan be changed. A larger range gate (i.e., a longer or more nar-row band transmit pulse) will improve the accuracy of the veloc-ity estimate and will improve the signal-to-noise ratio of thespectrogram but will obviously result in a less localized mea-surement. The operator can change the size of the range gate asappropriate.

Chapter 1 Ultrasound Physics 15

Figure 1.15 Example of a pulsed-wave Dopplerspectrogram of the left ventricular outflow tract.

As explained earlier, blood is relatively poorly echogenic mak-ing it appear dark in the B-mode image. As a result, echo sig-nals obtained from the blood are sensitive to, for example, sidelobe artifacts and may thus contain echo signals coming from thewall. In order to avoid displaying these “spilled over” velocitiesin the spectrogram, low velocities are removed from the spectro-gram using a high pass filter typically referred to as the “wall fil-ter.” The cut-off frequency of this high pass filter can be manuallyselected and should be chosen such that strong, low-amplitudevelocities from the wall are adequately removed without signifi-cantly impacting the velocity spectrum of the blood itself.

AliasingIf the velocity of the scattering object is relatively large and theshift between two subsequent acquisitions is larger than half awavelength, the PW Doppler system cannot differentiate this

high velocity (Figure 1.16a – red signal) from a low velocity(Figure 1.16a – green signal) because the extracted samples atthe range gate are identical. This effect is known as aliasing, and aclinical example is given in Figure 1.16b. To avoid aliasing, eitherthe PRF needs to be increased or the transmit frequency needsto be decreased. The latter option is not commonly used in prac-tice while the former can be adjusted as required. Depending onthe system, the PRF setting is referred to as “Nyquist velocity,”“Scale,” or “Velocity range.” Moreover, in case the direction offlow is known, most systems allow shifting the baseline of thespectrogram thereby increasing the maximal velocity that canbe measured without introducing aliasing.

Velocity resolutionA high PRF will ensure that aliasing will not occur when themaximal detectable velocity is high. However, as the ultrasound

TimeSample

Am

plitu

deA

mpl

itude

TimeSample

(a) (b)

Figure 1.16 Principle of aliasing in Doppler acquisitions (a) and a practical example (b).

16 Part I Introduction to Cardiac Ultrasound Imaging

system can only measure a fixed number of velocity ampli-tudes, the smallest difference between two velocity amplitudesdetectable by the system will decrease with increasing PRF. Assuch, a compromise needs to be made between velocity res-olution and maximal detectable velocity. For this reason, it isimportant in practice to keep the PRF as small as possible whilestill avoiding aliasing in order to have maximal accuracy of thevelocity measurement. When the maximum velocity is high, PWDoppler will not be able to detect the highest velocity withoutaliasing and CW Doppler will be required to measure true peakvelocities. Thus, in turbulent flow, both PW and CW Dopplerinterrogation is needed to make an accurate interpretation oflocation and maximum velocity. The velocity amplitude at whichPW Doppler is no longer capable of measuring the velocity with-out aliasing is dependent on the transmit frequency, the PRF, andthe depth at which these velocities occur (the closer to the trans-ducer, the higher the maximal velocity that can correctly be mea-sured).

Optimizing PW Doppler signals1 As for any Doppler technique, alignment with the direction of

the velocity is important.2 The gain control, compress, filter settings are similar com-

pared to CW Doppler.3 Baseline: shift of the baseline allows the whole display to be

used for either forward or reverse flow, which is useful if theflow is only in one direction.

4 The Nyquist limit should always be optimized and be set nohigher or lower than necessary to display the measured flowvelocities.

5 Sample volume: increase in sample volume increases thestrength of the signal and more velocity information at theexpense of a lower spatial resolution. In general the smallestsample volume that results in adequate signal-to-noise ratioshould be used.

Color flow imagingThe PW Doppler measurement can be implemented for severalrange gates along the image line and can be repeated for eachimage line in order to obtain velocity information within a 2Dregion of interest. However, many ultrasound pulses need to betransmitted to reconstruct a single image line (as explained ear-lier), therefore the temporal resolution of such a system wouldbe extremely poor (a maximal frame rate of a few Hertz isobtained). In order to overcome this problem, color flow (CF)imaging has been developed. CF imaging allows estimation ofthe flow velocity based on only two ultrasound transmissions inthe same direction. Although in theory two pulses are indeedsufficient, in practice more pulses are used to improve the qual-ity of the measurement. Indeed, similar to PW Doppler, motionof the scattering sites between acquisitions will result in a timedelay between both reflected signals if motion of the tissue ispresent (Figure 1.17). By measuring this time delay betweenboth reflections, the amount of tissue displacement betweenboth acquisitions is obtained. The local velocity is then sim-ply calculated as this displacement divided by the time intervalbetween both acquisitions (= 1/PRF).

This procedure can be applied along the whole RF line inorder to obtain local velocity estimates along the line (fromclose to the transducer to the deepest structures). Moreover, it isrepeated between subsequent image lines within the 2D image.In this way, CF imaging can visualize the spatial distribution ofthe velocities by means of color superimposed onto the grayscaleimage. By convention, red represents velocities toward the trans-ducer and blue represents velocities away from the transducerwhereby different shades of red/blue indicate different velocityamplitudes. High variance of the velocity estimate measured ina given pixel is typically encoded by adding yellow or green tothe color in this pixel. In this way, regions with high variance inthe velocity estimate are highlighted as they indicate disturbed(i.e., turbulent) flow.

Echo 1

Echo 2

D Time

T

Figure 1.17 Principle of color flow imaging.