Cardiooncology

CardioOncology
EDITOR-IN-CHIEF STATISTICAL EDITOR DIRECTOR, PRODUCT MANAGEMENT,

Bonnie Ky, MD, MSCE, Cindy Green, PhD, DIGITAL PUBLISHING
Perelman School of Medicine at Duke Clinical Research Institute, Nandhini Kuntipuram, MCA, PMP,
the University of Pennsylvania, Durham, NC American College of Cardiology,
Philadelphia, PA Washington, DC
DIVISIONAL SENIOR DIRECTOR,
PUBLISHING EDITORIAL ASSISTANT
Justine Varieur Turco, MA, Michelle McMullen, RN, BSN, MA
DEPUTY EDITORS American College of Cardiology, American College of Cardiology,
Washington, DC Washington, DC
Saro H. Armenian, DO, MPH,
City of Hope, Los Angeles, CA EXECUTIVE MANAGING EDITOR
SOCIAL MEDIA COORDINATOR
Daniel J. Lenihan, MD, Eileen Cavanagh, MPS
Washington University School of Medicine, Tamika Edaire, BS,
American College of Cardiology,
St. Louis, MO American College of Cardiology,
Washington, DC
Washington, DC
ASSOCIATE EDITORS GUEST EDITOR EDITORIAL CONSULTANTS
Gregory T. Armstrong, MD, MSCE, Anju Nohria, MD, R. Aplenc, MD, PhD, MSCE,
St. Jude Children’s Research Hospital, Brigham and Women’s Hospital, Boston, MA Children's Hospital of Philadelphia, Philadelphia, PA
Memphis, TN
SENIOR ADVISORS D. Cardinale, MD, PhD,
Ana Barac, MD, PhD, European Institute of Oncology, Milan, Italy
Eugene Braunwald, MD,
MedStar Health, Washington, DC
Brigham and Women’s Hospital, Boston, MA J.R. Carver, MD,
Anne Blaes, MD, Susan Faye Dent, MD, Perelman School of Medicine at the University of
University of Minnesota Health, Minneapolis, MN Duke Health, Durham, NC Pennsylvania, Philadelphia, PA
Paul W. Burridge, PhD, Carol Fabian, MD, L. Cheng, MD, PhD,

Northwestern Medicine, Chicago, IL University of Kansas Medical Center, Zhongshan Hospital, Fudan University,
Kansas City, KS Shanghai, China
Kathryn J. Ruddy, MD,
Mayo Clinic, Rochester, MN Thomas Marwick, MBBS, PhD, MPH, R.F. Cornell, MD, MS,
Baker Heart & Diabetes Institute, Vanderbilt Health, Nashville, TN
Ronald Witteles, MD, Melbourne, Australia
Stanford University Medical Center, C.T. Dang, MD,
Richard M. Steingart, MD, Memorial Sloan Kettering Cancer Center,
Palo Alto, CA
Memorial Sloan Kettering Cancer Center, New York, NY
New York, NY
SOCIAL MEDIA/CME DIRECTORS A. DeMichele, MD, MSCE,
Perelman School of Medicine at the University of
Juan C. Lopez-Mattei, MD, INTERNATIONAL ADVISORS Pennsylvania, Philadelphia, PA
MD Anderson Cancer Center, Ludhmila A. Hajjar, MD, PhD,
Houston, TX N. Denduluri, MD,
Universidade de São Paulo, São Paulo, Brazil
Virginia Cancer Specialists, Arlington, VA
Javid J. Moslehi, MD, Zaza Iakobishvili MD, PHD,
Vanderbilt University Medical Center, Nashville, TN Rabin Medical Center, Petah Tikva, Israel P. Desvigne-Nickens, MD,
National Institutes of Health, Bethesda, MA
Teresa López-Fernández, MD,
SOCIAL MEDIA/CME CONSULTANTS Hospital Universitario La Paz, Madrid, Spain M.F. DiCarli, MD,
Brigham and Women's Hospital, Boston, MA
Sherry-Ann Brown, MD, PhD, Alexander Lyon, MD, PhD,
Mayo Clinic, Rochester, MN Imperial College, London, United Kingdom S.M. Domchek, MD,
Perelman School of Medicine at the University of
Ryan P. Daly, MD, Paaladinesh Thavendiranathan, MD,
Pennsylvania, Philadelphia, PA
Franciscan Health, Indianapolis, IN Toronto General Hospital, Toronto, Canada
Carlo Gabriele Tocchetti, MD, PhD, P.S. Douglas, MD,

Naveen Pemmaraju, MD,
Federico II University, Naples, Italy Duke Health, Durham, NC
MD Anderson Cancer Center,
Houston, TX Yunlong Xia, MD, PhD, J.-B. Durand, MD,
Dalian Medical University, Dalian, China MD Anderson Cancer Center, Houston, TX
Nosheen Reza, MD,
Perelman School of Medicine at the University of Yuhui Zhang, MD, J.C. Fang, MD,
Pennsylvania, Philadelphia, PA Fuwai Hospital, Beijing, China University of Utah Health, Salt Lake City, UT
EDITORIAL C ONSULTANTS CONTINUED EDITORS-IN-CHIEF
JACC
G.M. Freedman, MD, S.K. Pal, MD, ASSISTANT EDITORS
Valentin Fuster, MD, PhD ,
Perelman School of Medicine at the City of Hope,
A.H. Asnani, MD, Mount Sinai School of Medicine,
University of Pennsylvania, Los Angeles, CA
Beth Israel Deaconess, New York, NY
Philadelphia, PA
J.C. Plana Gomez, MD, Boston, MA
M.M. Givertz, MD, Baylor College of Medicine, JACC: Cardiovascular
R.K. Cheng, MD,
Brigham and Women's Hospital, Houston, TX Interventions
University of Washington,
Boston, MA David J. Moliterno, MD ,
J.P. Plastaras, MD, Seattle, WA
N.B. Haas, MD, University of Kentucky,
Perelman School of Medicine at the
B. Demissei, MD, MSc, PhD, Lexington, KY
Perelman School of Medicine at the University of Pennsylvania,
University of Pennsylvania, Philadelphia, PA
University of Pennsylvania, JACC: Cardiovascular Imaging
Philadelphia, PA
C.B. Porter, MD, Philadelphia, PA Y. Chandrashekhar, MD, DM ,
J. Herrmann, MD, University of Kansas School of University of Minnesota/VAMC,
Mayo Clinic, Rochester, MN
M.J. Ehrhardt, MD, MS,
Medicine, Kansas City, KS Minneapolis, MN
St. Jude Children’s Research Hospital,
W.G. Hundley, MD, C.G. Robinson, MD, Memphis, TN
JACC: Heart Failure
VCU Health Pauley Heart Center,
Washington University in St. Louis, M. Fradley, MD,
Richmond, VA Christopher M. O’Connor, MD,
St. Louis, MO University of South Florida, Inova Heart and Vascular Institute,
J.L. Januzzi, MD, Tampa, FL Falls Church, VA
R.R. Russell, MD, PhD,
Massachusetts General Hospital,
Rhode Island Hospital, S. Ganatra, MD,
Boston, MA JACC: Clinical
Providence, RI Lahey Clinic, Burlington, MA
D.B. Johnson, MD, MSCI, Electrophysiology
J.-E. Salem, MD, PhD, K.D. Getz, MPH, PhD,
Vanderbilt Health, Nashville, TN David J. Wilber, MD ,
Hôpitaux de Paris, Perelman School of Medicine at the Loyola University Medical Center,
D.E. Lanfear, MD, Paris, France University of Pennsylvania, Chicago, IL
Henry Ford Health System, Philadelphia, PA
Detroit, MI D.B. Sawyer, MD, PhD,
JACC: Basic to Translational
Maine Health, Portland, ME J.D. Groarke, MD, MPH,
M.J. Leja, MD, Brigham and Women's Hospital, Science
Michigan Medicine, Ann Arbor, MI M. Scherrer-Crosbie, MD, PhD,
Boston, MA Douglas L. Mann, MD ,
J.E. Liu, MD, Washington University School of
University of Pennsylvania, B.C. Jensen, MD,
Memorial Sloan Kettering Cancer Medicine,
Philadelphia, PA University of North Carolina Hospitals,
Center, New York, NY St. Louis, MO
Chapel Hill, NC
N. Shelburne, MS, CRNP,
D.A. Mankoff, MD, PhD, National Cancer Institute, K.J. Leger, MD, MSc, JACC: Case Reports
Perelman School of Medicine at the Bethesda, MD Seattle Children’s, Seattle, WA Julia Grapsa, MD, PhD ,
University of Pennsylvania, Barts Health NHS Trust,
Philadelphia, PA K.M. Smith, MD, C. Melloni, MD, London, UK
Johns Hopkins Medicine, Duke Health, Durham, NC
K.B. Margulies, MD,
Baltimore, MD
Perelman School of Medicine at the J.D. Mitchell, MD,
University of Pennsylvania, W.H.W. Tang, MD, Washington University School of
Philadelphia, PA Cleveland Clinic, Medicine, St. Louis, MO
Cleveland, OH
M.J. Mauro, MD, H. Narayan, MD, MSCE,
Memorial Sloan Kettering Cancer E.J. Velazquez, MD, University of San Diego,
Center, New York, NY Yale School of Medicine, San Diego, CA
New Haven, CT T.M. Okwuosa, DO,
E.L. Mayer, MD, MPH,
Dana-Farber Cancer Institute, M.N. Walsh, MD, Rush University, Chicago, IL
Boston, MA St. Vincent Medical Group, C. Shenoy, MBBS,
L. Mehta, MD, Indianapolis, IN University of Minnesota Health,
Ohio State University Wexner Minneapolis, MN
C.W. Yancy, MD,
Medical Center, Dublin, OH Northwestern University Feinberg J. Upshaw, MD,
L. Minasian, MD, School of Medicine, Tufts Medical Center, Boston, MA
National Cancer Institute, Chicago, IL
E.H. Yang, MD,
Bethesda, MD
E.T. Yeh, MD, Ronald Reagan UCLA Medical Center,
T.G. Neilan, MD, University of Missouri School of Los Angeles, CA
Massachusetts General Hospital, Medicine, Columbia, MO
A.F. Yu, MD,
Boston, MA
V.G. Zaha, MD, PhD, Memorial Sloan Kettering
K.C. Oeffinger, MD, University of Texas Southwestern, Cancer Center,
Duke Health, Durham, NC Dallas, TX New York, NY
CardioOncology
2019-2020 OFFICERS
Richard J. Kovacs, MD, FACC,
President
Athena Poppas, MD, FACC,
Vice President
Howard “Bo” T. Walpole, Jr., MD, MBA, FACC,
Treasurer
Akshay Khandelwal, MD, FACC,
Secretary and Board of Governors Chair

Timothy W. Attebery, DSc, MBA, FACHE,
Chief Executive Officer
2019-2020 PUBLICATIONS AND EDITORIAL COORDINATION COMMITTEE
Viviany R. Taqueti, MD, MPH, FACC, Chair

Rhonda M. Cooper-DeHoff, MD, FACC, Annual Scientific Session Program Committee
Prasad C. Gunasekaran, MD, FIT Representative
Fadi G. Hage, MD, FACC
Fred M. Kusumoto, MD, FACC, Awards Committee
Renato D. Lopes, MD, PhD, FACC
Sandra M. Oliver-McNeil, DNP, ACNP-BC, AACC Committee
CORRESPONDENCE FOR
AMERICAN COLLEGE OF James E. Tcheng, MD, FACC, (Ex Officio) Chair, Digital Steering Committee
CARDIOLOGY
John U. Doherty, MD, FACC
All correspondence for the College, Syed Tanveer Rab, MBBS, MACC
other than that related to
Janice Sibley, MS, MA, ACC Executive Vice President, Education and Publishing Divisions
JACC: CardioOncology should be
sent to Resource Center, Justine Varieur Turco, MA, Divisional Senior Director, Publishing
American College of Cardiology,
2400 N Street,
NW, Washington, DC 20037
JACC: CARDIOONCOLOGY VOL. 2, NO. 1, 2020
ª 2020 THE AUTHORS. PUBLISHED BY ELSEVIER ON BEHALF OF THE AMERICAN
COLLEGE OF CARDIOLOGY FOUNDATION. THIS IS AN OPEN ACCESS ARTICLE UNDER
THE CC BY-NC-ND LICENSE (http://creativecommons.org/licenses/by-nc-nd/4.0/).
ORIGINAL RESEARCH
Cardiac Adverse Events in EGFR-Mutated

Non-Small Cell Lung Cancer Treated
With Osimertinib
Kei Kunimasa, MD, PHD,a,* Risa Kamada, MD,b,* Toru Oka, MD, PHD,b Makiko Oboshi, MD, PHD,b
Madoka Kimura, MD,a Takako Inoue, MD,a Motohiro Tamiya, MD,a Tatsuya Nishikawa, MD, PHD,b
Taku Yasui, MD, PHD,b Wataru Shioyama, MD, PHD,b Kazumi Nishino, MD, PHD,a Fumio Imamura, MD, PHD,a
Toru Kumagai, MD, PHD,a Masashi Fujita, MD, PHDb
ABSTRACT
OBJECTIVES The purpose of this study was to assess osimertinib-associated cardiac adverse events (AEs) in a
real-world setting, using a retrospective single-center cohort study in Japan.
BACKGROUND Cases of osimertinib-associated cardiac AEs have been reported but remain poorly understood.
METHODS A total of 123 cases of advanced non–small cell lung cancer (NSCLC) with confirmed EGFR mutations who
received osimertinib monotherapy from 2014 to 2019 at the Osaka International Cancer Institute (Osaka, Japan) were
evaluated. Cardiac AEs were defined according to Common Terminology Criteria for Adverse Events (CTCAE) version 5.0.
Changes in left ventricular ejection fraction (LVEF) and rates of cancer therapeutics-related cardiac dysfunction (CTRCD),
defined as a $10 % absolute decline in LVEF from baseline to a value of <53%, were further assessed in 36 patients in
whom serial measurements of LVEF were obtained before and during osimertinib treatment.
RESULTS Severe cardiac AEs (CTCAE grade 3 or higher) occurred in 6 patients (4.9%) after osimertinib administration.
These AEs included acute myocardial infarction (n ¼ 1), heart failure with reduced LVEF (n ¼ 3), and valvular heart disease
(n ¼ 2). Five of the 6 patients had a history of cardiovascular risk factors or disease. Myocardial biopsies in 2 of the
patients showed cardiomyocyte hypertrophy and lipofuscin deposition. In 36 patients assessed with serial LVEF, LVEF
declined from 69.4 4.2% to 63.4 10.5% with osimertinib therapy (p < 0.001). CTRCD occurred in 4 patients with a
nadir LVEF of 40.3 9.1% with osimertinib.
CONCLUSIONS In this retrospective cohort analysis, the incidence of cardiac AEs in patients treated with osimertinib
was 4.9%. Additional prospective data collected from patients with NSCLC treated with osimertinib will be important
in understanding the incidence, pathophysiology, and management of cardiac AEs with osimertinib.
(J Am Coll Cardiol CardioOnc 2020;2:1–10) © 2020 The Authors. Published by Elsevier on behalf of the
American College of Cardiology Foundation. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
From the aDepartment of Thoracic Oncology, Osaka International Cancer Institute, Osaka, Japan; and the bDepartment of Onco-
Cardiology, Osaka International Cancer Institute, Osaka, Japan. *Drs. Kunimasa and Kamada contributed equally to this work.
Supported by Japan Society for the Promotion of Science KAKEN grant JP16K09470 to Dr. Oka. Dr. Nishino has received speaker
honoraria from AstraZeneca, Chugai Pharmaceutical Co., Boehringer Ingelheim, and Novartis Pharmaceuticals; and has received
research support from Nippon Boehringer Ingelheim. Dr. Imamura has received honoraria and research support from AstraZeneca.
Dr. Kumagai has received speaker honoraria and research support from AstraZeneca, Chugai Pharmaceutical, Pfizer, Nippon
Boehringer Ingelheim, and Delta-Fly Pharma. All other authors have reported that they have no relationships relevant to the
contents of this paper to disclose.
ISSN 2666-0873 https://doi.org/10.1016/j.jaccao.2020.02.003

2 Kunimasa et al. JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020
Osimertinib-Associated Cardiac Adverse Events in NSCLC Patients MARCH 2020:1–10
O simertinib (Tagrisso, AstraZeneca)

ABBREVIATIONS
F I G U R E 1 Study Cohort
AND ACRONYMS is an orally available, third-
generation epidermal growth
ACE = angiotensin-converting
enzyme
factor receptor (EGFR)-tyrosine kinase inhib-
itor (TKI) that selectively inhibits both EGFR
AE = adverse event
TKI sensitizing and the EGFR T790M resis-
ARB = angiotensin II receptor
blocker tance mutation with less activity against
CTCAE = common terminology
wild-type EGFR (1). The AURA (A Phase I/II,
criteria for adverse event Open-Label, Multicentre Study to Assess the
CTRCD = cancer therapeutics- Safety, Tolerability, Pharmacokinetics and
related cardiac dysfunction Anti-tumour Activity of Ascending Doses of
EGRF = epidermal growth AZD9291 in Patients with Advanced Non
factor receptor
Small Cell Lung Cancer w have Progressed
HER = human epidermal Following Prior Therapy with an Epidermal
growth factor receptor
GrowthFactor Receptor Tyrosine Kinase In-
LVIDd = left ventricular
hibitor Agent) and FLAURA (AZD9291 Versus
internal end-diastolic diameter
Gefitinib or Erlotinib in Patients With Locally
LVIDs = left ventricular
internal end-systolic diameter Advanced or Metastatic Non-small Cell Lung
LVEF = left ventricular ejection
Cancer) trials have consistently demon-
fraction strated superior clinical activity and relative
MR = mitral regurgitation safety of osimertinib in advanced non–small
cell lung cancer (NSCLC) patients with EGFR CTRCD, as defined by a decrease in LVEF of more than 10 % to
NT-proBNP = N-terminal pro–
<53% LVEF, was observed in 4 patients (black box). The
B-type natriuretic peptide mutations regardless of their EGFR T790M
remaining 32 patients did not experience CTRCD (white box).
NSCLC = non–small cell lung mutation status (2,3). The incidence of Grade
CTRCD ¼ cancer therapeutics-related cardiac dysfunction;
cancer 3 or higher Common Terminology Criteria for LVEF ¼ left ventricular ejection fraction; NSCLC ¼ non–small
PASP = pulmonary artery Adverse Events (CTCAE) version 5.0 resulting cell lung cancer.
systolic pressure
in permanent discontinuation was lower
TKI = tyrosine kinase inhibitor
for osimertinib than in the first-generation
TR = tricuspid regurgitation EGFR-TKIs (4). Most recently, patients with
prolongation occurred in only 0.1% and other Grade 3
VEGF = vascular endothelial previously untreated, advanced EGFR
growth factor or higher cardiac adverse events (AEs) in 0.8% (6).
mutant NSCLC had longer survival with osi-
mertinib than those treated with EGFR TKIs, including SEE PAGE 11
gefitinib or erlotinib (3,5). With the success of these

Recently, severe osimertinib-associated cardiac
clinical trials, osimertinib has become the standard
dysfunction was observed in NSCLC patients with
care for patients with advanced NSCLC with EGFR
EGFR mutations who were treated with osimertinib.
mutations and has become the first-line therapy
These experiences prompted a detailed retrospective
for NSCLC patients with the EGFR T790M mutation
analysis in a single-center experience to further un-
who have progressed on previous EGFR-TKI
derstand cardiac AEs associated with osimertinib in a
treatment.
real-world setting in EGFR mutation-carrying NSCLC
Safety information for osimertinib from interna-
patients.
tional, large clinical trials have reported Grade 3 or
higher corrected QT (QTc) prolongation in 1% of pa- SUBJECTS AND METHODS
tients, any grade cardiac failure in 4%, and a decline
in left ventricular ejection fraction (LVEF) >10% from This was a retrospective, single-center, observational
baseline and to <50% in 3% of patients (3). However, study conducted at the Osaka International Cancer
in a post-marketing surveillance study of osimertinib Institute (OICI), Osaka, Japan. The study protocol and
in 3,578 Japanese NSCLC patients with EGFR consent forms were approved by the local ethics
mutations, the incidence of Grade 3 or higher QTc committee at OICI (approval 19017). The electronic
The authors attest they are in compliance with human studies committees and animal welfare regulations of the authors’
institutions and Food and Drug Administration guidelines, including patient consent where appropriate. For more information,
visit the JACC: CardioOncology author instructions page.
Manuscript received August 28, 2019; revised manuscript received January 31, 2020, accepted February 3, 2020.
JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020 Kunimasa et al. 3
MARCH 2020:1–10 Osimertinib-Associated Cardiac Adverse Events in NSCLC Patients
medical records of all NSCLC patients with an EGFR

T A B L E 1 Patient Characteristics of Osimertinib-Treated NSCLC
mutation treated with osimertinib between 2014 and Patients (N ¼ 123)
2019 at the OICI hospital were reviewed. Inclusion
Age, yrs 69 (33-86)
criteria were a diagnosis of histologically and cyto- Females 83 (67.5)
logically confirmed NSCLC; clinical stage IIIB-IV can- Smoking status
cer, according to the American Joint Committee for Former, current 35 (28.5)
Cancer staging system, 8th edition; presence of a Histological type
sensitizing, activating EGFR mutation; and treatment Adenocarcinoma 122 (99.2)

EGFR mutation status at osimertinib administration
history of osimertinib monotherapy. Patients were
Ex.19 del 26 (21.1)
excluded if they had discontinued osimertinib treat-
L858R 21 (17.1)
ment within 2 weeks for any reason or if their medical
T790M 1 (0.8)
records were unobtainable. All AEs associated with Ex.19 del þ T790M 36 (29.3)
osimertinib were graded according to CTCAE version L858R þ T790M 36 (29.3)
5.0 criteria. AEs were adjudicated by a cardiologist G719X 3 (2.4)
based upon detailed medical record review. Osimertinib treatment line
Within this cohort, there were 72 patients who 1st 33 (26.8)

2nd 32 (26.0)
underwent serial electrocardiography (ECG) and 36
3rd 19 (15.4)
who underwent echocardiography both before and
>4th 39 (31.7)
after osimertinib administration. QTc and echocar- Medical history
diographic parameters were compared. Cancer Hypertension 35 (28.5)
therapeutics-related cardiac dysfunction (CTRCD) Diabetes mellitus 3 (2.4)
was defined, according to the American Society of Dyslipidemia 11 (8.9)
Echocardiography/European Association of Cardio- Hyperuricemia 1 (0.8)

Arrhythmia 6 (4.9)
vascular Imaging, as a greater than 10 % decline in
Heart failure 1 (0.8)
LVEF from baseline to a value of <53% (7). Patients’
Vasospastic angina 1 (0.8)
characteristics and cardiac functions in the non-
Thoracic aortic aneurysm 2 (1.6)
CTRCD group (n ¼ 32) were compared with those in Atrial septal defect 1 (0.8)
the CTRCD group (n ¼ 4) (Figure 1). Valvular disease (moderate and over) 4 (3.3)
VTE 2 (1.6)
ELECTROCARDIOGRAPHY AND ECHOCARDIOGRAPHY.
Medication
Twelve-lead ECGs were obtained using a standard-
Beta-blocker 5 (4.1)
ized ECG machine (Fukuda Denshi, Tokyo, Japan) and CCB 17 (13.8)
QTc was performed using the Bazett formula. Echo- ACE inhibitor 1 (0.8)
cardiography was performed using an IE33 ultraso- ARB 15 (12.2)
nography machine (Philips, Stockton, California). LV MRA 3 (2.4)
end-diastolic and end-systolic diameters (LVIDd and Loop diuretic agent 3 (2.4)
LVIDs) were obtained by 2-dimensional (2D) imaging,

Values are median (interquartile range) or n (%).
and LVEF was obtained using the Teichholz method. ACE ¼ angiotensin-converting enzyme inhibitor; ARB ¼ angiotensin II receptor
Echocardiographic data were extracted from the re- blocker; CCB ¼ calcium channel blocker; MRA ¼ mineralocorticoid receptor
antagonist; VTE ¼ venous thromboembolism.
cords within 2 years before and 1 year after osi-
mertinib initiation. The lowest LVEF data
were extracted for analysis in 18 patients who un-
RESULTS
derwent repeated echocardiography after osimerti-
nib initiation.
OSIMERTINIB-ASSOCIATED CARDIAC ADVERSE
STATISTICAL ANALYSES. Within- and between- EVENTS IN NSCLC PATIENTS. From 2014 to 2019, 123
group comparisons of changes in cardiac parameters NSCLC patients with EGFR mutations were treated
were made and tested with the paired t-test, using with osimertinib (Table 1). Seven of those patients
Prism version 6.02 software (GraphPad, San Diego, were enrolled in clinical trials (AURA and FLAURA)
California). Nonparametric comparisons were per- (2,3). The median age of those patients was 69 years
formed using the Mann-Whitney U test. Results were (interquartile range 33 to 86); 83 patients (67.5%)
expressed as mean SD or median (interquartile were female; and 122 patients (99.2%) had adenocar-
range [IQR]), and a p value of <0.05 was considered cinoma (Table 1). The EGFR mutation profiles were as
statistically significant. follows: Exon (Ex.) 19 deletion (del), which occurred
T A B L E 2 Cases of Osimertinib-Induced Cardiac Adverse Events
Case 1 Case 2 Case 3 Case 4 Case 5 Case 6
Age, yrs 78 71 68 64 52 71
Sex Female Female Male Female Female Female
EGFR mutation L858R L858R＋T790M Ex.19 del＋T790M L858R＋T790M L858R L858R
Osimertinib line 2nd 3rd 3rd 3rd 1st 1st
Osimertinib effect PR PR PR PR PR NE
Tobacco No No No Yes (former) Yes (former) No
CVD risk/history HTN HTN Moderate AR Moderate MR Obesity HTN
Aortic aneurysm HU DM
Daily medications Carvedilol 20 mg; Amlodipine 5 mg Allopurinol 100 mg; Prednisolone 7.5 mg - Candesartan 8 mg
Nifedipine CR 40 mg; Prednisolone 5 mg
Azilsartan 40 mg;
Rosuvastatin 5 mg
Symptoms Exertional dyspnea; Fatigue Leg edema; Facial/leg edema - Chest pain;
leg edema fatigue dyspnea
Cardiac event Heart failure; MR progression; TR progression EF decline; HTN; EF decline Acute myocardial
QT prolongation mitral valve MR progression infarction
prolapse
CTCAE Grade 3 3 3 3 3 4
Time to event 3 months 3 months 1 month 9 months 2 weeks 2 months
NT-proBNP or BNP, pg/ml NT-proBNP 18,890 BNP 21.9 NT-proBNP 450 BNP 227.9 NT-proBNP 36 BNP 423.4
LVEF before osimertinib, % 61 82 74 72 63 69
LVEF after osimertinib 28 74 60 50 41 42
initiation, %
CTRCD Yes No No Yes Yes Yes
RV biopsy Yes No No No Yes No
Valvular disease MR; () MR; trace (0–1) TR; trace (0–1) MR; mild–moderate (2–3) - -
/ severe (3–4) / severe (3–4), / moderate–severe (3-4) / moderate (3)
prolapse
Osimertinib treatment Discontinued Discontinued Reduced dosage from 80 mg to Temporarily held and Discontinued Discontinued
40 mg every other day resumed at 80 mg daily
Subsequent cancer therapy Gefitinib Gefitinib Osimertinib rechallenge Osimertinib rechallenge Afatinib Erlotinib
Treatment for cardiac event Furosemide 40 mg; Furosemide 20 mg Furosemide 40 mg; Furosemide 20 mg; Candesartan 4 mg PCI for LAD 6
(daily medications) Spironolactone 50 mg; Tolvaptan 3.75 mg Spironolactone 25 mg
Candesartan 2 mg;
Carvedilol 5 mg;
Tolvaptan 7.5 mg
Return of LVEF to baseline No Yes Yes Yes Yes No
48% after 74% after 72% after 62% after 63% after 54% after
9 months 6 months 2 months 14 months 2 months 7 months
AR ¼ aortic valve regurgitation; BNP ¼ brain natriuretic peptide; CTCAE ¼ common terminology criteria of adverse event; CTRCD ¼ cancer therapeutics-related cardiac dysfunction; CVD ¼ cardiovascular
disease; DM ¼ diabetes mellitus; LVEF ¼ left ventricular ejection fraction; HTN ¼ hypertension; HU ¼ hyperuricemia; LAD ¼ left anterior descending coronary artery; MR ¼ mitral regurgitation; NA ¼ not
available; NE ¼ not evaluable; NT-proBNP ¼ N-terminal pro–B-type natriuretic peptide; PCI ¼ percutaneous coronary intervention; PR ¼ partial response; RV ¼ right ventricle; TR ¼ tricuspid regurgitation.
in 21.1% of patients; L858R, which occurred in 17.1% mutation was treated with osimertinib as the
of patients; and resistance mutation T790M, which second-line therapy. She had a history of
occurred in 59.4% of patients. Osimertinib was used hypertension for more than 30 years and had
as first-line therapy in 26.8% of patients. Grade 3 or undergone aortic graft replacement due to thoracic
higher CTCAE version 5.0 cardiac AEs were observed aortic aneurysm. Her daily medications were
in 6 patients (4.9%) (Table 2). Five of those patients azilsartan 40 mg; nifedipine 40 mg; carvedilol
were female, and all 5 had a history of cardiovascular 20 mg; and rosuvastatin 5 mg. After 3 months of
diseases and/or cardiovascular risk factors. Their osimertinib treatment, she presented with dyspnea.
cardiac AEs included acute myocardial infarction An ECG showed that the QTc was prolonged to
(n ¼ 1), heart failure with reduced LVEF (n ¼ 3), and 562 ms. N-terminal pro–B-type natriuretic peptide
valvular heart disease (n ¼ 2) (Table 2). The 6 CTCAE (NT-proBNP) and troponin I levels were elevated at
grade 3 or higher cardiac AEs are described in detail 18,890 pg/ml and 0.416 ng/ml, respectively.
below and presented in Table 2. Echocardiography revealed severely reduced LVEF
CASE SERIES. C a s e 1 . A 78-year-old female patient at 28% and an increased LVIDd/LVIDs at 44/38 mm,
with advanced NSCLC harboring an EGFR L858R with moderate to severe mitral regurgitation (MR).
F I G U R E 2 Histology Assessment of Myocardial Biopsy Specimens
Hematoxylin-eosin-stained specimens of myocardial biopsy are shown at an original magnification of 200. (A) In Case 1, cardiomyocytes
were moderately hypertrophied with disarray and deposition of lipofuscin. There was slight interstitial edema and fibrosis but no inflam-
matory cell infiltration. (B) In Case 5, cardiomyocytes were mildly hypertrophied with focal vacuolization and deposition of lipofuscin,
indicative of myocyte damage. Interstitial edema and fatty infiltration with partial fibrosis were observed around the vasculature, whereas
lymphocyte infiltration was modest.
Before osimertinib administration, her baseline LVEF 1 month, her symptoms improved, but the MR per-
was 61% and LVIDd/LVIDs were 44/29 mm. Coronary sisted. Given the severity of MR, 250 mg of gefitinib
angiography revealed no significant coronary artery was started as the fourth-line therapy. Four months
stenosis, and myocardial biopsy was performed after gefitinib therapy, her MR did not worsen, and
to further understand the pathophysiology of her advanced NSCLC did not progress.
her heart failure. Histopathology analysis of right C a s e 3 . A 68-year-old male was treated with osi-
ventricle biopsy specimens revealed moderately mertinib as the third-line therapy for advanced
hypertrophied cardiomyocytes with disarray and NSCLC harboring the EGFR Ex.19 del. and T790M
deposition of lipofuscin (Figure 2A). Mild interstitial mutations. He was previously treated with gefitinib
edema and fibrotic changes were observed, but and erlotinib for more than 7 years before osimerti-
no inflammatory cell infiltration was detected. Due nib. He had a history of hepatitis B but no history of
to her long-term history of hypertension, it was cardiovascular disease. After 3 months of osimertinib,
believed that her cardiac dysfunction was accelerated he presented with fatigue and pitting edema of the
with osimertinib in the setting of hypertensive lower extremities. ECG showed QTc prolongation,
heart disease. Osimertinib was discontinued, and and serum laboratory tests showed severe hypona-
furosemide 40 mg, spironolactone 50 mg, tolvaptan tremia (120 mEq/l). Echocardiography revealed
7.5 mg, carvedilol 5 mg, and candesartan 2 mg, daily severe tricuspid valve regurgitation (TR) and mild
were initiated and resulted in an improvement of her pulmonary hypertension (estimated pulmonary
heart failure symptoms. Her LVEF remained at 26% artery systolic pressure: 45 mm Hg), which had
after 3 months of osimertinib discontinuation but not been observed before osimertinib admini-
gradually improved to 48% after 9 months. stration. Osimertinib was discontinued, and tol-
C a s e 2 . A 72-year-old female patient was treated with vaptan 3.75 mg, and furosemide 40 mg daily, were
osimertinib as a third-line therapy for advanced initiated. After 1 month, fatigue and edema improved,
NSCLC harboring the EGFR L858R and T790M muta- but severe TR persisted. Because of the presence of
tions. She had had a history of hypertension for 15 the EGFR T790M mutation and lack of alternative
years, which was treated with amlodipine 5 mg daily. EGFR-TKIs to consider, low-dose osimertinib (40 mg
After 3 months of osimertinib, she presented with daily) was initiated with careful follow-up. After
fatigue. Echocardiography revealed severe MR with 6 months of re-administration of osimertinib, TR and
mitral valve prolapse and LVEF of 74%, which was NSCLC did not progress.
not detected by the echocardiography obtained C a s e 4 . A 64-year-old female was treated with osi-
before osimertinib treatment. Osimertinib was dis- mertinib as the third-line therapy for advanced
continued, and furosemide 20 mg, was started. After NSCLC harboring the EGFR L858R and T790M
QTc, and NT-proBNP and troponin I concentrations

T A B L E 3 Characteristics of 36 NSCLC Patients Treated With
Osimertinib with Serial Echocardiography Monitoring
were also within normal ranges at 36 pg/ml
and <0.01 ng/ml, respectively. Coronary angiography
All Non-CTRCD CTRCD
(N ¼ 36) (n ¼ 32) (n ¼ 4)
revealed no significant stenosis in her coronary ar-
Age, yrs 68 (63–74) 68 (63–74) 68 (55–76) teries, and right ventricular myocardial biopsy was
Females 26 (72.2) 22 (68.8) 4 (100) performed. Biopsy analysis revealed mildly hyper-
Smoking 12 (33.3) 10 (34.3) 2 (50) trophied cardiomyocytes with deposition of lip-
Medical history ofuscin, focal vacuolization, and focal degradation
Hypertension 17 (47.2) 15 (46.9) 2 (50) (Figure 2B). Edematous changes and fatty infiltration
Diabetes mellitus 1 (2.8) 0 1 (25)
with partial fibrosis around the vasculature were
Dyslipidemia 5 (13.9) 5 (15.6) 0
detected in the interstitial space, and a small amount
Hyperuricemia 1 (2.8) 1 (3.1) 0
Atrial fibrillation 2 (5.6) 2 (6.2) 0
of inflammatory cell infiltration was observed, but
Vasospastic angina 1 (2.8) 1 (3.1) 0 there were no signs of cardiomyocyte death. After
Thoracic aortic aneurysm 1 (2.8) 0 1 (25) osimertinib was discontinued and candesartan 4 mg
Atrial septal defect 1 (2.8) 1 (3.1) 0 daily, was initiated, the LVEF recovered to 63%. The
Valvular disease (moderate 4 (11.1) 3 (3.9) 1 (25) patient was then treated with afatinib, a second-
or greater)
generation of EGFR-TKI, and her cardiac function
Medication
Beta-blocker 2 (5.6) 1 (3.1) 1 (25)
remained stable.
CCB 9 (25) 8 (25.0) 1 (25) C a s e 6 . A 71-year-old woman was treated with osi-
ACE inhibitor 0 0 0 mertinib as the first-line therapy for advanced case of
ARB 8 (22.2) 6 (18.8) 2 (50) NSCLC harboring the EGFR L858R mutation. She had
MRA 1 (2.8) 1 (3.1) 0 a long-standing history of hypertension, which had
Loop diuretic 1 (2.8) 1 (3.1) 0
been treated with 8 mg of candesartan daily, and
Values are median (interquartile range) or n (%).
diabetes mellitus. After 2 months of osimertinib, she
ACE ¼ angiotensin-converting enzyme; ARB ¼ angiotensin II receptor blocker; presented with sudden dyspnea at rest following
CCB ¼ calcium channel blocker; CTRCD ¼ cancer therapeutics-related cardiac
dysfunction; LVEF ¼ left ventricular ejection fraction; MRA ¼ mineralocorticoid
chest pain. ECG showed QS pattern in V1–4 . Her crea-
receptor antagonist. tine phosphokinase concentration was 163 U/l, her
creatine kinase-myocardial band was 7.2 ng/ml, and
her troponin I level was elevated at 2.945 ng/ml.
mutations. She had a history of untreated moderate
Echocardiography showed anteroseptal hypokinesis
MR. Before osimertinib initiation, she had received
and an LVEF of 42%. Coronary angiography revealed
erlotinib and experienced erlotinib-associated inter-
99% stenosis with delayed flow in the left anterior
stitial pneumonitis, which was treated with prednis-
descending artery. Percutaneous coronary interven-
olone. After 9 months of osimertinib, she presented
tion of the left anterior descending artery was per-
with fatigue and pitting edema of her lower extrem-
formed. The patient was then treated with erlotinib
ities. NT-proBNP was elevated at 227.9 pg/ml and
instead of osimertinib.
LVEF decreased from 72% to 50%. Osimertinib was
discontinued, and she was treated with furosemide OSIMERTINIB RECHALLENGE OR DISCONTINUATION. In
20 mg; and spironolactone 25 mg daily. After 1 month each of the above cases, the decision regarding the
of close observation, her symptoms resolved, risks and benefits of continuing osimertinib was made
LVEF improved to 58%, and NT-proBNP decreased on an individual basis, after detailed consideration of
to 75.2 pg/ml. Osimertinib was restarted at 80 mg the patients’ histories and preferences. For the pa-
daily after LVEF improvement. Follow-up showed tient described in case 1, LVEF recovery was delayed,
there were no signs of additional cardiac dysfunction; and gefitinib was started 8 months after the AE.
LVEF was most recently reported at 62% after The patient in case 2 refused osimertinib re-
6 months. administration and also received gefitinib. Osimerti-
C a s e 5 . A 52-year-old woman was treated with osi- nib was re-administered to the patients described in
mertinib as the first-line therapy for her advanced cases 3 and 4, given the presence of the T790M mu-
NSCLC harboring the EGFR L858R mutation. She had tation, whereas patients were treated with other
no history of cardiovascular risk factors or disease EGFR-TKIs in cases 5 and 6. Altogether, osimertinib
except for obesity (body mass index: 29.7 kg/m 2). was discontinued in 4 of 6 patients and restarted in 2
After 2 weeks of osimertinib, screening echocardiog- patients; in one of those patients, osimertinib was re-
raphy revealed a reduction of LVEF from 63% to 41% administered at the reduced dose and at the original
without cardiac symptoms. An ECG showed normal dose in the other patient.
(between 450 and 480 ms) in 18 patients. No episodes

F I G U R E 3 Changes in LVEF Before and After Osimertinib
Administration
of fatal arrhythmia were documented in this cohort.
Among the 123 patients, 36 NSCLC patients were
identified who underwent echocardiography both
before and after osimertinib administration within a 1
years’ time period (Figure 1). Echocardiography was
performed in 10 patients as part of the protocol for the
FLAURA study or other clinical trials and in 12 pa-
tients for suspected cardiovascular complications.
The patient characteristics of these 36 patients with
serial echocardiography are displayed in Table 3. The
median age was 68 years old; 72.2% were female; and
47.2% had hypertension. Overall, LVEF significantly
decreased after osimertinib treatment from 69.4
4.2% at baseline to 63.4 10.5% (p < 0.001) (Figure 3,
Table 4), and LVIDd/LVIDs also significantly
increased from 42.6 4.5/26.3 3.3 mm to 44.5 5.2/
29.1 5.6 mm (p ¼ 0.005) (Table 4). Although the
changes in LVEF were relatively small and the
Echocardiography was performed within at 2 years before and average LVEF was still in the normal range, the de-
1 year after osimertinib initiation. The LVEF of each individual clines in LVEF in this subgroup of patients were sta-
patient is indicated as a round dot. Mean LVEFs are indicated tistically significant.
cross marks in the boxes. Overall, LVEF significantly
Among the 36 patients, 4 patients experienced
decreased after osimertinib (p < 0.001). Abbreviations as in
declines in LVEF of > 10% to <53% after osimertinib
Figure 1.
(mean LVEF: 40.3 9.1%) (Figure 1, Table 4), corre-
sponding to a widely accepted definition of CTRCD
(7). These patients were also categorized as CTCAE
ASSESSMENT OF SERIAL CHANGES QTc AND LVEF grade 3 of LVEF decreased. As expected, patients
IN NSCLC PATIENTS TREATED WITH OSIMERTINIB. with CTRCD had a significantly lower LVEF and larger
Among the 123 NSCLC patients treated with osi- LVIDd/lVIDs than those who did not have CTRCD
mertinib, 72 patients underwent serial ECG before (Table 4). However, the early mitral inflow velocity
and after osimertinib initiation. The average QTc was and mitral annular early diastolic velocity (E/e 0 ) ratio
significantly prolonged from 421.9 23.0 ms to was similar in both groups with analyzable diastolic
442.4 33.2 ms (p < 0.001) over a median of 116 days function parameters; this did not change with osi-
(interquartile range 16 to 851 days) after osimertinib mertinib (12.8 4.6% and 12.6 3.3%, respectively,
initiation. Grade 3 QTc prolongation >501 ms was in the non-CTRCD group [n ¼ 28] and 14.6 6.6% and
observed in 2 patients (562 and 517 ms, respectively); 15.7 7.9%, respectively, in the CTRCD group
Grade 2 (QTc between 481 and 500 ms), and Grade 1 [n ¼ 3]).
T A B L E 4 Comparison of Echocardiographic Parameters Before and After Osimertinib Administration
NSCLC (N ¼ 36) Non-CTRCD (n ¼ 32) CTRCD (n ¼ 4)
Before After p Value* Before After p Value† Before After p Value‡
LVEF, % 69.4 4.2 63.4 10.5 <0.001 69.8 4.0 67.9 5.5 0.004 66.3 5.1 40.3 9.1 <0.001
LVIDd, mm 42.6 4.5 44.5 5.2 0.005 42.4 4.4 43.9 4.8 0.003 44.0 4.9 48.8 7.2 0.344
LVIDs, mm 26.3 3.3 29.1 5.6 <0.001 26.0 3.1 27.8 3.9 <0.001 28.2 4.3 39.3 7.5 0.001
E peak, cm/s 66.7 18.0 73.4 19.2 0.017 67.3 18.2 73.1 17.9 0.035 62.4 18.1 75.7 31.4 0.858
Dct, ms 222.6 54.0 217.9 39.2 0.625 225.2 55.9 216.6 31.3 0.333 201.5 32.4 228.5 87.3 0.817
E/A ratio 0.90 0.26 1.04 0.36 0.013 0.90 0.28 1.04 0.38 0.047 0.70 0.14 0.98 0.26 0.870
HR, beats/min 74.2 10.8 75.1 14.6 0.600 74.3 11.3 74.3 14.5 0.875 73.5 7.0 81.6 16.5 0.442
Values are mean SD. *p value comparing echocardiographic parameters before and after osimertinib in 36 NSCLC patients with serial monitoring. †p value comparing parameters before and after in 32
patients without CTRCD. ‡p value comparing parameters between after in 32 non-CTRCD and after in 4 with CTRCD.
CTRCD ¼ cancer therapeutics-related cardiac dysfunction; Dct ¼ deceleration time; E peak ¼ peak velocity of early transmitral flow; E/A ratio ¼ peak velocities of early-to-late ratio of transmitral flow;
HR ¼ heart rate; LVEF ¼ left ventricular ejection fraction; LVIDd ¼ left ventricular end-diastolic dimension; LVIDs ¼ left ventricular end-systolic dimension.
C E NT R AL IL L U STR AT IO N Osimertinib and Adverse Cardiac Events
Kunimasa, K. et al. J Am Coll Cardiol CardioOnc. 2020;2(1):1–10.
In this retrospective analysis, the incidence of cardiac adverse events was 4.9% including cardiac systolic dysfunction, progression of valvular disease, myocardial
infarction, and QTc prolongation. Further investigations are necessary to clarify the associations between osimertinib and cardiac AEs in NSCLC patients harboring
EGFR mutations. AE ¼ adverse events; other abbreviations as in Figures 1 and 2.
DISCUSSION higher CTCAE version 5.0 cardiac AEs. Five of the 6

patients were female, and all 5 had a history of car-
A retrospective analysis of cardiac AEs was performed diovascular disease and/or cardiovascular risk factors.
in NSCLC patients harboring EGFR mutations treated Our observations suggest that cardiac AEs may occur
with osimertinib. This is the first study to describe more commonly in patients with pre-existing cardio-
osimertinib-associated cardiac AEs in a real-world vascular disease. However, the present cohort was
Japanese NSCLC population (Central Illustration). small, and these were single-center retrospective an-
Overall, we observed a 4.9% incidence of grade 3 or alyses where data were not collected in a
standardized, prospective manner. Further in- of case 5, typical myocyte degeneration/necrosis with
vestigations are necessary to further understand the adjacent mononuclear cell infiltration was not
associations between osimertinib and cardiac AEs in observed in either case 1 or 5. Nevertheless, the his-
NSCLC patients harboring EGFR mutations, including topathologic findings from these 2 cases suggest that
defining the predictors, natural history, and potential osimertinib does not result in a pattern of myocyte
mechanisms of toxicity. death seen with anthracyclines nor myocarditis seen
Commonly reported osimertinib-associated AEs in with immune checkpoint inhibitors and that cardiac
Phase 1 to 3 clinical trials include diarrhea, rash, dysfunction associated with osimertinib may result
nausea, and decreased appetite. Most concerning AEs from functional inhibition of myocyte contractility
reported to date include interstitial lung disease and without marked cell death or inflammation.
QTc prolongation (2–4). Generally, osimertinib has a Recent guidelines for heart failure suggest that early
more favorable toxicity profile than earlier generation detection of cardiac dysfunction and cardioprotection
EGFR-TKIs because it has a lower affinity for wild- in the setting of cardiac risk factors before symptom-
type EGFR (1,4). A post-marketing surveillance atic heart failure may be beneficial (12). This study
study of osimertinib in 3,578 Japanese NSCLC pa- raises the question of whether there is a need for
tients with EGFR mutations suggested that the inci- routine monitoring of cardiac function with osimerti-
dence of grade 3 or higher QTc prolongation occurred nib to detect asymptomatic declines in LVEF and
in only 0.1% and other grade 3 or higher cardiac AEs whether such declines should be treated with phar-
in only 0.8% of patients (6). However, U.S. Food and macologic intervention.
Drug Administration (FDA) reports suggest that car- Our findings also suggest statistically significant
diomyopathy associated with osimertinib occurred in increases in the QTc interval. Although a strong
1.9% of patients (16 of 833), and LVEF declined $ 10% emphasis has been placed on osimertinib-associated
to <50% in 4.0% of patients (26 of 655) (FDA Tagrisso QTc prolongation because of its reported frequency,
prescribing information [reference ID: 4077625; no fatal episodes of arrhythmia associated with
revised: March 2017]). osimertinib in this cohort were detected. Other
Our study indicates that osimertinib-associated recent reports have suggested an increased odds
cardiac AEs occur at a rate (4.9%) similar to that re- ratio for heart failure, atrial fibrillation, and QTc
ported by the FDA, but not by other large retrospec- prolongation in osimertinib compared with other
tive analyses (6). Because 5 of 6 cases had TKIs based on the FAERS (FDA Events Reporting
pre-existing cardiovascular risk factors or disease, System) (13). No new onset of atrial fibrillation
osimertinib may further worsen underlying cardiac associated with osimertinib was detected in the
pathology. Unfortunately, only a small subset of present cohort; however, as suggested by the FAERS
patients underwent serial echocardiography before study, we did observe an increased risk of heart
and after osimertinib administration, limiting a more failure.
detailed understanding of the changes in cardiac
function that occur with osimertinib. Functional STUDY LIMITATIONS. First, this is a retrospective
assessment before and at standardized intervals single-center study. As such, the sample size was
during treatment with osimertinib can help elucidate small. However, a detailed, careful review of each
the true rate of cardiac dysfunction. individual patient’s medical record was performed.
The common histopathological findings of cases 1 Second, there was no established, standardized pro-
and 5 were myocyte hypertrophy and lipofuscin tocol for cardiovascular monitoring and follow-up in
deposition, which is seen in heart failure (8). Both of this population. There may be selection bias in those
the cases had pre-existing cardiovascular risk factors, patients who underwent serial echocardiography or
including either hypertension or obesity, which might electrocardiogram evaluation, and as such, the true
have also worsened or resulted in cardiomyocyte incidence of cardiac dysfunction may also be under-
hypertrophy. Focal vacuolization and degeneration of estimated. The timing of follow-up echocardiography
myocytes were also observed in case 5, which is was also highly variable, from 5 to 327 days, poten-
suggestive of myocyte damage (9), potentially from tially resulting in additional detection bias. Third,
osimertinib. There are 2 published case reports of LVEF was quantified using a 2D Teichholz method
osimertinib-associated cardiac dysfunction (10,11), and not by biplane Simpson’s method. Fourth, the
and myocarditis with lymphocyte infiltration without case series of Grade 3 cardiac AEs was small, and bi-
myocyte hypertrophy has been described (10). opsy data were not available in all patients, limiting
Although a few lymphocytes infiltrated the specimen the interpretation of our findings.
CONCLUSIONS PERSPECTIVES
In this retrospective cohort study, the incidence of

COMPETENCY IN MEDICAL KNOWLEDGE: QTc
cardiac AEs in patients treated with osimertinib
prolongation is well described with osimertinib.
was similar to that observed in the clinical trials
However, the incidence of other cardiac AEs is less
but unexpectedly higher than that observed in
well described. Among 123 NSCLC patients treated
the post-marketing surveillance in Japan. Additional
with osimertinib, cardiac AEs occurred in 4.9%. These
prospective data collection of advanced NSCLC
included acute myocardial infarction, heart failure,
patients treated with osimertinib will be important
and valvular diseases. These AEs may occur more
in understanding the incidence, pathophysiology,
commonly in patients with pre-existing cardiovascular
and management of cardiac AEs with osimertinib.
disease or risk factors.
ACKNOWLEDGMENT The authors thank Dr. Hatsue
Ueda, Department of Pathology, National Cerebral TRANSLATIONAL OUTLOOK: Further studies
and Cardiovascular Center, Suita, Japan, for valuable should focus on elucidating the mechanisms under-
comments on myocardial biopsies. lying cardiac AEs associated with osimertinib. More-
over, prospective studies with standardized data
collection are needed to define the epidemiology of
ADDRESS FOR CORRESPONDENCE: Dr. Toru Oka,
osimertinib associated cardiac disease, and clinical
Department of Onco-Cardiology, Osaka International
studies are needed to understand strategies to pre-
Cancer Institute, 3-1-69 Otemae, Chuo-ku, Osaka
vent and treat cardiac AEs.
541-8567, Japan. E-mail: toruoka@oici.jp. Twitter:
@ToruOka1.
REFERENCES
1. Cross DA, Ashton SE, Ghiorghiu S, et al. results surveys final report. 2019 [in Japanese]. 11. Watanabe H, Ichihara E, Kano H, Ninomiya K,
AZD9291, an irreversible EGFR TKI, overcomes Available at: https://www.mhlw.go.jp/ Tanimoto M, Kiura K. Congestive heart failure
T790M-mediated resistance to EGFR inhibitors in content/11120000/000484135.pdf. Accessed during osimertinib treatment for epidermal
lung cancer. Cancer Discov 2014;4:1046–61. February 5, 2020. growth factor receptor (EGFR)-mutant non–small
cell lung cancer (NSCLC). Intern Med 2017;56:
2. Mok TS, Wu YL, Ahn MJ, et al. Osimertinib or 7. Plana JC, Galderisi M, Barac A, et al. Expert
2195–7.
platinum-pemetrexed in EGFR T790M-positive consensus for multimodality imaging evaluation of
lung cancer. N Engl J Med 2017;376:629–40. adult patients during and after cancer therapy: a 12. Goldberg LR, Jessup M. Stage B heart failure:
report from the American Society of Echocardi- management of asymptomatic left ventricular
3. Soria JC, Ohe Y, Vansteenkiste J, et al. Osi-
ography and the European Association of Cardio- systolic dysfunction. Circulation 2006;113:
mertinib in untreated EGFR-mutated advanced
vascular Imaging. J Am Soc Echocardiogr 2014;27: 2851–60.
non-small-cell lung cancer. N Engl J Med 2018;
378:113–25. 911–39. 13. Anand K, Ensor J, Trachtenberg B,
8. Kakimoto Y, Okada C, Kawabe N, et al. Bernicker EH. Osimertinib-induced cardiotoxicity:
4. Mezquita L, Varga A, Planchard D. Safety of
Myocardial lipofuscin accumulation in ageing and a retrospective review of the FDA adverse events
osimertinib in EGFR-mutated non–small cell lung
sudden cardiac death. Sci Rep 2019;9:3304. reporting system (FAERS). J Am Coll Cardiol Car-
cancer. Expert Opin Drug Saf 2018;17:1239–48.
dioOnc 2019;1:172–8.
5. Ramalingam SS, Vansteenkiste J, Planchard D, 9. Zuppinger C, Timolati F, Suter TM. Pathophys-
et al. Overall survival with osimertinib in un- iology and diagnosis of cancer drug induced car-
treated, EGFR-mutated advanced NSCLC. N Engl J diomyopathy. Cardiovasc Toxicol 2007;7:61–6.
KEY WORDS cardiac adverse events,
Med 2020;382:41–50.
10. Oyakawa T, Nakashima K, Naito T. Cardiac cardiac dysfunction, EGFR mutations,
6. AstraZeneca KK. Tagrisso tablets 40 mg, Dysfunction caused by osimertinib. J Thorac Oncol myocardial biopsy, non–small cell lung
Tagrisso tablets 80 mg, the result of the use- 2017;12:e159–60. cancer, osimertinib
EDITORIAL COMMENT
Targeting the Cardiotoxicity of

Epidermal Growth Factor
Receptor Inhibitors*
Dipesh Uprety, MBBS, Aaron S. Mansfield, MD
T argeted therapies have revolutionized the

treatment of non–small-cell lung cancer
(NSCLC). Despite a common histology, ade-
nocarcinomas of the lung result from a broad spec-
dacomitinib (4). For those reasons osimertinib was
developed to overcome resistance to EGFR T790M
mutations with minimal activity against the wild-
type EGFR (4).
trum of oncogenic drives including: KRAS, EGFR, Osimertinib was originally approved by the
ALK, ROS1, BRAF, and NTRK1 (1). Small molecule in- Food and Drug Administration for patients who
hibitors have been developed to target many of these developed EGFR T790M mutations while on treat-
drivers, and many have received Food and Drug ment with earlier generations of EGFR inhibitors (5).
Administration approval. Epidermal growth factor Subsequently, osimertinib was compared head-to-
receptor (EGFR) inhibitors have been available the head with oral gefitinib or erlotinib in patients
longest and at the vanguard of the targeted therapy with treatment naive EGFR-mutated stage IIIB or IV
approach to lung cancer. The first generation of NSCLC in the phase 3 clinical trial FLAURA (6).
EGFR inhibitors, erlotinib and gefitinib, were charac- Osimertinib resulted in significant improvement in
terized by significant dermatologic and gastrointes- progression-free survival and overall survival over
tinal toxicities (2). One common mechanism of gefitinib or erlotinib (6,7). Also, because osimertinib
resistance to these first-generation EGFR inhibitors can penetrate the central nervous system (CNS), the
is the development of T790M mutations that change rates of progression of disease in the CNS were
adenosine triphosphate affinity and result in steric significantly lower in the osimertinib arm. Although
hindrance of the inhibitors (3). Lung cancers with there is some debate about the optimal sequence of
EGFR T790M mutations are not sensitive to the targeted EGFR therapies, many oncologists use osi-
second-generation EGFR inhibitors afatinib and mertinib for the frontline treatment of EGFR mutant
NSCLC as per National Comprehensive Cancer
Network guidelines (8).
*Editorials published in JACC: CardioOncology reflect the views of EGFR is a receptor tyrosine kinase in the erythro-
the authors and do not necessarily represent the views of JACC: blastic leukemia viral oncogene homolog (ErbB)/hu-
CardioOncology or the American College of Cardiology.
man epidermal growth factor receptor (HER) family
From the Division of Medical Oncology, Mayo Clinic, Rochester, Minne- that includes (human epidermal growth factor re-
sota. Dr. Mansfield has received research support from Bristol-Myers
ceptor 2) HER2. Antibodies that target HER2 like
Squibb, National Institutes of Health, Novartis, and Verily; remunera-
tion to his institution for participation in advisory boards or other ac- trastuzumab and small molecule inhibitors of HER2
tivities from AstraZeneca, Bristol-Myers Squibb, F. Hoffmann-La Roche, like lapatinib have been approved for the treatment
Abbvie, and Genentech; travel support from F. Hoffmann-La Roche; and of HER2-amplified breast cancers. These agents have
is a non-remunerated director of the Mesothelioma Applied Research
been associated with cardiotoxicity, especially in pa-
Foundation. Dr. Uprety has reported that he has no relationships rele-
vant to the contents of this paper to disclose. tients also receiving anthracyclines (9). The reported
The authors attest they are in compliance with human studies commit- cardiotoxicity with these agents is often but not al-
tees and animal welfare regulations of the authors’ institutions and Food
ways reversible. HER2 is important in cardiac devel-
and Drug Administration guidelines, including patient consent where
appropriate. For more information, visit the JACC: CardioOncology opment and the maintenance of normal cardiac
author instructions page. structure and function under stress conditions (9).

12 Uprety and Mansfield JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020
Targeting the Cardiotoxicity of EGFR Inhibitors MARCH 2020:11–2
Osimertinib also has modest activity against HER2, so had a history of a cardiovascular disease or a risk factor
it is not unanticipated that cardiotoxicity has been for cardiovascular disease. In future prospective
observed with this agent (4). studies we may be able to validate this finding and
SEE PAGE 1 ultimately predict which individuals are at high risk for
untoward outcomes. Despite the limitations of this
In this issue of JACC: CardioOncology, Kunimasa
study, including the relatively small sample size and
et al. (10) report osimertinib-associated cardiac
the lack of baseline and sequential electrocardiograms
toxicity observed in a cohort of patients at the Osaka
and echocardiograms at standardized time intervals,
International Cancer Institute in Japan. Osimertinib-
these findings may help guide future studies for
associated cardiotoxicity is relatively uncommon, yet
identifying predictors of cardiotoxicity in patients
an important issue in clinical practice. In this retro-
with EGFR-mutant lung cancer treated with
spective study, the electronic medical records of 123
osimertinib.
patients with NSCLC and sensitizing EGFR mutations
Despite the concerns these results raise with osi-
treated with osimertinib monotherapy between 2014
mertinib, they should be viewed in the context of the
and 2019 were reviewed. Only 72 and 36 patients had
significant improvements in survival observed with
baseline electrocardiograms or echocardiograms
osimertinib over prior generations of EGFR-targeting
before osimertinib initiation, respectively. Six (4.9%)
therapies and the CNS penetration of osimertinib,
patients experienced grade $3 cardiac toxicity, which
which improves the treatment of brain metastases,
included acute myocardial infarction, heart failure
and the prevention of their development. For now,
with reduced left ventricular ejection fraction, and
oncologists should counsel patients with cardiac dis-
exacerbation of valvular heart disease. This study has
ease or with multiple risk factors for such that there
many important findings that are important to clinical
may be heightened risk for cardiotoxicity with this
care. First, the frequency of cardiotoxicity reported in
agent. Hopefully more sensitive predictors of car-
this study was higher than that reported in the original
diotoxicity will be discovered and incorporated into
clinical trials with osimertinib. Similar findings have
our clinical decision making.
been reported with trastuzumab. It is not entirely
surprising that adverse events are more common in the
real-world setting than in restricted clinical trial pop-
ulations (11,12). Second, patients can develop cardiac ADDRESS FOR CORRESPONDENCE: Dr. Aaron S.
toxicity as early as 2 weeks on treatment, suggesting Mansfield, Division of Medical Oncology, Mayo Clinic,
that cardiotoxicity may not represent a cumulative 200 First Street Southwest, Rochester, Minne-
dose-dependent phenomenon. Third, this study sug- sota 55905. E-mail: mansfield.aaron@mayo.edu.
gested that patients who developed cardiac toxicity Twitter: @AMansfieldMD.
REFERENCES
1. Hirsch FR, Scagliotti GV, Mulshine JL, et al. Lung 6. Soria JC, Ohe Y, Vansteenkiste J, et al. Osi- 10. Kunimasa K, Kamada R, Oka T, et al. Cardiac
cancer: current therapies and new targeted treat- mertinib in untreated EGFR-mutated advanced adverse events in EGFR-mutated non-small cell
ments. Lancet 2017;389:299–311. non-small-cell lung cancer. N Engl J Med 2018; lung cancer treated with osimertinib. J Am
378:113–25. Coll Cardiol CardioOnc 2020;2:1–10.
2. Burtness B, Anadkat M, Basti S, et al. NCCN Task
Force Report: management of dermatologic and 7. Ramalingam SS, Vansteenkiste J, Planchard D, 11. Bowles EJ, Wellman R, Feigelson HS, et al. Risk
other toxicities associated with EGFR inhibition in et al. Overall survival with Osimertinib in un- of heart failure in breast cancer patients after
patients with cancer. J Natl Compr Canc Netw treated, EGFR-mutated advanced NSCLC. N Engl J anthracycline and trastuzumab treatment: a
2009;7 Suppl 1:S5–21. quiz S2–4. Med 2020;382:41–50. retrospective cohort study. J Natl Cancer Inst
2012;104:1293–305.
3. Kobayashi S, Boggon TJ, Dayaram T, et al. EGFR 8. Ettinger DS, Wood DE, Aisner EL, et al. NCCN
mutation and resistance of non-small-cell lung clinical practice guidelines in oncology: non–small 12. Chen J, Long JB, Hurria A, Owusu C,
cancer to gefitinib. N Engl J Med 2005;352:786–92. cell lung cancer. Version 2.2020. Available at: Steingart RM, Gross CP. Incidence of heart failure
https://www.nccn.org/professionals/physician_ or cardiomyopathy after adjuvant trastuzumab
4. Cross DA, Ashton SE, Ghiorghiu S, et al.
gls/pdf/nscl.pdf. Accessed January 29, 2020. therapy for breast cancer. J Am Coll Cardiol 2012;
AZD9291, an irreversible EGFR TKI, overcomes
60:2504–12.
T790M-mediated resistance to EGFR inhibitors in
9. Kondapalli L. Cardiotoxicity: An unexpected
lung cancer. Cancer Discov 2014;4:1046–61.
consequence of HER2-targeted therapies. (2016).
5. Mok TS, Wu YL, Ahn MJ, et al. Osimertinib or Available at: https://www.acc.org/latest-in-
platinum-pemetrexed in EGFR T790M-positive cardiology/articles/2016/06/06/09/32/cardiotoxicity. KEY WORDS cardiotoxicity, EGFR, non–small-
lung cancer. N Engl J Med 2017;376:629–40. Accessed January 29, 2020. cell lung cancer, osimertinib
ORIGINAL RESEARCH
Early Detection and Prediction

of Anthracycline-Induced
Right Ventricular Cardiotoxicity by
3-Dimensional Echocardiography
Rui Zhao, MBBS,a,* Fang Shu, MD,b,* Chujie Zhang, MBBS,a Feiyan Song, MBBS,a Yuchen Xu, MBBS,a Ye Guo, MD,c
Kai Xue, MD,c Jinyi Lin, MD,d Xianhong Shu, MD, PHD,a David H. Hsi, MD,e Leilei Cheng, MD, PHDa
ABSTRACT
OBJECTIVES The purpose of this study was to assess the associations between 3-dimensional echocardiography (3DE)-
derived changes in right ventricular (RV) volumes and strains with subsequent RV cardiotoxicity in patients treated with
anthracyclines.
BACKGROUND Although early detection and prediction of left ventricular (LV) dysfunction has been widely studied in
patients receiving anthracyclines, little is known about the early changes in RV size and function in this population.
METHODS A total of 74 patients with diffuse large B-cell lymphoma who received 6 cycles of anthracycline-based
treatment were enrolled. Echocardiography was performed at baseline or before chemotherapy (pre-chemotherapy) (T0);
after 2 cycles (T1); after 4 cycles (T2); and at the end of 6 cycles of chemotherapy (T3). Right ventricular end-diastolic volume
(RVEDV), end-systolic volume (RVESV), ejection fraction (RVEF), longitudinal free wall strain (RVLFS), and longitudinal
septal strain (RVLSS) were quantified by 3DE. RV cardiotoxicity was defined as a relative reduction of >10% in 3D RVEF or a
relative reduction of >5% to a value of <45%. Volume status was assessed by inferior vena cava diameter (IVCD) and the
estimated right atrial pressure (RAP).
RESULTS Twenty-seven patients developed cardiotoxicity after 6 cycles of chemotherapy (T3). Compared to baseline,
increases in 3D RVEDV (58.5 7.7 ml vs. 64.2 7.0 ml; p < 0.001) and RVESV (27.8 4.2 ml vs. 31.3 4.2 ml; p < 0.001)
were observed by the end of the fourth cycle of chemotherapy (T2). 3D RVLFS (27.3 3.1% vs. 24.2 2.6%;
p < 0.001) was also decreased at T2 compared to baseline. Statistically significant declines in 3D RVLSS (26.1 2.5%
vs. 22.9 2.7%; p < 0.001) and RVEF (54.0 2.8% vs. 49.8 2.4%; p < 0.001) were only observed at T3. A relative
decrease in RVLFS of >12.4% (sensitivity, 78.6%; specificity, 82.6%; area under the curve (AUC), 0.80; p ＜ 0.001); and a
relative increase in RVESV of >13.2% (sensitivity, 71.4%; specificity, 71.7%; AUC, 0.76; p ＜0.001) from baseline to T2 -
predicted subsequent RV cardiotoxicity at T3. IVCD and RAP did not change significantly over time.
CONCLUSIONS 3DE-derived measurements of RV strain and volume were associated with subsequent changes in
RVEF. With further study, RVLFS and RVESV could potentially be used to predict subsequent declines in RVEF with
anthracyclines. (J Am Coll Cardiol CardioOnc 2020;2:13–22) © 2020 The Authors. Published by Elsevier on behalf of
the American College of Cardiology Foundation. This is an open access article under the CC BY-NC-ND license
From the aDepartment of Echocardiography, Zhongshan Hospital, Fudan University, Shanghai Institute of Cardiovascular Dis-
eases, Shanghai Institute of Medical Imaging, Shanghai, China; bDepartment of Ultrasound, Danyang Hospital of Traditional
Chinese Medicine, Jiangsu, China; cDepartment of Medical Oncology, Fudan University Shanghai Cancer Center, Shanghai, China;
d
Department of Cardiology, Zhongshan Hospital, Fudan University, Shanghai Institute of Cardiovascular Diseases, Shanghai,

14 Zhao et al. JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020
RV Cardiotoxicity After Anthracycline Exposure MARCH 2020:13–22
A nthracyclines are widely used in reproducible (11,12). Therefore, a prospective cohort

ABBREVIATIONS
AND ACRONYMS children and adults to treat hemato- study was designed to understand the potential
logic malignancies, breast cancer, role of 3DE in the early identification and prediction
2DE = 2-dimensional
echocardiography
and sarcoma (1). However, there is a poten- of RV cardiotoxicity induced by anthracycline
tial for cardiotoxicity, including dose- chemotherapy.
3DE = 3-dimensional
echocardiography dependent cardiomyopathy and heart
failure. Anthracycline-induced cardiotoxic- SUBJECTS AND METHODS
LV = left ventricle

ity has been observed to potentially be irre-
fraction versible in patients treated for lymphoma, STUDY POPULATION. A total of 74 consecutive
resulting in overall worse outcomes (2). Prior patients with pathologically confirmed diffuse large
RV = right ventricle
studies suggest that more than one-half of B-cell lymphoma at Fudan University Shanghai Can-
RVEDV = right ventricular end-
diastolic volume patients treated with anthracyclines exhibit cer Center were enrolled in the study between May
RVEF = right ventricular cardiac dysfunction to some extent, and an 2014 and May 2015. All patients received 6 cycles of
ejection fraction
estimated 5% to 65% develop over heart R-CHOP therapy (cyclophosphamide 750 mg/m 2;
RVESV = right ventricular end- failure, depending upon dose (3,4). vincristine 1.4 mg/m 2 up to a maximum dose of
systolic volume
However, more current studies have 2 mg/m 2; doxorubicin 50 to 70 mg/m 2 on day 1;
RVLFS = right ventricular
indicated that left ventricular ejection prednisone 100 mg on days 1 to 5; and rituximab
longitudinal free-wall strain
fraction (LVEF) recovery and a decrease in 375 mg/m 2 every 21 days). All therapy was given
RVLSS = right ventricular intravenously, except for prednisone, which was
longitudinal septal strain
adverse cardiac events may be achieved with
early institution of angiotensin-converting administered orally. Each patient underwent echo-
enzyme inhibitors and beta-blockers (5). These data cardiography examination before the commencement
would imply that it may be important to detect of chemotherapy (T0) and after the completion of
anthracycline-induced cardiotoxicity early. Echocar- every 2 cycles of chemotherapy (T1), for a total
diography has been recommended by European Soci- of 6 cycles (T3). Age >18 years, viral myocarditis,
ety of Cardiology as part of routine surveillance of hypertension with systolic blood pressure of
cardiotoxicity (6). Previous studies have provided a >140 mm Hg or diastolic pressure of >90 mm Hg,
variety of echocardiography parameters to detect and serious arrhythmia (atrial fibrillation, atrial flutter,
predict anthracycline-induced LV cardiotoxicity (7). ventricular tachycardia >3 beats and greater than the
However, limited data are available for parameters second-degree cardiac conduction block), a history of
which could reflect subclinical right ventricular (RV) heart failure, and/or coronary artery disease were
dysfunction and potential predictors of RV cardiotox- exclusion criteria for enrollment. All patients were
icity. Several animal experiments and clinical studies provided informed consent for participation in the
indicate that the RV may also be susceptible to study and anthracycline treatment. Fudan University
anthracycline-induced cardiotoxicity (8,9). Shanghai Cancer Center (1212117-6) and Zhongshan
Hospital Research Ethics Committee approved the
SEE PAGE 23
protocol (2011-117).
Due to the special geometric structure and unique ECHOCARDIOGRAPHY DATA ACQUISITION. All sub-
wall motion of the RV, routine techniques for echo- jects underwent standard 2DE and 3DE at the baseline
cardiographic assessment of the RV has challenges (T0), 8 to 15 days after the completion of 2 cycles (T1),
(10). However, 3-dimensional echocardiography (3DE) 8 to 15 days after the completion of 4 cycles (T2), and
tracks changes in RV volumes in real time, including at 50 to 60 days after 6 cycles (T3) of chemotherapy.
the inflow and outflow tracts and the apex Images were obtained with a commercially available
throughout the cardiac cycle. 3DE evaluation of RV ultrasonography system (iE33, Philips Medical Sys-
function and morphology has been compared to car- tems, Andover Massachusetts) equipped with S5-1
diac magnetic resonance and is accurate, feasible, and (1 to 5 MHz) and X3-1 (1 to 3 MHz), a fully sampled
China; and the eDepartment of Cardiology, Stamford Hospital and Vagelos Columbia University College of Physicians and Sur-
geons, Stamford, Connecticut. *Drs. Zhao and F. Shu are joint first authors. Supported by National Nature Science Foundation of
China grant N0.81771840 and Excellent Talent Training Program of Shanghai Health System grant NO.2017BR027. The authors
have reported that they have no relationships relevant to the contents of this paper to disclose.
The authors attest they are in compliance with human studies committees and animal welfare regulations of the authors’ in-
stitutions and Food and Drug Administration guidelines, including patient consent where appropriate. For more information, visit
the JACC: CardioOncology author instructions page.
Manuscript received August 28, 2019; accepted January 28, 2020.

JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020 Zhao et al. 15
MARCH 2020:13–22 RV Cardiotoxicity After Anthracycline Exposure
matrix transducer. Standard 2DE and 3DE were per- measurements showed ICCs of 0.961 (95% CI: 0.948
formed according to the recommendations of the to 0.975) for RVEDV; 0.960 (95% CI: 0.949 to 0.975)
American Society of Echocardiography (13). Five for RVESV; 0.933 (95% CI: 0.925 to 0.943) for RVLSS;
consecutive cycles for 2DE images and 6 consecutive and 0.937 (95% CI: 0.922 to 0.948) for RVLFS.
cardiac cycles for 3DE images were acquired for offline UNIVARIATE LOGISTIC REGRESSION ANALYSIS OF
analysis. Image parameters such as depth, sector size, PREDICTORS OF RV CARDIOTOXICITY. RV car-
angle, and focus were optimized to achieve a frame diotoxicity was defined in this study as a >10% rela-
rate range of 60 to 80 fps for 2DE and 30 to 45 fps for tive reduction in RVEF or a relative reduction
3DE analysis. Sonographers and off-line echocardiog- of >5% to an absolute value of <45% (7,13).
raphy readers were blinded to all clinical data. Changes in echocardiographic parameters over time
2D, M-MODE, AND DOPPLER ECHOCARDIOGRAPHY were included in the univariate logistic regression
ANALYSIS. Tricuspid annular plane systolic excur- analysis to evaluate possible predictors of subsequent
sion was acquired through the M-mode approach in RV cardiotoxicity. Baseline echocardiographic pa-
the apical 4-chamber view. Pulse wave Doppler vari- rameters and the relative percentage of changes in
ables (tricuspid/mitral inflow velocities: peak E ve- echocardiographic parameters between T0 and T1 and
locity of early diastolic filling [E]; peak A velocity of between T0 and T2 were included as variables in the
late diastolic filling [A]; tissue Doppler parameters univariate logistic regression analysis.
0
peak tricuspid/mitral annulus systolic velocity [S ]; STATISTICAL ANALYSIS. Continuous variables were
RV/LV peak early diastolic myocardial velocity [E 0 ]; expressed as mean SD. Categorical variables were
and RV/LV peak atrial contraction velocity [A 0 ]) were expressed as counts (percentages). Kolmogorov-
acquired according to a standard protocol (13). Pul- Smirnov Z tests were used to evaluate the normality
monary artery systolic pressure was estimated using of the data. Parametric tests were applied when
the continuous wave Doppler of the tricuspid regur- normality was satisfied, otherwise the Wilcoxon
gitation jet. Inferior vena cava diameter (IVCD) was signed rank and chi-square tests were used for
measured in the subcostal view with the patient in nonparametric data. Differences in echocardiographic
the supine position at 1.0 to 2.0 cm from the junction parameters at baseline and at the end of every 2 cy-
of the right atrium. The right atrial pressure was cles of chemotherapy were evaluated using linear
estimated based on IVCD and its collapse after sniff mixed-effects models with a random intercept to ac-
according to the ASE guidelines (13). count for intrapatient correlation of repeated mea-
3 DIMENSIONAL ECHO DATA ANALYSIS. Offline surements. An independent covariance structure was
analysis of the right ventricular 3-dimensional echo- assumed. The Bonferroni method was used to adjust
cardiography data was performed using the TomTec for multiple pairwise comparisons. Possible pre-
4D RV analysis workstation (version 4.6.0.411, TomTec dictors (changes over time in echocardiographic pa-
Imaging Systems, Unterschleißheim, Germany) rameters) of RV cardiotoxicity were analyzed by
(Figures 1A and 1B). Care was taken to include trabec- univariate logistic regression; odds ratio and 95% CIs
ulae during measurements. RV end-diastolic volume were calculated. Receiver operating characteristic
(RVEDV); RV end-systolic volume (RVESV); RV ejection (ROC) curves were obtained, and the associated cutoff
fraction (RVEF); LV end-diastolic volume (LVEDV); points with the greatest sensitivity and specificity
end-systolic volume (LVESV); and LVEF were were selected according to Youden’s index. Area un-
measured. RV longitudinal free wall strain (RVLFS); der the ROC curve (AUC) was calculated to determine
RV longitudinal septal strain (RVLSS); LV global lon- the discriminative ability of various echocardio-
gitudinal strain (LVGLS); and LV global circumferential graphic parameters between patients with and
strain (LVGCS) were acquired simultaneously. without RV cardiotoxicity. Intra- and interobserver
variability of RVLSS, RVLFS, RVEDV, and RVESV were
INTRAOBSERVER AND INTEROBSERVER VARIABILITY
assessed using ICCs. SPSS version 22 software (SPSS
ANALYSIS. Intraobserver and interobserver variability
Inc., Chicago, Illinois) and MedCalc software version
were assessed by randomly selecting 15 patients to be
15.6.1.0 (MedCalc, New York City, New York) were
measured by 1 observer twice and by another inde-
used for data processing. A p value of <0.05 was
pendent observer. Interobserver measurement
considered statistically significant.
showed intraclass correlation coefficients (ICCs) of
0.953 (95% confidence interval [CI]: 0.941 to 0.968) RESULTS
for RVEDV; 0.942 (95% CI: 0.932 to 0.956) for RVESV;
0.926 (95% CI: 0.912 to 0.939) for RVLSS; and 0.945 A total of 80 patients were screened. Two patients
(95% CI: 0.933 to 0.958) for RVLFS. Intraobserver were excluded due to uncontrolled hypertension
F I G U R E 1 Examples of Echocardiography Studies of 3DE Right Ventricular Parameters Using TomTec Offline Analysis Software
Four planes were displayed for tracing of the endocardium: the apical 4-chamber view, 3 short-axis planes near the apex, the mid-level, and the base of the right
ventricle. The right ventricular endocardial contour was traced semiautomatically frame by frame and with refinement by manual adjustment when necessary at end-
diastolic and end-systolic phases. 3DE ¼ 3-dimensional echocardiography.
and coronary artery disease. Four patients were

T A B L E 1 Baseline Clinical Characteristics and Cardiac Risk Factors in Patients With or
excluded because of poor image quality. The final Without Cardiotoxicity
analytic cohort included 74 patients, of whom 41 (55%)
All Patients Cardiotoxicity Noncardiotoxicity
were men, 33 were women, and ages ranged from 20 to (N ¼ 74) (n ¼ 27, 36%) (n ¼ 47, 64%) p Value
78 years old (mean age: 48.9 11.8 years). Patient Age, yrs 48.9 11.8 49.6 10.7 48.5 12.1 0.112
demographics and clinical characteristics are shown in Height, cm 166.9 6.1 165.7 6.9 167.6 5.7 0.201
Table 1. The mean cumulative anthracycline dose was Weight, kg 65.6 11.6 64.4 10.7 66.3 11.9 0.197
358.20 69.04 mg/m 2. No patients required acute HR, beats/min 78.6 11.9 79.3 11.1 78.2 12.4 0.226
treatment for dehydration or additional intravenous SBP, mm Hg 117.4 9.6 118.0 9.1 117.1 10.3 0.109
DBP, mm Hg 75.2 7.4 74.8 6.3 75.4 8.1 0.318
fluids beyond the standard chemotherapy. None of the
Women 33 11 22 0.092
patients received other cardiotoxic therapy, radiation
Men 41 16 25 0.114
therapy, or cardioprotective medications during the Diabetes mellitus 5 (7) 2 (7) 3 (6) 0.252
study period, and none developed signs or symptoms Smoker 20 (27) 8 (29) 12 (25) 0.102
of heart failure. Cumulative anthracycline 358.20 69.04 367.30 63.10 352.90 71.14 0.097
dose, mg/m2
EFFECTS OF ANTHRACYCLINE CHEMOTHERAPY ON
2DE AND DOPPLER PARAMETERS. 2DE and Doppler Values are mean SD, n, or n (%). The p values compare cardiotoxicity with noncardiotoxicity.
DBP ¼ diastolic blood pressure; HR ¼ heart rate; SBP ¼ systolic blood pressure.
parameters are shown in Table 2. RV fractional area did
not change significantly during follow-up. Pulmonary
artery systolic pressure and IVCD remained stable and nor were changes in these measurements between T0
within normal limits at baseline and at each follow-up and T1. Relative or percent changes (D ) between T0
point of the chemotherapy. There were no significant and T2 of measurements of RV size and function were
changes in Doppler parameters of RV or LV diastolic also evaluated as potential predictors of RV car-
function during anthracycline chemotherapy. diotoxicity (Table 4).
CHANGES IN 3DE PARAMETERS WITH ANTHRACYCLINE

CHEMOTHERAPY. 3DE-derived RV and LV volumes, EF, T A B L E 2 Vital Signs and 2DE Parameters Before and After Chemotherapy
and strain parameters are shown in Table 3. Compared T0 T1 T2 T3

to T0 (baseline, pre-chemotherapy), a significant Weight, kg 65.6 11.6 65.2 11.8 65.3 12.4 65.5 12.2
decrease in RVEF was only observed at T3. A total of 27 HR, beats/min 78.6 11.9 79.1 11.3 80.5 10.7 79.3 11.6
patients (36%) met the criteria of RV cardiotoxicity at SBP, mm Hg 117.4 9.6 116.8 9.8 116.1 11.0 115.9 9.5
T3 (RVEF decreased from 54.8 3.5% at baseline to DBP, mm Hg 75.2 7.4 75.4 7.3 74.8 6.8 74.0 7.0
48.3 2.6%). Of the 27 patients, 4 patients experi- RVFAC, % 43.2 4.6 43.4 4.9 44.6 4.2 44.3 5.5
TAPSE, mm 21.6 2.3 21.4 2.0 21.3 1.7 20.9 1.9
enced a decline in RVEF to <45%. Two of those pa-
RV E, cm/s 55.2 8.8 56.4 9.9 54.9 7.6 56.5 8.3
tients had a decline in RVEF percentage of >10%, and 2
RV A, cm/s 50.7 10.8 52.5 12.0 52.3 12.0 48.8 12.0
patients had a decline in RVEF between 5% and 10%. RV S0 , cm/s 14.5 3.4 14.6 3.0 14.0 2.6 14.6 3.1
RVEDV increased significantly at T2 and T3 compared RV E0 , cm/s 12.3 4.0 11.4 3.1 11.2 2.6 11.2 3.0
to T0 or T1. RV A0 , cm/s 14.6 4.2 15.4 4.8 14.7 3.8 14.7 3.4
A statistically significant change in RVESV was RV E/E0 ratio 5.3 1.5 5.2 1.4 5.1 1.4 4.9 1.5
detected at T2 (p < 0.001) and persisted at T3. Sta- PASP, mm Hg 30.9 3.8 30.7 4.0 30.4 3.6 30.5 3.9
LV E, cm/s 64.9 11.8 64.1 12.6 62.7 10.9 60.6 13.3
tistically significant strain abnormalities in RVLFS
LV A, cm/s 63.0 14.9 65.6 12.1 64.7 16.2 66.3 14.1
and LVGLS were observed at T2 (p < 0.001). However,
LV E/A ratio 1.01 0.21 0.99 0.30 0.97 0.35 0.94 0.25
RVLSS and LVGCS were significantly decreased
LV S0 , cm/s 12.2 2.4 12.0 2.5 11.5 2.9 11.6 1.8
only at T3 (p < 0.001), suggestive of a late change. LV E0 , cm/s 10.0 1.4 9.7 1.3 9.6 1.8 9.4 1.4
3DE-derived LVEDV (77.6 10.0 ml vs. 80.4 12.0 ml; LV A0 , cm/s 8.8 2.4 8.7 1.9 8.4 1.4 8.2 1.6
p ¼ 0.12), LVESV (30.3 7.6 ml vs. 33.2 8.0 ml; LV E/E0 ratio 7.0 1.4 6.6 1.7 6.4 2.2 6.5 1.5
p ¼ 0.078), and LVEF (62.1 5.7% vs. 59.3 5.9%; IVCD, mm 16.7 1.1 16.4 1.3 17.1 1.5 17.7 1.3
p ¼ 0.083) were not significantly different from the RAP, mm Hg 3.9 1.8 4.0 1.7 3.9 1.8 4.1 2.0
values at T0 to T3.
Values are mean SD.
UNIVARIABLE LOGISTIC REGRESSION AND ROC A ¼ peak inflow velocity of late diastolic filling; A’ ¼ peak late diastolic myocardial velocity;
DBP ¼ diastolic blood pressure; E ¼ peak inflow velocity of early diastolic filling; E’ ¼ peak early
CURVE ANALYSIS OF PREDICTORS OF RV CARDIO- diastolic myocardial velocity; HR ¼ heart rate; IVCD ¼ inferior vena cava diameter; LV ¼ left
ventricle; PASP ¼ pulmonary artery systolic pressure; RAP ¼ right atrial pressure; RV ¼ right
TOXICITY. There were no differences in clinical char- ventricle; RVFAC ¼ right ventricular fractional area change; S0 ¼ systolic velocity of tricuspid
acteristics between patients with and without RV annulus; SBP ¼ systolic blood pressure; T0 ¼ baseline before chemotherapy; T1 ¼ after 2 cycles of
chemotherapy; T2 ¼ after 4 cycles of chemotherapy; T3 ¼ after 6 cycles of chemotherapy;
cardiotoxicity (Table 1). Baseline values of RV strain TAPSE ¼ tricuspid annular plane systolic excursion.
or RV volumes were not predictors of cardiotoxicity,
DISCUSSION
T A B L E 3 3DE Parameters Before and at Each Follow-Up Point During Chemotherapy
T0 T1 T2 T3 In this study, changes in RV structure and function

RVEDV, ml 58.5 7.7 60.2 7.7 64.2 7.0*† 66.0 6.6*† in lymphoma patients treated with anthracycline-
RVESV, ml 27.8 4.2 28.9 4.3 31.3 4.2*† 34.1 3.7*†‡
based chemotherapy were carefully detailed using
RVEF, % 54.0 2.8 53.6 2.9 52.8 3.1 49.8 2.4*†‡
3DE. Compared with 3DE-derived LVEF, 3DE-derived
RVLFS, % 27.3 3.1 25.4 3.3 24.2 2.6* 21.9 2.8*†‡
RVLSS, % 26.1 2.5 25.6 2.9 24.7 2.9 22.9 2.7*†
RVEF was impaired at an earlier stage. Of the sub-
LVEDV, ml 77.6 10.0 78.2 11.2 79.4 9.7 80.4 12.0 clinical changes in RV structure and function, the
LVESV, ml 30.3 7.6 31.5 6.9 32.0 6.4 33.2 8.0 changes in RVLFS and RVESV occurred after the
LVEF, % 62.1 5.7 61.5 5.1 60.5 6.2 59.3 5.9 completion of the fourth cycle of chemotherapy (T2).
LVGLS, % 24.2 3.6 22.3 4.0 20.5 3.3* 19.1 4.2*† These changes were associated with a subsequent
LVGCS, % 26.6 5.3 25.7 4.9 24.9 4.6 23.7 5.6*
decline of RVEF at the end of the entire chemo-
therapy (T3).
Values are mean SD. *Compared with T0, p < 0.05. †Compared with T1, p < 0.05. ‡Compared with T2,
p < 0.05.
LVEDV ¼ left ventricular end-diastolic volume; LVEF ¼ left ventricular ejection fraction; LVESV ¼ left ven-
RVEF, LVEF, AND ANTHRACYCLINE-INDUCED
tricular end-systolic volume; LVGCS ¼ left ventricular global circumferential strain; LVGLS ¼ left ventricular CARDIOTOXICITY. The definition of LV cardiotox-
global longitudinal strain; RVEDV ¼ right ventricular end-diastolic volume; RVEF ¼ right ventricular ejection
fraction; RVESV ¼ right ventricular end-systolic volume; RVLFS ¼ right ventricular longitudinal free-wall strain; icity in previously published studies has been vari-
RVLSS ¼ right ventricular longitudinal septal strain; other abbreviations are as in Table 2. able, which has been a limitation for the field of
cardio-oncology. Some studies defined LV car-
diotoxicity as a reduction of the LVEF of $5%
DRVESV and DRVLFS between T0 and T2 were to <55% with symptoms of heart failure or an
associated with subsequent RV cardiotoxicity asymptomatic reduction of the LVEF of $10%
(p ¼ 0.002 and p ¼ 0.001, respectively). However, to <55% (7), whereas other studies have chosen an
DRVEDV, DRVLSS, DLVGLS, and DLVGCS were not LVEF decline of <45% or 15 points from baseline as
predictors of RV cardiotoxicity. As shown in the ROC LV cardiotoxicity (14). The 2016 European Society of
curve analysis (Figures 2A and 2B), a relative decrease Cardiology position paper on cancer treatments and
in D RVLFS of >12.4% (sensitivity: 78.6%; specificity: cardiovascular toxicity recommends an absolute
82.6%; AUC: 0.80; p ＜0.001) and a relative increase decrease of 10% to a value below a lower limit of
in D RVESV of >13.2% (sensitivity: 71.4%; specificity: normal (6). Due to the paucity of data, the definition
71.7%; AUC: 0.76; p ＜0.001) were able to discriminate RV cardiotoxicity is not completely established in
between patients with and without RV cardiotoxicity cardio-oncology. Based on the aforementioned
(Figure 2A). When D RVLFS and D RVESV were com- guidance in defining LV cardiotoxicity from previous
bined, the sensitivity decreased to 69.6%, the speci- studies, the present authors found it reasonable to
ficity increased to 93.5%, and the AUC increased to use a percentage or relative decline in RVEF as a
0.82 (Figure 2B). The Central Illustration displays the definition of cardiotoxicity. The lower limit of
change sequence of 3DE-based RV parameters during normal in the present study was defined as 45%,
the chemotherapy. RV volumes and strains changed which has been recommended by the ASE guide-
earlier than RVEF. Of these volume and strain indices, lines (13).
RVESV and RVLFS could be predictors of RV car- Of note, 3DE-based LVEF did not show a significant
diotoxicity defined as the decline of RVEF at the end change during chemotherapy, whereas RVEF
of the follow-up. decreased significantly at T3. The reasons accounting
for this phenomenon may be the structural differ-
ences between the 2 ventricles and time period in
T A B L E 4 Univariable Analysis of Potential Predictors of
RV Cardiotoxicity
which cardiotoxicity may manifest. Anthracycline-
induced cardiotoxicity tends to display a regional
Odds Ratio 95% CI p Value
pattern in which the subendocardial component of
DRVESV 1.526 1.087–1.972 0.002
the ventricle is most susceptible to the toxicity (15).
DRVEDV 0.849 0.641–1.173 0.225
At an early stage, the impaired contractility of the
DRVLFS 1.393 1.093–1.692 0.001
DRVLSS 1.059 0.921–1.182 0.120
subendocardial layer could be compensated by the
DLVGLS 1.146 0.986–1.671 0.092 relatively normal epicardial layer, thus leading to a
DLVGCS 0.983 0.649–2.143 0.116 preserved LVEF. If the toxic effects of chemotherapy
persist, cardiac decompensation and symptomatic
D ¼ percentages of change between T0 and T2; CI ¼ confidence interval; other
abbreviations are as in Tables 2 and 3.
heart failure could occur. It is possible that the
thinner RV wall has less reserve for compensation
F I G U R E 2 ROC Curves for 3-Dimensional Echocardiographic Parameters in Predicting Right Ventricular Cardiotoxicity
(A) ROC curve analysis of DRVLFS and DRVESV respectively. The point corresponding to the cutoff value were marked as yellow circles on each curve respectively.
There were no significant differences between the area under each curve (p ¼ 0.45). (B) ROC curve analysis of combined DRVLFS and DRVESV (curve). When DRVLFS
and DRVESV were combined in the ROC curve analysis, the sensitivity declined while the specificity increased, and the AUC increased. D ¼ percentage change of measure
between T0 (pre-chemotherapy) and T2 (after 4 cycles); AUC ¼ area under the curve; ROC ¼ receiver operating characteristic; RVESV ¼ right ventricular end-systolic
volume; RVLFS ¼ right ventricular longitudinal free wall strain.
which might cause the earlier manifestation of a consistent with the results in the present study.
decline in RVEF (16). Given that RV dysfunction has They also found that longitudinal strain-RV free wall
been associated with worse outcomes in other pop- was impaired 3 months after the initiation of
ulations (17), it is important to further define and chemotherapy, which was similar to the time frame
understand the changes in RV function with observed in the present study. The present longitu-
anthracyclines. dinal data, however, uniquely demonstrates the
CHANGES IN RV STRAIN PARAMETERS AND ASSOCIATIONS temporal change in RV strain, RV volumes, and
WITH RV CARDIOTOXICITY. EF is an index frequently RVEF. After analyzing the relationship between the
used to reflect chemotherapy-related systolic changes of RV strain indices and RVEF, it was found
dysfunction. However, due to its lack of sensitivity that a relative decrease of >12.4% in RVLFS between
in early detection of subclinical cardiac injury in baseline and the completion of 4 cycles of anthra-
cancer patients receiving chemotherapy, the assess- cycline chemotherapy (T2) was associated with a
ment of EF alone may be limited (7). Strain derived subsequent RVEF decline after 6 cycles (T3). Our
from speckle tracking imaging is a technique which data may suggest that, with further study, RVLFS
has been used to evaluate subclinical cardiac could be a useful index to aid in the prediction of
dysfunction across multiple disease states (18). subsequent RV cardiotoxicity.
Several studies have demonstrated the use of LV However, whether there is an earlier time point at
strain parameters in the early detection and predic- which RVLFS could predict subsequent declines in
tion of LV cardiotoxicity induced by cancer (19,20). RVEF is still unknown. The changes in RVLSS, the
However, there are few studies in RV strain as it longitudinal septal strain of RV, demonstrated a
relates to cancer cardiotoxicity. delayed response compared with that in RVLFS. The
Boczar et al. (21) reported that longitudinal strain- present authors believe that the irregular geometry
RV free wall decreased significantly in patients with and the potential resultant uneven stress of the RV
breast cancer treated with anthracyclines, which was may have contributed to the asynchronous pattern in
C E NT R AL IL L U STR AT IO N 3DE RV Parameters to Predict RV Cardiotoxicity
Zhao, R. et al. J Am Coll Cardiol CardioOnc. 2020;2(1):13–22.
RV volumes, RV strain parameters and RVEF were determined during anthracycline chemotherapy. RVESV, RVEDV, and RVLFS changed significantly by T2. RVEF
declined significantly at T3. RV cardiotoxicity was defined as a relative reduction in RVEF of >10% or a 5% relative reduction in RVEF to <45%. A relative decrease in
RVLFS of >12.4% and a relative increase in RVESV of >13.2% from T0 to T2 was associated with subsequent RV cardiotoxicity. 3DE ¼ 3-dimensional echocardiography;
AUC ¼ area under the curve; DLBCL ¼ diffuse large B-cell lymphoma; R-CHOP ¼ cyclophosphamide, 750 mg/m2; vincristine, 1.4 mg/m2 up to a maximum dose of
2 mg/m2; doxorubicin, 50 to 70 mg/m2 on day 1; prednisone, 100 mg on days 1 to 5; and rituximab, 375 mg/m2; RV ¼ right ventricular; RVEDV ¼ right ventricular
end-diastolic volume; RVEF ¼ right ventricular ejection fraction; RVESV ¼ right ventricular end-systolic volume; RVLFS ¼ right ventricular longitudinal free-wall
strain; RVLSS ¼ right ventricular longitudinal septal strain; T0 ¼ before chemotherapy; T1 ¼ after the completion of 2 cycles of chemotherapy; T2 ¼ after the
completion of 4 cycles of chemotherapy; T3 ¼ after the completion of 6 cycles of chemotherapy.
the changes in strain (22). RVLSS mainly reflects the cycles of chemotherapy (T2) compared to baseline
longitudinal strain of the septal myocardium, (T0). Analogous to studies of LV size (24), RVESV
whereas RVLFS reflects the longitudinal strain of the increased before RVEF decline. The percent change in
free wall. According to Laplace’s law, wall stress RVESV between T0 and T2 was associated with a
correlates positively with pressure and radius and decrease in RVEF.
inversely with wall thickness. It is tempting to spec- Because the RV is more dependent on preload than
ulate that, given a free wall that is thinner in com- the LV, one might assume that preload resulted in the
parison to the septum, there was a greater degree of observed changes in RV function. However, the pa-
wall stress and thus worse strain (22,23) Moreover, as tient’s body weight, blood pressure, pulmonary arte-
the septum is shared by both LV and RV, LV defor- rial systolic pressures, IVCD, and estimated right atrial
mation may result in some compensatory effects. pressure did not change significantly during the
RV VOLUME PARAMETERS IN PREDICTING RV course of the study. Moreover, each echocardiogram
CARDIOTOXICITY. The volumetric alterations in 3DE was obtained 8 to 15 days after T1 and T2 and 50 to
RV were assessed, and it was determined that both 60 days after T3 to avoid the influence of acute preload
RVESV and RVEDV increased significantly after 4 changes. Of note, LV volumes did not change
significantly during follow-up. It remains unknown the present study, the decrease in RVEF preceded
why the RV may be more susceptible than the LV in the LVEF changes. If confirmed in external cohorts,
setting. At least to some extent, this may be related to future monitoring of anthracycline toxicity may
differences in LV wall thickness and wall stress. potentially include parameters based on the 3DE
assessment of the right ventricle.
STUDY LIMITATIONS. One limitation of the present
study was the absence of a comparison between 3DE
ADDRESS FOR CORRESPONDENCE: Dr. Leilei Cheng,
and cardiac magnetic resonance imaging or computed
Department of Echocardiography, Zhongshan Hospi-
tomography. Additionally, due to the small size and
tal, Fudan University, Shanghai Institute of Cardio-
short follow-up period of our study, a larger sample
vascular Diseases, Shanghai Institute of Medical
with longer follow-up period is required in future
Imaging, 180 Fenglin Road, Shanghai 200032, China.
studies and to draw more definitive conclusions.
E-mail: cheng.leilei@zs-hospital.sh.cn. OR Dr. David
Moreover, the sample size was relatively small, which
H. Hsi, Department of Cardiology, Stamford Hospital,
also limited the ability to adjust for confounders.
Vagelos Columbia University College of Physicians
Admittedly, the present definition of RV cardiotox-
and Surgeons, One Hospital Plaza, Stamford, Con-
icity was somewhat arbitrary and based on relative,
necticut 06904. E-mail: dhsi@stamhealth.org.
instead of absolute changes of small magnitude.
Twitter: @StamfordHosp.
Standardized, clinically meaningful definitions of RV
cardiotoxicity need to be developed. The specific in-
PERSPECTIVES
clusion and exclusion criteria also affect the gener-
alizability of the study. Although we tried to maintain
COMPETENCY IN MEDICAL KNOWLEDGE: In patients
stable volume status within patients and carefully
treated with anthracycline-based chemotherapy, the decrease
monitor preload conditions, parameters such as RV
in 3DE RVEF during anthracycline treatment occurred
volumes and strain are preload-dependent. Finally,
earlier than changes in LVEF. With further study, early changes
the reference cutoff points presented in our AUC an-
in 3DE RV longitudinal free-wall strain and RV end-systolic
alyses need to be externally validated and studied in
volumes could be promising predictors of cardiotoxicity.
much larger datasets.
TRANSLATIONAL OUTLOOK: More studies are needed to

CONCLUSIONS
determine the relationship between RVEF declines and adverse
clinical outcomes in patients exposed to anthracyclines. There is
Subclinical changes in RV function and size, including
also an important need for standardized definitions of cardio-
in 3DE-derived RV strain and volume, occur after the
toxicity. This study provides early data to support the potential
anthracycline-based chemotherapy. With further
use of 3DE-derived measurements of RV function and size in
study, RV longitudinal free-wall strain and RV end-
cardio-oncology.
systolic volumes could be useful indices to predict
subsequent declines in RVEF with anthracyclines. In
REFERENCES
1. Early Breast Cancer Trialists Collaborative Group relevance and response to pharmacologic therapy. 9. Oliveira GH, Dupont M, Naftel D, et al.
(EBCTCG). Effects of chemotherapy and hormonal J Am Coll Cardiol 2010;55:213–20. Increased need for right ventricular support in
therapy for early breast cancer on recurrence and patients with chemotherapy-induced cardiomy-
6. Zamorano JL, Lancellotti P, Rodriguez Muñoz D,
15-year survival: an overview of the randomized opathy undergoing mechanical circulatory sup-
et al. 2016 ESC position paper on cancer treat-
trials. Lancet 2005;365:1687–717. port. J Am Coll Cardiol 2014;63:240–8.
ments and cardiovascular toxicity developed un-
2. Hershman DL, McBride RB, Eisenberger A, et al. der the auspices of the ESC Committee for Practice 10. Surkova E, Muraru D. The use of multi-
Doxorubicin, cardiac risk factors, and cardiac Guidelines: the task force for cancer treatments modality cardiovascular imaging to assess right
toxicity in elderly patients with diffuse B-cell non- and cardiovascular toxicity of the European Soci- ventricular size and function. Int J Cardiol 2016;
Hodgkin’s lymphoma. J Clin Oncol 2008;26: ety of Cardiology (ESC). Eur Heart J 2016;37: 214:54–69.
3159–65. 2768–801.
11. Grewal J, Majdalany D, Syed I, et al. Three-
3. Pai VB, Nahata MC. Cardiotoxicity of chemo- 7. Sawaya H, Sebag IA, Plana JC, et al. Early dimensional echocardiographic assessment of
therapeutic agents. Incidence, treatment and detection and prediction of cardiotoxicity in right ventricular volume and function in adult
prevention. Drug Safety 2000;22:263–302. chemotherapy-treated patients. Am J Cardiol patients with congenital heart disease: comparison
4. Swain SM, Whaley FS, Ewer MS. Congestive 2011;107:1375–80. with magnetic resonance imaging. J Am Soc
heart failure in patients treated with doxorubicin: Echocardiogr 2010;23:127–33.
8. Bing R, Naoum C, Kritharides L. The vulnerable
a retrospective analysis of three trials. Cancer
right ventricle: recurrent, transient right ventric- 12. vanderZwaan HB, Helbing WA, McGhie JS,
2003;97:2869–79.
ular failure on a background of systemic sclerosis et al. Clinical value of real-time three-dimensional
5. Cardinale D, Colombo A, Lamantia G, et al. and previous anthracycline exposure. Int J Cardiol echocardiography for right ventricular quantifica-
Anthracycline-induced cardiomyopathy: clinical 2015;15:223–5. tion in congenital heart disease: validation with
cardiac magnetic resonance imaging. J Am Soc for Ewing’s sarcoma. J Cardiovasc Magn Reson cardiotoxicity. J Am Soc Echocardiogr 2013;26:
Echocardiogr 2010;23:134–40. 2006;8:789–91. 493–8.
13. Rudski LG, Lai WW, J, et al. Guidelines for the 17. Nagata Y, Wu VC, Kado Y, et al. Prognostic 21. Emery Boczar K, Aseyev O, Sulpher J, et al.
echocardiographic assessment of the right heart in value of right ventricular ejection fraction Right heart function deteriorates in breast cancer
adults: a report from the American Society of assessed by transthoracic 3D echocardiography. patients undergoing anthracycline-based chemo-
Echocardiography endorsed by the European As- Circ Cardiovasc Imaging 2017;10:1–11. therapy. Echo Res Pract 2016;2:79–84.
sociation of Echocardiography, a registered branch 22. Haddad F, Hunt SA, Rosenthal DN, et al. Right
18. Plana JC, Galderisi M, Barac A, et al. Expert
of the European Society of Cardiology, and the ventricular function in cardiovascular disease, part
consensus for multimodality imaging evaluation of
Canadian Society of Echocardiography. J Am Soc I anatomy, physiology, aging, and functional
adult patients during and after cancer therapy: a
Echocardiogr 2010;23:685–713. assessment of the right ventricle. Circulation
report from the American Society of Echocardi-
14. Ryberg M, Nielsen D, Cortese G, et al. ography and the European Association of Cardio- 2008;117:1436–48.
New insight into epirubicin cardiac toxicity: vascular Imaging. Eur Heart J Cardiovasc Imaging 23. Regen DM. Calculation of left ventricular wall
competing risks analysis of 1097 breast 2014;15:1063–93. stress. Circ Res 1990;67:245–52.
cancer patients. J Natl Cancer Inst 2008;100:
19. Stoodley PW, Boyd A, Hui R, et al. Left ven- 24. Mousavi N, Tan TC, Ali M, et al. Echocardio-
1058–67.
tricular systolic function in HER2/neu negative graphic parameters of left ventricular size and
15. Ho E, Brown A, Barrett P, et al. Subclinical breast cancer patients treated with anthracycline function as predictors of symptomatic heart failure
anthracycline- and trastuzumab-induced car- chemotherapy: a comparative analysis of left ven- in patients with a left ventricular ejection fraction
diotoxicity in the long-term follow up of asymp- tricular ejection fraction and myocardial strain im- of 50-59% treated with anthracyclines. Eur Heart
tomatic breast cancer survivors: a speckle tracking aging over 12 months. Eur J Cancer 2013;49: J Cardiovasc Imaging 2015;16:977–84.
echocardiographic study. Heart 2010;96:701–7. 3396–403.
16. Perel RD, Slaughter RE, Strugnell WE. Sub- 20. Negishi K, Negishi T, Hare JL, et al. Inde- KEY WORDS 3-dimensional
endocardial late gadolinium enhancement in two pendent and incremental value of deformation echocardiography, anthracycline, cardio-
patients with anthracycline following treatment indices for prediction of trastuzumab-induced oncology, cardiotoxicity, right ventricle
ª 2020 THE AUTHOR. PUBLISHED BY ELSEVIER ON BEHALF OF THE AMERICAN
EDITORIAL COMMENT
Anthracycline-Induced Cardiotoxicity
Remembering the Forgotten Ventricle*
Jennifer E. Liu, MD
S ince the 1960s, anthracyclines have been an in-

tegral part of cancer treatment, and they
remain as first-line therapy for a wide variety
of solid tumors and hematological malignancies,
walls and a normal or dilated LV cavity size. Given the
poor prognosis of advanced heart failure, many ef-
forts have focused on cardiac monitoring because
early detection of cardiac injury may facilitate timely
including breast cancer, lymphoma, leukemia, and therapeutic measures and mitigate cardiac damage.
sarcoma. However, dose-related cardiotoxicity pre- Two-dimensional (2D) transthoracic echocardiogra-
senting as left ventricular (LV) systolic dysfunction phy (TTE) has been the mainstay of imaging for the
with or without symptomatic heart failure is a assessment of cardiac function and detection of car-
limiting factor in the use of this highly effective diotoxicity during cancer treatment (4). Thus far,
chemotherapeutic agent. The cumulative incidence attention has been directed to the left side of the
of cardiac events has been estimated at 32%, 54%, heart, particularly LV ejection fraction and strain.
and 65% for cumulative doxorubicin doses of 400, Myocardial strain imaging is often incorporated into
500, and 550 mg/m 2 (1). Clinical evidence of cardio- monitoring for detection of subclinical LV dysfunc-
toxicity may become apparent within 1 to 2 years after tion. Right ventricular (RV) dysfunction, however,
treatment but can also manifest years later. Long- has not been considered in the definition of car-
term follow-up data from the Childhood Cancer diotoxicity. There have been limited investigations
Survivorship Study showed that adult survivors of into the effects of cancer chemotherapy on RV func-
childhood cancer treated with anthracyclines were tion and remodeling. The incidence and prognosis of
5 to 6 times as likely as their siblings to experience RV dysfunction during cardiotoxic treatment are
heart failure (2). largely unknown. Until recently, the right ventricle
Anthracyclines induce ultrastructural changes in had been relatively neglected, mainly due to the
the cardiomyocyte that lead to cellular apoptosis and technical challenges in the accurate assessment of the
necrosis. Early animal and human studies have sug- complex crescent-shaped right ventricle by 2D TTE
gested that the cardiotoxic effect of anthracyclines plus the impression that it may not be as important in
provokes global injury extending to both the left and cardiovascular disease.
right ventricles (3). The phenotype of anthracycline- SEE PAGE 13
induced cardiomyopathy is characterized by impair-
Advanced echocardiographic techniques such as
ment of systolic and diastolic function with thinned
three-dimensional TTE and strain imaging have led to
an improved assessment of RV anatomy and function
(5). This modality has emerged as a valuable tool
*Editorials published in JACC: CardioOncology reflect the views of the
providing critical information regarding the unique
authors and do not necessarily represent the views of JACC:
physiological properties of the right ventricle in
health and disease. Multiple studies have now shown
From the Memorial Sloan Kettering Cancer Center, Weill Cornell Medical
Center, New York, New York. Dr. Liu serves on the advisory board for
the incremental value of RV function over clinical risk
Pfizer; and is a consultant for Bay Labs. factors and other parameters of LV dysfunction for
The authors attest they are in compliance with human studies commit- predicting outcome in patients with various cardio-
vascular pathologies (6). In this issue of JACC:
appropriate. For more information, visit the JACC: CardioOncology author
CardioOncology, Zhao et al. (7) reported their inves-
instructions page. tigation using 2D and 3-dimensional (3D) TTE with

24 Liu JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020
Anthracycline-Induced Cardiotoxicity MARCH 2020:23–5
strain imaging to describe the sequence of changes in childhood cancers treated with anthracyclines have
RV size and function in patients with lymphoma impaired RV function or reduced functional reserve
during the course of their doxorubicin-based during exercise compared with matched control
chemotherapy. The results showed an increase in subjects (11,12). It would be important to establish if
RV end-systolic and end-diastolic volume followed by changes in RV function in the current study persist
a decrease in RV ejection fraction (RVEF), as well as a over time or predict subsequent declines in LVEF or
worsening in RV strain over time. There was a con- heart failure. Another limitation is that the RV
current worsening in LV global longitudinal strain but remodeling observed may still be related to fluctua-
no change in LV volume or left ventricular ejection tions in pre-load, even though the volume status was
fraction (LVEF) during the follow-up period. The deemed stable, as even a small change in fluid status
study identified changes in RV end-systolic volume could lead to a modest change in RV volume. Re-
and RV free wall longitudinal strain as predictors of measuring the RV parameters after therapy and sta-
RV cardiotoxicity defined by a reduction in RVEF bilization would help clarify and confirm the presence
from baseline. of RV dysfunction. Furthermore, the clinical signifi-
This paper (7) adds to the existing literature and cance of “RV cardiotoxicity” is unknown as it was
further informs our understanding of the impact of arbitrarily defined based on a relative change in
anthracycline therapy on the right ventricle. Although RVEF. Whether RV dysfunction assessed by echocar-
a few studies have shown a decline in RV function diography is associated with any adverse clinical
during cancer treatment (8), this study found that RV outcome requires further evaluation. Finally, this was
remodeling and functional impairment occurred a small single-center study, and the validity of the
before any changes in LVEF or LV volume. This is an findings must be confirmed in larger studies. Ideally,
important finding not previously reported, suggesting cardiac magnetic resonance imaging, the reference
the possibility that the right ventricle may be affected standard for RV volume and function assessment, at
earlier than the left ventricle. Given its distinct baseline and end of therapy, would improve the
morphological features and myofiber architecture, the reliability of the observations. The addition of tissue
right ventricle could conceivably be more vulnerable characterization by cardiac magnetic resonance im-
than the left ventricle to injury. It is intuitive that the aging, such as myocardial edema and inflammation,
thinner RV walls with fewer myofibrils may be more could provide further insight into the underlying
sensitive to the toxic effects of chemotherapy than the histopathological myocardial changes (13).
thicker, more muscular left ventricle. Furthermore, Despite the uncertain impact of these findings, this
the RV free wall is composed predominantly of longi- study (7) should be recognized for expanding the
tudinally arranged, apex to base subendocardial fibers application and feasibility of 3D TTE for cardiotox-
that account for the majority of RV pump function icity monitoring in patients during cancer treatment.
under normal conditions (9). The subendocardial layer Previous studies have reported abnormalities in 3D
of the myocardium, responsible for the longitudinal LV mechanics in anthracycline-induced cardiotox-
shortening, has been shown to be most vulnerable to icity (14). This study provides novelty with mea-
the toxic effects of chemotherapy in histological surement of 3D RV mechanics and volume in patients
studies (10). Unlike the left ventricle, the mid-layer receiving anthracycline treatment. However, whether
containing circumferential fibers in the RV myocar- 3D TTE is a viable modality to monitor RV structure
dium is absent and thus cannot compensate for the loss and function outside the research setting remains to
of longitudinal shortening of the subendocardial fi- be seen. Longitudinal studies are needed to assess the
bers. Thus, longitudinal shortening assessed by strain impact of echocardiography-derived indices of RV
imaging should have the potential to detect early function on prognostic outcome and whether they
changes in RV myocardial dysfunction. The current represent clinically meaningful markers of car-
study detected impairment of longitudinal strain of diotoxicity that should be incorporated as part of
the RV free wall before RVEF decline which can be cardiotoxicity monitoring during cancer therapy.
explained by RV mechanics. However, whether
impairment occurs in the right ventricle before the left
ventricle warrants further evaluation. ADDRESS FOR CORRESPONDENCE: Dr. Jennifer E.
There are several limitations to the study by Zhao Liu, Memorial Sloan Kettering Cancer Center, Weill
et al. (7). The most important one is the lack of follow- Cornell Medical Center, 1275 York Avenue, New York,
up data to place the RV changes in context. Studies New York 10065. E-mail: liuj1234@mskcc.org. Twitter:
have shown that long-term adult survivors of @sloan_kettering.
JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020 Liu 25
MARCH 2020:23–5 Anthracycline-Induced Cardiotoxicity
REFERENCES
1. Swain SM, Whaley FS, Ewer MS. Congestive heart focused update on the right ventricle in cardiovas- of childhood lymphoma and acute lymphoblastic
failure in patients treated with doxorubicin: a retro- cular disease. J Am Coll Cardiol HF 2018;6:891–903. leukaemia. Eur Heart J Cardiovasc Imaging 2016;
spective analysis of three trials. Cancer 2003;97: 17:735–41.
7. Zhao R, Shu F, Zhang C, et al. Early detection
2869–79.
and prediction of anthracycline-induced right 12. Li VWY, Liu APY, Wong WHS, et al. Left and
2. Oeffinger KC, Mertens AC, Sklar CA, et al. ventricular cardiotoxicity by 3-dimensional echo- right ventricular systolic and diastolic functional
Chronic health conditions in adult survivors of cardiography. J Am Coll Cardiol CardioOnc 2020; reserves are impaired in anthracycline-treated
childhood cancer. N Engl J Med 2006;355: 2:13–22. long-term survivors of childhood cancers. J Am
1572–82. Soc Echocardiogr 2019;32:277–85.
8. Tadic M, Cuspidi C, Hering D, Venneri L,
3. Shan K, Lincoff AM, Young JB. Anthracycline- Danylenko O. The influence of chemotherapy on 13. Jordan JH, Todd RM, Vasu S, Hundley WG.
induced cardiotoxicity. Ann Intern Med 1996;125: the right ventricle: did we forget something? Clin Cardiovascular magnetic resonance in the
47–58. Cardiol 2017;40:437–43. oncology patient. J Am Coll Cardiol Img 2018;11:
1150–72.
4. Liu J, Banchs J, Mousavi N, et al. Contemporary 9. Ho SY, Nihoyannopoulos P. Anatomy, echocar-
role of echocardiography for clinical decision diography, and normal right ventricular di- 14. Zhang KW, Finkelman BS, Gulati G, et al.
making in patients during and after cancer ther- mensions. Heart 2006;92 Suppl 1:i2–13. Abnormalities in 3-dimensional left ventricular
apy. J Am Coll Cardiol Img 2018;11:1122–31. mechanics with anthracycline chemotherapy
10. Mortensen SA, Olsen HS, Baandrup U. Chronic
are associated with systolic and diastolic
5. Lang RM, Addetia K, Narang A, Mor-Avi V. anthracycline cardiotoxicity: haemodynamic and
dysfunction. J Am Coll Cardiol Img 2018;11:
3-Dimensional echocardiography: latest de- histopathological manifestations suggesting a
1059–68.
velopments and future directions. J Am Coll Cardiol restrictive endomyocardial disease. Br Heart J
Img 2018;11:1854–78. 1986;55:274–82.
6. Amsallem M, Mercier O, Kobayashi Y, 11. Christiansen JR, Massey R, Dalen H, et al. Right KEY WORDS anthracyclines, cardiotoxicity, 3D
Moneghetti K, Haddad F. Forgotten no more: a ventricular function in long-term adult survivors echocardiography, right ventricle
ORIGINAL RESEARCH
Longitudinal Changes in
Echocardiographic Parameters
of Cardiac Function in
Pediatric Cancer Survivors
William L. Border, MBCHB, MPH,a,b Ritu Sachdeva, MBBS,a,b Kayla L. Stratton, MS,c Saro H. Armenian, DO, MPH,d
Aarti Bhat, MD,e David E. Cox, RDCS,b Kasey J. Leger, MD, MS,e Wendy M. Leisenring, SCD,c
Lillian R. Meacham, MD,a,f Karim T. Sadak, MD,g Shanthi Sivanandam, MD,g Paul C. Nathan, MD, MSC,h
Eric J. Chow, MD, MPHc,e
ABSTRACT
OBJECTIVES The purpose of this study was to assess the timing of changes in serial echocardiographic parameters in
pediatric cancer survivors and to evaluate their associations with cardiomyopathy development.
BACKGROUND Pediatric cancer survivors undergo serial echocardiograms to screen for cardiotoxicity. It is not clear
whether small longitudinal changes in functional or structural parameters over time have clinical significance.
METHODS This is a multicenter, retrospective, case-control study of $1-year survivors following the end of cancer
therapy. Cardiomyopathy cases (fractional shortening [FS] #28% or ejection fraction [EF] #50% on $2 occasions) were
matched to control subjects (FS $30%, EF $55%, not on cardiac medications) by cumulative anthracycline and chest
radiation dose, follow-up duration, and age at diagnosis. Digitally archived clinical surveillance echocardiograms were
quantified in a central core laboratory, blinded to patient characteristics. Using mixed models with interaction terms
between time and case status, we estimated the least square mean differences of 2-dimensional, M-mode, pulsed wave
Doppler, and tissue Doppler imaging–derived parameters over time between cases and control subjects.
RESULTS We identified 50 matched case-control pairs from 5 centers. Analysis of 412 echocardiograms (cases, n ¼ 181;
control subjects, n ¼ 231) determined that indices of left ventricular systolic function (FS, biplane EF), diastolic function
(mitral E/A ratio), and left ventricular size (end-diastolic dimension z-scores) were significantly different between cases
and control subjects, even 4 years prior to the development of cardiomyopathy.
CONCLUSIONS Longitudinal changes in cardiac functional parameters can occur relatively early in pediatric cancer
survivors and are associated with the development of cardiomyopathy. (J Am Coll Cardiol CardioOnc 2020;2:26–37)
© 2020 The Authors. Published by Elsevier on behalf of the American College of Cardiology Foundation. This is an open
access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
From the aEmory University School of Medicine, Atlanta, Georgia; bChildren’s Healthcare of Atlanta, Sibley Heart Center, Atlanta,
Georgia; cClinical Research and Public Health Sciences Divisions, Fred Hutchinson Cancer Research Center, Seattle, Washington;
d
Department of Population Sciences, City of Hope, Duarte, California; eDepartment of Pediatrics, University of Washington,
Seattle Children’s Hospital, Seattle, Washington; fChildren’s Healthcare of Atlanta Aflac Cancer Center, Atlanta, Georgia; gDe-
partment of Pediatrics, University of Minnesota, Masonic Children’s Hospital, Minneapolis, Minnesota; and the hDepartment of
Pediatrics, University of Toronto, The Hospital for Sick Children, Toronto, Ontario, Canada. This work was supported by a
generous grant from The Rally Foundation, Atlanta, Georgia, and the National Institutes of Health (R01 CA211996). Dr. Leger has

JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020 Border et al. 27
MARCH 2020:26–37 Longitudinal Echo Changes in Pediatric Cancer Survivors
S urvival rates in pediatric cancer continue to centrally analyzed serial echocardiograms ABBREVIATIONS
improve with current 5-year overall survival previously obtained as part of routine care AND ACRONYMS
rates at 80%, and over 430,000 pediatric can- prior to the onset of cardiomyopathy and
EF = ejection fraction
cer survivors living in the United States (1). However, over a similar time period for matched control
FS = fractional shortening
cardiovascular disease is one of the leading contribu- subjects. Cases who developed cardiac
PW = pulsed wave
tors to late morbidity and mortality in this population dysfunction were matched to control sub-
(2). More specifically, those treated with anthracy- jects, and detailed evaluation of echocardio- TDI = tissue Doppler imaging
cline chemotherapy and/or chest radiation are at an graphic indices of left ventricular (LV) systolic
increased risk of developing heart failure and cardio- function, diastolic function, size, and geometry were
myopathy (3,4). Many of these patients will have undertaken. By taking advantage of serial echocar-
detectable cardiac dysfunction prior to the develop- diograms within a multi-institutional case-control
ment of overt clinical heart failure (5). As a result, na- study design, we sought to address the question of
tional and international groups have developed which longitudinal changes in echocardiographic pa-
recommendations regarding echocardiographic sur- rameters are associated with the subsequent devel-
veillance to detect cardiac dysfunction early (6,7). opment of cardiac dysfunction in pediatric cancer
survivors.
SEE PAGE 38
The adult cardio-oncology community has pub- METHODS

lished echocardiographic guidelines for the detailed
evaluation of cardiac function in adults during and PATIENTS. Eligibility for this retrospective case-
after cancer therapy, including non-ejection fraction control study was limited to individuals previously
(EF)–based parameters such as global longitudinal diagnosed with cancer at age <21 years and followed
strain (8). However, specific guidelines do not exist at 1 of 5 participating centers (City of Hope, Children’s
for pediatric patients, and there is a paucity of lon- Healthcare of Atlanta/Emory University, Hospital for
gitudinal data to facilitate comparable recommenda- Sick Children, Seattle Children’s Hospital, and
tions. With rare exceptions (9), most research studies Masonic Children’s Hospital/University of Minne-
examining echocardiographic changes in pediatric sota). All participants had to survive at least 1 year
patients have been cross-sectional single-institution beyond the end of initial cancer therapy. Cardiomy-
studies not enriched for cardiomyopathy cases, which opathy cases were defined by echocardiography as
makes it difficult to evaluate trajectories over time having either a fractional shortening (FS) #28% or
(10,11). The relative rarity at single centers of ejection fraction (EF) #50%, on at least 2 occasions,
pediatric-age patients developing cardiac dysfunc- with at least 1 of those measurements occurring after
tion has hampered the ability to rigorously study completion of cancer therapy. Cases where these FS
echocardiographic parameters longitudinally. Greater or EF criteria were met only once were also accepted
insight into these trajectories could help identify the if the qualifying measurement occurred after the
subpopulation of patients who will develop subse- completion of cancer therapy and led to the initiation
quent, true cardiac dysfunction, and could poten- of medical treatment for cardiomyopathy. Control
tially allow for earlier medical intervention. subjects were individually matched to cases based on
We therefore undertook a retrospective multi- the following criteria (in order of descending prior-
institutional study involving five pediatric cancer ity): 1) cumulative anthracycline dose and type
centers aimed at assessing the longitudinal trajec- (50 mg/m 2 increments for doxorubicin and daunoru-
tories of cardiac function and structure in pediatric bicin; 25 mg/m2 for idarubicin and mitoxantrone); 2)
cancer survivors prior to the onset of cardiomyopathy chest radiotherapy (none, <15 Gy, 15 to 34 Gy, $35
and contrasting these with closely matched survivor Gy); 3) follow-up duration (1, 2, and 3 years to
control subjects. Specifically, we collected and within 5, 5 to 9, and $10 years of cancer diagnosis,
received research funding from Abbott (and upcoming Jazz funding); and has served on the advisory board of Boston Scientific.
All other authors have reported that they have no relationships relevant to the contents of this paper to disclose. Amil Shah, MD,
served as Guest Associate Editor for this paper.
Manuscript received September 20, 2019; revised manuscript received February 10, 2020, accepted February 11, 2020.
28 Border et al. JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020
Longitudinal Echo Changes in Pediatric Cancer Survivors MARCH 2020:26–37
respectively); 4) age at cancer diagnosis (5 years); quantitation. This was felt to be an inadequate sam-
and 5) sex. Furthermore, control subjects had to have, ple, and thus GLS was not included in the analysis.
by their institutional echocardiographic reports,
STATISTICAL ANALYSES. Categorical patient char-
FS $30% and EF $55%, without known qualitative
acteristics and treatments are presented as frequency
changes concerning for cardiomyopathy during the
(percent) and were compared between cases and
same follow-up time interval as their matched case.
control subjects using chi-square tests. For continuous
To avoid issues with misclassification, control sub-
measures, normality was evaluated based on the
jects also could not be treated with antihypertensive
Kolmogorov-Smirnov statistic, and measures are pre-
medications during the follow-up time interval. We
sented as mean SD if normally distributed or median
also excluded patients with known congenital heart
(interquartile 25th and 75th percentiles) if not, with p
anomalies (except patent foramen ovale) or underly-
values based on either Student’s t-tests or Wilcoxon
ing genetic syndrome associated with abnormal car-
tests, respectively. Because distributions of the echo-
diovascular development. The institutional review
cardiographic parameters were normally distributed,
boards at all participating institutions approved the
mixed models were used to compare cases and control
study procedures with waiver of consent.
subjects over time prior to cardiomyopathy diagnosis
EXPOSURES. The medical records of cases and con- or the same time for control subjects (case index time).
trols were reviewed. Information on patient de- Models included a random effect to account for
mographics, the original cancer diagnosis, and any matched pairs, and an autoregressive correlation
relapse, if applicable, was abstracted. Lifetime structure was assumed for repeated measurements
anthracycline and anthraquinone doses, by individ- within individuals. Categorical time variables were
ual agent, were summed based on doxorubicin created for echocardiogram timing prior to the index
equivalency, as was any radiotherapy field poten- diagnosis, mapping continuous times to integer year
tially affecting the heart, using the Children’s categories as follows: index, <1, 1 to <2, 2 to <3, 3 to <4,
Oncology Group version 5 guideline definitions (12), 4 to <5, 5 to <6, 6 to <7, and $7 years prior to index
with the exception that mitoxantrone was considered time. Mixed models included indicator variables for all
10 times more cardiotoxic than doxorubicin. Age and time categories, indicator of case status, and the
vital status at last follow-up were also recorded. interaction terms between time and case variables.
OUTCOMES. All available archived echocardio- Least square means (LSMs) with 95% confidence in-
graphic images for selected cases and matched con- tervals (CIs) were estimated for each time category and
trol subjects were deidentified and submitted in displayed for cases and control subjects. Additional
DICOM format to a central core laboratory at Emory comparisons within specific time intervals (e.g., more
University for offline analysis, using a vendor-neutral than 2 years before the index time) were calculated
software analysis package (TOMTEC Corporation, using differences between LSMs, constructed from the
Chicago, Illinois). The submitted studies were ob- relevant parameter and variance covariance estimates
tained for routine clinical purposes from multiple from the model.
institutions over a 13-year period (2004 to 2017). A During the time period more than 2 years prior to
single blinded reviewer (D.E.C.) analyzed all studies. the index time, all endpoints were fit as a linear
Study parameters included those related to LV size model as a function of time. To illustrate the trajec-
and geometry (posterior wall thickness, end-diastolic tory of echocardiogram endpoints over this time for
dimension [EDD], and thickness-to-dimension ratio), cases and control subjects, linear mixed models with
LV systolic function (FS, EF derived by Biplane continuous time, case status, and an interaction be-
Simpson’s method, mitral Sʹ, and septal Sʹ), LV dia- tween time and case status were also fit for each
stolic function (mitral E/A ratio, mitral Eʹ, mitral E/Eʹ, endpoint; predicted means with 95% confidence
septal Eʹ, and septal E/Eʹ), and combined LV systolic/ bands were displayed.
diastolic function (myocardial performance index Most echocardiograms were assessed post-therapy.
[MPI], derived from both pulsed wave Doppler and However, an additional set of mixed models
tissue Doppler). Both 2-dimensional (2D) and M-mode compared cases and control subjects with time cate-
measurements were made, but unless otherwise gorized into 3 exclusive categories: 1) at the time of
specified, all reported values are based on 2D views. cancer diagnosis; 2) on cancer therapy; and 3) post-
We attempted to derive indices of deformation such cancer therapy. These models all included an inter-
as global longitudinal strain (GLS). However, only action term between case status and time, with re-
10% of the submitted echocardiograms had adequate sults presented as LSMs (95% CIs) and p values. Data
apical 2-, 3-, and 4- chamber views for accurate were analyzed using SAS (version 9.4, SAS Institute
Inc., Cary, North Carolina), and a 2-sided p

T A B L E 1 Demographic and Treatment Characteristics of Matched Cardiomyopathy Cases
value <0.05 was considered statistically significant. and Control Subjects
Cases Control Subjects

RESULTS (n ¼ 50) (n ¼ 50) p Value*
Female 18 (36) 22 (44) 0.414

By design, demographic and treatment characteristics Race/ethnicity
among the 50 cases and 50 matched control subjects White, non-Hispanic 26 (56) 27 (63) 0.547†
were similar (Table 1). Notably, 64% of cases were Black 7 (15) 4 (9)
men (56% of control subjects; p ¼ 0.414) and Asian 7 (15) 7 (16)

Hispanic 4 (9) 5 (12)
approximately 40% of the sample was of a minority
Multiracial 2 (4) 0
racial/ethnic background. Among cases, the mean
Unknown 4 (8) 7 914)
time interval from cancer diagnosis to the cardiomy- Mean age at cancer diagnosis, yrs 8.0 5.5 7.2 4.6 0.431
opathy index time point was 6.4 5.3 years. Most Year of cancer diagnosis 0.773
cases and control subjects received only 1 anthracy- 1991–1999 8 (16) 7 (14)
cline or anthraquinone (76% or 78%, respectively); 2000–2005 17 (34) 15 (30)
only 4% received 3 agents. 2006–2009 17 (34) 22 (44)
2010–2015 8 (16) 6 (12)
Detailed, core laboratory quantified measure-
Cancer diagnosis 0.127
ments were performed on 412 echocardiograms
Leukemia 14 (28) 14 (28)
(cases, 181 studies; control subjects, 231 studies),
Lymphoma 11 (22) 15 (30)
with a median number of 5 echocardiograms for Sarcoma 14 (28) 18 (36)
cases and 4 echocardiograms for control subjects, Other solid tumor 11 (22) 3 (6)
and a mean follow-up time since cancer diagnosis Doxorubicin exposure 41 (82) 42 (84) 0.790
of 5.4 5.0 years for cases and 6.2 4.4 years for Median dose, mg/m2 260 (225–375) 300 (200–375) 0.934
control subjects. We then examined differences in Daunorubicin exposure 17 (34) 14 (28) 0.516
Median dose, mg/m2 120 (95–300) 100 (100–300) 0.732
LV systolic function, diastolic function, MPI, size,
Mitoxantrone exposure 5 (10) 5 (10) 1.000
and geometry across categorical time periods. We
Median dose, mg/m2 48 (48–50) 48 (48–48) n-e
observed significant differences between cases
Idarubicin exposure 1 (2) 2 (4) 0.557
versus control subjects for FS, EF, mitral E/A ratio, Cumulative anthracycline dose, mg/m2‡ 280 (200–450) 300 (200–450) 0.994
and LVEDD across multiple time periods, whereas Radiotherapy affecting the heart 22 (44) 21 (42) 0.839
significant differences in MPI and LV thickness- Years of follow-up (cancer diagnosis 0.588
to last follow-up)
dimension ratio were only seen at the index time
<10 23 (46) 29 (58)
point (Table 2, Figure 1). For example, even 2 to 3.9
10þ 27 (54) 21 (42)
years prior to the index point, significant differ- Years of follow-up from cancer diagnosis N/A
ences in FS, EF, mitral E/A ratio, and LVEDD were to first abnormal echo
observed, whereas the difference in MPI, derived by <2 17 (34) N/A
2–9 17 (34) N/A
pulsed wave Doppler, was of borderline significance
10þ 16 (32) N/A
(p ¼ 0.055). Combining estimates from all time pe-
Age at first abnormal echo, yrs
riods other than the 2 years within the index time
0–9 7 (14) N/A
point, differences in FS, EF, mitral E/A ratio, and 10–14 17 (34) N/A
LVEDD remained significant. Results for LV FS and 15–19 20 (40) N/A
EDD by M-mode were similar to 2D results (data not 20–24 4 (8) N/A
shown). 25–35 2 (4) N/A
Given our interest in the trajectory of change Treated for cardiomyopathy 27 (54) N/A
History of heart transplant 1 (2) N/A
prior to cardiomyopathy diagnosis, we also assessed
Mean age at last follow-up, yrs 18.3 5.2 17.2 4.4 0.224
differential model shapes and slopes over time,
excluding data within 2 years of the index time Values are n (%), mean SD, or median (interquartile 25th and 75th percentiles). *Comparison of cases and
point. In these models, which included an interac- control subjects based on chi-square test for categorical values; Student’s t-test, or Wilcoxon for continuous
values. †White non-Hispanic vs. other. ‡Based on the following conversion for doxorubicin equivalence:
tion of case status with time, significant nonline- daunorubicin 0.5, epirubicin 0.67, idarubicin 3.0, and mitoxantrone 10.0.
arity was not detected, and therefore, linear models N/A ¼ not applicable; n-e ¼ not estimable.
were used (Central Illustration). The slope of biplane

EF (as well as septal Eʹ and septal E/Eʹ; data not LVEDD, and thickness-dimension ratio did not have
shown) was significantly different between cases significantly differing trajectories.
versus control subjects over this time period, but Finally, we evaluated differences in central, core
FS, mitral E/A ratio, pulsed wave Doppler MPI, laboratory quantified echocardiographic parameters
T A B L E 2 Differences, Derived by LSM (95% Confidence Intervals), of Centrally Quantified Echocardiographic Parameters Between Cases and Control Subjects as a
Function of Time Prior to the Index Time Point (i.e., Cardiomyopathy Diagnosis)*
$6 Yrs 4 to <6 Yrs 2 to <4 Yrs <2 Yrs All Times >2 Yrs
Prior to Index Prior to Index Prior to Index Prior to Index Index Timepoint Prior to Index Prior to Index
LV systolic
function
2D-derived 4.0 3.1 2.8 3.0 8.4 3.2 3.3
FS, % (1.2 to 6.7) (0.7 to 5.5) (0.7 to 4.8) (1.0 to 4.8) (6.7 to 10.1) (1.9 to 4.5) (1.7 to 4.8)
p ¼ 0.005 p ¼ 0.013 p ¼ 0.008 p ¼ 0.003 p < 0.001 p < 0.001 p < 0.001
Biplane 2.8 2.8 4.8 4.6 12.4 3.7 3.5
EF, % (3.2 to 8.8) (1.4 to 7.0) (1.5 to 8.1) (1.6 to 7.6) (9.8 to 15.0) (1.3 to 6.1) (0.5 to 6.4)
p ¼ 0.364 p ¼ 0.194 p ¼ 0.006 p ¼ 0.004 p < 0.001 p ¼ 0.003 p ¼ 0.023
Mitral Sʹ 0.005 0.003 0.007 0.007 0.027 0.003 0.002
TDI, m/s (0.025 to 0.015) (0.013 to 0.019) (0.007 to 0.021) (0.004 to 0.018) (0.017 to 0.038) (0.005 to 0.011) (0.009 to 0.012)
p ¼ 0.602 p ¼ 0.700 p ¼ 0.316 p ¼ 0.193 p < 0.001 p ¼ 0.469 p ¼ 0.745
Septal Sʹ 0.005 0.003 0.003 0.003 0.012 0.001 0.000
TDI, m/s (0.014 to 0.003) (0.004 to 0.010) (0.004 to 0.009) (0.008 to 0.003) (0.007 to 0.017) (0.004 to 0.003) (0.004 to 0.005)
p ¼ 0.234 p ¼ 0.406 p ¼ 0.421 p ¼ 0.366 p < 0.001 p ¼ 0.778 p ¼ 0.952
LV diastolic
function
Mitral PW 0.59 0.07 0.45 0.24 0.56 0.34 0.37
E/A ratio (0.21 to 0.96) (0.27 to 0.40) (0.16 to 0.74) (0.02 to 0.50) (0.32 to 0.80) (0.16 to 0.51) (0.16 to 0.58)
p ¼ 0.002 p ¼ 0.699 p ¼ 0.002 p ¼ 0.069 p < 0.001 p < 0.001 p < 0.001
Mitral Eʹ, 0.014 0.004 0.013 0.001 0.045 0.007 0.010
m/s (0.018 to 0.046); (0.022 to 0.029) (0.010 to 0.036) (0.017 to 0.016) (0.029 to 0.061) (0.007 to 0.021) (0.007 to 0.027)
p ¼ 0.405 p ¼ 0.784 p ¼ 0.268 p ¼ 0.930 p < 0.001 p ¼ 0.307 p ¼ 0.257
Mitral E/Eʹ 0.10 0.38 0.06 0.26 0.77 0.12 0.08
ratio (1.62 to 1.81) (1.69 to 0.93) (1.23 to 1.35) (1.15 to 0.62) (1.63 to 0.08) (0.92 to 0.67) (1.04 to 0.89)
p ¼ 0.913 p ¼ 0.567 p ¼ 0.927 p ¼ 0.560 p ¼ 0.078 p ¼ 0.762 p ¼ 0.878
Septal Eʹ 0.018 0.010 0.008 0.002 0.018 0.000 0.000
TDI, m/s (0.035 to 0.002) (0.004 to 0.023) (0.004 to 0.021) (0.0130.009) (0.009 to 0.028) (0.008 to 0.007) (0.009 to 0.009)
p ¼ 0.032 p ¼ 0.150 p ¼ 0.192 p ¼ 0.751 p < 0.001 p ¼ 0.918 p ¼ 0.991
Septal E/Eʹ 1.74 0.46 0.26 0.16 0.07 0.21 0.34
ratio (0.33 to 3.16) (1.61 to 0.69) (1.35 to 0.83) (1.09 to 0.76) (0.90 to 0.76) (0.46 to 0.89) (0.47 to 1.15)
p ¼ 0.016 p ¼ 0.431 p ¼ 0.639 p ¼ 0.729 p ¼ 0.867 p ¼ 0.533 p ¼ 0.413
LV combined
systolic/
diastolic
function
PW Doppler- 0.006 0.030 0.032 0.026 0.061 0.021 0.019
derived (0.036 to 0.048) (0.065 to 0.004) (0.064 to 0.001) (0.055 to 0.003) (0.088 to 0.034) (0.039 to 0.003) (0.041 to 0.003)
MPI p ¼ 0.788 p ¼ 0.088 p ¼ 0.055 p ¼ 0.084 p < 0.001 p ¼ 0.026 p ¼ 0.093
TDI-derived 0.020 0.042 0.032 0.001 0.063 0.013 0.018
MPI (0.043 to 0.084) (0.091 to 0.008) (0.075 to 0.011) (0.041 to 0.043) (0.101 to 0.026) (0.042 to 0.016) (0.053 to 0.017)
p ¼ 0.536 p ¼ 0.100 p ¼ 0.144 p ¼ 0.962 p ¼ 0.001 p ¼ 0.374 p ¼ 0.314
LV size and
geometry
Posterior 0.26 0.11 0.16 0.03 0.10 0.14 0.18
wall (0.81 to 0.30) (0.58 to 0.35) (0.57 to 0.25) (0.39 to 0.33) (0.21 to 0.41) (0.42 to 0.14) (0.52 to 0.16)
thickness, p ¼ 0.364 p ¼ 0.628 p ¼ 0.437 p ¼ 0.867 p ¼ 0.532 p ¼ 0.330 p ¼ 0.305
z-score
End- 0.28 0.87 1.09 0.81 1.18 0.77 0.75
diastolic (1.04 to 0.47) (1.49 to 0.25) (1.64 to 0.55) (1.31 to 0.32) (1.62 to 0.74) (1.19 to 0.34) (1.24 to 0.26)
dimension, p ¼ 0.462 p ¼ 0.006 p < 0.001 p ¼ 0.001 p < 0.001 p < 0.001 p ¼ 0.003
z-score
Thickness to 0.001 0.005 0.005 0.006 0.013 0.004 0.003
dimension (0.016 to 0.014) (0.008 to 0.018) (0.005 to 0.016) (0.004 to 0.016) (0.004 to 0.022) (0.004 to 0.011) (0.006 to 0.012)
ratio p ¼ 0.864 p ¼ 0.439 p ¼ 0.321 p ¼ 0.231 p ¼ 0.004 p ¼ 0.323 p ¼ 0.507
*Differences were based on comparisons of LSMs from longitudinal mixed models for each measure, including an interaction between case status and time period, using an autoregressive correlation structure
between echocardiograms for the same patient. For example, at the index time point, the estimated FS is 8.4% greater in control subjects compared with cases; at the <2-year time point, the estimated FS is
3.0% greater in control subjects compared with cases.
2D ¼ 2-dimensional; EF ¼ ejection fraction; FS ¼ fractional shortening; LSM ¼ least square means; LV ¼ left ventricular; MPI ¼ myocardial performance index; PW ¼ pulsed wave Doppler derived;
TDI ¼ tissue Doppler imaging.
F I G U R E 1 Least Square Means for Echocardiographic Functional Parameters Across Time Periods Prior to Cardiomyopathy
Continued on the next page

between cases and control subjects during 3 clinically the past, making it difficult to be able to intervene
defined time periods: at time of cancer diagnosis, before the abnormal finding is reported. This has
while on cancer therapy, and after therapy up to and started to change in more recent guidelines in adult
including the time of the index case echocardiogram cancer patients with an effort to account for trajec-
(Table 3). Notably, significant differences between tory of change in these indices (8).
cases and control subjects were seen during the “on TRADITIONAL ECHOCARDIOGRAPHIC MEASURES
therapy” period. OF LV SYSTOLIC FUNCTION AND REMODELING.
Early studies showed progressive cardiac dysfunction
DISCUSSION
in pediatric cancer survivors primarily by document-
ing the development of abnormalities in systolic
This retrospective multi-institutional case-control
function as measured by fractional shortening and
study of pediatric cancer survivors demonstrated that
geometric changes, such as reduced LV mass and wall
there was a measurable decline in echocardiographic
thinning (9). Similar changes have been demon-
parameters of cardiac function and remodeling prior
strated even in pediatric patients exposed to low
to developing overt cardiac dysfunction. This held
doses of anthracycline (#100 mg/m2), manifesting as
true for indices of systolic function, diastolic func-
decreased posterior wall thickness and decreased FS
tion, and cardiac size. Even as far back as 4 years prior
compared with control subjects (15). Interestingly,
to the development of overt cardiac dysfunction,
our study did not show significant differences in
cases exhibited a decline in LV systolic function (both
posterior wall thickness z-score between cases and
FS and EF) and diastolic function (mitral E/A)
control subjects, and only the LVEDD z-score was
compared with matched control subjects. This
significantly greater in cases across multiple
trajectory has been difficult to demonstrate in the
time intervals.
past, given the relative paucity of pediatric cancer
LVEF has surpassed FS as a more reliable measure
survivors with cardiac dysfunction in individual
of LV systolic function (16). However, the detection of
centers.
abnormal FS or EF is felt to be a relatively late finding,
ECHOCARDIOGRAPHIC SCREENING FOR CARDIAC occurring after injury and pathological remodeling
DYSFUNCTION. Increased survival in pediatric can- have already occurred (17). This has led investigators
cer survivors has led to an associated increase in long- to search for other more sensitive early markers of
term cardiovascular complications such as cardiac cardiac dysfunction. However, this has perhaps
dysfunction and heart failure (13,14). Select chemo- ignored the value of studying the rate of change in
therapy agents (e.g., anthracyclines) and chest these more traditional indices over time, prior to
radiotherapy are associated with this increased risk. crossing the threshold into dysfunction. This has also
Early detection of cardiac dysfunction would poten- been difficult to document, given the relative lack of
tially provide the opportunity to intervene to affect longitudinal data available in pediatric age cancer
pathological remodeling of the heart prior to the survivors. Lipshultz et al. (9) have previously pub-
development of sustained abnormalities. This has led lished their experience among high-risk leukemia
to the development of consensus-based guidelines survivors, but among these participants, only 18
that attempt to standardize the frequency and mo- developed clinical cardiomyopathy. By evaluating
dality to best surveil for the development of cardiac serial echocardiograms in a larger group of cardio-
dysfunction (7,12). Given the safety, convenience, myopathy cases prior to the onset of LV dysfunction,
and ubiquity of echocardiography, this modality has we were able to examine this trajectory of change and
formed the backbone of this screening effort. How- identify significant changes in FS and EF relative to
ever, many of the echocardiographic indices of car- control subjects up to 4 years prior. This trajectory of
diac function (such as EF and FS) have been change raises the potential possibility of deriving a
considered binary variables (normal/not normal) in predictive equation to calculate risk of developing
F I G U R E 1 Continued
Mean estimates (red open circles ¼ cases; black filled circles ¼ control subjects) with 95% confidence intervals (red bars ¼ cases; black bars
¼ control subjects) for categorical time periods prior to cardiomyopathy diagnosis (or index time for control subjects; denoted as time “0”) of
(A) left ventricular fractional shortening (LVFS), (B) biplane Simpson’s derived ejection fraction (EF), (C) pulsed wave Doppler mitral E/A ratio,
(D) pulsed wave Doppler-derived myocardial performance index (MPI), (E) left ventricular end-diastolic dimension (LVEDD), and (F) left
ventricular wall thickness to dimension ratio. Shaded areas represent the same time period included in each category, with estimates and
bars placed on the x-axis at the median time within the given period.
C ENTR AL I LL U STRA T I O N Longitudinal Changes in Echocardiographic Functional Parameters

Excluding Data From the 2 Years Prior to Cardiomyopathy
A B
36
70
34
Biplane EF (%)
LVFS (%)
32 65
30
60
28
26 55
–9 –8 –7 –6 –5 –4 –3 –2 –1 0 –9 –8 –7 –6 –5 –4 –3 –2 –1 0
Time (Years) Prior to Index Time (Years) Prior to Index
C D
.36
2.25 .34
Mitral E/A Ratio
.32
2.0
LV MPI
.30
1.75
.28
1.5 .26
.24
1.25
–9 –8 –7 –6 –5 –4 –3 –2 –1 0 –9 –8 –7 –6 –5 –4 –3 –2 –1 0
E F
1.5
Thickness to Dimension Ratio
.16
1.0
LV EDD z-score
.15
0.5
.14
0
.13
–0.5
–9 –8 –7 –6 –5 –4 –3 –2 –1 0 –9 –8 –7 –6 –5 –4 –3 –2 –1 0
Case Control
Border, W.L. et al. J Am Coll Cardiol CardioOnc. 2020;2(1):26–37.
Continued on the next page

T A B L E 3 Estimates, Derived by LSM of Centrally Quantified Echocardiographic Parameters for Cases and Control Subjects During 3 Time
Periods: at the Time of Cancer Diagnosis; While on Cancer Therapy, and After Therapy up to and Including the Time of the Index Case
Echocardiogram *
Case Control
Echo Time LSM Estimate Echo Time LSM Estimate

n Points, n (95% CI) n Points, n (95% CI) p Value
2D-derived FS, %
At cancer diagnosis 15 15 31 (29 to 33) 11 11 35 (33 to 38) 0.010
On therapy 22 44 27 (26 to 29) 14 3 34 (32 to 35) <0.001
Post-therapy 37 121 26 (25 to 27) 50 181 31 (30 to 32) <0.001
Biplane EF, %
At cancer diagnosis 6 6 62 (57 to 66) 3 3 64 (58 to 70) 0.523
On therapy 10 13 55 (52 to 59) 7 15 64 (61 to 68) <0.001
Post-therapy 28 67 55 (53 to 56) 41 99 63 (61 to 64) <0.001
Mitral PW E/A ratio
At cancer diagnosis 13 13 1.61 (1.28 to 1.94) 9 9 2.14 (1.75 to 2.53) 0.035
On therapy 18 34 1.65 (1.42 to 1.88) 13 33 1.78 (1.53 to 2.02) 0.419
Post-therapy 37 114 1.84 (1.69 to 1.99) 50 170 2.24 (2.12 to 2.37) <0.001
PW Doppler-derived MPI
On therapy 14 26 0.33 (0.31 to 0.36) 13 32 0.29 (0.26 to 0.31) 0.006
LV end-diastolic dimension, z-score
On therapy 17 33 0.95 (0.53 to 1.38) 14 31 0.10 (0.33 to 0.53) 0.006
LV thickness-dimension ratio
On therapy 22 44 0.13 (0.12 to 0.14) 14 35 0.15 (0.14 to 0.16) 0.003
Post-therapy 37 121 0.13 (0.13 to 0.14) 50 182 0.14 (0.14 to 0.15) 0.027
*LSM, 95% confidence intervals (CIs), and p values for case-control comparisons were from longitudinal mixed models for each measure, including an interaction between case
status and time period, and using an autoregressive correlation structure between echocardiograms for the same patient. As an example, the estimated FS is 31% in cases at
cancer diagnosis as compared to 35% in controls at cancer diagnosis.
Abbreviations as in Table 2.
cardiac dysfunction in individual patients. However, impaired relaxation and/or filling (21–23). Standard
our study was not sufficiently powered to be able to methods have included evaluating LV inflow by
do this. pulsed wave (PW) Doppler and reporting the E/A ratio
ECHOCARDIOGRAPHIC MEASURES OF DIASTOLIC as a measure of relaxation. In addition, tissue Doppler
FUNCTION. Diastolic dysfunction can occur in the imaging (TDI) has been used to measure tissue ve-
setting of preserved systolic function, and hence, locities of the mitral annulus (free wall and septum),
echocardiographic guidelines have recommended and report Eʹ velocities as a measure of relaxation.
incorporating measures of diastolic function into When combined with the mitral PW Doppler as the E/
standard practice (8,18). Adult studies have shown Eʹ ratio, this has been felt to reflect LV compliance and
a high prevalence of diastolic abnormalities in be a marker of restrictive physiology. In our study,
anthracycline-exposed cancer survivors (19). Howev- when evaluated across multiple time points, only the
er, the picture has been far more mixed in pediatric mitral E/A ratio showed significant differences be-
age cancer survivors, with some studies showing tween cases and control subjects, even years before
normal diastolic function (20) and others showing cardiomyopathy diagnosis.
C E N T R A L IL L U ST R A T I O N Continued
Fitted line plots up to 2 years prior to cardiomyopathy, with 95% confidence intervals of (A) left ventricular (LV) fractional shortening, (B)
biplane Simpson’s derived ejection fraction, (C) pulsed wave Doppler mitral E/A ratio, (D) pulsed wave Doppler-derived myocardial per-
formance index, (E) LV end-diastolic dimension, and (F) left ventricular wall thickness to dimension ratio. Cardiomyopathy cases denoted by
blue line and controls denoted by the red line, both based on a linear model. Time “0” denotes time point when cardiomyopathy was
diagnosed.
ECHOCARDIOGRAPHIC MEASURES OF COMBINED monitored for the issue of left truncated data (i.e.,
SYSTOLIC-DIASTOLIC FUNCTION. MPI is a measure patients only having echocardiographic data starting
that incorporates both isovolumic contraction and after the time point at which their institution began
isovolumic relaxation, and hence, has been used as a archiving digitally). We attempted to minimize any
measure of combined systolic and diastolic function. bias that this could introduce among cases and con-
It can be measured using both PW Doppler and TDI trol subjects by matching on echocardiogram follow-
techniques. Studies in adult cancer survivors have up duration and by institution.
shown the usefulness of MPI in detecting early car- Due to the retrospective nature of this study, we
diotoxicity (24). These findings have been replicated relied on the post hoc analysis of clinically obtained
in pediatric age cancer survivors (25–27). In our series, echocardiograms for routine surveillance from mul-
differences in PW Doppler MPI and TDI MPI between tiple institutions. We found significant variability in
cases and control subjects were noted at the time of quality and completeness of the echocardiograms
cardiomyopathy diagnosis, but we only saw border- because they were not obtained with a specific
line differences in Doppler MPI in the years prior to research protocol in mind. Thus, not every echocar-
cardiomyopathy diagnosis. diographic measure could be obtained in each
ADVANCED ECHOCARDIOGRAPHIC MEASURES OF patient.
DEFORMATION. Given the potential for regional CONCLUSIONS
myocardial abnormalities in pediatric cancer survi-
vors and the limitations around the geometric and This multi-institutional study of 50 case-control
uniform contractility assumptions made by FS and pairs of pediatric cancer survivors demonstrated
EF, GLS has been recognized as a new method for that there is a measurable longitudinal decline in
early cardiotoxicity detection (28,29). In adult breast many standard echocardiographic parameters of
cancer patients, screening that includes EF and GLS cardiac function prior to crossing the traditionally
has nearly twice the predictive power for future defined threshold of LV dysfunction (FS #28% or
symptomatic disease compared with EF alone (30). EF #50%). Cases and control subjects exhibited dif-
The St. Jude Lifetime Cohort Study has provided ferences in indices of systolic function, diastolic
some of the strongest data supporting a comprehen- function, and cardiac geometry. Although power
sive approach utilizing EF, indices of diastolic func- limitations of the present sample size preclude
tion, and GLS. In a cohort of 1,820 adult cancer development of cardiomyopathy risk prediction al-
survivors, 5.8% had abnormal 3-dimensional LVEF, gorithms, these data indicate the potential role for
but in those with normal EF, 28% had abnormal GLS an expanded analysis, which involves a larger cohort
and 8.7% had diastolic dysfunction (31). In the pre- of cardiomyopathy cases to facilitate development of
sent study, we attempted to perform post hoc mea- predictive models incorporating more sensitive
surements of GLS at the core echocardiography echocardiographic indices, to identify those at risk of
laboratory. However, this requires excellent quality cardiomyopathy.
apical 2-, 3-, and 4-chamber views to allow for accu-
ACKNOWLEDGMENTS The authors would like to
rate calculation of GLS. Due to the retrospective na-
thank the following research coordinators for their
ture of our study, many of the images our study had
considerable work on this study: Cardiovascular Im-
access to (90%), especially in the earlier eras, did not
aging Research Core at Emory/Children’s Healthcare
have the required views or were not of an acceptable
of Atlanta (Heather Friedman, Nicole Krupa, Kelsey
quality, and thus we had to abandon this component
Zinck, and Cortlin Yancey); Children’s Healthcare of
of the study.
Atlanta Aflac Cancer Center (Rebecca Lewis); City of
STUDY LIMITATIONS. The study required that echo- Hope Medical Center (Lanie Lindenfeld); Fred
cardiograms were stored in DICOM format and could Hutchinson Cancer Research Center (Nancy Blythe);
be uploaded to the central core laboratory for post The Hospital for Sick Children (Emily Lam); and
hoc review and analysis. This meant that certain cases University of Minnesota Masonic Children’s Hospital
or control subjects could not be included in the study (Susan C. Anderson).
and may have biased us toward more of a contem-
porary cohort. In addition, this included less than
complete echocardiographic data at earlier time ADDRESS FOR CORRESPONDENCE: Dr. William L.
points, such as time of cancer diagnosis, which Border, Emory University School of Medicine, 1405
limited our ability to determine the significance of Clifton Road, Atlanta, Georgia 30322. E-mail:
any differences during those time periods. We also borderw@kidsheart.com. Twitter: @childrensatl.
PERSPECTIVES
COMPETENCY IN MEDICAL KNOWLEDGE: This TRANSLATIONAL OUTLOOK: Future studies should

study demonstrates that there is a measurable decline in consider an expanded analysis involving a larger cohort of
echocardiographic parameters of cardiac systolic func- pediatric cancer survivors with cardiomyopathy to facili-
tion, diastolic function, and size in pediatric cancer tate the development of predictive models, incorporating
survivors prior to the development of overt cardiac echocardiographic indices, to allow for accurate identifi-
dysfunction. Even 4 years prior to overt dysfunction, cation of those at risk prior to the development of overt
cases exhibited a decline in LV systolic function (both FS cardiac dysfunction.
and EF), diastolic function (mitral E/A), and size (LVEDD)
compared with matched control subjects.
REFERENCES
1. Robison LL, Hudson MM. Survivors of childhood 10. Hudson MM, Rai SN, Nunez C, et al. Noninva- 19. Nagy AC, Cserep Z, Tolnay E, Nagykalnai T,
and adolescent cancer: life-long risks and re- sive evaluation of late anthracycline cardiac Forster T. Early diagnosis of chemotherapy-
sponsibilities. Nat Rev Cancer 2014;14:61–70. toxicity in childhood cancer survivors. J Clin Oncol induced cardiomyopathy: a prospective tissue
2007;25:3635–43. Doppler imaging study. Pathol Oncol Res 2008;
2. Lipshultz SE, Adams MJ, Colan SD, et al. Long-
14:69–77.
term cardiovascular toxicity in children, adoles- 11. van der Pal HJ, van Dalen EC, Hauptmann M,
cents, and young adults who receive cancer ther- et al. Cardiac function in 5-year survivors of 20. Harahsheh A, Aggarwal S, Pettersen MD,
apy: pathophysiology, course, monitoring, childhood cancer: a long-term follow-up study. L’Ecuyer T. Diastolic function in anthracycline-
management, prevention, and research directions: Arch Intern Med 2010;170:1247–55. treated children. Cardiol Young 2015;25:1130–5.
a scientific statement from the American Heart
12. Children’s Oncology Group. Long-term follow- 21. Karakurt C, Kocak G, Ozgen U. Evaluation of
Association. Circulation 2013;128:1927–95.
up guidelines for survivors of childhood, adoles- the left ventricular function with tissue tracking
3. Mulrooney DA, Yeazel MW, Kawashima T, et al. cent, and young adult cancers, version 5.0. and tissue Doppler echocardiography in pediatric
Cardiac outcomes in a cohort of adult survivors of October 2018. Available at: http://www. malignancy survivors after anthracycline therapy.
childhood and adolescent cancer: retrospective survivorshipguidelines.org. Accessed February Echocardiography 2008;25:880–7.
analysis of the Childhood Cancer Survivor Study 18, 2020.
22. Santin JC, Deheinzelin D, Junior SP, Lopes LF,
cohort. BMJ 2009;339:b4606.
13. Bloom MW, Hamo CE, Cardinale D, et al. Can- de Camargo B. Late echocardiography assessment
4. Oeffinger KC, Mertens AC, Sklar CA, et al. cer therapy-related cardiac dysfunction and heart of systolic and diastolic function of the left
Chronic health conditions in adult survivors of failure: part 1: definitions, pathophysiology, risk ventricle in pediatric cancer survivors after
childhood cancer. N Engl J Med 2006;355: anthracycline therapy. J Pediatr Hematol Oncol
factors, and imaging. Circ Heart Fail 2016;9:
1572–82. 2007;29:761–5.
e002661.
5. Adams MJ, Lipshultz SE. Pathophysiology of 23. Stapleton GE, Stapleton SL, Martinez A, et al.
14. Hamo CE, Bloom MW, Cardinale D, et al.
anthracycline- and radiation-associated cardio- Evaluation of longitudinal ventricular function
Cancer therapy-related cardiac dysfunction and
myopathies: implications for screening and pre- with tissue Doppler echocardiography in children
heart failure: part 2: prevention, treatment,
vention. Pediatr Blood Cancer 2005;44:600–6. treated with anthracyclines. J Am Soc Echo-
guidelines, and future directions. Circ Heart Fail
6. Shankar SM, Marina N, Hudson MM, et al. 2016;9:e002843. cardiogr 2007;20:492–7.
Monitoring for cardiovascular disease in survivors
15. Leger K, Slone T, Lemler M, et al. Subclinical 24. Dodos F, Halbsguth T, Erdmann E, Hoppe UC.
of childhood cancer: report from the Cardiovas-
cardiotoxicity in childhood cancer survivors Usefulness of myocardial performance index and
cular Disease Task Force of the Children’s
exposed to very low dose anthracycline therapy. biochemical markers for early detection of
Oncology Group. Pediatrics 2008;121:e387–96.
Pediatr Blood Cancer 2015;62:123–7. anthracycline-induced cardiotoxicity in adults. Clin
7. Armenian SH, Hudson MM, Mulder RL, et al. Res Cardiol 2008;97:318–26.
16. Yahalom J, Portlock CS. Long-term cardiac and
Recommendations for cardiomyopathy surveil-
pulmonary complications of cancer therapy. 25. Eidem BW, Sapp BG, Suarez CR, Cetta F. Use-
lance for survivors of childhood cancer: a report
Hematol Oncol Clin North Am 2008;22:305–18. fulness of the myocardial performance index for
from the International Late Effects of Childhood
vii. early detection of anthracycline-induced cardiotox-
Cancer Guideline Harmonization Group. Lancet
icity in children. Am J Cardiol 2001;87:1120–2. A9.
Oncol 2015;16:e123–36. 17. Ewer MS, Ewer SM. Long-term cardiac safety
of dose-dense anthracycline therapy cannot be 26. Ishii M, Tsutsumi T, Himeno W, et al.
8. Plana JC, Galderisi M, Barac A, et al. Expert
predicted from early ejection fraction data. J Clin Sequential evaluation of left ventricular myocar-
consensus for multimodality imaging evaluation of
Oncol 2009;27:6073–5. dial performance in children after anthracycline
adult patients during and after cancer therapy: a
therapy. Am J Cardiol 2000;86:1279–81. A9.
report from the American Society of Echocardi- 18. Zamorano JL, Lancellotti P, Rodriguez
ography and the European Association of Cardio- 27. Ruggiero A, De Rosa G, Rizzo D, et al. Myocar-
Munoz D, et al. 2016 ESC position paper on cancer
vascular Imaging. J Am Soc Echocardiogr 2014;27: dial performance index and biochemical markers
treatments and cardiovascular toxicity developed
911–39. for early detection of doxorubicin-induced car-
under the auspices of the ESC Committee for
diotoxicity in children with acute lymphoblastic
9. Lipshultz SE, Lipsitz SR, Sallan SE, et al. Chronic Practice Guidelines: The Task Force for Cancer
leukaemia. Int J Clin Oncol 2013;18:927–33.
progressive cardiac dysfunction years after doxo- Treatments and Cardiovascular Toxicity of the
rubicin therapy for childhood acute lymphoblastic European Society of Cardiology (ESC). Eur Heart J 28. Al-Biltagi M, Abd Rab Elrasoul Tolba O, El-
leukemia. J Clin Oncol 2005;23:2629–36. 2016;37:2768–801. Shanshory MR, Abd El-Aziz El-Shitany N, El-Sayed
El-Hawary E. Strain echocardiography in early systematic review. J Am Coll Cardiol 2014;63: treatment-related cardiac dysfunction in adult
detection of Doxorubicin-induced left ventricular 2751–68. survivors of childhood cancer: results from the St.
dysfunction in children with acute lymphoblastic Jude Lifetime Cohort Study. J Am Coll Cardiol
30. Negishi K, Negishi T, Hare JL, Haluska BA,
leukemia. ISRN Pediatr 2012;2012:870549. 2015;65:2511–22.
Plana JC, Marwick TH. Independent and incre-
29. Thavendiranathan P, Poulin F, Lim KD, mental value of deformation indices for prediction
Plana JC, Woo A, Marwick TH. Use of myocar- of trastuzumab-induced cardiotoxicity. J Am Soc
dial strain imaging by echocardiography for the Echocardiogr 2013;26:493–8.
KEY WORDS cancer survivorship,
early detection of cardiotoxicity in patients 31. Armstrong GT, Joshi VM, Ness KK, et al. cardiomyopathy, echocardiography,
during and after cancer chemotherapy: a Comprehensive echocardiographic detection of longitudinal
ª 2020 PUBLISHED BY ELSEVIER ON BEHALF OF THE AMERICAN COLLEGE OF
CARDIOLOGY FOUNDATION. THIS IS AN OPEN ACCESS ARTICLE UNDER THE
CC BY-NC-ND LICENSE (http://creativecommons.org/licenses/by-nc-nd/4.0/).
EDITORIAL COMMENT
Childhood Cancer Survivors

Screening Little Hearts for Big Problems*
Ming Hui Chen, MD, MMSC
T he American Cancer Society predicts that

more than 11,000 children will be diagnosed
with cancer during 2020; fortunately, child-
hood cancer survival has improved significantly in
and help mold a model for future studies in this
growing field of pediatric cardio-oncology.
In the past several years, there has been a rapid
expansion in the field of adult cardio-oncology
the past few decades, with 84% surviving to 5 years and standardization of noninvasive imaging for sur-
(1). However, with better survival comes an increased veillance (6–9). The adult survivorship field has
burden of short- and long-term complications in defined specific recommendations for indices of car-
these childhood cancer survivors. Cardiovascular diac function in noninvasive imaging of survivors of
complications leading to heart failure are well- adult cancers, including novel indices, such as global
known in long-term survivors, particularly in those longitudinal strain for subclinical left ventricular (LV)
who received chest radiation or anthracycline chemo- dysfunction (6). Pediatric cardio-oncology is now
therapy (2–4), which have been the mainstay of anti- similarly focusing on further defining surveillance by
cancer therapy for children. However, there are noninvasive imaging and correlation with cardiotox-
significant challenges in assessing the noninvasive icity (10,11). Echocardiography is usually preferred
imaging predictors of these complications in children, over other modalities in children because of its
because there are simply fewer childhood cancer sur- noninvasive nature—requiring no needle sticks for
vivors than adult cancer survivors. Single-center children or radiation exposure to developing tissues—
studies, which form the bulk of the pediatric cardio- and its ready availability. Cardiac magnetic resonance
oncology published data, can be limited by the small has greater reproducibility in LV volume and ejection
number of childhood cancer survivors. Consequently, measurements, but is more costly, usually requires
there is an enduring difficulty in timely identification anesthesia in children, and is less accessible, and
of those children who are at greatest risk for devel- therefore is reserved for those with inadequate
oping cardiac complications and in pinpointing the echocardiographic images. Radionuclide imaging also
optimal time for cardiac intervention. In this issue allows quantification of left ventricular ejection
of JACC: CardioOncology, Border et al. (5) highlight fraction (LVEF) but requires radiation exposure and is
an innovative, collaborative, multicenter study not portable. 2-dimensional echocardiographic
approach with the use of longitudinal echocardio- assessment of fractional shortening (FS) and LVEF
graphic data in children to examine this problem have been the traditional indices used in children, but
as in adults, are subject to significant variability. The
wider clinical use of 3-dimensional echocardiography
in childhood cancer survivors promises potentially
greater accuracy and reproducibility of echocardio-
graphic indices (9).
From the Departments of Cardiology and Pediatrics, Boston Children’s SEE PAGE 26
Hospital, Boston, Massachusetts. Dr. Chen is supported by National

Cancer Institute RO1 CA196854. One of the most significant obstacles to studying
The authors attest they are in compliance with human studies commit- the importance of surveillance and cardiac effects of
cancer treatment in pediatric cancer survivors is the
appropriate. For more information, visit the JACC: CardioOncology author long lag time from treatment exposure to the devel-
instructions page. opment of overt cardiac dysfunction. Because cardiac

JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020 Chen 39
MARCH 2020:38–40 Childhood Cancer Survivors
events and cardiomyopathy typically develop months apparent disease. They compared longitudinal
or even years after curative cancer therapy, an indi- changes in echocardiographic parameters that were
vidual center will often not have a sufficiently large present in 50 children who eventually developed
sample to evaluate the relationship between echo- cardiomyopathy, with a matched cohort of 50
cardiographic indices and cardiac outcomes. The childhood cancer survivors who did not develop car-
resulting lag time may contribute to the perception by diomyopathy, after anthracycline and radiation
clinicians that few children develop cardiomyopathy exposure. These cohorts were matched for cumula-
after cancer treatment; childhood cancer survivors tive anthracycline and chest radiation dose, duration
may develop cardiomyopathy and heart failure in of follow-up, and age at cancer diagnosis. All echo-
adulthood, after they are no longer followed by their cardiograms were retrospectively analyzed by a single
pediatric providers, thereby reducing the likelihood core laboratory. The authors concluded that there
that the causal relationships will be recognized. This were significant differences in several traditional
transition also exacerbates the loss of patients to systolic and some diastolic parameters, including FS,
follow-up and, therefore, the loss of opportunity to LVEF, LV end-diastolic dimension, mitral E/A, and LV
initiate early therapy for cardiotoxicity. Few clinical myocardial performance index, between the 2 groups.
centers have integrated programs of pediatric and Intriguingly, the authors found all of these echo
adult cardiac care or research for childhood cancer indices, except myocardial performance index,
survivors. remained significant between the 2 groups, as far
The small number of patients and the multisystem back as 2 years prior to the recognition of cardiomy-
nature of potential toxicities also make it difficult to opathy. Given the era of the echocardiograms avail-
carry out longitudinal screening for adverse outcomes able, retrospective analysis of global longitudinal
in this population outside of a research setting. As a strain analysis for these patients was not possible,
result, most published clinical studies of cardiac out- and only traditional systolic and diastolic indices
comes in children, even large ones, rely on cross- could be analyzed. Also, with the overlapping range
sectional imaging data (4,12). However, longitudinal of the measurements between the 2 groups, it was not
studies provide critical insight into functional changes possible to prospectively identify individual patients
in the heart over time and the optimal timepoint at as at-risk from echocardiographic indices alone.
which a preventative intervention should be Importantly, this study underscores the importance
employed, issues that are seldom addressable using of examining longitudinal trends of systolic and dia-
cross-sectional data (3). Thus, multicenter, collabora- stolic echocardiographic indices on serial studies,
tive, longitudinal studies ultimately may be necessary instead of focusing just on the binary categorization
for the optimal study of childhood cancer survivors, of indices (e.g., LVEF, FS, LV end-diastolic dimen-
and be necessary to obtain the desired number of pa- sion) as normal or abnormal.
tients to correlate cardiac imaging indices with long- Measurements by a single observer in an echocar-
term cardiac effects. Recently, a Canadian study on diographic core laboratory reduce interobserver error
cardiac assessment of pediatric cancer survivors uti- of any measurement and, therefore, increase the po-
lized this multicenter approach to recruit a sample of wer to detect changes between groups. Unfortu-
over 500 patients (13). In addition, the National Cancer nately, because of the large range of LVEF between
Institute is funding a collaborative, multicenter na- cardiomyopathies and control subjects in the study, a
tional study with a central echo core laboratory to clinician, unlike a core laboratory, may not find that
prospectively assess childhood cancer survivors pre- FS or LVEF alone helps identify an individual patient
viously treated with high-dose anthracycline, who are at risk for ensuing cardiomyopathy. Both LVEF and FS
randomized to receive either placebo or car- are indices that have traditionally been used, but they
dioprotective therapy with carvedilol (14). This issue are very sensitive to afterload and preload changes as
of JACC: CardioOncology presents a pioneering multi- can be seen during cancer treatment (3). Therefore,
center longitudinal study in children using data pooled the ability of the physician to clinically alter therapy
by 5 participating centers, which further supports this based on a single LVEF or FS value would be limited.
model for future collaborations and efforts (5). However, analyzing and plotting trends over time in
The multicenter, retrospective study by Border an individual patient, even those with a “normal
et al. (5) published in this issue demonstrates the LVEF,” might be potentially useful, but even this
utility of serial echocardiograms, analyzed in a core approach warrants prospectively evaluation of its
laboratory, to retrospectively detect cardiomyopathy- efficacy.
related changes in children who are cancer survivors, The ability to identify the patients at risk at the
often several years before the onset of clinically time of their echocardiographic screening would be
40 Chen JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020
Childhood Cancer Survivors MARCH 2020:38–40
important because it represents the opportunity for oncology patients will increase the reproducibility
early introduction of cardioprotective agents such as and sensitivity of echocardiographic measures, and
beta-blockers or ACEI while the patient continues the incorporation of newer echocardiographic indices
anticancer therapy. However, because we know that into routine practice will potentially allow for earlier
even a single dose of anthracycline is cardiotoxic (3), detection for cardiac dysfunction in children.
consideration of cardioprotection for all patients at In the future, more prospective, multicenter
the time of anticancer therapy may be studied in a cardio-oncology studies in children will be needed to
prospective manner. assess—and ultimately, predict—cardiovascular
events. Prediction models that include patient- and
FUTURE DIRECTIONS
treatment-specific variables, noninvasive imaging
indices, and biomarkers will continue to be refined
High-quality right and left heart functional imaging,
(10,15), and help inform future recommendations for
which is easily accessible, is critical to the manage-
cardiac surveillance. In addition, greater integration
ment of pediatric cancer survivors. Therefore, routine
of cardiac care for childhood cancer survivors will be
measurement of LV volumes from 3-dimensional
crucial as these patients reach adulthood (16). In
echocardiography and the integration of novel
short, the paper by Border et al. (5) underscores the
indices, such as global longitudinal strain assess-
importance of examining trends in echocardiographic
ment, may be useful and need further investigation.
indices over time in pediatric cardiology, and exciting
Furthermore, in older children who are near adult
opportunities for multicenter collaboration and early
size, the acquisition of biplane Simpson’s LVEF, in
detection of cardiotoxicity await.
addition to the traditional 5/6 area-length method,
also may improve correlation of assessments from
pediatric and adult centers. As a child grows to adult ADDRESS FOR CORRESPONDENCE: Dr. Ming Hui
size, the configuration of the heart in the chest typi- Chen, Cardiovascular Health for Cancer Survivors Pro-
cally changes and the LV apex moves both more gram, Boston Children’s Hospital, 300 Longwood
laterally and toward the feet. Therefore, further Avenue, Boston, Massachusetts 02115. E-mail: minghui.
standardization of functional assessment in pediatric chen@cardio.chboston.org. Twitter: @BostonChildrens.
REFERENCES
1. American Cancer Society. Cancer facts & fig- adult patients during and after cancer therapy: a toxicity in childhood cancer survivors. J Clin Oncol
ures 2020. Available at: https://www.cancer.org/ report from the American Society of Echocardi- 2007;25:3635–43.
content/dam/cancer-org/research/cancer-facts- ography and the European Association of Cardio-
13. Slieker MG, Fackoury C, Slorach C, et al.
and-statistics/annual-cancer-facts-and-figures/ vascular Imaging. Eur Heart J Cardiovasc Imaging
Echocardiographic assessment of cardiac function
2020/cancer-facts-and-figures-2020.pdf. Accessed 2014;15:1063–93.
in pediatric survivors of anthracycline-treated
February 18, 2020.
8. Armenian SH, Lacchetti C, Barac A, et al. Pre- childhood cancer. Circ Cardiovasc Imaging 2019;
2. Chen MH, Colan SD, Diller L. Cardiovascular vention and monitoring of cardiac dysfunction in 12:e008869.
disease: cause of morbidity and mortality in adult survivors of adult cancers: American Society of 14. Armenian SH, Hudson MM, Chen MH, et al.
survivors of childhood cancers. Circ Res 2011;108: Clinical Oncology clinical practice guideline. J Clin Rationale and design of the Children’s Oncology
619–28. Oncol 2017;35:893–911. Group (COG) study ALTE1621: a randomized,
3. Lipshultz SE, Lipsitz SR, Sallan SE, et al. Chronic 9. Armstrong GT, Joshi VM, Ness KK, et al. placebo-controlled trial to determine if low-
progressive cardiac dysfunction years after doxo- Comprehensive echocardiographic detection of dose carvedilol can prevent anthracycline-
rubicin therapy for childhood acute lymphoblastic treatment-related cardiac dysfunction in adult related left ventricular remodeling in childhood
leukemia. J Clin Oncol 2005;23:2629–36. survivors of childhood cancer: results from the St. cancer survivors at high risk for developing
Jude Lifetime Cohort Study. J Am Coll Cardiol heart failure. BMC Cardiovasc Disord 2016;16:
4. Mulrooney DA, Armstrong GT, Huang S, et al.
2015;65:2511–22. 187.
Cardiac outcomes in adult survivors of childhood
cancer exposed to cardiotoxic therapy: a cross- 15. Chow EJ, Chen Y, Kremer LC, et al.
10. Armenian SH, Hudson MM, Mulder RL, et al. for
sectional study. Ann Intern Med 2016;164:93–101. Individual prediction of heart failure among
the International Late Effects of Childhood Cancer
Guideline Harmonization Group. Recommendations childhood cancer survivors. J Clin Oncol 2015;33:
5. Border WL, Sachdeva R, Stratton KL, et al.
for cardiomyopathy surveillance for survivors of 394–402.
Longitudinal changes in echocardiographic
parameters of cardiac function in pediatric age childhood cancer: a report from the International 16. Ryan TD, Border WL, Baker-Smith C, et al. The
childhood cancer survivors. J Am Coll Cardiol Late Effects of Childhood Cancer Guideline Harmo- landscape of cardiovascular care in pediatric can-
CardioOnc 2020;2:26–37. nization Group. Lancet Oncol 2015;16:e123–36. cer patients and survivors: a survey by the ACC
pediatric cardio-oncology work group. Cardio-
6. Liu J, Banchs J, Mousavi N, et al. Contemporary 11. Lipshultz SE, Law YM, Asante-Korang A, et al.
Oncology 2019;5:16.
role of echocardiography for clinical decision Cardiomyopathy in children: classification and
making in patients during and after cancer ther- diagnosis: a scientific statement from the American
apy. J Am Coll Cardiol Img 2018;11:1122–31. Heart Association. Circulation 2019;140:e9–68.
KEY WORDS cancer survivorship,
7. Plana JC, Galderisi M, Barac A, et al. Expert 12. Hudson MM, Rai SN, Nunez C, et al. Noninva- cardiomyopathy, children, echocardiography,
consensus for multimodality imaging evaluation of sive evaluation of late anthracycline cardiac pediatrics
ORIGINAL RESEARCH
Repeated Remote Ischemic Conditioning

Reduces Doxorubicin-Induced
Cardiotoxicity
Quan He, PHD, Fangfei Wang, MD, Thomas D. Ryan, PHD, MD, Meghana Chalasani, BS, Andrew N. Redington, MD
ABSTRACT
OBJECTIVES This study investigated the cardioprotective effect of repeated remote ischemic preconditioning (rRIC) on
doxorubicin-induced cardiotoxicity in mice.
BACKGROUND Doxorubicin is an effective chemotherapeutic agent for a wide range of tumor types but its use and
dosing are limited by acute and chronic cardiotoxicity. Remote ischemic conditioning (RIC) is cardioprotective in multiple
cardiovascular injury models, but the effectiveness of rRIC in doxorubicin-induced cardiotoxicity has not been fully
elucidated.
METHODS rRIC was performed on mice before and after doxorubicin administration. Cardiac function was assessed by
echocardiography and myocardial biology was tested by molecular approaches.
RESULTS Doxorubicin administration induced acute cardiotoxicity, as indicated by reduced cardiac function, reduced
myocyte cross-section area and increased extracellular collagen deposition, increased circulating cardiac muscle damage
markers, and decreased heart weight. Doxorubicin also adversely affected other organs, including the kidney, liver, and
spleen, as evaluated by circulating markers or organ weight loss. rRIC not only abrogated doxorubicin-induced cardio-
toxicity (left ventricular ejection fraction, doxorubicin 47.5 1.1%, doxorubicin þ rRIC 51.6 0.7%, p ¼ 0.017), but also
was associated with multiorgan protection. Within the myocardium, rRIC attenuated doxorubicin-induced cardiomyocyte
apoptosis, reduced inflammation, and increased autophagy signaling.
CONCLUSIONS rRIC may be a promising approach to reduce doxorubicin-induced cardiotoxicity.

A nthracyclines are a class of chemotherapeutic

drugs used to treat a wide range of hemato-
logical malignances and
including leukemias, lymphomas, neuroblastoma,
solid tumors,
exact antiproliferative mechanisms underlying their
clinical effectiveness remain to be determined. How-
ever, it is well described that doxorubicin leads
(either directly or via iron accumulation) to overpro-
soft tissue and bone sarcomas, breast carcinoma, duction of reactive oxygen species, which damage vi-
and ovarian carcinoma. Despite decades of use, the tal subcellular components, including DNA, protein,
From The Heart Institute, Cincinnati Children’s Hospital Medical Center, Cincinnati, Ohio. Financial support was from the Cin-
cinnati Children’s Research Fund. The authors have reported that they have no relationships relevant to the contents of this paper
to disclose.
Manuscript received December 19, 2019; revised manuscript received January 22, 2020, accepted January 28, 2020.

42 He et al. JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020
Remote Ischemic Conditioning and Doxorubicin-induced Cardiotoxicity MARCH 2020:41–52
ABBREVIATIONS and lipids, leading to cell death (1,2). reduce doxorubicin cardiotoxicity remain a clinical
AND ACRONYMS Furthermore, because of its planar structure, need.
doxorubicin intercalates between adjacent Remote ischemic conditioning (RIC), induced by
BUN = blood urea nitrogen
GC base pairs, probably due to specific repeated cyclic occlusion of blood flow and reperfu-
rRIC = repeated remote
hydrogen-bond formation between doxoru- sion of a limb, protects the heart in various animal
ischemic conditioning
bicin and guanine (3), leading to the forma- models and clinical scenarios of ischemia/reperfusion
tion of DNA adducts with resultant torsional stress injury (14), postinfarction ventricular remodeling
and nucleosome destabilization (4). This intercala- (15), and sepsis-induced cardiomyopathy (16). Besides
tion also inhibits progression of topoisomerase II the heart, RIC has been shown to be protective in
involved in DNA replication, transcription, and DNA other organs, including liver, brain, kidney, lung,
repair (5). Interfered topoisomerase II results in gastrointestinal tract, skeletal muscle, and even skin
double-strand DNA breaks and cell death. Other flaps.
studies have shown that doxorubicin increases intra- We therefore performed this proof-of-principle
cellular ceramides that arrest the cell cycle via pro- study to test the effect of repeated RIC (rRIC) on
teolytic activating transcription factor CREB3L1 doxorubicin-induced cardiotoxicity in mice. We
(cAMP response element binding protein 3-like 1) (6). additionally performed preliminary studies to assess
the potential effects of rRIC on doxorubicin-induced
SEE PAGE 53
multiorgan dysfunction.
Although the anticancer mechanisms continue to
be investigated, the antiproliferative effects of METHODS
doxorubicin clearly also target proliferative non-
cancer cells, such as stem cells in the bone marrow ANIMAL EXPERIMENT. All animal protocols were
and hair follicles, commonly leading to hematologic approved by the Institutional Animal Care and Use
side effects and hair loss during treatment. However, Committees of the Cincinnati Children’s Hospital
cardiotoxicity has emerged as one of the most con- Medical Center in accordance with the Animal Wel-
cerning adverse effects of the use of doxorubicin in fare Act and Public Health Service Policy on Humane
cancer therapy. The risk of cardiomyopathy has been Care and Use of Laboratory Animals. Mice were
noted to increase with dose, and is estimated to be housed in a fully equipped animal facility, with free
7%, 18%, to 65% with cumulative anthracycline doses access to food and water ad libitum, and on a 12-h
of 150, 350, and 550 mg/m 2, respectively (7). There is dark/light cycle. Male 9-week-old C57 mice were or-
considerable interindividual variability, however, dered from Charles River Laboratories (Wilmington,
with adverse effects at much lower doses seen in Massachusetts). Baseline echocardiography was per-
some patients. Clinically, cardiac function is usually formed after a 48-h acclimation. Animals with normal
assessed by echocardiography before and after heart function (baseline ejection fraction between
doxorubicin administration. Strain echocardiography 50% and 65%) were randomized to receive doxoru-
may be a more sensitive measure to detect early bicin 10 mg/kg (doxorubicin group, n ¼ 8) or an
cardiotoxicity of doxorubicin (8), and cardiac mag- equivalent volume of saline (vehicle group, n ¼ 8)
netic resonance imaging also can be used to monitor injected via intraperitoneal injection. RIC was per-
cardiac function. Previously, serial radionuclide im- formed 30 min before doxorubicin injection and then
aging had been used to assess left ventricular ejection repeated at the same time every day for the following
fraction, and endomyocardial biopsies were some- 5 days (doxorubicin þ rRIC group, n ¼ 8). The mice
times obtained to characterize the myocardial dam- were euthanized after echocardiography on day 6.
age with anthracyclines (9,10). Heparinized blood was collected from the inferior
Several approaches to mitigate cardiotoxicity have caval vein, and subsequently plasma was isolated by
been suggested, including: 1) limiting cumulative centrifugation, aliquoted, snap frozen in liquid ni-
dose; 2) extended infusion time to reduce peak blood trogen, and kept at 80 C for biomarker and inflam-
levels; 3) liposomal doxorubicin to increase tumor matory factor assays. Vital organs, including heart,
tissue distribution (11); 4) concomitant treatment liver, kidney, and spleen, were harvested and
with dexrazoxane (12), the only drug approved spe- weighed. The heart was sliced into 3 pieces trans-
cifically to prevent doxorubicin-induced cardiotox- versely after removal of the atriums, and the middle
icity; and 5) primary or secondary prevention of section was embedded in optimal cutting tempera-
worsening cardiomyopathy with neurohormonal an- ture compound for histology, apoptosis, and other
tagonists. The merits of these approaches have been immune stains. The remainder was homogenized and
reviewed elsewhere (13), but improved approaches to prepared for protein analysis.
JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020 He et al. 43
MARCH 2020:41–52 Remote Ischemic Conditioning and Doxorubicin-induced Cardiotoxicity
rRIC. We used a protocol of 4 cycles of 5 min of hin- 40 ,6-diamidino-2-phenylindole under 20 objective
dlimb ischemia, induced by tourniquet, followed by using Nikon (Tokyo, Japan) Eclipse confocal micro-
5 min of reperfusion using the method we have scopy equipped with NIS-Elements software. The 3
described before (17). Ischemia was confirmed by color images were merged together and apoptotic
pallor and cyanosis of the lower limb below the nuclei expressed as number per field.
tourniquet. As noted previously, this was performed
MYOCARDIAL FIBROSIS. Frozen section slides were
30 min before doxorubicin injection and then
air dried for 30 min, then fixed in Bonin’s solution,
repeated at the same time every day for 5 days.
and stained with Pico-Sirius for 60 min. The slides
WESTERN BLOT. Heart tissue was homogenized in a were washed with acid water and mounted with a
lysis buffer containing 150 mM NaCl, 50 mM Tris$Cl resinous medium after dehydration and cleared with
(pH 7.5), 0.5% deoxycholate, 0.1% sodium dodecyl xylene. Five images per slide were obtained
sulfate, 1% nonidet P-40, PhosSTOP, and Complete under 20 objective and collagen area was analyzed
Mini. Protein concentration was determined using a with Image J software, as previously described in
Coomassie protein assay kit and bovine serum albu- detail (18).
min was used as the standard. Cardiac protein levels
BIOMARKER ASSAYS. Circulating biomarkers of or-
were determined by Western blot analysis (mini-gel,
gan injury were assayed with the following commer-
100 m g loading weight) as described previously (18)
cially available kits: uric acid with a Uric Acid
and glyceraldehyde-3-phosphate dehydrogenase
Colorimetric/Fluorometric Assay kit; aspartate
(GAPDH) was used as the loading control.
aminotransferase activity with a Colorimetric Assay kit
ECHOCARDIOGRAPHY. Cardiac function was evalu- from Biovision (Milpitas, California); Creatine kinase
ated by echocardiography by using the Vevo2100 with an EnzyChrom Creatine Kinase Assay kit from
mouse echocardiography system equipped with a 40- BioAssay System (Hayward, California); and troponin I
MHz high-frequency transducer. Studies were per- with a Mouse High Sensitivity Cardiac Troponin I Elisa
formed under anesthesia induced by isoflurane kit from MyBioSource (San Diego, California).
inhalation. Anesthetic depth was indicated by heart
CIRCULATING INFLAMMATORY MARKERS. The con-
rate, and all images were obtained between 400 and
centration of inflammatory factors in plasma was
600 beats/min. Measurements and calculations were
determined with a Mouse Inflammation Array Q1,
made offline by investigators blinded to treatments.
sandwich enzyme-linked immunosorbent assay–
CARDIOMYOCYTE CROSS-SECTIONAL AREA. After based quantitative array platform from RayBiotech
air drying and fixation in ice-cold acetone, the slides Life (Peachtree Corners, Georgia) following the man-
were treated with 3.3 U/ml neuraminidase type v for ufacturer’s directions. The array detects 40 inflam-
60 min at room temperature. The slides then were matory factors and each antibody is arrayed on
incubated with fluorescein-labeled peanut agglutinin standard glass slide in quadruplicate.
and rhodamine-labeled Griffonia simplicifolia lectin I
STATISTICAL ANALYSIS. Data are expressed as the
for 2 h at room temperature. The slides were covered
mean SE unless otherwise noted. Differences be-
with coverslip in presence of antifade fluorescent
tween 2 groups were analyzed by a 2-tailed t-test,
mounting medium and sealed with nail polish oil.
and multiple groups were compared using 1-way
Four radially oriented microscope fields were
analysis of variance and post hoc Tukey’s or
selected from each section and photographed under
Bonferroni-Holm correction. A value of p < 0.05 was
a 40 objective. Myocyte cross-section area was
regarded as a significant. Analysis was carried out
measured with Image J (National Institutes of Health,
with SAS software version 9.4 (SAS Institute, Cary,
Bethesda, Maryland) and 4 images from the same
North Carolina) and Excel 2016 (Microsoft,
sample were averaged.
Redmond, Washington).
APOPTOSIS. Frozen sections were air dried for 30
min, then fixed in ice-cold acetone for 10 min. The RESULTS
slides were immune-stained with a -smooth muscle
actin antibody and a Texas-red conjugated second rRIC REDUCES THE INFLAMMATORY RESPONSE TO
antibody. Apoptotic nuclei were stained with a TACS DOXORUBICIN THERAPY. Doxorubicin is well known
TDT-Fluor in situ apoptosis detection kit from to induce a systemic inflammatory reaction (19),
RnDSystem. The total nuclei count was made after which, in turn, can be associated with multiorgan
staining with 4 0 ,6-diamidino-2-phenylindole. Five dysfunction. Previous data from our laboratory
images were obtained with 3 filters: green for showed that RIC improved lipopolysaccharide-
apoptosis, red for a -smooth muscle actin, and blue for induced cardiomyopathy and reduces systemic
inflammatory responses in a mouse sepsis model (16).

T A B L E 1 rRIC Globally Antagonized Doxorubicin-induced Alteration of
Inflammatory Factors
As expected, doxorubicin induced a systemic in-
flammatory reaction in our animals, with 39 of 40 of
Vehicle Dox Dox/rRIC
(n ¼ 8) (n ¼ 8) (n ¼ 8) p Value
the inflammatory markers showing proinflammatory
BLC 23.9 (15.2–28.1) 52.1 (34.8–108)* 47.6 (33.9–60.4)* 0.004 directional change (12 of them reaching statistical
CD30L 0.39 (0–1.6) 0.68 (0–3.0) 0.22 (0–1.7) 0.74 significance). rRIC reversed these responses in
Eotaxin 601 22.6 692 17.7 746 38.2† 0.009 most of the inflammatory factors, although only
Eotaxin-2 89.6 (50–127) 132 (71.5–512) 129 (118–136) 0.28 macrophage inflammatory protein-1 g and tumor ne-
Fas L 1.9 (0.28–3.6) 3.5 (1.2–9.1) 1.8 (0–5.3) 0.40 crosis factor receptor-1 reached statistical signifi-
G-CSF 29.4 (17.0–74.0) 56.1 (18.7–131) 29.1 (10.4–114) 0.69
cance, when compared with doxorubicin alone
GM-CSF 3.0 0.46 3.5 1.1 2.4 0.74 0.51
(Table 1).
ICAM-1 959 144 1436 40.9† 1425 114* 0.016
IFNg 8.1 (5.6–9.7) 4.6 (3.3–6.8) 5.7 (3.2–9.1) 0.26
rRIC PRESERVES ORGAN WEIGHT. We found that
IL-1a 1.4 (0.27–2.5) 1.4 (0.76–2.7) 1.7 (1.4–6.4) 0.39
rRIC had little effect on doxorubicin-induced body
IL-1b 13.8 (6.2–16.1) 6.4 (4.7–12.3) 11.2 (5.2–17.5) 0.68
IL-2 5.1 (4.1–8.9) 4.9 (1.0–7.4) 4.5 (2.5–8.3) 0.72 weight loss (Figure 1A), but significantly attenuated
IL-3 0.32 (0.19–0.50) 0.16 (0.02–0.25) 0.02 (0–0.16)* 0.024 the doxorubicin-induced decrease in heart weight
IL-4 1.8 0.20 2.2 0.38 1.6 0.43 0.85 corrected for body weight (Figure 1B). Although
IL-5 5.4 (0–12) 0.22 (0–7.62) 1.98 (0–10.8) 0.78 doxorubicin reduced kidney weight, it failed to reach
IL-6 3.6 (2.6–6.8) 4.6 (3.4–14.9) 3.5 (0.68–6.68) 0.42 statistical significance compared with vehicle; how-
IL-7 0 (0–0) 5.1 (0–57.2)* 0 (0–1.35) 0.050
ever, kidney weight was significantly improved by
IL-10 13.2 (7.5–21.9) 7.3 (2.4–13.3) 12.8 (7.7–18.6) 0.34
rRIC when the doxorubicin and doxorubicin þ rRIC
IL-12p70 5.7 (2.6–7.0) 2.4 (0–4.1) 1.9 (0–7.4) 0.14
IL-13 3.6 (1.4–10.8) 10.5 (0–19.9) 11.7 (2.5–14.5) 0.67
groups were compared (Figure 1C). The significant
IL-15 0 (0–2.9) 0 (0–13.1) 0.78 (0–12.7) 0.47 effect of doxorubicin on spleen weight was also
IL-17 2.6 (2.0–4.1) 5.5 (3.7–14.6) 3.8 (3.3–4.6) 0.08 partially abrogated by rRIC (Figure 1D), but the liver
IL-21 0 (0–0.65) 1.3 (0.56–3.4) 0.43 (0–2.1) 0.07 was unaffected by doxorubicin and, as such, it was
KC 1.9 (1.6–5.4) 3.4 (2.4–4.6) 2.6 (2.0–5.0) 0.55 unchanged in the doxorubicin þ rRIC group (Figure 1E).
Leptin 283 51.6 173 47.5 269 41.0 0.14
LIX 67.5 (36.5–88.9) 57 (49.1–74.4) 69.4 (49.1–112.0) 0.80 rRIC IMPROVES CIRCULATING BIOMARKER
MCP-1 4.9 (3.5–8.4) 9.6 (6.9–13.4) 7.6 (6.1–9.2) 0.18
EVIDENCE OF MULTIORGAN DYSFUNCTION. We
MCP-5 8.1 (5.4–9.5) 30.3 (17.5–63.0)† 21.3 (15.6–24.9)† 0.001
next examined circulating markers of organ dysfunc-
MCSF 0.38 (0.24–0.65) 0.37 (0.31–0.63) 0.66 (0.47–1.00) 0.28
tion. These data are shown in Figure 2. As a marker of
MIG 348 93.8 273.3 45.4 278 55.6 0.82
MIP-1a 2.6 0.22 2.0 0.16 3.3 0.48 ‡
0.023 predominantly skeletal muscle damage, creatinine
MIP-1g 1,200 (1,120–1,250) 1,060 (985–1,130) 1,050 (946–1,160) 0.050 kinase levels after doxorubicin treatment were
PF4 3,436 105 3,212 211 3,360 170 0.64 increased up to 4-fold compared with vehicle, and this
RANTES 2.9 (2.6–3.2) 3.2 (2.7–3.4) 2.6 (0.67–3.0) 0.40 increase was completely abrogated by rRIC (Figure 2A).
TARC 4.8 (3.5–6.0) 7.9 (6.7–8.7)* 5.7 (5.1–7.7) 0.007 Cardiac troponin I release showed a similar profile to
TCA-3 10.8 0.36 21.2 3.35* 15.5 2.6 0.015
that of creatinine kinase. Doxorubicin led to signifi-
TIMP-1 365 (301–430) 1,130 (923–1,260)† 828 (786–1,100)† <0.001
cantly increased troponin I levels and, again, rRIC
TNF-a 3.8 (1.5–4.8) 3.7 (1.8–5.4) 2.1 (1.5–5.3) 0.95
TNF RI 998 32.4 1,366 46.2† 1,112 74.7 ‡
0.005
completely abrogated this effect (Figure 2B). Doxoru-
TNF RII 69.3 (58.4–81.0) 115 (102–169)* 144 (94.3–186.0)† 0.009 bicin also significantly increased circulating aspartate
aminotransferase, and rRIC restored levels to near
Values are median (interquartile range [Q1 and Q3]) or mean SE. Multiplex inflammatory array quantified 40 normal (Figure 2C). The changes in blood urea nitrogen
inflammatory factors, and in 28 factors rRIC abrogated doxorubicin-induced alteration. Concentrations in pg/ml.
*p < 0.05. †p < 0.01 versus vehicle. ‡p < 0.05 versus doxorubicin alone. (BUN) (Figure 2D) were perhaps unexpected. Although
BLC ¼ B lymphocyte chemoattractant (CXCL13); CD30L ¼ tumor necrosis factor ligand superfamily member 8 there was no difference between controls and the
(TNFSF8, CD30 ligand); Fas L ¼ Fas ligand (CD95L, CD178); G-CSF ¼ granulocyte colony-stimulating factor;
GM-CSF ¼ granulocyte-macrophage colony-stimulating factor; ICAM ¼ intercellular adhesion molecule (CD54); doxorubicin þ rRIC groups, BUN in the doxorubicin
IFN ¼ interferon; IL ¼ interleukin; KC ¼ keratinocyte-derived chemokine (C-X-C motif chemokine receptor 1,
CXCL1); LIX ¼ lipopolysaccharide-induced CXC chemokine; MCP ¼ monocyte chemoattractant protein (CCL2);
treated animals fell. We speculate that this may relate
MCSF ¼ macrophage colony-stimulating factor; MIG ¼ monokine induced by gamma (CXCL9); MIP ¼ macrophage to a decrease in skeletal muscle mass evidenced by the
inflammatory protein (CCL3); PF ¼ platelet factor (CXCL4); RANTES ¼ regulated upon activation, normal T cell
expressed and secreted (CCL5); TARC ¼ thymus and activation regulated chemokine (CCL17); TCA-3 ¼ CCL1, raised creatine kinase levels, and plasma BUN being
I-309; TIMP ¼ tissue inhibitor matrix metalloproteinase; TNF ¼ tumor necrosis factor; TNFR ¼ tumor necrosis closely related to muscle mass in the absence of renal
factor receptor.
dysfunction. Overall, these data suggest that besides
cardiotoxicity, doxorubicin also has detrimental
F I G U R E 1 rRIC Reduced Doxorubicin-Induced Multiorgan Weight Loss
At end of the experiment, body weight of the animals was measured (A), organs were harvested and weighed, and corrected by body weight.
The tested organs include heart (B), kidney (C), spleen (D), and liver (E). *p < 0.05 and **p < 0.01 versus vehicle, whereas †p < 0.05 and
††p < 0.01 versus doxorubicin alone. n ¼ 8 in each of the groups (vehicle, doxorubicin, doxorubicin þ rRIC). BW ¼ body weight; Dox
¼ doxorubicin; Dox/RIC ¼ doxorubicin þ remote ischemic conditioning; HW ¼ heart weight; KW ¼ kidney weight; LW ¼ liver weight; rRIC ¼
repeated remote ischemic conditioning; SW ¼ spleen weight; Veh ¼ vehicle.
effects on skeletal muscle, liver, and spleen, which (doxorubicin 47.5 1.1%, doxorubicin þ rRIC 51.6
was reduced by rRIC. 0.7%, p ¼ 0.017) and fractional shortening (vehicle
27.9 0.7, doxorubicin 23.6 0.7, doxorubicin þ rRIC
rRIC PRESERVES CARDIAC FUNCTION. High-dose 26.0 0.5, p < 0.02) at 5 days (Figure 3).
doxorubicin (10 mg/kg) decreased cardiac function
as evaluated by echocardiography. Both ejection
fraction and fractional shortening were significantly rRIC REDUCES MYOCARDIAL FIBROSIS. Myocyte
reduced compared with vehicle-treated animals. rRIC cross-sectional area was reduced by doxorubicin,
significantly abrogated this decline in cardiac func- and this reduction was significantly reversed by rRIC
tion, as indicated by improved ejection fraction (Figure 4). Doxorubicin also led to a 4-fold increase in
F I G U R E 2 rRIC Inhibited Doxorubicin-Induced Alteration of Circulating Markers of Multiorgan Dysfunction
Circulating multiorgan markers were tested with commercially available assay kits. The tested markers include creatinine kinase (CK) (A),
troponin I (B), aspartate aminotransferase (AST) (C), and blood urea nitrogen (BUN) (D). *p < 0.05 and **p < 0.01 versus vehicle, whereas
†p < 0.05 versus doxorubicin alone. n ¼ 8 in each of the groups (vehicle, doxorubicin, doxorubicin þ rRIC). Abbreviations as in Figure 1.
F I G U R E 3 rRIC Abrogated Doxorubicin-Induced Cardiac Function Decline
Echocardiography was performed under isoflurane anesthesia. Representative m-mode 2-dimensional images (A). Ejection fraction (EF)
(B) and fraction shortening (FS) (C). **p < 0.01 versus vehicle, †p< 0.05 versus doxorubicin alone. n ¼ 8 in each of the groups (vehicle,
doxorubicin, doxorubicin þ rRIC). Abbreviations as in Figure 1.
F I G U R E 4 rRIC Protected Against Doxorubicin-induced Cardiotoxicity at the Tissue Level
Extracellular collagen was stained with Sirius Red (A), quantified with Image J, and expressed as fold change versus vehicle (C). Represen-
tative picture for myocyte cross-section area (MCSA) (B) and quantified in (D). *p < 0.05 versus vehicle, whereas †p < 0.05 versus
doxorubicin alone. Scale bar in A is 20 mm and 25 mm in B. n ¼ 8 in each of the groups (vehicle, doxorubicin, doxorubicin þ rRIC). Abbreviations
as in Figure 1.
collagen content in the extracellular matrix, and this We therefore examined autophagy signaling in our
was significantly reduced by rRIC. animals. As shown in Figure 6, doxorubicin had mini-
mal effect on autophagy proteins LC3 and p62/
rRIC REDUCES DOXORUBICIN-INDUCED MYOCYTE SQSTM1, but rRIC significantly increased their
APOPTOSIS. Figure 5 shows that doxorubicin led expression. The other measured autophagy protein
to a marked increase in cardiac myocyte apoptosis Atg5 was not affected by doxorubicin or concurrent
when compared with vehicle. This was significantly rRIC. This enhanced autophagy signaling may
reduced by rRIC (Figures 5A and 5B). Western contribute to the beneficial effects of rRIC on
blot analysis showed that rRIC also blocked doxorubicin-induced cardiotoxicity and warrants
doxorubicin-induced Bax overexpression (Figure 5C), future study.
but Bcl2 was not affected (Figure 5D), suggesting
that modification of Bax signaling is, at least in
part, involved in the beneficial effect of rRIC on DISCUSSION
doxorubicin-induced cardiomyocyte apoptosis.
This study demonstrates, for the first time, that rRIC
rRIC INCREASES MYOCARDIAL PRO-AUTOPHAGY ameliorates acute doxorubicin-induced cardiotoxicity
SIGNALING. Autophagy flux has been shown in vivo. rRIC not only improves cardiac function,
impaired by doxorubicin (20,21), and previous studies but our data also suggest multiorgan protection.
from our laboratory and others showed that myocar- In terms of the cardioprotection, rRIC: 1) reduced
dial autophagy signaling is enhanced by RIC (22–24). doxorubicin-induced troponin I elevations,
F I G U R E 5 rRIC Abrogated Doxorubicin-Induced Cardiomyocyte Apoptosis
Apoptosis was assayed by TUNEL (terminal deoxynucleotidyl transferase dUTP nick end labeling) stain (A) and quantified (B) by Image J expressed as apoptotic nuclei
per image. Apoptotic protein Bax (C) and Bcl2 (D) were quantified by Western blot and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as loading
control. *p < 0.05 and **p < 0.01 versus vehicle, whereas †p < 0.05 versus doxorubicin alone. Scale bar is 25 mm. n ¼ 8 in each of the groups (vehicle, doxorubicin,
doxorubicin þ rRIC). Abbreviations as in Figure 1.
suggesting reduced myocardial damage; 2) improved of injury that do not involve ischemia-reperfusion
cardiac function evaluated by echocardiography; 3) injury, such as contrast-induced nephropathy
modified myocardial structural responses as indi- (28,29) and sepsis-induced cardiomyopathy (16,30).
cated by improved cardiomyocyte cross-sectional The mechanisms underlying these latter effects also
area and reduced fibrosis; and 4) increased cardiac remain to be fully elucidated, but almost certainly
pro-autophagy signaling and reduced cardiomyocyte involve reduction of oxidative stress and inflamma-
apoptosis (Central Illustration). tion (30,31).
As discussed earlier, the mechanisms of the Although it is difficult to provide a comprehensive
beneficially cytotoxic effects of doxorubicin on mechanistic explanation for the benefits of rRIC in
cancer cells continues to be explored and debated doxorubicin-induced cardiotoxicity, our data suggest
(25–27). Similarly, the exact mechanisms of the car- substantial benefits that, given the potential clinical
dioprotective effects of RIC remain to be fully impact, would warrant future investigation. One of
delineated. That said, it is clear that the effects of the most notable effects of rRIC was the complete
RIC go beyond simple induction of ischemia resis- abrogation of the 4-fold increase in cardiomyocyte
tance, or modification of reperfusion injury, apoptosis induced by doxorubicin. Doxorubicin-
as it has been shown to be effective in other models induced apoptosis has been well described in
both cardiac myocytes (32) and cancer cells (33). Our

F I G U R E 6 rRIC Increased Cardiac Autophagy Proteins
data showed that doxorubicin significantly induced
Bax, the earliest identified pro-apoptotic member of
the Bcl-2 protein family. This increase in Bax
was rescued completely by rRIC and consequently
the number of apoptotic nuclei was markedly
reduced by rRIC. It is worth noting, however, that
rRIC only partially reduced the number of
doxorubicin-induced apoptotic nuclei, despite
rescuing Bax, suggesting Bax is only one of several
apoptotic mechanisms at play. Interestingly, neither
doxorubicin nor rRIC had an effect on Bcl2 expression,
but the effect of rRIC to reduce apoptosis in the
absence of changes in Bcl2 signaling is compatible
with previous observations of the effect of RIC
to reduce apoptosis after ischemia-reperfusion
injury (34).
Similarly, the current data are consistent with our
previous studies showing that rRIC enhances auto-
phagy signaling (14,22). In this study, we showed that
rRIC enhanced autophagy protein expression
including LC3 I/II and p62/SGSTM1. Autophagy is a
highly regulated process of cell death. It is a multistep
process including recruitment to the autophagosomal
membrane of p62/SQSTM1 ubiquitinated denatured
protein, LC3 II, and other autophagy proteins.
Elevation of LC3 and p62/SQSTM1 in general indicates
increased autophagosome formation, suggesting
ether increased autophagy activity or blockade of
autophagosome clearance. The process of autophagy
can be inhibited at multiple steps and leads to
cell stress and death. Doxorubicin has been
showed to interfere with the late steps of autophagy
leading to transient autophagosome accumulation,
which returns to normal levels by 7 days after
doxorubicin injection (20,35). Our data, showing no
residual autophagy activation 6 days after doxoru-
bicin injection, are compatible with these prior
observations; however, continued RIC leads to
persistent pro-autophagy signaling, perhaps contrib-
uting to the cardioprotection we observed. The
involvement of apoptosis and autophagy in rRIC-
diminished doxorubicin-induced cardiotoxicity is
consistent with the recent report by Gertz et al. (36).
In their study, a higher dose of doxorubicin was
used (20 mg/kg) and only 1 episode of RIC, consisting
of 3 cycles of 5-min occlusion and 5-min reperfusion,
was used, 1 h before therapy. Their data showed that
this RIC abrogated doxorubicin-induced apoptosis,
Autophagy proteins were quantified by Western blot analysis as described
and enhanced autophagy signaling (36). Interest-
under Methods. Cardiac LC3 (A), p62/SQSTM1 (B), and Atg5 (C) were
ingly, in contrast to our findings, despite the higher quantified and expressed as the ratio over loading control glyceraldehyde-3-
dose of doxorubicin, there was no significant change phosphate dehydrogenase (GAPDH). *p < 0.05 versus vehicle and ††p <
in cardiac function in their animals, although long- 0.01 versus doxorubicin alone. n ¼ 8 in each of the groups (vehicle,
term survival was enhanced, perhaps reflecting doxorubicin, doxorubicin þ rRIC). Abbreviations as in Figure 1.
C E N T R A L IL L U ST R A T I O N Effects of Remote Ischemic Conditioning on

Doxorubicin-Induced Cardiotoxicity
He, Q. et al. J Am Coll Cardiol CardioOnc. 2020;2(1):41–52.
Injection of doxorubicin caused decline of cardiac function, increased cardiac fibrosis, decreased myocyte cross-section area, and increased
cardiomyocyte apoptosis. Remote ischemic conditioning abrogated doxorubicin-induced cardiotoxicity. rRIC ¼ repeated remote ischemic
conditioning.
some of the multiorgan benefits that we additionally autophagocytic and apoptotic cell death we observed
showed in our study. must remain speculative. Cardiac dysfunction can be
The relationship between the modification of sys- induced as part of the systemic inflammatory
temic inflammatory responses and the reduction in response, and we have previously shown that
mortality associated with this “sepsis-induced” car- concern of the effect of rRIC on the beneficial
diomyopathy can be reduced by rRIC (16). Sepsis- cytotoxicity that underlies its benefits as a cancer
induced cardiomyopathy tends to be transient, how- therapeutic. Further studies of this potential adverse
ever, and sustained myocardial structural changes are effect of rRIC in a tumor-bearing model are clearly
rare, unlike what is seen with doxorubicin-induced needed, and careful attention is also needed to un-
cardiotoxicity. Conversely, widespread cell death derstand the potential effects of rRIC on tumor effi-
from any cause can induce a systemic inflammatory cacy in any early-phase clinical trials. Future
reaction, and it may be reasonable to suggest that experimental studies also should be designed to
our observations are more consistent with rRIC, elucidate the mechanisms of rRIC in preventing
primarily affording protection against cytotoxic cell doxorubicin cardiotoxicity, as they may identify new
death, with resulting reduced systemic inflammatory targets for the development of specific therapeutic
responses, rather than vice versa. agents.
Finally, we showed that rRIC not only provided
myocardial protection, but also positively influenced CONCLUSIONS
other organs, in terms of preventing reductions in
size and improving some aspects of their function. These data show that rRIC reduces doxorubicin car-
The underlying mechanisms of, and potential clinical diotoxicity in a mouse model of acute injury. These
impact of, these “off-target” benefits remain to be preliminary data suggest that further mechanistic
seen, but again may be important areas for future studies of the acute and chronic effects of rRIC are
investigation. warranted.
STUDY LIMITATIONS. Given the lack of prior studies
ACKNOWLEDGMENTS The authors acknowledge
regarding the potential role of this strategy of rRIC in
Victoria Moore and Chrissy Schulte from the Cardio-
doxorubicin-induced cardiotoxicity, this “proof-of-
vascular Imaging Core Research Laboratory of Cin-
principle” study has necessary limitations. We
cinnati Children’s Hospital for their support in
only measured the effects of rRIC at a single
echocardiography.
time point, 6 days after administration. We
therefore have no data regarding any effect on early
responses, and no longer-term data showing that the ADDRESS FOR CORRESPONDENCE: Dr. Andrew N.
improvements we observed are sustained beyond the Redington, Cincinnati Children’s Hospital Medical
period of rRIC administration. The improved struc- Center, 3333 Burnet Avenue, Cincinnati, Ohio 45229.
tural integrity of the myocardium, with improved E-mail: andrew.redington@cchmc.org. Twitter:
cardiomyocyte cross-sectional area, and reduced @DrThomasRyan.
fibrosis, are encouraging, but clearly further
studies, focused on improving our mechanistic un- PERSPECTIVES
derstanding, and confirming long-term benefit,
are required.
COMPETENCY IN MEDICAL KNOWLEDGE: Dose-dependent
Given that rRIC can be induced noninvasively
cardiotoxicity of doxorubicin limits its use in patients with cancer.
using a blood pressure cuff, its translation into
The mechanisms of doxorubicin cardiotoxicity are not fully un-
clinical trials has been quite rapid, but the impact
derstood and effective approaches to reduce cardiotoxicity are
despite encouraging preclinical and proof-of-
an important clinical need. Our data in mice show rRIC is an
principle clinical trials has been disappointing in
effective noninvasive strategy to protect the heart and other
larger studies (37). Multiple factors for lack of
organs from doxorubicin toxicity.
clinical effect have been postulated, including older
patient age, the effect of comorbidities, and the
TRANSLATIONAL OUTLOOK: We demonstrated that rRIC
concomitant effect of drugs that may inhibit the
protects doxorubicin-induced cardiotoxicity in a mouse model.
rRIC effect or themselves induce a “conditioned”
Further studies of the acute and chronic effects of rRIC, and
state (38–40). These factors may be less of a
confirmation that the anticancer effects of doxorubicin are un-
problem in the sometimes younger population of
affected by rRIC, are necessary prior to clinical translation.
patients in whom doxorubicin is used for chemo-
therapy, but our data cannot address a potential
REFERENCES
1. Ichikawa Y, Ghanefar M, Bayeva M, et al. Car- elevation myocardial infarction. J Am Coll Cardiol 29. Er F, Nia AM, Dopp H, et al. Ischemic pre-
diotoxicity of doxorubicin is mediated through Intv 2015;8:178–88. conditioning for prevention of contrast medium-
mitochondrial iron accumulation. J Clin Invest induced nephropathy: randomized pilot RenPro
16. Honda T, He Q, Wang F, Redington AN. Acute
2014;124:617–30. Trial (Renal Protection Trial). Circulation 2012;126:
and chronic remote ischemic conditioning atten-
296–303.
2. Kim SY, Kim SJ, Kim BJ, et al. Doxorubicin- uate septic cardiomyopathy, improve cardiac
induced reactive oxygen species generation and output, protect systemic organs, and improve 30. Kim YH, Yoon DW, Kim JH, Lee JH, Lim CH.
intracellular Ca2þ increase are reciprocally mortality in a lipopolysaccharide-induced sepsis Effect of remote ischemic post-conditioning on
modulated in rat cardiomyocytes. Exp Mol Med model. Basic Res Cardiol 2019;114:15. systemic inflammatory response and survival
2006;38:535–45. rate in lipopolysaccharide-induced systemic
17. Kharbanda RK, Mortensen UM, White PA, et al.
inflammation model. J Inflamm (Lond) 2014;11:
3. Jawad B, Poudel L, Podgornik R, Steinmetz NF, Transient limb ischemia induces remote ischemic
16.
Ching WY. Molecular mechanism and binding free preconditioning in vivo. Circulation 2002;106:
energy of doxorubicin intercalation in DNA. Phys 2881–3. 31. Lotfollahi H, Mohammadi M, Ghaffari S, et al.
Chem Chem Phys 2019;21:3877–93. Effect of remote ischemic post-conditioning on
18. He Q, Wang F, Honda T, James J, Li J,
oxidative stress in blood of STEMI patients treated
4. Yang F, Teves SS, Kemp CJ, Henikoff S. Doxo- Redington A. Loss of miR-144 signaling interrupts
with primary angioplasty. J Cardiovasc Thorac Res
rubicin, DNA torsion, and chromatin dynamics. extracellular matrix remodeling after myocardial
2016;8:113–8.
Biochim Biophys Acta 2014;1845:84–9. infarction leading to worsened cardiac function.
Sci Rep 2018;8:16886. 32. Zhao L, Zhang B. Doxorubicin induces car-
5. Lyu YL, Kerrigan JE, Lin CP, et al. Topoisomer-
diotoxicity through upregulation of death re-
ase IIbeta mediated DNA double-strand breaks: 19. Wang L, Chen Q, Qi H, et al. Doxorubicin-
ceptors mediated apoptosis in cardiomyocytes. Sci
implications in doxorubicin cardiotoxicity and induced systemic inflammation is driven by upre-
Rep 2017;7:44735.
prevention by dexrazoxane. Cancer Res 2007;67: gulation of toll-like receptor TLR4 and endotoxin
8839–46. leakage. Cancer Res 2016;76:6631–42. 33. Pilco-Ferreto N, Calaf GM. Influence of
doxorubicin on apoptosis and oxidative stress in
6. Denard B, Lee C, Ye J. Doxorubicin blocks pro- 20. Li DL, Wang ZV, Ding G, et al. Doxorubicin
breast cancer cell lines. Int J Oncol 2016;49:
liferation of cancer cells through proteolytic acti- blocks cardiomyocyte autophagic flux by inhibiting
753–62.
vation of CREB3L1. Elife 2012;1:e00090. lysosome acidification. Circulation 2016;133:
1668–87. 34. Nakamura M, Wang NP, Zhao ZQ, et al. Pre-
7. Swain SM, Whaley FS, Ewer MS. Congestive
conditioning decreases Bax expression, PMN
heart failure in patients treated with doxorubicin: 21. Abdullah CS, Alam S, Aishwarya R, et al.
accumulation and apoptosis in reperfused rat
a retrospective analysis of three trials. Cancer Doxorubicin-induced cardiomyopathy associated
heart. Cardiovasc Res 2000;45:661–70.
2003;97:2869–79. with inhibition of autophagic degradation process
and defects in mitochondrial respiration. Sci Rep 35. Shi J, Surma M, Wei L. Disruption of ROCK1
8. Charbonnel C, Convers-Domart R, Rigaudeau S,
2019;9:2002. gene restores autophagic flux and mitigates
et al. Assessment of global longitudinal strain at
doxorubicin-induced cardiotoxicity. Oncotarget
low-dose anthracycline-based chemotherapy, for 22. Rohailla S, Clarizia N, Sourour M, et al. Acute,
2018;9:12995–3008.
the prediction of subsequent cardiotoxicity. Eur delayed and chronic remote ischemic conditioning
Heart J Cardiovasc Imaging 2017;18:392–401. is associated with downregulation of mTOR and 36. Gertz ZM, Cain C, Kraskauskas D, et al. Remote
enhanced autophagy signaling. PLoS One 2014;9: ischemic pre-conditioning attenuates adverse
9. Choi BW, Berger HJ, Schwartz PE, et al. Serial
e111291. cardiac remodeling and mortality following doxo-
radionuclide assessment of doxorubicin car-
rubicin administration in mice. J Am Coll Cardiol
diotoxicity in cancer patients with abnormal 23. Xie Y, Jiang D, Xiao J, et al. Ischemic pre-
CardioOnc 2019;1:221–34.
baseline resting left ventricular performance. Am conditioning attenuates ischemia/reperfusion-
Heart J 1983;106:638–43. induced kidney injury by activating autophagy via 37. Hausenloy DJ, Kharbanda RK, Moller UK, et al.
the SGK1 signaling pathway. Cell Death Dis 2018; Effect of remote ischaemic conditioning on clinical
10. Friedman MA, Bozdech MJ, Billingham ME,
9:338. outcomes in patients with acute myocardial
Rider AK. Doxorubicin cardiotoxicity. Serial endo-
infarction (CONDI-2/ERIC-PPCI): a single-blind
myocardial biopsies and systolic time intervals. 24. Wang Y, Shen J, Xiong X, et al. Remote
randomised controlled trial. Lancet 2019;394:
JAMA 1978;240:1603–6. ischemic preconditioning protects against liver
1415–24.
11. Wong AD, Ye M, Ulmschneider MB, Searson PC. ischemia-reperfusion injury via heme oxygenase-1-
induced autophagy. PLoS One 2014;9:e98834. 38. Hausenloy DJ, Botker HE. Why did remote
Quantitative analysis of the enhanced permeation
ischaemic conditioning not improve clinical out-
and retention (EPR) effect. PLoS One 2015;10: 25. Taymaz-Nikerel H, Karabekmez ME, Eraslan S, comes in acute myocardial infarction in the
e0123461. Kirdar B. Doxorubicin induces an extensive tran- CONDI-2/ERIC-PPCI trial? Cardiovasc Res 2019;
12. Narayan HK, Putt ME, Kosaraju N, et al. Dex- scriptional and metabolic rewiring in yeast cells.
115:e161–3.
razoxane preferentially mitigates doxorubicin Sci Rep 2018;8:13672.
39. Ludman AJ, Yellon DM, Hausenloy DJ. Cardiac
cardiotoxicity in female children with sarcoma. 26. Renu K, Abilash VG, Pichiah PBT, preconditioning for ischaemia: lost in translation.
Open Heart 2019;6:e001025. Arunachalam S. Molecular mechanism of Dis Model Mech 2010;3:35–8.
13. Vejpongsa P, Yeh ET. Prevention of doxorubicin-induced cardiomyopathy - An update.
Eur J Pharmacol 2018;818:241–53. 40. Abete P, Testa G, Cacciatore F, et al. Ischemic
anthracycline-induced cardiotoxicity: challenges
preconditioning in the younger and aged heart.
and opportunities. J Am Coll Cardiol 2014;64:
27. Patel AG, Kaufmann SH. CANCER How does Aging Dis 2011;2:138–48.
938–45.
doxorubicin work? Elife 2012;1.
14. Li J, Rohailla S, Gelber N, et al. MicroRNA-144
28. Zhou F, Song W, Wang Z, et al. Effects of
is a circulating effector of remote ischemic pre-
remote ischemic preconditioning on contrast
conditioning. Basic Res Cardiol 2014;109:423. KEY WORDS apoptosis, autophagy, cardiac
induced nephropathy after percutaneous coro-
15. White SK, Frohlich GM, Sado DM, et al. Remote nary intervention in patients with acute coro- function, doxorubicin, inflammation,
ischemic conditioning reduces myocardial infarct nary syndrome. Medicine (Baltimore) 2018;97: multiorgan toxicity, repeated remote
size and edema in patients with ST-segment e9579. ischemic conditioning
EDITORIAL COMMENT
Repeated Remote Ischemic Conditioning

Protects Against Doxorubicin
Cardiotoxicity
Never Too Much of a Good Thing*
Derek J. Hausenloy, MD, PHD,a,b,c,d,e,f Choon Ta Ng, MD,g Jun Hua Chong, MDb,g
T he anthracycline chemotherapeutic, doxoru-

bicin (DOX), is one of the most efficacious
anticancer agents for treating a number of
solid and hematological malignancies, but its use
early or late-onset cardiomyopathy, and heart failure
in cancer survivors. Postulated mechanisms underly-
ing DOX cardiotoxicity include oxidative stress,
iron overload, mitochondrial injury, topoisomerase
has been limited by its dose-dependent cardiotoxic IIb inhibition, and DNA intercalation (1). Therapeutic
effects, which can increase the risk of developing interventions to mitigate DOX cardiotoxicity, such
as limiting the cumulative dose of DOX, use of
liposomal DOX, dexrazoxane therapy, and cardiopro-
*Editorials published in JACC: CardioOncology reflect the views of the tective therapies (such as beta-blockers and renin-
authors and do not necessarily represent the views of JACC: angiotensin antagonists), have shown some benefit,
CardioOncology or the American College of Cardiology. but cardiovascular morbidity and mortality from
From the aNational Heart Research Institute Singapore, National Heart cardiotoxicity in cancer survivors is still significant
b
Centre Singapore, Singapore; Cardiovascular & Metabolic Disorders despite these measures (2). Therefore, new therapeu-
Program, Duke-National University of Singapore Medical School,
c
tic strategies are needed to protect the myocardium
Singapore; Yong Loo Lin School of Medicine, National University
d
Singapore, Singapore; The Hatter Cardiovascular Institute, University
against DOX cardiotoxicity to improve cardiac out-
College London, London, United Kingdom; eCardiovascular Research comes in oncology patients undergoing anthracycline
Center, College of Medical and Health Sciences, Asia University, Taichung chemotherapy. In this regard, the endogenous cardio-
City, Taiwan; fTecnologico de Monterrey, Centro de Biotecnologia-
g
protective strategy, “remote ischemic conditioning”
FEMSA, Nuevo Leon, Mexico; and the Department of Cardiology, Na-
(RIC), has the therapeutic potential to protect the
tional Heart Centre Singapore, Singapore. Dr. Hausenloy has been sup-
ported by the British Heart Foundation (CS/14/3/31002), the National myocardium against DOX cardiotoxicity.
Institute for Health Research University College London Hospitals RIC refers to the phenomenon in which brief
Biomedical Research Centre, Duke-National University Singapore Medical
nonlethal cycles of ischemia and reperfusion to an
School, Singapore Ministry of Health’s National Medical Research Council
organ or tissue remote from the heart confer car-
under its Clinician Scientist-Senior Investigator scheme (NMRC/CSA-SI/
0011/2017) and Collaborative Centre Grant scheme (NMRC/ dioprotection against an episode of lethal acute
CGAug16C006), and the Singapore Ministry of Education Academic myocardial ischemia/reperfusion injury (IRI) (3).
Research Fund Tier 2 (MOE2016-T2–2–021). This paper is based on work
Importantly, the RIC stimulus can be applied to the
from COST Action EU-CARDIOPROTECTION CA16225 supported by COST
(European Cooperation in Science and Technology). Dr. Chong has been
limb by simply inflating and deflating a pneumatic
supported by the Singapore Ministry of Health’s National Medical cuff placed on the upper arm or thigh (termed limb
Research Council Research Training Fellowship (FLWSHP19may-0013) RIC) (4), making it a low-cost and noninvasive car-
and the National Medical Research Council Collaborative Centre Grant
dioprotective strategy that can be easily tested in the
Seed Funding (NHCS-CCGSF/2019/002). All other authors have reported
that they have no relationships relevant to the contents of this paper to
clinical setting. Although a single-limb RIC stimulus
disclose. has been reported to reduce myocardial infarct size in
The authors attest they are in compliance with human studies commit- patients with acute myocardial infarction (AMI) (5), a
large multicenter study failed to demonstrate any
appropriate. For more information, visit the JACC: CardioOncology author beneficial effects on clinical outcomes (6). The
instructions page. mechanisms underlying the cardioprotective effect of

54 Hausenloy et al. JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020
Repeat Remote Ischemic Conditioning and Doxorubicin Cardiotoxicity MARCH 2020:53–5
limb RIC remain unclear, but involve the release of dysfunction, hypertension, chronic heart failure, and
cytoprotective factor(s) into the systemic circulation, post-AMI LV remodeling (9).
and the activation of prosurvival signaling pathways In this issue of JACC: CardioOncology, a study by
in the heart, that target the detrimental effects of He et al. (13) tested the cardioprotective efficacy of
acute IRI, including oxidative stress, apoptosis, rRIC in a mouse model of doxorubicin-induced car-
inflammation, and mitochondrial dysfunction (3). The diotoxicity. They found that limb RIC (four 5-min
fact that many of these factors have also been impli- cycles of limb ischemia and reperfusion using a
cated as mediators of DOX cardiotoxicity, raised the tourniquet placed on the hindlimb), initiated 30 min
possibility that limb RIC may be cardioprotective in before DOX administration and applied daily for
patients undergoing anthracycline chemotherapy. 5 days, had the following beneficial effects: it reduced
In this regard, Gertz et al. (7) recently reported the DOX-induced myocardial apoptosis, myocardial
potential cardioprotective effects of limb RIC against injury (measured by troponin release), myocardial
DOX cardiotoxicity. Using a mouse model, they inflammation, and interstitial fibrosis; it enhanced
demonstrated that a single-limb RIC stimulus (three cellular autophagy and maintained cardiomyocyte
5-min cycles of femoral artery ligation and reflow) size and heart weight; and it partially attenuated the
reduced mortality, preserved left ventricular (LV) DOX-induced depression in cardiac function,
mass, decreased myocardial fibrosis and apoptosis, although the effect was modest (assessed by echo-
and upregulated autophagy signaling. However, no determined LV ejection fraction [DOX 47.5% vs.
beneficial effects were observed on either mitochon- DOX þ rRIC 51.6%] and fractional shortening). As no
drial or cardiac function, and so the mechanisms sham limb RIC group was included in the study, the
underlying the observed improvement in survival potential confounding effects of daily episodes of
with RIC remain unclear. The findings from this study anesthesia over 5 days (which would be required to
suggested that a single-limb RIC stimulus has the administer the limb RIC stimulus) on the observed
therapeutic potential to prevent or attenuate DOX effects of rRIC, cannot be excluded. Given that the
cardiotoxicity, and provide a low-cost and noninva- salutary effects on cardiac function of rRIC were
sive cardioprotective strategy for testing in patients modest, perhaps a more prolonged course of rRIC
undergoing anthracycline chemotherapy. In this re- (as used in the studies demonstrating less post-AMI
gard, the ongoing ERIC-ONC study (Effect of Remote adverse LV remodeling with rRIC applied for
Ischemic Conditioning in Oncology Patients; 28 days [10,11]) may have had a greater car-
NCT02471885) is currently investigating the cardioprotective effect against DOX cardiotoxicity.
dioprotective effects of a single-limb RIC stimulus However, a prolonged rRIC protocol would raise is-
(four 5-min inflations and deflations of a pneumatic sues of patient compliance in future clinical studies.
cuff placed on the upper arm) applied immediately
SEE PAGE 41
before each chemotherapy cycle in adult oncology
patients newly referred for anthracycline chemo- Interestingly, the authors also demonstrated that
therapy, with the primary endpoint being myocardial rRIC had cytoprotective effects in other organs, such
injury (measured by high-sensitive troponin T at 6 as the kidney, spleen, and liver, in terms of preser-
and 24 h) (8). ving organ weight and function (13). The multiorgan
Interestingly, emerging data have suggested that benefits of rRIC may be explained by the release into
repeated daily episodes of limb RIC for 28 days, the systemic circulation of, so far unidentified, cyto-
termed repeated RIC (rRIC) or chronic remote protective factor(s) by the limb RIC stimulus (3). In
ischemic conditioning (9), may have beneficial effects future studies, it would be important to determine
over and above a single-limb RIC stimulus in experi- whether rRIC can protect mitochondrial respiratory
mental small animal models of AMI in terms of pre- function against the damaging effects of DOX in the
venting adverse post-AMI LV remodeling via heart and other organs, although a single RIC stim-
potential anti-apoptotic, anti-inflammatory, auto- ulus did not have any mitoprotective effects against
phagic effects, and exosome-mediated improvements DOX cardiotoxicity in a recent study (7). Furthermore,
in intercellular communication (10,11). Furthermore, studies of rRIC in a tumor-bearing animal model are
in patients with intracranial arterial stenosis, rRIC needed before clinical testing to ensure rRIC does not
administered for 365 days was shown to reduce the attenuate the cytotoxic potency and anticancer effi-
risk of stroke (12), and rRIC applied for 7 to 28 days cacy of DOX.
has been shown to have vascular and cytoprotective In summary, recent studies have demonstrated
effects that may benefit patients with endothelial potential cardioprotective effects of limb rRIC against
JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020 Hausenloy et al. 55
MARCH 2020:53–5 Repeat Remote Ischemic Conditioning and Doxorubicin Cardiotoxicity
DOX cardiotoxicity, providing a low-cost noninvasive cytotoxic actions of DOX). Identification of the
treatment strategy for improving cardiac outcomes in circulating cytoprotective factor(s) released by limb
oncology patients undergoing anthracycline chemo- RIC, and the elucidation of the molecular pathways
therapy. Further work is needed to determine the underlying RIC-induced multiorgan protection, may
optimal rRIC protocol for cardioprotection in terms of result in the discovery of novel therapeutic targets
the duration of the rRIC protocol (as this will have and treatments for improving clinical outcomes in
implications for patient compliance); the number of oncology patients at risk of DOX cardiotoxicity.
ischemia and reperfusion cycles that make up the
optimum limb RIC stimulus (this has not been
adequately defined); whether it is efficacious in pa- ADDRESS FOR CORRESPONDENCE: Dr. Derek J.
tients with cardiovascular morbidities such as dia- Hausenloy, Cardiovascular and Metabolic Disorders
betes, hypertension, and obesity, factors that may Program, Duke-National University of Singapore, 8
confound RIC-induced cardioprotection; and its College Road, Singapore 169857. E-mail: derek.
safety (in relation to whether it interferes with the hausenloy@duke-nus.edu.sg. Twitter: @dukenus.
REFERENCES
1. McGowan JV, Chung R, Maulik A, Piotrowska I, infarction (CONDI-2/ERIC-PPCI): a single-blind heart failure after myocardial infarction. Int J
Walker JM, Yellon DM. Anthracycline chemo- randomised controlled trial. Lancet 2019;394: Cardiol 2015;178:239–46.
therapy and cardiotoxicity. Cardiovasc Drugs Ther 1415–24.
11. Wei M, Xin P, Li S, et al. Repeated remote
2017;31:63–75.
7. Gertz ZM, Cain C, Kraskauskas D, et al. Remote ischemic postconditioning protects against
2. Hayek SS, Ganatra S, Lenneman C, et al. Pre- ischemic pre-conditioning attenuates adverse adverse left ventricular remodeling and improves
paring the cardiovascular workforce to care for cardiac remodeling and mortality following doxo- survival in a rat model of myocardial infarction.
oncology patients: JACC Review Topic of the rubicin administration in mice. J Am Coll Cardiol Circ Res 2011;108:1220–5.
Week. J Am Coll Cardiol 2019;73:2226–35. CardioOnc 2019;1:221. 12. Meng R, Asmaro K, Meng L, et al. Upper limb
3. Sivaraman V, Pickard JMJ, Hausenloy DJ. 8. Chung R, Maulik A, Hamarneh A, et al. Effect of ischemic preconditioning prevents recurrent
Remote ischaemic conditioning: cardiac protection Remote ischaemic conditioning in oncology pa- stroke in intracranial arterial stenosis. Neurology
from afar. Anaesthesia 2015;70:732–48. tients undergoing chemotherapy: rationale and 2012;79:1853–61.
4. Kharbanda RK, Mortensen UM, White PA, et al. design of the ERIC-ONC study–a single-center, 13. He Q, Wang F, Ryan TD, Chalasani M,
Transient limb ischemia induces remote ischemic blinded, randomized controlled trial. Clin Cardiol Redington AN. Repeated remote ischemic
preconditioning in vivo. Circulation 2002;106: 2016;39:72–82. conditioning reduces doxorubicin-induced car-
2881–3. 9. Chong J, Bulluck H, Ho AFW, Boisvert WA. diotoxicity. J Am Coll Cardiol CardioOnc 2020;2:
Hausenloy DJ Chronic remote ischemic condition- 41–52.
5. Chong J, Bulluck H, Yap EP, Ho AF, Boisvert WA,
Hausenloy DJ. Remote ischemic conditioning in ing for cardiovascular protection. Cond Med 2019;
ST-segment elevation myocardial infarction - an 2:164–9.
update. Cond Med 2018;1:13–22.
10. Yamaguchi T, Izumi Y, Nakamura Y, et al.
6. Hausenloy DJ, Kharbanda RK, Moller UK, et al. Repeated remote ischemic conditioning attenu- KEY WORDS anthracycline cardiotoxicity,
Effect of remote ischaemic conditioning on clinical ates left ventricular remodeling via exosome- cardioprotection, chronic remote ischemic
outcomes in patients with acute myocardial mediated intercellular communication on chronic conditioning, repeat remote ischemic conditioning
ORIGINAL RESEARCH
Plasma Hepatocyte Growth Factor for

Diagnosis and Prognosis in Light Chain
and Transthyretin Cardiac Amyloidosis
Kathleen W. Zhang, MD,a Jennifer Miao, MD,b Joshua D. Mitchell, MD,a Jose Alvarez-Cardona, MD,a
Kelsey Tomasek, BS,b Yan Ru Su, MD,b Mary Gordon, RN,b R. Frank Cornell, MD,c Daniel J. Lenihan, MDa
ABSTRACT
OBJECTIVES This study determined the diagnostic and prognostic usefulness of hepatocyte growth factor (HGF) in
light chain and transthyretin cardiac amyloidosis.
BACKGROUND Delays in diagnosis of cardiac amyloidosis are common, usually resulting from nonspecific findings on
clinical examination and testing. A discriminatory plasma biomarker could result in earlier diagnosis and improve prog-
nosis assessment.
METHODS A total of 188 patients with cardiac amyloidosis, amyloidosis without cardiac involvement, symptomatic
heart failure with left ventricular hypertrophy (LVH), or heart failure with a reduced ejection fraction (HFrEF) were
enrolled prospectively. Serum biomarkers were measured at study enrollment, and all patients with amyloidosis were
followed for all-cause mortality, cardiac transplantation, or left ventricular assist device implantation. Multinomial lo-
gistic regression and Kaplan-Meier survival estimates tested the association of biomarker levels with cardiac amyloidosis
and clinical outcomes, respectively. Harrell’s C-statistic and the likelihood ratio test compared the prognostic accuracy of
plasma biomarkers.
RESULTS HGF was significantly higher in patients with cardiac amyloidosis (p < 0.001). An HGF level of 205 pg/ml
discriminated cardiac amyloidosis from LVH and HFrEF with 86% sensitivity, 84% specificity, and an area under the curve
of 0.88 (95% confidence interval: 0.83 to 0.94). In patients with amyloidosis, elevated HGF levels were associated
with worse event-free survival over a median follow-up of 2.6 years (p < 0.001) with incremental prognostic accuracy
over N-terminal pro-B-type natriuretic peptide and troponin T (p < 0.001).
CONCLUSIONS HGF discriminates light chain and transthyretin cardiac amyloidosis from patients with symptomatic
heart failure with LVH or HFrEF and is associated with worse cardiac outcomes. Confirmation of these findings in a
larger, multicenter study that is enrolling suspected cases of cardiac amyloidosis is underway.
From the aCardio-Oncology Center of Excellence, Division of Cardiology, Washington University School of Medicine, St. Louis,
Missouri; bDepartment of Internal Medicine, Vanderbilt University Medical Center, Nashville, Tennessee; and the cDivision of
Hematology and Oncology, Vanderbilt University Medical Center, Nashville, Tennessee. This study was supported by the National
Center for Advancing Translational Sciences of the National Institutes of Health (Award Number UL1TR002345). Dr. Mitchell has
received research support from Pfizer. Dr. Lenihan has been a consultant for Lilly, Roche, Pfizer, Prothena, and Acorda; and has
received research funding from Myocardial Solutions, Inc. All other authors have reported that they have no relationships relevant
to the contents of this paper to disclose. Frederick L. Ruberg, MD, served as Guest Editor for this paper.
Manuscript received October 21, 2019; revised manuscript received January 20, 2020, accepted January 21, 2020.

JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020 Zhang et al. 57
MARCH 2020:56–66 Hepatocyte Growth Factor in Cardiac Amyloidosis
C ardiac amyloidosis is a restrictive cardiomy- Fibrosis and inflammation are additional ABBREVIATIONS
opathy that results from the intramyocardial processes that may mediate the pathophysi- AND ACRONYMS
deposition of amyloid fibrils, which are ological effects of amyloid fibril deposition
AL = light chain amyloidosis
composed of misfolded light chain or transthyretin (1,19). Galectin-3 and interleukin (IL)-6 are
ATTR = transthyretin
proteins in >95% of cases. Although light chain known biomarkers of fibrosis and inflamma-
amyloidosis
amyloidosis (AL) and hereditary transthyretin tion, respectively, which may have diagnostic
ATTR-CA = transthyretin
amyloidosis (ATTR) are uncommon conditions, wild- and prognostic usefulness for cardiac cardiac amyloidosis
type ATTR is likely underdiagnosed, particularly amyloidosis. CI = confidence interval
among patients with heart failure with preserved The primary objective of this study was to
CV = coefficient of variation
ejection fraction and calcific aortic stenosis (1–5). assess the diagnostic performance of HGF for
eGFR = estimated glomerular
Traditionally, treatment options for AL and ATTR AL and ATTR cardiac amyloidosis among pa- filtration rate
have been limited, and median survival has been tients with symptomatic heart failure, and to HFrEF = heart failure with
only 2 to 4 years from time of diagnosis (6,7). determine the prognostic usefulness of HGF reduced ejection fraction
Recently, several therapeutics were approved for for adverse cardiac events in patients with AL HGF = hepatocyte growth
the treatment of ATTR, whereas multiple other and ATTR amyloidosis. We also explored the factor
agents are in late stage drug development for both diagnostic usefulness of galectin-3 and IL-6 IL = interleukin
AL and ATTR (8). Accordingly, earlier diagnosis of for AL and ATTR cardiac amyloidosis. IVSd = interventricular septal
cardiac amyloidosis is an important clinical priority. diameter in diastole
The presenting signs, symptoms, and findings on METHODS LVAD = left ventricular assist
device
routine cardiac testing for patients with cardiac
LVEF = left ventricular
amyloidosis are nonspecific. As a result, a potential STUDY PARTICIPANTS. Consecutive patients
ejection fraction
delay of up to 36 months is common for patients with referred to the Vanderbilt Amyloidosis
LVH = left ventricular
AL and ATTR cardiac amyloidosis, with more Multidisciplinary Program for initial evalua- hypertrophy
advanced disease documented at presentation in tion of cardiac amyloidosis were enrolled
NT-proBNP = N-terminal pro-
those with longer delays in diagnosis (9,10). Trans- prospectively into the Main Heart Registry B-type natriuretic peptide
thoracic echocardiography (TTE) is widely available from 2010 to 2017. The diagnosis of cardiac ROC = receiver-operating
and may be useful to differentiate cardiac amyloid- amyloidosis was established by 1 of 2 criteria: characteristic
osis from other causes of symptomatic heart failure. 1) positive endomyocardial biopsy by Congo TTE = transthoracic
Specifically, the ratio of left ventricular ejection red stain; or 2) positive noncardiac biopsy by echocardiography
fraction (LVEF) to global longitudinal strain and a Congo red stain with increased wall thickness on TTE,
relative apical sparing pattern in longitudinal strain low voltage or pseudo-infarction on electrocardio-
have been shown to be diagnostically useful (11,12). gram, or characteristic delayed myocardial enhance-
However, a plasma biomarker that discriminates ment on cardiac magnetic resonance imaging. All
cardiac amyloidosis among patients with symptom- positive biopsy specimens were sent to the Mayo
atic heart failure has yet to be identified. Clinic reference laboratory for mass spectrometry
subtyping of the amyloid protein. Radionuclide bone
SEE PAGE 67
scintigraphy was not used for diagnosis because it
Hepatocyte growth factor (HGF) is a pro- was not routinely available over the entire study.
angiogenic, mitogenic, and morphogenic cytokine TTE, electrocardiography, and cardiac magnetic
that is primarily expressed in mesenchymal cells (13). resonance studies were performed as part of routine
Small studies, including preliminary data from our patient care. Increased wall thickness was defined
own group, have shown that HGF is elevated in pa- according to American Society of Echocardiography
tients with amyloidosis, particularly in those with guidelines for left ventricular hypertrophy (LVH) as a
cardiac involvement (14–16). The b-pleated sheet left ventricular mass index of >115 g/m 2 in men and
structure of amyloid fibrils has been shown to acti- >95 g/m 2, in women with relative wall thickness of
vate tissue-type plasminogen activator, which in- >0.44 (20). Low voltage on electrocardiography was
creases production of HGF due to structural defined as QRS amplitude <5 mV in all limb leads
homology of pro-HGF with plasminogen (17,18). In or <10 mV in all precordial leads (21). Late gadolinium
addition, HGF mediates mesenchymal-epithelial in- enhancement on cardiac magnetic resonance was
teractions during the process of tissue repair, which evaluated using standard sequences.
may be triggered by amyloid fibril deposition (13,14). Detailed clinical and demographic data were ob-
The usefulness of HGF to differentiate cardiac tained at the time of study entry using a standardized
amyloidosis from other causes of symptomatic heart questionnaire administered by study personnel, with
failure is not known. verification via medical records. All patients with
58 Zhang et al. JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020
Hepatocyte Growth Factor in Cardiac Amyloidosis MARCH 2020:56–66
amyloidosis were followed for the combined outcome measured in the clinical laboratory using commercial
of all-cause mortality, cardiac transplantation, or left fluorescence immunoassays as part of routine clinical
ventricular assist device (LVAD) implantation. care.
As a control or comparator group, consecutive pa-
STATISTICAL ANALYSIS. Baseline data were sum-
tients referred to the Heart Failure Clinic at the Van-
marized as median (25th and 75th percentile range
derbilt Heart and Vascular Institute from 2010 to 2012
[interquartile range]) for continuous variables and
with symptomatic heart failure and either LVH or
counts (percentages) for categorical data. Categorical
heart failure with reduced ejection fraction (HFrEF)
variables among the 4 groups were compared using
were enrolled into the Main Heart Registry. LVH was
the chi-square test or Fisher’s exact test. Continuous
defined using the previously described American So-
variables with approximately normal distributions
ciety of Echocardiography guidelines, and reduced
were compared with 1-way analysis of variance, ac-
ejection fraction was defined as LVEF #30% and New
counting for unequal variances and using the Tukey-
York Heart Association functional class III or IV. All
Kramer adjustment for between group comparisons.
control patients underwent a baseline assessment
The Kruskal-Wallis test was used to compare contin-
and were deemed to be free of amyloidosis by clinical
uous variables where the normality assumption was
evaluation.
not met. Approximate normality was determined
This study was approved by the Vanderbilt Uni-
graphically by reviewing the histograms and Q-Q
versity Medical Center Institutional Review Board. All
plots for each variable’s distribution.
participants provided written informed consent.
Multinomial logistic regression tested the associa-
LABORATORY ANALYSES. Blood samples were tion of biomarker levels with cardiac amyloidosis
collected at the time of study enrollment, and plasma compared with the other groups (amyloidosis without
was separated by centrifugation for 15 min at 1,500g cardiac involvement, LVH, and HFrEF) after adjusting
at 4 C. Plasma aliquots were stored at 80 C before for age, left ventricular interventricular septal diam-
use in assays. eter at end-diastole (IVSd), and LVEF, which were
HGF and IL-6 were measured using the Milliplex selected based on clinical judgment. Log trans-
Map Kit (EMD Millipore Corporation, Billerica, Mas- formation of continuous variables was used for line-
sachusetts), which is a 2-step chemiluminescent arity assumption. Receiver-operating characteristic
microsphere immunoassay. The limit of detection of (ROC) analysis was performed to identify an optimal
the HGF assay is 6.8 pg/ml, with intra- and inter-assay HGF cutoff for the diagnosis of cardiac amyloidosis
coefficients of variation (CV) of 8.7% and 8.8%, using Youden’s index. Further ROC analysis explored
respectively. The limit of detection of the IL-6 assay is the relative discriminative ability of HGF, NT-
0.2 pg/ml with intra-assay and interassay CVs of 9.3% proBNP, and troponin-T to predict cardiac involve-
and 5.5%, respectively. ment in patients with amyloidosis. All biomarker data
Galectin-3 was measured using a 2-step, sandwich were subjected to natural logarithmic transformation
enzyme chemiluminescent immunoassay using Quan- for the regression models, including the ROC
tikine technology (BioTek Instruments, Winooski, analyses.
Vermont). The limit of detection of galectin-3 ranged Kaplan-Meier survival estimates for the combined
from 0.003 to 0.085 ng/ml (mean: 0.016 ng/ml). The outcome of all-cause mortality, cardiac trans-
intra assay and interassay CVs ranged from 3.0% to plantation, or LVAD implantation were compared
4.4% and 6.8% to 8.6%, respectively. between those with normal and abnormal biomarker
N-terminal pro-B-type natriuretic peptide (NT- values using the log-rank test. We determined
proBNP) and troponin-T were measured on the Cobas abnormal levels of HGF through our ROC analysis,
e 411 analyzer using Elecsys immunoassays (Roche whereas abnormal levels of previous biomarkers were
Diagnostics, Indianapolis, Indiana), which uses 2- defined by cutoffs that were used in published
step, sandwich electrochemoluminescence technol- staging models for AL and ATTR cardiac amyloidosis
ogy. The limit of detection of the NT-proBNP assay is (NT-pro-BNP: 332 pg/ml; troponin-T: 35 ng/l, and
5.0 pg/ml with intra- and interassay CVs ranging from estimated glomerular filtration rate [eGFR]: 45 ml/
1.8% to 4.6% and 1.3% to 4.2%, respectively. The limit min/1.73 m2 ) (7,22).
of detection of the troponin-T assay is 5 ng/l with Harrell’s C-statistic was used to compare the pre-
intra- and inter-assay CVs ranging from 1.4% to 10.3% dictive ability of abnormal biomarker values for the
and 0.7% to 5.6%, respectively. combined outcome by area under the curve analysis,
For patients with symptomatic heart failure and accounting for censoring (23). The discriminative
either LVH or HFrEF, BNP and troponin-I were ability of NT-proBNP, troponin-T, and HGF for
T A B L E 1 Baseline Clinical and Echocardiographic Characteristics
Amyloidosis Heart Failure Heart Failure

Without Cardiac With Left Ventricular With Reduced
Cardiac Amyloidosis Involvement Hypertrophy Ejection Fraction
(n ¼ 72) (n ¼ 30) (n ¼ 44) (n ¼ 42) p Value*
Age (yrs) 65 (5671)§ 59 (5270) 60 (4670) 56 (4967) 0.005

Male (%) 60 50 59 57 0.83
Systolic BP (mm Hg) 112 (101123)†‡ 119 (110132) 122 (104132) 107 (100128) 0.009
Diastolic BP (mm Hg) 70 (6278)† 77 (7282) 70 (6278) 68 (6472) 0.001
eGFR <45 ml/min/1.73 m2 (%) 35 37 30 31 0.38
NYHA functional class III/IV (%) 44‡§ 0 16 100 <0.001
Medications (%)
ACEI/ARB 36 37 32 57 0.073
Aldosterone antagonist 58†‡ 13 11 50 <0.001
Beta-blocker 38‡§ 27 73 86 <0.001
Loop diuretic 78†‡ 27 25 81 <0.001
Digoxin 11§ 0 2 36 <0.001
Echocardiography
LVEF (%) 55 (4060)†‡§ 60 (5562) 60 (6065) 20 (1525) <0.001
IVSd (cm) 1.5 (1.21.7)†‡§ 1.1 (1.01.2) 1.6 (1.42.1) 1.1 (0.91.2) <0.001
Values are median (interquartile range) or %. Categorical variables among the 4 groups were compared using the chi-square test or Fisher’s exact test. Continuous variables
with approximately normal distributions were compared with 1-way analysis of variance, accounting for unequal variances and using the Tukey-Kramer adjustment for between
group comparisons. The Kruskal-Wallis test was used to compare continuous variables when the normality assumption was not met. *Compared across all groups. †p < 0.05
compared with amyloidosis without cardiac involvement. ‡p < 0.05 compared with left ventricular hypertrophy. §p < 0.05 compared with heart failure with reduced ejection
fraction.
ACEI ¼ angiotensin-converting enzyme inhibitor; ARB ¼ aldosterone receptor blocker; BP ¼ blood pressure; eGFR ¼ estimated glomerular filtration rate;
IVSd ¼ interventricular septal thickness in diastole; LVEF ¼ left ventricular ejection fraction; NYHA ¼ New York Heart Association.
prediction of the combined outcome was further IL-6, and galectin-3 at baseline. We then used the 10
evaluated using Akaike information criterion, with imputed datasets to confirm the association of HGF
smaller values indicating better fit, and the likelihood with the diagnosis of cardiac amyloidosis.
ratio test to compare the difference between models All statistical analyses were performed using SAS
(24). version 9.4 (SAS Institute, Cary, North Carolina) with
All analyses were performed using complete case a 2-sided alpha set at 0.05.
analysis. As a sensitivity analysis, we conducted 10
iterations of multiple imputation to estimate missing RESULTS
HGF values using fully conditional specification.
Covariates in the regression model were selected PATIENT CHARACTERISTICS. Baseline clinical and
based on clinical judgment and included age, sex, use echocardiographic characteristics of the 188 patients
of cardiac medications, LVEF, IVSd, blood pressure, included in this study are shown in Table 1. Of the 102
T A B L E 2 Baseline Biomarker Characteristics
Amyloidosis Without Heart Failure With Heart Failure With

Cardiac Amyloidosis Cardiac Involvement Left Ventricular Hypertrophy Reduced Ejection Fraction
(n ¼ 72) (n ¼ 30) (n ¼ 44) (n ¼ 42) p Value*
HGF (pg/ml) 455 (323735)†‡§ 247 (124490) 105 (66151) 131 (89234) <0.001
Galectin-3 (ng/ml) 14 (1018)‡§ 13 (1121) 19 (1429) 22 (1526) <0.001
IL-6 (pg/ml) 4.1 (3.25.5)‡ 4.3 (2.74.7) 3.0 (1.54.3) 3.8 (2.26.2) <0.003
NT-proBNP (pg/ml) 2,363 (7355,634)† 204 (53559) d d <0.001
BNP (pg/ml) d d 357 (135775) 945 (2561,495) d
Troponin-T (ng/l) 43 (1472)† 12 (923) d d <0.001
Troponin-I (ng/ml) d d 0.04 (0.020.14) 0.04 (0.030.51) d
Values are median (interquartile range). The Kruskal-Wallis test was used to compare biomarker values across groups. *Compared across all groups. †p < 0.05 compared with
amyloidosis without cardiac involvement. ‡p < 0.05 compared with left ventricular hypertrophy. §p < 0.05 compared with heart failure with reduced ejection fraction.
HGF ¼ hepatocyte growth factor; IL ¼ interleukin; NT-proBNP ¼ N-terminal pro-B-type natriuretic peptide.
F I G U R E 1 HGF Is Elevated in Light Chain and Transthyretin Cardiac Amyloidosis
Box and whisker plots show median and interquartile range for normalized hepatocyte growth factor (HGF) values for patients with light chain
cardiac amyloidosis, transthyretin cardiac amyloidosis, left ventricular hypertrophy (LVH), and heart failure with reduced ejection fraction
(HFrEF). HGF was significantly higher in light chain and transthyretin cardiac amyloidosis compared with LVH and HFrEF (p < 0.001).
*p < 0.001 compared with LVH and HFrEF.
patients with amyloidosis, 69 had AL and 33 had BASELINE BIOMARKER CHARACTERISTICS. Base-
ATTR; 90 were white and 12 were African American. line biomarker levels for the study population are
Seventy-two patients had amyloidosis with cardiac shown in Table 2. HGF, galectin-3, and IL-6 were
involvement (40 diagnosed by endomyocardial bi- measured in $90% of patients with cardiac amyloid-
opsy, 32 diagnosed by extracardiac biopsy with sup- osis and $83% of patients with amyloidosis without
portive cardiac imaging findings) and 30 patients had cardiac involvement. NT-proBNP and troponin-T
amyloidosis without cardiac involvement (26 with were measured in 97% of patients with cardiac
AL, 4 with ATTR). In our control group, 44 patients amyloidosis and 90% of patients with amyloidosis
had LVH and 42 had HFrEF. There were no significant without cardiac involvement. HGF and galectin-3
differences in sex or prevalence of renal dysfunction were measured in the entire control group.
across the 4 groups. HGF was significantly higher in patients with car-
Patients with cardiac amyloidosis tended to be diac amyloidosis compared with patients with
older than those with LVH (p ¼ 0.052) and HFrEF amyloidosis without cardiac involvement, LVH, and
(p < 0.05). The LVEF was lower compared with those HFrEF (p < 0.001) (Figure 1, Table 2). In multinomial
with amyloidosis without cardiac involvement and logistic regression analysis, HGF remained signifi-
LVH and higher compared with those with HFrEF cantly associated with cardiac amyloidosis after
(p < 0.05). Patients with cardiac amyloidosis had adjusting for age, IVSd, and LVEF (p < 0.001). In
increased IVSd compared with those with amyloidosis sensitivity analysis using 10 imputed datasets,
without cardiac involvement and HFrEF but had this multivariable association remained significant
decreased IVSd compared with those with LVH (p < 0.001). There was no significant difference in
(p < 0.05). HGF between the AL and ATTR cardiac amyloidosis
groups (p ¼ 0.87). Compared with the LVH and HFrEF

F I G U R E 2 HGF Has Diagnostic Usefulness for Cardiac Amyloidosis Compared With
control subjects, HGF showed good discrimination for Symptomatic Heart Failure With LVH or HFrEF
cardiac amyloidosis by ROC analysis with an area
under the curve of 0.88 (95% CI: 0.83 to 0.94). An
HGF cutoff of 205 pg/ml was optimal for the diagnosis
of cardiac amyloidosis in this population, with 86%
(95% CI: 78% to 94%) sensitivity and 84% (95% CI:
76% to 92%) specificity (Figure 2).
Galectin-3 was lower in patients with cardiac
amyloidosis compared with the control subjects with
LVH and HFrEF in the univariable (p < 0.001) and
multivariable analyses (p ¼ 0.002). There was no
difference in galectin-3 compared with those with
amyloidosis without cardiac involvement (p ¼ 0.99).
A cutoff of 18.4 ng/ml was optimal in discriminating
cardiac amyloidosis from the control subjects with
LVH and HFrEF with an area under the curve of 0.71
(95% CI: 0.64 to 0.78). IL-6 was higher in patients
with cardiac amyloidosis compared with the control
subjects with LVH in the univariable (p < 0.01) and
multivariable analysis (p ¼ 0.014), with no difference
compared with those with amyloidosis without car-
diac involvement (p ¼ 0.66) and HFrEF (p ¼ 0.36).
Among patients with amyloidosis, NT-proBNP and
troponin-T were significantly higher in those with
cardiac involvement in univariable (p < 0.001 for
Receiver-operator characteristic analysis was used to evaluate the diagnostic perfor-
both) and multivariable analyses (p ¼ 0.002 and mance of HGF for cardiac amyloidosis compared with symptomatic heart failure with LVH
0.008, respectively). NT-proBNP and troponin-T data hypertrophy or HFrEF. An HGF level of 205 pg/ml was optimal and identified light chain
were not available for the LVH and HFrEF groups, or transthyretin cardiac amyloidosis with 86% sensitivity, 84% specificity, and area
under the curve (AUC) of 0.88. Other abbreviations as in Figure 1.
although BNP and troponin-I were provided as a
reference (Table 2).
CONFIRMATORY USEFULNESS OF HGF FOR CARDIAC

102 patients with amyloidosis. All patients had at
INVOLVEMENT IN AMYLOIDOSIS. Among the
least 18 months of follow-up. We evaluated the
102 patients with amyloidosis, we evaluated the
prognostic usefulness of abnormal HGF, NT-proBNP,
confirmatory usefulness of elevated HGF for
troponin-T, and eGFR for all-cause mortality, car-
cardiac involvement with amyloidosis. In this
diac transplantation, and LVAD implantation within
cohort, the optimal HGF cutoff for discriminating
this group. Cutoffs used were HGF 320 pg/ml as
cardiac involvement was 320 pg/ml, which yielded
derived from our amyloidosis cohort, and NT-proBNP
77% (95% CI: 67% to 87%) sensitivity and 69% (95%
332 pg/ml, troponin-T 35 ng/l, and eGFR 45 ml/
CI: 51% to 87%) specificity with an area under the
min/1.73 m 2 cutoffs that were used in published
curve of 0.69 (95% CI: 0.55 to 0.83). An HGF level
staging models for AL and ATTR cardiac amyloid-
>320 pg/ml predicted cardiac involvement with
osis (7,22).
amyloidosis in univariable (p < 0.001) and multivar-
By Kaplan-Meier analysis, elevated HGF, NT-
iable (p ¼ 0.004) analyses. In this cohort with known
proBNP, and troponin-T were associated with worse
amyloidosis, NT-proBNP and troponin-T showed
event-free survival (p < 0.001 for all), whereas
better discrimination for cardiac involvement by the
abnormal eGFR was not (Figure 4). An HGF level of
c-statistic with areas under the curve of 0.89 (95% CI:
676 pg/ml optimally discriminated those with the
0.80 to 0.97) and 0.83 (95% CI: 0.73 to 0.93), respec-
combined outcome with 69% (95% CI: 52% to 86%)
tively (Figure 3).
sensitivity and 79% (95% CI: 70% to 87%) specificity
PROGNOSTIC USEFULNESS OF HGF IN AMYLOIDOSIS. by ROC analysis.
Over a median of 2.6 years (interquartile range: 1.9 to Individually, elevated HGF, NT-proBNP, and
3.1 years) of follow-up, there were 27 deaths, 2 cardiac troponin-T above their respective cutoffs each
transplantations, and 1 LVAD implantation among the showed modest predictive accuracy for the combined
of HGF (p < 0.001). In the 33 patients with ATTR, the

F I G U R E 3 NT-proBNP, Troponin-T, and HGF Are Useful for Confirming Cardiac
Involvement in Patients With Light Chain and Transthyretin Amyloidosis
addition of HGF to a model of NT-proBNP and eGFR
improved the area under the curve from 0.61 (95% CI:
0.47 to 0.74) to 0.65 (95% CI: 0.52 to 0.78), although
the discriminatory ability of the model was not
significantly improved by the likelihood ratio test
(p ¼ 0.14).
DISCUSSION
In this single-center prospective cohort study, we

found that HGF was discriminatory for AL and ATTR
cardiac amyloidosis among patients with symptom-
atic heart failure. Among patients with known
amyloidosis, NT-proBNP was a stronger indicator of
cardiac involvement than HGF or troponin-T. In
addition, elevated HGF was associated with the
combined outcome of all-cause mortality, cardiac
transplantation, and LVAD implantation with incre-
mental prognostic value in addition to NT-proBNP
and troponin-T (Central Illustration).
We found that HGF identified cardiac amyloidosis
in a population with symptomatic heart failure, both
with preserved and reduced ejection fraction. To
date, no other circulating biomarker has shown
Receiver-operator characteristic analysis was used to evaluate the confirmatory useful-

diagnostic usefulness in this setting. These data
ness of biomarkers for cardiac involvement with light chain and transthyretin amyloid- extended our previously reported analysis in a larger
osis. Curves are shown for N-terminal pro-B-type natriuretic peptide (NT-proBNP) (red), sample size and validated the previous findings (16).
troponin-T (green), and HGF (blue). NT-proBNP had the strongest confirmatory use- Additional studies are required in a broad population
fulness for cardiac involvement with amyloidosis (AUC 0.89), followed by troponin-T
of patients with symptomatic HF to further validate
(AUC 0.83) and HGF (AUC ¼ 0.69). Abbreviations as in Figures 1 and 2.
the usefulness of HGF for diagnostic purposes in
cardiac amyloidosis. In contrast, NT-proBNP
appeared to be indicative of cardiac involvement in
the setting of known amyloidosis based on our find-
outcome with areas under the curves of 0.68 (95% CI: ings and published reports in AL amyloidosis (25). In
0.60 to 0.75), 0.64 (95% CI: 0.59 to 0.69), and 0.65 the heart failure setting, relative apical sparing of
(95% CI: 0.56 to 0.74), respectively (Table 3). In longitudinal strain and cardiac magnetic resonance
combination, NT-proBNP and troponin-T had similar are currently used as noninvasive screening tools for
predictive accuracy (area under the curve: 0.69; 95% cardiac amyloidosis (12,26), although image quality,
CI: 0.61 to 0.77). The addition of HGF to NT-proBNP cost, and accessibility represent important barriers to
and troponin-T significantly improved the predictive the use of these techniques as screening tools. With
model with an area under the curve of 0.75 (95% CI: additional validation, HGF may represent a reliable,
0.68 to 0.83) (p < 0.001 compared with NT-proBNP inexpensive, and accessible initial screening test for
and troponin-T by the likelihood ratio test). cardiac amyloidosis that would direct the use of
As a sensitivity analysis, we evaluated the incre- downstream confirmatory testing, including TTE with
mental prognostic value of elevated HGF for the strain, cardiac magnetic resonance, technetium py-
combined outcome in the AL and ATTR groups rophosphate bone scintigraphy, or endomyocardial
separately, using established staging systems for each biopsy.
(7,22). In the 69 patients with AL, the addition of HGF Although HGF showed good discrimination for
to a model of NT-proBNP and troponin-T improved cardiac amyloidosis compared with symptomatic
the area under the curve from 0.74 (95% CI: 0.64 to heart failure, it was less useful in differentiating
0.84) to 0.81 (95% CI: 0.71 to 0.91). By the likelihood cardiac amyloidosis from amyloidosis without
ratio test, the discriminatory ability of the predictive cardiac involvement. This supported the hypothesis
model was significantly improved with the addition that HGF elevation in amyloidosis was specific to the
F I G U R E 4 Elevated HGF, NT-proBNP, and Troponin-T Are Associated With Worse Event-Free Survival in Light Chain and Transthyretin Cardiac Amyloidosis
Patients with amyloidosis were followed for a median of 2.6 years (interquartile range: 1.9 to 3.1 years) for the combined clinical outcome of all-cause mortality, cardiac
transplantation, or left ventricular assist device implantation. The prognostic usefulness of (A) HGF, (B) NT-proBNP, (C) troponin-T, and (D) estimated glomerular
filtration rate (eGFR) were evaluated using the following cutoffs: HGF 320 pg/ml as derived from our amyloidosis cohort, and NT-proBNP 332 pg/ml, troponin-T 35 ng/l,
and eGFR 45 ml/min/1.73 m2 as used in published staging models for light chain and transthyretin cardiac amyloidosis. By Kaplan-Meier analysis, elevated HGF, NT-
proBNP, and troponin-T were associated with worse event-free survival (p < 0.001 for all), whereas abnormal eGFR was not. Abbreviations as in Figures 1 to 3.
process of extracellular amyloid fibril deposition.

T A B L E 3 Predictive Ability of Elevated HGF, NT-proBNP, and Troponin-T for All-Cause
Mechanistically, HGF elevation may result from a role
Mortality, Cardiac Transplantation, and LVAD Implantation in Patients With Light Chain
in tissue repair because the mitogenic HGF cytokine is and Transthyretin Amyloidosis
produced by mesenchymal cells, whereas its receptor,
Biomarker C-Statistic (95% CI) AIC LRT p Value
c-met, is expressed on epithelial cells (13,14). Alter-
HGF 0.68 (0.600.75) 181.3 d
natively, elevated HGF may result because the b -
NT-proBNP 0.64 (0.590.69) 221.3 d
pleated sheet structure of amyloid proteins activates
Troponin-T 0.65 (0.560.74) 223.6 d
tissue-type plasminogen activator, which then in- NT-proBNP þ troponin-T 0.69 (0.610.77) 219.8 0.061*
creases production of HGF due to structural homol- NT-proBNP þ troponin-T þ HGF 0.75 (0.680.83) 175.0 <0.001†
ogy of pro-HGF and plasminogen (17,18). Consistent
Biomarkers were incorporated into models using cutoff values derived from our amyloidosis cohort (HGF of 320
with the findings by Abraham et al. (15), we observed pg/ml) or used in published staging models for light chain and transthyretin cardiac amyloidosis (NT-proBNP of
higher HGF levels in patients with amyloidosis with 332 pg/ml, troponin-T of 35 ng/l). *Compared with NT-proBNP alone. †Compared to NT-proBNP þ troponin-T.
AIC ¼ Akaike information criterion; CI ¼ confidence interval; LRT ¼ likelihood ratio test; LVAD ¼ left ventricular
cardiac involvement compared with patients without assist device; other abbreviations as in Table 2.
cardiac involvement. It remains unclear why cardiac
C E NT R A L IL L U ST R A T I O N Clinical Usefulness of Biomarkers in Light Chain and Transthyretin Cardiac Amyloidosis
Zhang, K.W. et al. J Am Coll Cardiol CardioOnc. 2020;2(1):56–66.
Our findings indicate that plasma biomarkers including hepatocyte growth factor (HGF), N-terminal pro-B-type natriuretic peptide (NT-proBNP), and troponin-T may
have clinical usefulness for diagnosis and prognosis in light chain and transthyretin cardiac amyloidosis. (A) Compared with patients with symptomatic heart failure
due to left ventricular hypertrophy (LVH) or heart failure with reduced ejection fraction (HFrEF), HGF is higher in patients with cardiac amyloidosis. (B) Among patients
with known systemic amyloidosis, NT-proBNP and troponin-T are better indicators of cardiac involvement with amyloidosis compared with HGF. (C) NT-proBNP,
troponin-T, and HGF are associated with worse event-free survival in patients with cardiac amyloidosis, whereas the estimated glomerular ejection fraction (eGFR) is
not. HF ¼ heart failure; IVSd ¼ interventricular septal thickness (diastole).
involvement with amyloidosis would lead to higher associated with the combined outcome in our study;
HGF levels. this might be due to our inclusion of patients with AL
Our results indicated that low galectin-3 might amyloidosis, as well as disproportionate enrollment of
discriminate cardiac amyloidosis from symptomatic whites compared with African Americans. By com-
heart failure with LVH or HFrEF, whereas IL-6 did parison, the population studied by Gillmore et al. (7)
not. This suggested that the processes of tissue consisted of 36% of patients with hereditary ATTR-CA,
fibrosis and ventricular remodeling that characterize with 64% of those patients having the V122I mutation.
chronic systolic heart failure are less active in cardiac Furthermore, eGFR is not specific to the disease pro-
amyloidosis (27), whereas inflammation is equally cess of cardiac amyloidosis and may reflect differences
prominent in both disease processes. Additional in comorbidities or ethnicity as opposed to cardiac
study in larger patient cohorts is needed to confirm amyloidosis disease stage. Future efforts to establish
these findings. prognostic models for cardiac amyloidosis should
We found that elevated HGF was associated with include the cardiac-specific biomarkers NT-proBNP
adverse cardiovascular outcomes with incremental and troponin-T and confirm our findings regarding
prognostic value in addition to NT-proBNP and HGF in larger patient populations with AL and ATTR.
troponin-T in a mixed population with AL and ATTR
amyloidosis. Additional study in larger populations is STUDY LIMITATIONS. First, this was a single-center
needed to confirm the prognostic usefulness of HGF study with a limited sample size and small group as-
in AL and ATTR, individually. Current staging models signments that raise the risk of type I error. Our
for AL incorporate circulating free light chains, findings should therefore be seen as preliminary.
troponin-T, and NT-proBNP (22,28), highlighting the Confirmation of our findings in a larger, multicenter
clinical importance of cardiac involvement with prospective study that enrolled patients with
amyloidosis. Published staging models for ATTR car- confirmed and suspected cardiac amyloidosis is un-
diac amyloidosis (ATTR-CA) use NT-proBNP and derway as part of the ACE-MOVE (Prospective Char-
troponin-T in wild-type ATTR-CA, and NT-proBNP acterization of Patients with Amyloidosis Focusing
and eGFR in mixed populations of wild-type and he- on Diagnosis and Innovative Treatment to Improve
reditary ATTR-CA (7,29). Abnormal eGFR was not Outcomes) Registry. Similarly, we were limited in our
adjustment for potential confounders. Second, only cardiac transplantation, and LVAD implantation with
56% of cardiac amyloidosis cases were diagnosed by incremental prognostic value over NT-proBNP and
endomyocardial biopsy, which remains the gold troponin-T. Additional study in a larger, multicenter
standard for diagnostic testing. Nonbiopsy diagnostic cohort of patients with confirmed and suspected car-
criteria used for cardiac amyloidosis were in accor- diac amyloidosis is underway to validate our findings.
dance with clinically accepted and published stan-
dards (29,30). Third, we combined AL and ATTR
ADDRESS FOR CORRESPONDENCE: Dr. Kathleen W.
cardiac amyloidosis in our outcome analyses due to
Zhang, Cardio-Oncology Center of Excellence, Car-
sample size limitations. Additional studies in sepa-
diovascular Division, Washington University School
rate AL and ATTR populations are needed to confirm
of Medicine, Campus Box 8086, 660 South Euclid
our findings regarding the prognostic usefulness of
Avenue, St. Louis, Missouri 63110-1093. E-mail:
HGF. Finally, NT-proBNP and troponin-T were not
kwzhang@wustl.edu. Twitter: @DLenihanICOS_MD.
available for the LVH or HFrEF control groups, which
precluded comparison to HGF for the diagnosis of
cardiac amyloidosis in this population. The diagnostic PERSPECTIVES
performance of HGF compared with NT-proBNP
and troponin-T for cardiac amyloidosis will be COMPETENCY IN MEDICAL KNOWLEDGE: Elevated HGF is
explored in the prospective, previously mentioned suggestive of the diagnosis of light chain or transthyretin cardiac
ACE-MOVE Registry. amyloidosis as well as adverse clinical outcomes in this
population.
CONCLUSIONS
TRANSLATIONAL OUTLOOK: Further study is underway to
HGF is elevated in patients with AL and ATTR cardiac validate the role of HGF as a biomarker. Future studies are also
amyloidosis and differentiates patients with cardiac needed to define the pathophysiological basis of HGF elevation
amyloidosis from those with other etiologies of in cardiac amyloidosis.
symptomatic heart failure. Furthermore, elevated
HGF was associated with risk of all-cause mortality,
REFERENCES
1. Mohammed SF, Mirzoyev SA, Edwards WD, et al. 7. Gillmore JD, Damy T, Fontana M, et al. A new and human hepatocyte growth factor. Proc Natl
Left ventricular amyloid deposition in patients staging system for cardiac transthyretin amyloid- Acad Sci U S A 1991;88:7001–5.
with heart failure and preserved ejection fraction. osis. Eur Heart J 2017;44:1–8.
14. Shikano M, Kushimoto H, Hasegawa H, et al.
J Am Coll Cardiol HF 2014;2:113–22.
8. Zhang KW, Stockerl-Goldstein KE, Lenihan DJ. Usefulness of serum hepatocyte growth factor for
2. González-López E, Gallego-Delgado M, Guzzo- Emerging therapeutics for the treatment of light the diagnosis of amyloidosis. Intern Med 2000;39:
Merello G, et al. Wild-type transthyretin amyloidosis chain and transthyretin amyloidosis. J Am Coll 715–9.
as a cause of heart failure with preserved ejection Cardiol Basic Trans Sci 2019;4:1–11. 15. Abraham J, Desport E, Rigaud C, et al. Hepa-
fraction. Eur Heart J 2015;36:2585–94. tocyte growth factor measurement in AL
9. Bishop E, Brown EE, Fajardo J, Barouch LA,
3. Scully PR, Treibel TA, Fontana M, et al. Preva- Judge DP, Halushka MK. Seven factors predict a amyloidosis. Amyloid 2015;22:112–6.
lence of cardiac amyloidosis in patients referred delayed diagnosis of cardiac amyloidosis. Amyloid 16. Swiger KJ, Friedman EA, Brittain EL, et al.
for transcatheter aortic valve replacement. J Am 2018;25:174–9. Plasma hepatocyte growth factor is a novel marker
Coll Cardiol 2018;71:463–4. of AL cardiac amyloidosis. Amyloid 2016;23:242–8.
10. Lousada I, Comenzo RL, Landau H, Guthrie S,
4. Treibel TA, Fontana M, Gilbertson JA, et al. Merlini G. Light chain amyloidosis: patient expe- 17. Mars WM, Zarnegar R, Michalopoulos GK.
Occult transthyretin cardiac amyloid in severe rience survey from the Amyloidosis Research Activation of hepatocyte growth factor by the
calcific aortic stenosis: prevalence and prognosis Consortium. Adv Ther 2015;32:920–8. plasminogen activators uPA and tPA. Am J Pathol
in patients undergoing surgical aortic valve 11. Pagourelias ED, Mirea O, Duchenne J, et al. 1993;143:949–58.
replacement. Circ Cardiovasc Imaging 2016:1–11. Echo parameters for differential diagnosis in car- 18. Kranenburg O, Bouma B, Kroon-
5. Castano A, Narotsky DL, Hamid N, et al. diac amyloidosis: a head-to-head comparison of Batenburg LMJ, et al. Tissue-type plasminogen
Unveiling transthyretin cardiac amyloidosis and its deformation and nondeformation parameters. Circ activator is a multiligand cross-b structure recep-
predictors among elderly patients with severe Cardiovasc Imaging 2017;10:e005588. tor. Curr Biol 2002;12:1833–9.
aortic stenosis undergoing transcatheter aortic 12. Phelan D, Collier P, Thavendiranathan P, et al. 19. Siegismund CS, Escher F, Lassner D, et al.
valve replacement. Eur Heart J 2017;38:2879–87. Relative apical sparing of longitudinal strain using Intramyocardial inflammation predicts adverse
two-dimensional speckle-tracking echocardiogra- outcome in patients with cardiac AL amyloidosis.
6. Cibeira MT, Sanchorawala V, Seldin DC, et al.
phy is both sensitive and specific for the diagnosis Eur J Heart Fail 2018;20:751–7.
Outcome of AL amyloidosis after high-dose
of cardiac amyloidosis. Heart 2012;98:1442–8.
melphalan and autologous stem cell trans- 20. Lang RM, Badano LP, Mor-Avi V, et al. Rec-
plantation: long-term results in a series of 421 13. Weidner KM, Arakaki N, Hartmann G, et al. ommendations for cardiac chamber quantification
patients. Blood 2011;118:4346–52. Evidence for the identity of human scatter factor by echocardiography in adults: an update from the
American Society of Echocardiography and the 24. Allison PD. Logistic Regression Using the SAS 28. Kumar S, Dispenzieri A, Lacy MQ, et al. Revised
European Association of Cardiovascular Imaging. System: Theory and Application. 1st edition. prognostic staging system for light chain
Eur Heart J Cardiovasc Imaging 2015;16:233–71. Hoboken, NJ: Wiley-SAS, 2008. amyloidosis incorporating cardiac biomarkers and
serum free light chain measurements. J Clin Oncol
21. Murtagh B, Hammill SC, Gertz MA, Kyle RA, 25. Aljama MA, Sidiqi MH, Dispenzieri A, et al.
2012;30:989–95.
Tajik AJ, Grogan M. Electrocardiographic findings Comparison of different techniques to identify
in primary systemic amyloidosis and biopsy- cardiac involvement in immunoglobulin light 29. Grogan M, Scott CG, Kyle RA, et al. Natural
proven cardiac involvement. Am J Cardiol 2005; chain (AL) amyloidosis. Blood Adv 2019;3: history of wild-type transthyretin cardiac amyloid-
95:535–7. 1226–9. osis and risk stratification using a novel staging
system. J Am Coll Cardiol 2016;68:1014–20.
22. Dispenzieri A, Gertz MA, Kyle RA, et al. 26. Baggiano A, Boldrini M, Martinez-Naharro A,
Serum cardiac troponins and N-terminal pro- et al. Noncontrast magnetic resonance for the 30. Grogan M, Dispenzieri A, Gertz MA. Light-
brain natriuretic peptide: a staging system for diagnosis of cardiac amyloidosis. J Am Coll Cardiol chain cardiac amyloidosis: strategies to promote
primary systemic amyloidosis. J Clin Oncol Img 2020;13:69–80. early diagnosis and cardiac response. Heart 2017;
2004;22:3751–7. 103:1065–72.
27. French B, Wang L, Ky B, et al. Prognostic
23. Harrell FEJ, Califf RM, Pryor DB, Lee KL, value of galectin-3 for adverse outcomes in
Rosati RA. Evaluating the yield of medical tests. chronic heart failure. J Card Fail 2016;22: KEY WORDS biomarkers, cardiac
JAMA 1982;247:2543–6. 256–62. amyloidosis, hepatocyte growth factor
EDITORIAL COMMENT
Hepatocyte Growth Factor and

Cardiac Amyloidosis*
Frederick L. Ruberg, MD, Omar K. Siddiqi, MD
O nce the subject of case reports and rare dis-

ease clinical-pathological conferences, it
seems that cardiac amyloidosis is now
everywhere. Awareness of this restrictive cardiomy-
suspicion to recognize “red flag” features of symp-
toms and laboratory or echocardiographic
rangements (4). Next, appropriate testing must be
ordered and correctly interpreted. Recent multi-
de-
opathy caused by the interstitial deposition of mis- societal consensus recommendations have been
folded protein fragments has been fueled by developed that standardize both test acquisition and
revolutionary advances in diagnostic testing and interpretation, as well as frame diagnostic testing
innovative therapies. Although AL (misfolded immu- within appropriate clinical contexts (5,6). Although
noglobulin light-chain) amyloidosis remains a rare the capacity to diagnose ATTR amyloidosis non-
disease affecting 10,000 to 15,000 patients in the invasively through imaging has undoubtedly
United States (1), ATTR amyloidosis (misfolded trans- contributed to its increased recognition (7), there
thyretin or prealbumin) resulting from an inherited remains considerable heterogeneity in adherence to
variation (referred to as hATTR or ATTRv) or from guideline recommendations, resulting in test misuse
genetically normal (wild-type) TTR protein (referred and inaccurate diagnoses. A highly discriminative
blood biomarker test would represent a significant
to as ATTRwt) may affect 100,000 people or more
advancement in the approach to cardiac amyloidosis
(2). Previously untreatable, therapies for AL cardiac
diagnosis by serving as a gatekeeper whereby
amyloidosis can extend survival in most patients for
follow-on imaging testing could be triggered or
many years and for some for well over a decade (3).
avoided.
Similarly, ATTR amyloidosis, once treatable only by
Circulating biomarkers are attractive as screening
organ transplantation until a few years ago, now is
tests for cardiac amyloidosis because they lack the
primarily managed by highly effective pharmaceu-
inherent subjectivity of imaging and are easy to ac-
tical therapies that stabilize the TTR protein or
quire and easy to interpret. Free light-chain concen-
silence its production (2), resulting in symptomatic
tration in AL amyloidosis identifies systemic disease
improvement and increased survival.
but does not specify organ involvement. Cardiac-
Diagnosis of cardiac amyloidosis still requires that
specific biomarkers including natriuretic peptides
the astute clinician exercise a high degree of
(B-type natriuretic peptide [BNP] and N-terminal pro-
B type natriuretic peptide [NTpro-BNP]) and tropo-
nins (troponin I and T) are useful for prognostication
authors and do not necessarily represent the view of JACC: in both AL (8,9) and ATTR (10), but are not specific
CardioOncology or the American College of Cardiology. diagnostically. In ATTR, there is evidence that pre-
From the Section of Cardiovascular Medicine, Department of Medicine, albumin (or TTR) concentration may also prove
Amyloidosis Center, Boston University School of Medicine, Boston prognostic (11) and appears to be a means to follow
Medical Center, Boston, Massachusetts. Dr. Ruberg is supported by the
response to ATTR-specific therapy (12), but its role as
National Institutes of Health R01 HL139671; and has received research
grants from Pfizer and Eidos Therapeutics. Dr. Siddiqi has reported that
a diagnostic screening test for cardiac amyloidosis
he has no relationships relevant to the contents of this paper to disclose. remains unproven. At present, only retinol binding
The authors attest they are in compliance with human studies commit- protein 4 appears useful as a specific biomarker
screening test for hATTR amyloidosis, but one that
appropriate. For more information, visit the JACC: CardioOncology author
functions best when incorporated into a prediction
instructions page. model that requires other diagnostic test results

68 Ruberg and Siddiqi JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020
Hepatocyte Growth Factor and Cardiac Amyloidosis MARCH 2020:67–9
(electrocardiogram and echocardiography) (13). No and class 3 to 4 heart failure; n ¼ 42). Baseline plasma
biomarker appears to confer both reliable diagnostic markers were then tested for association with a
and prognostic information. composite clinical outcome endpoint of all-cause
In this context, one can consider the report by mortality, cardiac transplantation, or left ventricular
Zhang et al. (14) in this issue of JACC: CardioOncology assist device implantation over a median observation
as a potential major advancement if validated in of 2.6 years (interquartile range: 1.9 to 3.1 years). A
larger studies. Hepatocyte growth factor (HGF) was total of 30 events occurred, nearly entirely mortality
first described as a mitogen for hepatic cells but is (n ¼ 27).
expressed in mesenchymal cell derivatives (including The authors found that HGF concentrations were
cardiomyocytes) as, among other actions, a pro- higher in cardiac amyloidosis as compared with all
angiogenic cytokine. Increased circulating concen- other groups, and that this association persisted after
trations of HGF, attributed to response to injury, have adjustment for baseline differences in age (cardiac
been associated with mortality in heart failure (15), amyloidosis patients were older), septal thickness,
whereas transfection of HGF into myocytes in animal and LVEF. Using receiver operator characteristic
models of post-infarction cardiomyopathy improved analysis, an HGF of >205 pg/ml conferred an 86%
left ventricular (LV) remodeling (16). Zhang et al. (14) sensitivity (95% confidence interval [CI]: 78% to
postulated that in the context of amyloid fibril 94%), 84% specificity (95% CI: 76% to 92%), and area
deposition, HGF would be upregulated and measur- under the curve (AUC) of 0.88 (95% CI: 0.83 to 0.94)
able in the circulation, as has been reported in smaller for the differentiation of cardiac amyloidosis from
studies (17), thus potentially representing a specific the 2 non-amyloid control groups. HGF did not differ
marker for cardiac amyloidosis. between cardiac amyloidosis owing to AL or ATTR.
SEE PAGE 56
In respect to predictive capacity, a baseline HGF
at a threshold of 320 pg/ml was associated with
In this retrospective, single-center cohort study, the composite outcomes similar to that of baseline
plasma biomarkers including natriuretic peptides, NT-proBNP and troponin T using the Mayo 2004
troponins, HGF, galectin-3, and interleukin-6 were classification thresholds. The addition of HGF to NT-
measured in 102 patients with systemic amyloidosis proBNP and troponin T significantly improved the
and 86 controls. The systemic amyloidosis cohort was predictive model for outcomes. Galectin-3, although
further categorized as cardiac amyloidosis (n ¼ 72 of lower in cardiac amyloidosis as compared with the 2
whom 55% had endomyocardial biopsy) and non-amyloid control groups, proved a poorer
amyloidosis without cardiac involvement (n ¼ 30). discriminator as compared with HGF.
Noncardiac biopsy assessment of cardiac involvement As a means to identify cardiac involvement in the
required an extra-cardiac biopsy showing amyloid in systemic amyloidosis group, measurement of the
conjunction with either increased wall thickness established biomarkers NT-proBNP and troponin-T
based on echocardiography, low-voltage electrocar- discriminated cardiac from non-cardiac amyloidosis
diogram pattern or characteristic late gadolinium as predicted. HGF also similarly discriminated cardiac
enhancement noted based on cardiac magnetic reso- involvement (AUC: 0.69; 95% CI: 0.55 to 0.83), but
nance. The looseness of this definition comprises a proved inferior to either NT-proBNP (AUC: 0.89; 95%
relative weakness of the study and would tend to CI: 0.80 to 0.97) and troponin-T (AUC: 0.83; 95% CI:
equilibrate the two amyloid subgroups, reducing the 0.73 to 0.93). As a means to identify outcomes, the
capacity to identify differences. The amyloid patient authors performed a sensitivity analysis to evaluate
group was mixed with respect to type but predomi- the capacity of HGF to improve established prediction
nantly consisted of light-chain disease in 67% of the models using NT-proBNP and troponin T among the
total cohort, while 60% of the cardiac amyloidosis patients with amyloidosis. Significant improvement
subgroup had AL amyloidosis. It is notable that 88% in the predictive model was observed for AL but not
of the patients with amyloidosis were Caucasian, ATTR, perhaps related to lack of adequate power.
suggesting that the conclusions of this study may not Interestingly, lower eGFR did not associate with
be generalizable to the population with pV142I he- worse outcomes, likely owing to the smaller compo-
reditary ATTR amyloidosis (an allele noted in 3.4% of nent of patients with ATTR.
African Americans). The control groups were There are a number of important limitations to this
composed of patients with normal LV systolic func- study that merit discussion. First, and foremost, this
tion but with hypertrophy (by mass index and relative is a relatively small single-center study with results
wall thickness; n ¼ 44) and heart failure with reduced that should be seen as hypothesis-generating,
left ventricular ejection fraction (LVEF) (LVEF #30% requiring validation in larger cohorts (as the authors
JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020 Ruberg and Siddiqi 69
MARCH 2020:67–9 Hepatocyte Growth Factor and Cardiac Amyloidosis
are intending). Second, because this was clearly a hypertrophy control group is well selected, heart
dataset comprised through retrospective review and failure with reduced ejection fraction as characterized
subject to variation in clinical practice (different here is really not a phenotypic match for cardiac
troponin and BNP assays, for example), there was a amyloidosis. Finally, the rationale for selecting HGF
considerable amount of missing data. To address this as a marker in cardiac amyloidosis, although briefly
shortcoming, the authors repeated analyses with touched upon by the authors, is unclear with only
multiple imputation datasets demonstrating consis- speculative mechanisms. Although well beyond the
tent results. Third, AL and ATTR are diseases with scope of this study, pathophysiological evidence of
distinct courses and treatments. It is difficult to HGF relevance to cardiac amyloidosis from cellular or
interpret these survival results without treatment animal-based studies would support these findings.
response information, and, furthermore, because Will HGF prove to be that elusive biomarker that
clinical courses are different, outcomes for these 2 might specifically identify cardiac amyloidosis and
amyloid types cannot be equitably compared. Fourth, confer information regarding prognosis? We eagerly
various HGF thresholds are presented (205 pg/ml to await the answer.
differentiate cardiac amyloidosis from other causes of
heart failure, 320 pg/ml to identify cardiac involve- ADDRESS FOR CORRESPONDENCE: Dr. Frederick L.
ment in systemic amyloidosis, 676 pg/ml to best Ruberg, Section of Cardiovascular Medicine, Boston Med-
discriminate outcomes) creating confusion as to what ical Center, 72 East Concord Street, Boston, Massachu-
to use clinically. As larger datasets become available, setts 02118. E-mail: frruberg@bu.edu. Twitter:
a single value or range affording clinical utility may @BUMedicine, @omarsiddiqi, @frederickruberg,
become evident. Fifth, while the left ventricular @BU_Amyloidosis.
REFERENCES
1. Quock TP, Yan T, Chang E, Guthrie S, Broder MS. 7. Lane T, Fontana M, Martinez-Naharro A, et al. 13. Arvanitis M, Koch CM, Chan GG, et al. Iden-
Epidemiology of AL amyloidosis: a real-world study Natural history, quality of life, and outcome in tification of transthyretin cardiac amyloidosis
using US claims data. Blood Adv 2018;2:1046–53. cardiac transthyretin amyloidosis. Circulation using serum retinol-binding protein 4 and a
2019;140:16–26. clinical prediction model. JAMA Cardiol 2017;2:
2. Ruberg FL, Grogan M, Hanna M, Kelly JW,
305–13.
Maurer MS. Transthyretin amyloid cardiomyopa- 8. Lilleness B, Ruberg FL, Mussinelli R, Doros G,
thy: JACC state-of-the-art review. J Am Coll Car- Sanchorawala V. Development and validation of a 14. Zhang KW, Miao J, Mitchell JD, et al. Plasma
diol 2019;73:2872–91. survival staging system incorporating BNP in pa- hepatocyte growth factor for diagnosis and
tients with light chain amyloidosis. Blood 2019; prognosis in light chain and transthyretin cardiac
3. Merlini G, Dispenzieri A, Sanchorawala V, et al.
133:215–23. amyloidosis. J Am Coll Cardiol CardioOnc 2020;2:
Systemic immunoglobulin light chain amyloidosis.
56–66.
Nat Rev Dis Primers 2018;4:38. 9. Kumar S, Dispenzieri A, Lacy MQ, et al. Revised
prognostic staging system for light chain 15. Rychli K, Richter B, Hohensinner PJ, et al.
4. Witteles RM, Bokhari S, Damy T, et al. Screening
amyloidosis incorporating cardiac biomarkers and Hepatocyte growth factor is a strong predictor of
for transthyretin amyloid cardiomyopathy in
serum free light chain measurements. J Clin Oncol mortality in patients with advanced heart failure.
everyday practice. J Am Coll Cardiol HF 2019;7:
2012;30:989–95. Heart 2011;97:1158–63.
709–16.
10. Gillmore JD, Damy T, Fontana M, et al. A new 16. Li Y, Takemura G, Kosai K, et al. Postinfarction
5. Dorbala S, Ando Y, Bokhari S, et al. ASNC/AHA/
staging system for cardiac transthyretin amyloid- treatment with an adenoviral vector expressing
ASE/EANM/HFSA/ISA/SCMR/SNMMI expert
osis. Eur Heart J 2018;39:2799–806. hepatocyte growth factor relieves chronic left
consensus recommendations for multimodality
ventricular remodeling and dysfunction in mice.
imaging in cardiac amyloidosis: part 1 of 2- 11. Hanson JLS, Arvanitis M, Koch CM, et al. Use of
Circulation 2003;107:2499–506.
evidence base and standardized methods of im- serum transthyretin as a prognostic indicator and
aging. J Card Fail 2019;25:e1–39. predictor of outcome in cardiac amyloid disease 17. Abraham J, Desport E, Rigaud C, et al. Hepa-
associated with wild-type transthyretin. Circ Heart tocyte growth factor measurement in AL
6. Dorbala S, Ando Y, Bokhari S, et al. ASNC/AHA/
Fail 2018;11:e004000. amyloidosis. Amyloid 2015;22:112–6.
ASE/EANM/HFSA/ISA/SCMR/SNMMI expert
consensus recommendations for multimodality 12. Judge DP, Heitner SB, Falk RH, et al. Trans-
imaging in cardiac amyloidosis: part 2 of 2- thyretin stabilization by AG10 in symptomatic
diagnostic criteria and appropriate utilization. transthyretin amyloid cardiomyopathy. J Am Coll KEY WORDS cardiac amyloidosis, hepatocyte
J Card Fail 2019;25:854–65. Card 2019;74:285–95. growth factor
ORIGINAL RESEARCH
Cardiovascular Safety of Degarelix Versus

Leuprolide for Advanced Prostate Cancer
The PRONOUNCE Trial Study Design
Chiara Melloni, MD, MHS,a Susan F. Slovin, MD, PHD,b Allan Blemings, MS,c Shaun G. Goodman, MD, MSC,d,e
Christopher P. Evans, MD,f Jan Nilsson, MD,g Deepak L. Bhatt, MD, MPH,h Konstantin Zubovskiy, MD,i
Tine K. Olesen, MS, MBA,i Klaus Dugi, MD,j Noel W. Clarke, MBBS, CHM,k Celestia S. Higano, MD,l
Matthew T. Roe, MD, MHS,a on behalf of the PRONOUNCE Investigators
ABSTRACT
OBJECTIVES This study will compare the incidence of major adverse cardiovascular events (MACEs) with androgen
deprivation therapy (ADT) among men with advanced prostate cancer who are being treated with a gonadotropin-
releasing hormone (GnRH) antagonist versus a GnRH agonist.
BACKGROUND Treatment of advanced prostate cancer with ADT might increase the risk of subsequent cardiovascular
events among men with known atherosclerotic cardiovascular disease (ASCVD), but a recent meta-analysis suggested
that this risk might be lower with ADT using a GnRH antagonist versus a GnRH agonist.
METHODS PRONOUNCE is a multicenter, prospective, randomized, open, blinded endpoint trial that will enroll approxi-
mately 900 patients with advanced prostate cancer and pre-existing ASCVD who will be treated with ADT. Participants will
be randomized to receive the GnRH antagonist degarelix or the GnRH agonist leuprolide as ADT for 12 months. The primary
endpoint is time from randomization to first confirmed, adjudicated occurrence of a MACE, which is defined as a composite of
all-cause death, nonfatal myocardial infarction, or nonfatal stroke through 12 months of ADT treatment. Baseline cardio-
vascular biomarkers (high-sensitivity C-reactive protein, high-sensitivity troponin T, and N-terminal pro-brain natriuretic
peptide), as well as serial inflammatory and immune biomarkers, will be evaluated in exploratory analyses.
RESULTS As of October 1, 2019, a total of 364 patients have been enrolled. The mean age is 74 years, 90% are white, 80%
have hypertension or dyslipidemia, 30% diabetes mellitus, 40% have had a previous myocardial infarction, and 65% have
had previous revascularization. Regarding prostate cancer features at randomization, 48% of the patients had localized
disease, 23% had locally advanced disease, and 18% had metastatic disease.
CONCLUSIONS PRONOUNCE is the first prospective cardiovascular outcomes trial in advanced prostate cancer
that will delineate whether the risk of subsequent cardiovascular events associated with ADT is lower with a
GnRH antagonist versus a GnRH agonist for men with pre-existing ASCVD. (A Trial Comparing Cardiovascular Safety of
Degarelix Versus Leuprolide in Patients With Advanced Prostate Cancer and Cardiovascular Disease [PRONOUNCE];
NCT02663908) (J Am Coll Cardiol CardioOnc 2020;2:70–81) © 2020 The Authors. Published by Elsevier on behalf of
From the aDepartment of Medicine, Division of Cardiology, Duke Clinical Research Institute, Duke University Medical Center,
Durham, North Carolina; bDepartment of Medicine, Division of Medical Oncology, Memorial Sloan Kettering Cancer Center, New
York, New York; cFerring Pharmaceuticals A/S, Copenhagen, Denmark; dDepartment of Medicine, Division of Cardiology, St.
Michael’s Hospital, University of Toronto, Toronto, Ontario, Canada; eCanadian VIGOUR Centre, University of Alberta, Edmonton,
Alberta, Canada; fDepartment of Urologic Surgery, University of California, Davis, Sacramento, California; gDepartment of Clinical
Sciences Malmö, Lund University, Lund, Sweden; hDivision of Cardiovascular Medicine, Brigham and Women’s Hospital Heart
and Vascular Center, Harvard Medical School, Boston, Massachusetts; iFerring Pharmaceuticals A/S, Parsippany, New Jersey;

JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020 Melloni et al. 71
MARCH 2020:70–81 The PRONOUNCE Design and Rationale
T he known concurrence of cancer and athero- particularly among older cancer survivors ABBREVIATIONS
sclerotic cardiovascular disease (ASCVD) (4,5,7–9). More specifically, prostate cancer AND ACRONYMS
in certain patient populations has been is the most common form of cancer among
ASCVD = atherosclerotic
informed by improved methods of early cancer detec- older men and has an increasing occurrence cardiovascular disease
tion and integrated treatment approaches that have with older age; similar age-related trends
ADT = androgen deprivation
resulted in significant cancer-related survival gains have been found in the prevalence and con- therapy
over the past few decades (1–3). With increasing sur- sequences of ASCVD (10,11). As a result, CI = confidence interval
ASCVD has become the second most com- DSMB = Data Safety
SEE PAGE 82
mon cause of death among men with pros- Monitoring Board
vival trends associated with cancer, the competing tate cancer (12,13). GnRH = gonadotropin-
risks of downstream morbidity and mortality for pa- For locally advanced, relapsed, or meta- releasing hormone
tients with cancer may be more influenced by static prostate cancer, androgen deprivation MACE = major adverse
cardiovascular event
concomitant ASCVD (when present) than from the therapy (ADT) is the backbone of treatment,
incident type of cancer (4–6). As a result, ASCVD alone or in combination with radiation therapy or
has emerged as the predominant cause of mortality, other agents because of the role of testosterone in the
j
Ferring Pharmaceuticals A/S, Saint-Prex, Switzerland; kDivision of Urology, Institute of Cancer Sciences, University of Man-
chester, United Kingdom; and the lDivision of Medical Oncology, University of Washington and Fred Hutchinson
Cancer Research Center, Seattle, Washington. The PRONOUNCE trial is supported by Ferring Pharmaceuticals Inc. Dr. Melloni
has received research grant funding through Duke University for conduct of the PRONOUNCE trial; and has received research grant
or contract from Amgen, AstraZeneca, Bristol-Myers Squibb, Ferring Pharmaceuticals, GlaxoSmithKline, IntraCellular Therapies,
Luitpold Pharmaceuticals, Merck, Roche, Sanofi, St. Jude Medical, and Pfizer. Dr. Slovin has been on the advisory boards for Amgen,
Bayer, OncLive, and PER. Dr. Goodman has received research grant support (e.g., steering committee or data monitoring committee)
and/or speaker/consulting honoraria (e.g., advisory boards) from Ferring Pharmaceuticals, Amgen, AstraZeneca, Bayer, Boehringer
Ingelheim, Bristol-Myers Squibb, CSL Behring, Daiichi-Sankyo, Eli Lilly, Esperion, Fenix Group International, GlaxoSmithKline, HLS
Therapeutics, Janssen/Johnson & Johnson, Luitpold Pharmaceuticals, Matrizyme, Merck, Novartis, Novo Nordisk A/C, Pfizer,
Regeneron, Sanofi, Servier, Tenax Therapeutics, Heart and Stroke Foundation of Ontario/University of Toronto, Canadian Heart
Research Centre and MD Primer, Canadian VIGOUR Centre, Duke Clinical Research Institute, and PERFUSE. Dr. Evans has received
research support from Astellas/Pfizer, Ferring, Sanofi, Janssen, National Institutes of Health, SU2C/AACR (Stand Up for cancer/
American Association for Cancer Research), Department of Defense, and Prostate Cancer Foundation; has been a consultant for
Astellas/Pfizer and Janssen; has been a member of the Speakers Bureau for Astellas/Pfizer, Janssen; has received honoraria from
Astellas/Pfizer, Janssen, and MDxHealth; and has been a member of the Scientific Advisory Board for Astellas, Janssen, and
MDxHealth. Dr. Nilsson has received unrestricted research support from Ferring Pharmaceuticals, Novo Nordisk A/S, Follicum AB,
and Medimmune. Dr. Bhatt has been a member of the Advisory Board for Cardax, Cereno Scientific, Elsevier Practice Update Car-
diology, Medscape Cardiology, PhaseBio, and Regado Biosciences; has been a member of the Board of Directors for Boston VA
Research Institute, Society of Cardiovascular Patient Care, and TobeSoft; has been the chair of the American Heart Association
Quality Oversight Committee; has been a member of the Data Monitoring Committees for Baim Institute for Clinical Research
(formerly Harvard Clinical Research Institute, for the PORTICO trial, funded by St. Jude Medical, now Abbott), Cleveland Clinic
(including for the ExCEED trial, funded by Edwards), Duke Clinical Research Institute, Mayo Clinic, Mount Sinai School of Medicine
(for the ENVISAGE trial, funded by Daiichi Sankyo), and Population Health Research Institute; has received honoraria from the
American College of Cardiology (Senior Associate Editor, Clinical Trials and News, ACC.org; Vice-Chair, ACC Accreditation Com-
mittee), Baim Institute for Clinical Research (formerly Harvard Clinical Research Institute, RE-DUAL PCI clinical trial steering
committee funded by Boehringer Ingelheim, AEGIS-II executive committee funded by CSL Behring), Belvoir Publications (Editor in
Chief, Harvard Heart Letter), Duke Clinical Research Institute (clinical trial steering committees, including for the PRONOUNCE trial,
funded by Ferring Pharmaceuticals), HMP Global (Editor in Chief, Journal of Invasive Cardiology), Journal of the American College of
Cardiology (Guest Editor; Associate Editor), Medtelligence/ReachMD (CME steering committees), Population Health Research
Institute (for the COMPASS operations committee, publications committee, steering committee, and USA national co-leader, funded
by Bayer), Slack Publications (Chief Medical Editor, Cardiology Today’s Intervention), Society of Cardiovascular Patient Care (Sec-
retary/Treasurer), WebMD (CME steering committees); and has received other funding from Clinical Cardiology (Deputy Editor),
NCDR-ACTION Registry Steering Committee (Chair), VA CART Research and Publications Committee (Chair); has received research
funding from Abbott, Afimmune, Amarin, Amgen, AstraZeneca, Bayer, Boehringer Ingelheim, Bristol-Myers Squibb, Chiesi, CSL
Behring, Eisai, Ethicon, Ferring Pharmaceuticals, Forest Laboratories, Idorsia, Ironwood, Ischemix, Lilly, Medtronic, PhaseBio,
Pfizer, Regeneron, Roche, Sanofi Aventis, Synaptic, The Medicines Company; has received royalties from Elsevier (Editor, Cardio-
vascular Intervention: A Companion to Braunwald’s Heart Disease); has been the site co-investigator for Biotronik, Boston Scientific,
St. Jude Medical (now Abbott), and Svelte; has been a trustee for the American College of Cardiology; and has received unfunded
research from FlowCo, Fractyl, Merck, Novo Nordisk, PLx Pharma, and Takeda. Dr. Higano has been a member of the Advisory Boards
for Aptevo, Asana, Astellas, Bayer, Blue Earth Diagnostics, Churchill Pharma, Clovis Oncology, Dendreon, Endocyte, Ferring, Hinova,
Janssen, Myriad, Orion Corporation, and Pfizer; has been a consultant for Astellas, Bayer, Blue Earth Diagnostics, Carrick Thera-
peutics, Clovis, Ferring Pharmaceuticals, Janssen, Hinova, Merck, Novartis, Pfizer, Asana, Aptevo, Churchill Pharma, Dendreon,
Myriaad, and Orion; and has done sponsored research for Aptevo, Aragon, Astellas, AstraZeneca, Bayer, Clovis, Dendreon,
eFFECTOR Therapeutics, Emergent, Ferring, Genentech, Hoffman-Laroche, Medivation, Sanofi, and Pfizer. Dr. Roe has received
research grant funding from Sanofi-Aventis, AstraZeneca, Patient Centered Outcomes Research Institute, Ferring Pharmaceuticals,
72 Melloni et al. JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020
The PRONOUNCE Design and Rationale MARCH 2020:70–81
F I G U R E 1 Mechanism of Action
Mechanism of action of GnRH agonist and antagonist. FSH ¼ follicle-stimulating hormone; GnRH ¼ gonadotropin-releasing hormone; LH ¼ luteinizing hormone.
development and progression of prostate cancer (14). “testosterone flare.” Long-term exposure to a GnRH
Medical castration with ADT is accomplished with agonist eventually shuts down luteinizing hormone
either a gonadotropin-releasing hormone (GnRH) and follicle-stimulating hormone; consequently,
agonist or a GnRH antagonist because both methods testicular production of testosterone is stopped. In
effectively achieve a castration level of testicular contrast, GnRH antagonists inhibit pituitary GnRH
testosterone suppression. Whether the different receptors, which immediately shuts down luteinizing
mechanisms of action between these 2 agents might hormone secretion, which leads to subsequent sup-
have differing effects on off-target tissues (e.g., pression of testosterone production without an
atherosclerotic plaque) remains unknown (15). Initial associated testosterone flare (Figure 1).
administration of a GnRH agonist causes luteinizing Over the past 2 decades, several observational
hormone and follicle-stimulating hormone release, studies have demonstrated an association between
which results in an increase in serum testosterone, or the use of ADT and an increased risk of thrombotic
Myokardia, Familial Hypercholesterolemia Foundation, Bayer; and has been a consultant and received honoraria from AstraZeneca,
Amgen, Cytokinetics, Eli Lilly, Roche-Genentech, Janssen Pharmaceuticals, Regeneron, Novo Nordisk, Pfizer, Sanofi-Aventis, Signal
Path, and Elsevier Publishers. All other authors have reported that they have no relationships relevant to the contents of this paper to
disclose.
The authors attest they are in compliance with human studies committees and animal welfare regulations of the authors’ institutions
and Food and Drug Administration guidelines, including patient consent where appropriate. For more information, visit the JACC:
CardioOncology author instructions page.
Manuscript received September 10, 2019; revised manuscript received January 15, 2020, accepted January 16, 2020.
T A B L E 1 Inclusion Criteria
Main inclusion criteria for prostate cancer

Histologically confirmed adenocarcinoma of the prostate
Tumor, node, metastasis staging available before treatment start (bone scan and/or CT scan and/or MRI) <12 weeks before study start.
If no radiographic image is available at the time of screening, a bone scan should be performed
Investigator judgement to initiate continued ADT therapy with intended duration of 12 months or longer.
Patients with metastatic prostate cancer at time of diagnosis
Patients with prostate cancer who develop metastases after local therapy
Patients with prostate cancer with very high-risk, high-risk, or intermediate risk disease with feature of unfavorable prognosis who will be treated
with definitive radiation therapy in combination with at least 12 months of neoadjuvant/adjuvant ADT
Patients must be treatment-naive (ADT)
If patients received previous ADT for neoadjuvant or adjuvant therapy, then the last dose of therapy must be at least 12 months before
randomization
Any additional hormonal therapy upfront (i.e., abiraterone) is prohibited in the study; however, anti-androgen use for initial flare protection is
allowed for a maximum period of up to 28 days after randomization
Main cardiovascular inclusion criteria
Pre-existing ASCVD (confirmed diagnosis, documented) according to at least 1 of the following criteria
Previous myocardial infarction $30 days before randomization
Previous revascularization procedure $30 days before randomization
Coronary artery: stent placement/balloon angioplasty or coronary artery bypass graft surgery
Carotid artery: stent placement/balloon angioplasty or endarterectomy surgery
Iliac, femoral, popliteal arteries: stent placement/balloon angioplasty or vascular bypass surgery
At least 1 vascular stenosis $50% at any time point before randomization by angiography or CT angiography
Coronary artery
Carotid artery
Iliac, femoral, or popliteal arteries
Carotid ultrasound results that documented a vascular stenosis $50% at any time point before randomization
Anklebrachial pressure index <0.9 at any time point before randomization
ADT ¼ androgen deprivation therapy; ASCVD ¼ atherosclerotic cardiovascular disease; CT ¼ computed tomography; MRI ¼ magnetic resonance imaging.
cardiovascular events, including pulmonary embo- provided from a meta-analysis of pooled data from 6
lism, myocardial infarction, and cardiovascular- prospective randomized clinical trials that compared
related mortality (16–18). Findings from the 2 initial short-term (up to 12 months) treatment of advanced
studies that demonstrated this association analyzed prostate cancer with a GnRH antagonist (degarelix)
data from the Surveillance Epidemiology and End versus a GnRH agonist (leuprolide). In the overall
Results-Medicare linked database and identified an population, the risk of cardiovascular events was
increased risk of incident coronary heart disease, similar between the treatment arms. However, in
myocardial infarction, and cardiovascular death those with pre-existing ASCVD before the start of
among men with prostate cancer treated with a GnRH ADT, the frequency of cardiovascular events was
agonist (19,20). Although this signal was not uni- substantially lower with the GnRH antagonist versus
formly observed in other studies published at that the GnRH agonist (6.5% vs. 14.7%, respectively) with
time, these new data led to the 2010 publication of a a separation of event curves that occurred approxi-
joint scientific statement from the American Heart mately 3 to 6 months after ADT initiation that lasted
Association, American Cancer Society, and American throughout the first year of treatment (23).
Urological Association that suggested a possible as- Prompted by this additional evidence that
sociation between ADT and risk of cardiovascular demonstrated a potential differential risk of cardio-
events (16,21). Shortly after this publication, the U.S. vascular events by type of ADT when treatment was
Food and Drug Administration and Health Canada administered for up to 12 months, the PRONOUNCE (A
asked manufacturers of GnRH agonists to add extra Trial Comparing Cardiovascular Safety of Degarelix
safety information to drug labels with a warning Versus Leuprolide in Patients With Advanced Pros-
about the possible increased risks for cardiovascular tate Cancer and Cardiovascular Disease) trial was
events. Although the Food and Drug Administration designed to rigorously and prospectively evaluate the
did not request a label change for GnRH antagonists cardiovascular safety of a GnRH receptor antagonist
(22), the European Medical Agency requested similar (degarelix) versus a GnRH receptor agonist (leupro-
label warnings for both GnRH agonists and GnRH lide) among men with prostate cancer who had pre-
antagonists. Thereafter, further insights were existing ASCVD. To our knowledge, the
STUDY POPULATION. Eligible patients must have a

T A B L E 2 Main Exclusion Criteria
pathological diagnosis of adenocarcinoma of the
Main prostate cancer exclusion criteria
prostate with newly diagnosed localized disease,
Previous or current hormonal management of prostate cancer
biochemical recurrence after definitive therapy, or
Surgical castration
Any hormonal manipulation
hormone-sensitive metastatic disease. If a participant
Any previous neoadjuvant/adjuvant hormonal therapy, unless treatment terminated received previous ADT for neoadjuvant or adjuvant
>12 months before study start therapy, then the last dose of therapy must have been
Main cardiovascular exclusion criteria
at least 12 months before randomization. Serum
Uncontrolled type 1 or type 2 diabetes mellitus (defined as HbA1c >10%) at time of
randomization
testosterone level must be in the non-castration
Uncontrolled hypertension (SBP >180 mm Hg or DBP >110 mm Hg) at time of range, and the duration of planned ADT must be at
randomization least 12 months. In addition, eligible patients must
A history of congenital long QT syndrome or risk factors for Torsade de pointes
have pre-existing ASCVD defined as: a history of
ventricular arrhythmias (e.g., heart failure, hypokalemia, concomitant medication
known to cause QT prolongation) myocardial infarction; previous percutaneous or
Within 30 days before randomization: surgical revascularization of the carotid, coronary,
Myocardial infarction iliac, femoral, or popliteal arteries; previous docu-
Stroke (hemorrhagic/ischemic)
mentation of a stenosis of >50% in these vessels by
Coronary, carotid, or peripheral artery revascularization
angiography or carotid ultrasound; or peripheral
Planned or scheduled cardiac surgery or PCI procedure that is known at the time of
randomization arterial disease (PAD) confirmed with a diminished
ankle-brachial pressure index (Tables 1 and 2,
DBP ¼ diastolic blood pressure; HbA1c ¼ glycosylated hemoglobin A1c; PCI ¼ percutaneous coronary interven-
tion; SBP ¼ systolic blood pressure.
Supplemental Appendix).
During screening, potentially eligible patients are
evaluated by a local cardiovascular specialist to
PRONOUNCE trial is the first prospective cardiovas- ensure that baseline secondary prevention medica-
cular outcomes trial in advanced prostate cancer to tions for ASCVD are optimized according to guideline
compare 2 different types of cancer treatment, recommendations and to provide verification of the
thereby representing an important development in ASCVD inclusion criteria for the trial (26,27).
the field of cardio-oncology (24). Furthermore, to support sites in properly confirming
the ASCVD inclusion criteria, cardiovascular
METHODS disease information and source medical documents
for the first series of patients screened by each site are
PRONOUNCE is a phase IIIb, multicenter, prospective, reviewed centrally by a cardiologist at the Duke
randomized, open, blinded endpoint trial designed to Clinical Research Institute (Durham, North Carolina),
compare the occurrence of major adverse cardiovas- which is the academic coordinating center for the
cular events (MACEs) in patients with advanced trial. The investigators are also required to ensure
prostate cancer and pre-existing ASCVD who will that a cardiovascular specialist is treating the patients
receive either a GnRH antagonist (degarelix) or a GnRH during their trial participation to ensure optimization
agonist (leuprolide) as ADT for 12 months (25). The of secondary prevention medications for ASCVD for
trial plans to enroll approximately 900 patients at the duration of the trial.
approximately 100 sites in North America and Europe. RANDOMIZATION AND TREATMENT. Eligible pa-
The first patient was randomized in April 2016. tients are randomized 1:1 (with a fixed block size of 4)
This trial is being conducted in compliance with to degarelix or leuprolide acetate in an open-label
the study protocol, the Declaration of Helsinki, and fashion. Randomization lists are prepared by an in-
Good Clinical Practice, as defined by the International dependent statistician not involved with the trial and
Conference on Harmonization. Before patient enroll- sent to an external electronic Case Report Form
ment, written informed consent is obtained from vendor for upload to an eCRF/online randomization.
each patient, and approval is obtained from appro- Each patient receives a unique randomization num-
priate institutional review boards and ethics com- ber. Randomization is stratified by baseline age group
mittees for participating sites. The steering and (younger than 75 years vs. 75 years or older) and re-
operations committees, which include academic gion (North America vs. other geographic regions).
members and sponsor representatives, oversee the Patients randomized to degarelix receive a starting
medical, scientific, and operational conduct of the dose of 240 mg degarelix (2 subcutaneous injections,
study. The PRONOUNCE trial is supported by the each of 120 mg) followed by 11 subcutaneous in-
manufacturer of degarelix, Ferring Pharmaceuticals jections of 80 mg degarelix given at 28-day intervals.
(Parsippany, New Jersey). Patients randomized to leuprolide receive 22.5 mg
F I G U R E 2 Trial Design
The trial design, from the initial screening period through treatment and/or follow-up, and the end of the trial. CVD ¼ cardiovascular disease;
i.m. ¼ intramuscularly; MACE ¼ major adverse cardiovascular event; s.c. ¼ subcutaneously.
administered intramuscularly every 84 days, for 4 had a previous myocardial infarction, and 65% have
doses. Each patient is treated with 12 months of ADT had previous revascularization. Regarding prostate
(Figure 2). Because of the different frequency (once a cancer features at randomization, 48% of the patients
month vs. once every 3 months), modality of dosing had localized disease, 23% had locally advanced dis-
(subcutaneous vs. intramuscular), and the known ease, and 18% had metastatic disease.
difference in injection site reactions of the 2 ran- SCHEDULE OF ASSESSMENTS. A detailed schedule
domized treatment regimens, a double-blind, pla- of study assessments is delineated in Supplemental
cebo-controlled treatment design was determined to Tables 1 and 2. Suspected cardiovascular events,
be too challenging to successfully implement, as well adverse events, and use of concomitant medications
as too difficult and uncomfortable for patients are assessed at the baseline visit and during
because of the need for multiple sham injections. pre-specified monthly visits. In addition, at the
Nonetheless, several mechanisms and approaches are baseline visit, blood samples are obtained to
used to minimize bias in the ascertainment, classifi- evaluate cardiovascular biomarkers (high-sensitivity
cation, adjudication, and confirmation of suspected C-reactive protein, high-sensitivity troponin T, and
cardiovascular events as detailed in subsequent sec- N-terminal pro-brain natriuretic peptide). At each
tions (Central Illustration). Any additional hormonal monthly trial-related visit, each participant is
therapy upfront (i.e., abiraterone) is prohibited in the administered a detailed questionnaire that captures
study; however, anti-androgen use for initial flare information on potential cardiovascular events,
protection is allowed for a maximum period such as hospitalizations for cardiovascular causes,
of #28 days after randomization. If a patient pro- occurrence of angiographic and revascularization
gresses and requires additional hormonal therapy, procedures, use of brain imaging procedures, and
these subjects are subsequently excluded from the so forth. Patient responses to this monthly question-
per protocol analysis. naire prompt sites to report potential cardiovascular
As of October 1, 2019, a total of 364 patients have events and submit requisite source documents
been enrolled with 60% from North America and 40% through a standard cardiovascular endpoint reporting
from Europe. The mean age is 74 years, 90% are process. Each patient has a final clinic visit
white, approximately 80% have hypertension or 1 month after the last dose of their assigned treat-
dyslipidemia, 30% have diabetes mellitus, 40% have ment regimen.
C E N T R A L IL L U ST R A T I O N PRONOUNCE Study Population, Allocation, Follow-Up, and Analysis
Assessed for eligibility

Histologically confirmed adenocarcinoma
of the prostate
Pre-existing ASCVD
Randomized
1:1
(Approximately 900 pts)
Allocation
GnRH antagonist GnRH agonist

Degarelix Leuprolide
12 months
Follow-up
Discontinue treatment Discontinue treatment

Lost to follow up Lost to follow up
Analysis
Primary Endpoint Primary Endpoint

MACE (Death, MI and Stroke) MACE (Death, MI and Stroke)
Melloni, C. et al. J Am Coll Cardiol CardioOnc. 2020;2(1):70–81.
The study population for the PRONOUNCE trial for initial inclusion, allocation, follow-up, and analysis. ASCVD ¼ atherosclerotic cardio-
vascular disease; GnRH ¼ gonadotropin-releasing hormone; MACE ¼ major adverse cardiovascular event.
PRIMARY ENDPOINTS. The primary objective of this 336, to evaluate the progression-free survival failure
trial is to evaluate the impact of degarelix versus rates (defined as either death, radiographic disease
leuprolide on the first occurrence of a MACE progression, introduction of additional prostate can-
(all-cause death, nonfatal myocardial infarction, or cer therapies for progression, or prostate-specific
nonfatal stroke) through 12 months of ADT. antigen failure, whichever is first), and to compare
SECONDARY ENDPOINTS. The main secondary car- urinary and prostate cancer-related symptoms with
diovascular objectives are to assess the frequency of the International Prostate Symptom Score question-
the individual components of the composite MACE naire. Health economics and patient-reported
primary endpoint (myocardial infarction, stroke, and outcome objectives include comparing healthcare
all-cause death); cardiovascular-related death; and a resource use, health status through the EuroQol
composite of cardiovascular-related death, myocar- Group 5 Dimensions 5 Levels Questionnaire, func-
dial infarction, or stroke; and unstable angina. Defi- tional capacity and quality of life through the Duke
nitions of cardiovascular endpoints are standardized Activity Status Index, and heart-focused anxiety
based upon published regulatory recommendations through the Cardiac Anxiety Questionnaire. Adverse
(28). The main prostate cancer-related objectives are events are collected on a monthly basis and graded
to monitor testosterone levels on days 28, 168, and according to the National Cancer Institute Common
Terminology Criteria for Adverse Events. These received at least 1 dose of study drug will be included
events will be reported as part of the secondary in the safety analysis set according to the actual
safety assessment. treatment received.
EXPLORATORY ENDPOINTS. Exploratory endpoints The analysis of the primary MACE endpoint will be
for the comparison of degarelix with leuprolide performed for the intention-to-treat and the per-
include: 1) time from first adjudicated nonfatal MACE protocol analysis sets. The time from randomization
to a second confirmed (adjudicated) occurrence of the to the first confirmed occurrence of the composite
composite MACE endpoint in the subgroup of pa- MACE endpoint in the 2 treatment groups will be
tients that survived the first cardiovascular event, analyzed based on the Kaplan-Meier estimate of the
including an analysis of all (total) MACE events by survival function and the log-rank test stratified by
treatment arm; 2) the regression coefficient associ- age group and geographic region. The null hypothesis
ated with the interaction between treatment and of equal hazard functions between the 2 treatment
baseline cardiovascular biomarker status with respect groups will be rejected if the inverse normal test
to a Cox regression model of the time from initial statistics exceed the critical level for a 2-sided hy-
dosing to the first confirmed occurrence of the com- pothesis test with a type I error level of 5%. Unless
posite MACE endpoint; 3) the area under the receiver- otherwise specified, time-to-event endpoints will be
operating characteristics curve for cardiovascular censored at the time when a patient initiates new
biomarkers based on the prediction of event-free and/or different ADT, is lost to follow-up and/or
survival; and 4) the difference in the area under the withdraws from the study or day 336, whichever oc-
receiver-operating characteristics curves based on curs first. In addition, all hypothesis tests will be 2-
Cox regression models, including the traditional sided at a significance level of 5%, and missing data
baseline cardiovascular risk factors, and the tradi- will not be imputed.
tional cardiovascular risk factors plus cardiovascular SAFETY MONITORING. The independent, external
biomarkers (high-sensitivity C-reactive protein, high- Data and Safety Monitoring Board (DSMB) is
sensitivity troponin T, and N-terminal pro-brain composed of 1 chairperson (cardiologist), 1 indepen-
natriuretic peptide) as covariates. dent statistician, 1 cardiologist, and 1 urologist
(Supplemental Appendix). The major roles of the
STATISTICAL CONSIDERATIONS: SAMPLE SIZE
DSMB are to periodically evaluate safety data and to
CALCULATION AND ASSUMPTIONS FOR THE
perform the pre-planned, unblinded interim analysis
PRIMARY MACE ENDPOINT. Based on the aforemen-
after approximately 33 positively adjudicated MACEs
tioned published data from the meta-analysis of
have been observed. Based on pre-specified criteria,
pooled randomized clinical trial data (23), the 1-year
the DSMB will make recommendations to the Steering
MACE event rates for sample size calculations were
Committee on whether to continue the trial as is,
set to 5.1% and 10.2% for degarelix versus leuprolide,
modify (increase the sample size), or stop the trial
respectively (23). With a hypothesized hazard ratio of
due to futility according to the pre-determined
0.49 based upon previous observations, 66 MACE
criteria. The interim analysis will be based on data
events will be required at final analysis, correspond-
from the intention-to-treat analysis set.
ing to a sample size of 876 patients, to reject the null
hypothesis of equal hazards at the 2-sided 5% type I STEERING COMMITTEE. The Steering Committee
error level with 80% power. One interim analysis is consists of external clinical and scientific experts,
planned after 50% of the expected number of adju- including cardiologists, oncologists, and urologists,
dicated MACE endpoints have been collected. The as well as a Sponsor representative (Supplemental
objective at the interim analysis is to test whether Appendix). The Steering Committee will be respon-
there is any reason to stop the trial early for futility sible for overseeing trial integrity and making de-
purposes. In the event that the stopping boundaries cisions related to the trial conduct, such as potential
are not crossed for futility, the required sample size protocol amendments and decisions based on interim
will be reassessed based upon MACE event rates recommendations from the DSMB, as previously
observed at the time of the interim analysis, to ach- described. The Duke Clinical Research Institute is the
ieve a conditional power of 80% at trial conclusion. academic coordinating center for the trial and sup-
All randomized patients will be included in the ports and organizes the Steering Committee.
intention-to-treat analysis set at trial conclusion. All CLINICAL EVENT CLASSIFICATION COMMITTEE. An
analyses will be performed based on the planned independent, firewalled, blinded clinical events adju-
(randomized) treatment. The data of all patients who dication committee from the Duke Clinical Research
Institute will adjudicate all potential cardiovascular that new hormonal agents (e.g., enzalutamide, dar-
endpoints with the endpoint definitions listed in the olutamide, and others) are associated with improve-
Supplemental Appendix. Members of the adjudication ment in terms of metastases-free survival, but these
committee have no other role in the trial conduct. Po- agents come with a higher grade risk of cardiovascu-
tential MACEs will be evaluated by applying a specific lar events (36). Because new hormonal agents (e.g.,
clinical events classification process to ensure blinding abiraterone, docetaxel, and enzalutamide) all have
of the adjudicators, which includes extensive redac- their own added toxicities, including cardiovascular
tion of treatment-related details from source docu- toxicity, we found that co-administration of these
ments used during the adjudication processes. Each agents is not always in the best interest of patients
potential event will be adjudicated independently by 2 with pre-existing cardiovascular disease or diabetes.
physician adjudicators, and disagreements will be In this study, we want to minimize any outside tox-
reviewed by a committee of the adjudicators. Cardiol- icities that can be contributed by these drugs; there-
ogists will adjudicate death, myocardial infarction, fore, new hormonal agents are prohibited in the
and unstable angina events; neurologists will review PROUNOUNCE trial.
stroke events; and oncologists will participate in the Recently, results from a small, investigator-
cause of death adjudication activities together initiated, open-label trial that randomized 80 men
with cardiologists. with prostate cancer and pre-existing ASCVD to
receive degarelix versus leuprolide for 12 months
DISCUSSION demonstrated that the number of cardiovascular
events was lower with degarelix versus leuprolide (a
The primary goal of the PRONOUNCE study is to pre-specified secondary endpoint) (37). Although the
prospectively evaluate the occurrence of MACE results from this study are hypothesis-generating and
events in patients with advanced prostate cancer and not definitive, these findings further support the
concomitant ASCVD who are treated with a GnRH rationale for an adequately powered, prospective trial
antagonist degarelix compared with the GnRH such as PRONOUNCE to definitively ascertain the
agonist leuprolide over the initial 12 months of ADT. relative cardiovascular safety of ADT with a GnRH
In addition, several non-cardiovascularrelated sec- antagonist versus a GnRH agonist.
ondary endpoints will be studied to further investi- Mechanisms of increased cardiovascular risk with
gate the potential differences in the riskbenefit ADT appear to be multifactorial and are believed to be
profile of degarelix versus leuprolide. Our main hy- related to both metabolic and immunomodulatory
pothesis is that patients treated with degarelix will changes that may destabilize pre-existing athero-
have a lower risk of cardiovascular events than those sclerotic plaques and potentially accelerate the pro-
treated with leuprolide. The PRONOUNCE trial is the gression of atherosclerosis.
first prospective cardiovascular outcomes trial to Pre-clinical studies of androgen-receptor knockout
evaluate different cancer treatments. and orchiectomized low-density lipoprotein
Although ADT is the mainstay treatment of receptor knockout models demonstrated that andro-
advanced prostate cancer, post hoc analyses have gens could exert both favorable direct and indirect
demonstrated an association between treatment with effects on the development and progression of
a GnRH receptor agonist and an increased risk of atherosclerotic lesions (38). The administration of
downstream cardiovascular events (20,29,30). For the testosterone to animal and cell models of athero-
most part, cardiovascular events occurred early after sclerosis showed a decrease in the expression of
ADT initiation (typically after 1 to 4 months of expo- vascular cell adhesion molecules (e.g., VCAM-1) and
sure), which could suggest a short-term, treatment- pro-inflammatory cytokines (tumor necrosis factor-
related risk for aggravation or destabilization of alpha, interleukin-1) (39,40).
existing atherosclerotic plaques (16,29–31). At the After post hoc analyses indicated that GnRH re-
same time, other analyses have shown that although ceptor agonists and GnRH receptor antagonists
the risk seems to peak in the first 6 months of treat- might have different levels of cardiovascular risk,
ment, the event curve continues to diverge over the qualitative difference in the mechanism of action
longer follow-up (20,32). In line with these observa- between GnRH receptor antagonists and GnRH re-
tions, studies have consistently shown that a history ceptor agonists, including the effect on the follicle-
of ASCVD is strongly associated with subsequent stimulating hormone, as well as potentially func-
cardiovascular complications during ADT therapy tional GnRH receptors identified in peripheral tis-
(33–35). A recent meta-analysis has also demonstrated sues, was investigated. These studies raised the
possibility that GnRH receptor agonists and GnRH mechanisms that may underlie the increased cardio-
receptor antagonists might have different profiles vascular risk observed with ADT.
with respect to short-term cardiovascular safety in
patients with established ASCVD (23,41). For CONCLUSIONS
example, T cells present in atherosclerotic plaque
may express GnRH receptors, and, consequently, Cardio-oncology involves caring for patients with
may be stimulated by a GnRH agonist, thereby cancer who have concomitant cardiovascular disease
potentially promoting fibrotic cap disruption and at the time of cancer diagnosis or who develop car-
plaque destabilization (23). diovascular disease during increasingly more com-
Although the pathophysiological mechanisms of plex and efficacious, but also potentially more cardio-
potential differential cardiovascular risk with a GnRH toxic, cancer treatment. Due to the frequent co-
antagonist versus GnRH agonist remain to be fully existence of cardiovascular disease and cancer, the
elucidated, putative differences in mechanisms of field of cardio-oncology is rapidly expanding. Effec-
action of these agents may underlie the observed tive communication and collaboration between
findings previously mentioned. different specialty providers (oncologists, urologists,
There are several distinctive features that make the and cardiologists) as essential partners in a care team
PRONOUNCE trial unique. First, a multispecialty is critical to balance cancer care and cardiovascular
group of cardiologists, urologists, and oncologists outcomes toward optimal survival. Cooperation
work together on the Steering Committee and DSMB among specialties also has regulatory repercussions,
(42). Second, numerous strategies have been imple- where different divisions at the Food and Drug
mented to support investigators (mainly urology/ Administration and European Medical Agency must
oncology investigators) and to help in the assessment collaborate to oversee and approve these types of
of cardiovascular inclusion criteria. Several training trials. In this context, the PRONOUNCE trial is the first
sessions covering cardiovascular disease definitions randomized trial that is designed to prospectively
and cardiovascular inclusion criteria have been capture cardiovascular outcomes as a primary study
conducted during the course of the study and are endpoint comparing different treatments for prostate
offered on an as-needed basis to all sites. At the site cancer.
level, local urology/oncology site investigators are ACKNOWLEDGMENT The authors acknowledge Erin
supported by their cardiovascular specialists to Campbell, MS (Duke Clinical Research Institute), for
confirm cardiovascular inclusion criteria and optimal her editorial contributions.
background cardiovascular medication treatment.
Furthermore, a cardiologist at Duke Clinical Research ADDRESS FOR CORRESPONDENCE: Dr. Chiara Mel-
Institute (C.M.) reviews and confirms cardiovascular loni, Department of Medicine, Division of Cardiology,
eligibility criteria for at least the first 3 patients Duke Clinical Research Institute, Duke University
screened at each site and is available thereafter based Medical Center, 200 Morris Street, Durham, North
on a site’s needs. Clinical trial educators who are Carolina 27701. E-mail: chiara.melloni@duke.edu.
trained nurses with previous cardiovascular trial Twitter: @DukeHealth, @dlbhattmd.
experience visit sites on a regular basis to support
investigators in the screening and enrollment process
PERSPECTIVES
and to assist sites with performing well in the trial.
Third, structured questions to ensure full and com-
COMPETENCY IN MEDICAL KNOWLEDGE: Contemporary
plete cardiovascular endpoint ascertainment have
management of advanced prostate cancer involves the use of
been created and are administered by urology/
ADT, which is associated with several immunomodulatory and
oncology site research personnel to patients every
metabolic changes that may increase the risk of subsequent
month to inquire about potential cardiovascular
cardiovascular events among men with known ASCVD.
events in a process that likely has not been used by
these sites in previous clinical trials. Finally, several
TRANSLATIONAL OUTLOOK: The PRONOUNCE trial will
exploratory biomarkers (cardiovascular and immune)
delineate the relative cardiovascular safety of degarelix (GnRH
will be collected and assessed to determine how these
antagonist) versus leuprolide (GnRH agonist) and is the first
biomarkers may relate to putative differences in car-
prospective cardiovascular outcomes trial comparing treatments
diovascular events between the GnRH antagonist
for patients with prostate cancer.
(degarelix) and the GnRH agonist (leuprolide); these
analyses will also explore several potential biological
REFERENCES
1. Jemal A, Ward E, Hao Y, Thun M. Trends in the Comprehensive Cancer Network guidelines for the Cardiovascular Nurses Association, Society for
leading causes of death in the United States, management of prostate cancer. Urology 2003; Cardiovascular Angiography and Interventions,
1970-2002. JAMA 2005;294:1255–9. 61:14–24. and Society of Thoracic Surgeons. J Am Coll Car-
diol 2014;64:1929–49.
2. Jemal A, Ward E, Thun M. Declining death rates 15. Sharifi N, Gulley JL, Dahut WL. Androgen
reflect progress against cancer. PLoS One 2010;5: deprivation therapy for prostate cancer. JAMA 27. Task Force Members, Montalescot G,
e9584. 2005;294:238–44. Sechtem U, et al. 2013 ESC guidelines on the
management of stable coronary artery disease: the
3. Howlader N, Ries LA, Mariotto AB, Reichman ME, 16. Tsai HK, D’Amico AV, Sadetsky N, Chen MH,
Task Force on the management of stable coronary
Ruhl J, Cronin KA. Improved estimates of cancer- Carroll PR. Androgen deprivation therapy for
artery disease of the European Society of Cardi-
specific survival rates from population-based data. localized prostate cancer and the risk of cardio-
ology. Eur Heart J 2013;34:2949–3003.
J Natl Cancer Inst 2010;102:1584–98. vascular mortality. J Natl Cancer Inst 2007;99:
1516–24. 28. Hicks KA, Mahaffey KW, Mehran R, et al. 2017
4. Chapman JA, Meng D, Shepherd L, et al.
cardiovascular and stroke endpoint definitions for
Competing causes of death from a randomized 17. Hu JC, Williams SB, O’Malley AJ, Smith MR,
clinical trials. J Am Coll Cardiol 2018;71:1021–34.
trial of extended adjuvant endocrine therapy for Nguyen PL, Keating NL. Androgen-deprivation
breast cancer. J Natl Cancer Inst 2008;100: therapy for nonmetastatic prostate cancer is 29. Smith MR, Lee H, Fallon MA, Nathan DM.
252–60. associated with an increased risk of peripheral Adipocytokines, obesity, and insulin resistance
arterial disease and venous thromboembolism. Eur during combined androgen blockade for prostate
5. Hanrahan EO, Gonzalez-Angulo AM,
Urol 2012;61:1119–28. cancer. Urology 2008;71:318–22.
Giordano SH, et al. Overall survival and cause-
specific mortality of patients with stage T1a, 18. Ehdaie B, Atoria CL, Gupta A, et al. Androgen 30. Basaria S. Androgen deprivation therapy, in-
bN0M0 breast carcinoma. J Clin Oncol 2007;25: deprivation and thromboembolic events in men sulin resistance, and cardiovascular mortality: an
4952–60. with prostate cancer. Cancer 2012;118:3397–406. inconvenient truth. J Androl 2008;29:534–9.
6. Lloyd-Jones DM, Leip EP, Larson MG, et al. 19. Keating NL, O’Malley AJ, Smith MR. Diabetes 31. Lester-Coll NH, Goldhaber SZ, Sher DJ,
Prediction of lifetime risk for cardiovascular dis- and cardiovascular disease during androgen D’Amico AV. Death from high-risk prostate cancer
ease by risk factor burden at 50 years of age. deprivation therapy for prostate cancer. J Clin versus cardiovascular mortality with hormonal
Circulation 2006;113:791–8. Oncol 2006;24:4448–56. therapy: a decision analysis. Cancer 2013;119:
20. Saigal CS, Gore JL, Krupski TL, Hanley J, 1808–15.
7. Schairer C, Mink PJ, Carroll L, Devesa SS.
Probabilities of death from breast cancer and Schonlau M, Litwin MS. Androgen deprivation 32. O’Farrell S, Garmo H, Holmberg L, Adolfsson J,
other causes among female breast cancer patients. therapy increases cardiovascular morbidity in men Stattin P, Van Hemelrijck M. Risk and timing of
J Natl Cancer Inst 2004;96:1311–21. with prostate cancer. Cancer 2007;110:1493–500. cardiovascular disease after androgen-deprivation
21. Bolla M, de Reijke TM, Van Tienhoven G, et al. therapy in men with prostate cancer. J Clin Oncol
8. Patnaik JL, Byers T, Diguiseppi C, Dabelea D,
Duration of androgen suppression in the treatment 2015;33:1243–51.
Denberg TD. Cardiovascular disease competes
of prostate cancer. N Engl J Med 2009;360: 33. Nanda A, Chen MH, Braccioforte MH,
with breast cancer as the leading cause of death
2516–27. Moran BJ, D’Amico AV. Hormonal therapy use for
for older females diagnosed with breast cancer: a
retrospective cohort study. Breast Cancer Res 22. United States Food and Drug Administration prostate cancer and mortality in men with coro-
2011;13:R64. (FDA). FDA drug safety communication: ongoing nary artery disease-induced congestive heart fail-
safety review of GnRH antagonists and possible ure or myocardial infarction. JAMA 2009;302:
9. Colzani E, Liljegren A, Johansson AL, et al.
increased risk of diabetes and certain cardiovas- 866–73.
Prognosis of patients with breast cancer: causes of
cular diseases. . Updated March 8, 2016. Available 34. D’Amico AV, Denham JW, Crook J, et al. In-
death and effects of time since diagnosis, age, and
at: https://www.fda.gov/drugs/postmarket-drug- fluence of androgen suppression therapy for
tumor characteristics. J Clin Oncol 2011;29:
safety-information-patients-and-providers/fda- prostate cancer on the frequency and timing of
4014–21.
drug-safety-communication-ongoing-safety- fatal myocardial infarctions. J Clin Oncol 2007;25:
10. Epstein MM, Edgren G, Rider JR, Mucci LA, review-gnrh-agonists-and-possible-increased-risk. 2420–5.
Adami HO. Temporal trends in cause of death Accessed August 16, 2019.
among Swedish and US men with prostate cancer. 35. Ly-Yao G, Nikita N, Keith SW, et al. Mortality
23. Albertsen PC, Klotz L, Tombal B, Grady J,
J Natl Cancer Inst 2012;104:1335–42. and hospitalization risk following oral androgen
Olesen TK, Nilsson J. Cardiovascular morbidity
signaling inhibitors among men with advanced
11. Driver JA, Djoussé L, Logroscino G, Gaziano JM, associated with gonadotropin releasing hormone
prostate cancer by pre-existing cardiovascular
Kurth T. Incidence of cardiovascular disease and agonists and an antagonist. Eur Urol 2014;65:
comorbidities. Eur Urol 2019;77:158–66.
cancer in advanced age: prospective cohort study. 565–73.
BMJ 2008;337:a2467. 36. Di Nunno V, Mollica V, Santoni M, et al. New
24. Bhatt DL. Birth and maturation of cardio-
hormonal agents in patients with nonmetastatic
12. Studer UE, Whelan P, Albrecht W, et al. Im- oncology. J Am Coll Cardiol CardioOnc 2019;1:
castration-resistant prostate cancer: meta-
mediate or deferred androgen deprivation for pa- 114–6.
analysis of efficacy and safety outcomes. Clin
tients with prostate cancer not suitable for local
25. Hansson L, Hedner T, Dahlöf B. Prospective Genitourin Cancer 2019;17:e871–7.
treatment with curative intent: European Organi-
randomized open blinded end-point (PROBE)
sation for Research and Treatment of Cancer 37. Margel D, Peer A, Ber Y, et al. Cardiovascular
study. A novel design for intervention trials. Pro-
(EORTC) Trial 30891. J Clin Oncol 2006;24: morbidity in a randomized trial comparing GnRH-
spective Randomized Open Blinded End-Point.
1868–76. agonist and GnRH-antagonist among patients with
Blood Press 1992;1:113–9.
advanced prostate-cancer and pre-existing cardio-
13. Calais da Silva FE, Bono AV, Whelan P, et al.
26. Fihn SD, Blankenship JC, Alexander KP, et al. vascular disease. J Urol 2019;202:1199–208.
Intermittent androgen deprivation for locally
2014 ACC/AHA/AATS/PCNA/SCAI/STS focused
advanced and metastatic prostate cancer: results 38. Bourghardt J, Wilhelmson AS, Alexanderson C,
update of the guideline for the diagnosis and
from a randomised phase 3 study of the South et al. Androgen receptor-dependent and inde-
management of patients with stable ischemic
European Uroncological Group. Eur Urol 2009;55: pendent atheroprotection by testosterone in male
heart disease: a report of the American College of
1269–77. mice. Endocrinology 2010;151:5428–37.
Cardiology/American Heart Association Task Force
14. Scherr D, Swindle PW, Scardino PT. National on Practice Guidelines, and the American Associ- 39. Hatakeyama H, Nishizawa M, Nakagawa A,
Comprehensive Cancer Network. National ation for Thoracic Surgery, Preventive Nakano S, Kigoshi T, Uchida K. Testosterone
inhibits tumor necrosis factor-alpha-induced 41. Smith MR, Klotz L, van der Meulen E, Colli E,
vascular cell adhesion molecule-1 expression in Tankó LB. Gonadotropin-releasing hormone KEY WORDS cardiovascular safety,
human aortic endothelial cells. FEBS Lett 2002; blockers and cardiovascular disease risk: analysis outcomes, prostate cancer
530:129–32. of prospective clinical trials of degarelix. J Urol
2011;186:1835–42.
40. Corcoran M, Meydani M, Lichtenstein A,
Schaefer E, Dillard A, Lamon-Fava S. Sex hormone 42. Campia U, Moslehi JJ, Amiri-Kordestani L, A PP END IX For a list of the members of the
modulation of proinflammatory cytokine and C- et al. Cardio-oncology: vascular and metabolic Data and Safety Monitoring Board and the
reactive protein expression in macrophages from perspectives: a scientific statement from the Steering Committee, an expanded Methods
older men and postmenopausal women. American Heart Association. Circulation 2019;139: section, and supplemental tables, please see
J Endocrinol 2010;206:217–24. e579–602. the online supplement.
EDITORIAL COMMENT
Addressing Cardiovascular Risk of

Prostate Cancer Hormonal Therapy*
Charles J. Ryan, MD,a Alicia K. Morgans, MD, MPHb
P rostate cancer is the most common noncuta-

neous malignancy among American men and
the second-leading cause of cancer death in
the United States. Although the leading cause of
are currently under treatment with ADT, most of
whom are receiving treatment with GnRH agonists.
Several population-based studies with long-term
outcomes have implicated ADT for the observed
death in men with prostate cancer is prostate cancer, increased risk of adverse cardiovascular outcomes,
the second-leading cause of death is cardiovascular whereas meta-analyses of randomized trials are
disease (1). Because of the competing risks of these conflicted on this point (2,3).
two causes of death in an aging male population, it An important missing link in unraveling the con-
is no surprise that controversy exists as to the troversy is a relative lack of a foundation of pre-
strength of the causal link between testosterone sup- clinical data to support the mechanism of action
pressive therapy and adverse cardiovascular out- that may drive heart disease in the prostate cancer
comes. However, it is agreed that metabolic insults population. GnRH receptors are expressed in cardiac
induced by androgen deprivation therapy (ADT) may and vascular tissue (4), but their role and the impact
incur deleterious effects on cardiovascular health, of systemic GnRH agonism remain unknown. Nguyen
including alterations in lipid levels and glycemic con- et al. (5) assessed the effect of GnRH agonist plus
trol, and direct effects of low testosterone on cardiac androgen receptor antagonist treatments in men with
and vascular tissue. The development of strategies to prostate cancer on the vasculature, expecting that
minimize cardiovascular risks while maximizing can- they would find vasoconstriction to be a cause of
cer control benefits is critically important. increased cardiovascular risk. Conversely, they found
Androgen deprivation is the foundation of sys- reversible endothelium-dependent vasodilation at
temic therapy for men with advanced prostate cancer 3 months. Understanding the complex role of the
and is achieved using gonadotropin-releasing hor- androgen receptor and GnRH signaling throughout
mone (GnRH) agonist or antagonist drugs or using the cardiovascular system at the most basic level will
surgical orchiectomy. Over 400,000 American men be a necessary aspect of fully understanding the car-
diovascular effects of GnRH agonists and antagonists
in men with prostate cancer.
In addition to effects on the vasculature, there is a
need to better understand the effects of GnRH agonist
and antagonist drugs on the heart. Where and in what
concentration does the heart express these receptors
From the aUniversity of Minnesota and Masonic Cancer Center, Minne-
apolis, Minnesota; and b
Northwestern University Feinberg School of and what role, if any, do they play in the cardiovas-
Medicine, Chicago, Illinois. Dr. Morgans is a consultant for Astellas, cular health of an aging male? How could a GnRH
Sanofi, AstraZeneca, Bayer, and Janssen; and has received research antagonist affect this and are these effects different
funding from Seattle Genetics, Genentech, and Bayer. Dr. Ryan has re-
from those exerted by a GnRH agonist? Is there a
ported that he has no relationships relevant to the contents of this paper
to disclose. potential beneficial effect of blocking these receptors
The authors attest they are in compliance with human studies commit- on the heart or is it that the antagonist approach
merely represents the absence of an agonist
appropriate. For more information, visit the JACC: CardioOncology author approach? Surgical orchiectomy likely raises GnRH
instructions page. levels in the blood, so may be expected to confer

JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020 Ryan and Morgans 83
MARCH 2020:82–3 CV Risk in Prostate Cancer Hormonal Therapy
similar benefits as those of GnRH agonists (e.g., leu- characterization of potential differences between the
prolide or goserelin). How do these effects differ from therapies that may emerge with extended follow-up.
simply performing an orchiectomy to reduce testos- Finally, the study will not provide any data on the
terone levels? Is the cardiac risk highest during the relative effect of orchiectomy, because it is not
drop in testosterone from normal to the castrate included in the study design. Despite these limita-
range or is the morbidity a result of long-term sup- tions, this is the first prospective study that is
pression? What is the effect of cardioprotective appropriately powered to assess a cardiovascular
measures, such as exercise, diet, beta blockers, and primary endpoint in men with prostate cancer.
statins on these potential adverse outcomes? In addition to the need to complete prospective
trials that assess cardiovascular outcomes in men
SEE PAGE 70
with prostate cancer, there is an urgent need to
These and other questions merit consideration in understand the array of effects that androgen
the modern era as cardiology and oncology blend deprivation and other therapies have on prostate
their expertise in the study of this population that is cancer survivors. In response to this, we have
vulnerable to heightened morbidity. The prospective established, in conjunction with the Prostate Cancer
PRONOUNCE (A Trial Comparing Cardiovascular Foundation, the SURECAP (Survivorship Research in
Safety of Degarelix Versus Leuprolide in Patients Carcinoma of the Prostate) initiative (7). This initia-
With Advanced Prostate Cancer and Cardiovascular tive establishes a framework for survivorship
Disease) trial, which is highlighted in this issue of research in prostate cancer and seeks the involve-
JACC: CardioOncology (6), is representative of this ment and contribution of interested cardiologists
potential. PRONOUNCE is a phase IIIb, multicenter, wishing to join the oncology and urology commu-
prospective, randomized, open, trial designed to nities in addressing these issues. Caring for the
compare the occurrence of major adverse cardiovas- growing population of prostate cancer survivors
cular events in patients with advanced prostate can- around the world requires an understanding of the
cer and pre-existing atherosclerotic cardiovascular effects of our treatments; understanding the car-
disease receiving either a GnRH antagonist (degar- diovascular effects through the PRONOUNCE trial is
elix) or a GnRH agonist (leuprolide) as ADT for a clear step in the right direction.
12 months. This prospective study, which has a pri-
mary endpoint to assess cardiovascular morbidity as
opposed to anticancer efficacy, is rare in modern ADDRESS FOR CORRESPONDENCE: Dr. Charles J.
oncology; however, more such approaches are Ryan, Division of Hematology, Oncology and Trans-
needed. Notably, the study enrolls only patients who plantation, University of Minnesota Medical School,
are at risk at baseline due to previous cardiac events, Department of Medicine, 14-106A Phillips-
enriching the risk pool, but making causation linkages Wangensteen Building, 516 Delaware Street SE, Min-
difficult. In addition, the study only monitors neapolis, Minnesota 55455. E-mail: Ryanc@umn.edu.
the first 12 months during treatment, precluding Twitter: @charlesryanmd.
REFERENCES
1. Epstein MM, Edgren G, Rider JR, et al. Temporal patients. Prostate Cancer Prostatic Dis 2016;19: advanced prostate cancer: the PRONOUNCE trial
trends in cause of death among Swedish and US 333–9. study design. J Am Coll Cardiol Oncol 2020;2:
men with prostate cancer. J Natl Cancer Inst 2012; 70–81.
4. Kakar SS, Jennes L. Expression of
104:1335–42.
gonadotropin-releasing hormone and 7. Narayan V, Harrison M, Cheng H, et al.
2. Nguyen PL, Je Y, Schutz FAB, et al. Association gonadotropin-releasing hormone receptor mRNAs Improving research for prostate cancer survi-
of androgen deprivation therapy with cardiovas- in various non-reproductive human tissues. Cancer vorship: a statement from the Survivorship
cular death in patients with prostate cancer: a Lett 1995;98:57–62. Research in Prostate Cancer (SuRECaP) working
meta-analysis of randomized trials. JAMA 2011; group. Urol Oncol 2019 Nov 13 [E-pub ahead of
5. Nguyen PL, Jarolim P, Basaria S, et al.
306:2359–66. print].
Androgen deprivation therapy reversibly increases
3. Jin C, Fan Y, Meng Y, et al. A meta-analysis endothelium-dependent vasodilation in men
of cardiovascular events in intermittent with prostate cancer. J Am Heart Assoc 2015;4:1–8.
androgen-deprivation therapy versus continuous 6. Melloni C, Slovin SF, Blemings A, et al. Cardio- KEY WORDS cardio-oncology, hormonal
androgen-deprivation therapy for prostate cancer vascular safety of degarelix versus leuprolide for therapy, prostate cancer survivorship
PUBLISHED BY ELSEVIER ON BEHALF OF THE AMERICAN COLLEGE OF
STATE-OF-THE-ART REVIEW
Using Behavioral Economics and

Technology to Improve Outcomes
in Cardio-Oncology
Kimberly J. Waddell, PHD,a,b Payal D. Shah, MD,c Srinath Adusumalli, MD, MSC,b,c Mitesh S. Patel, MD, MBA, MSa,b,c,d
ABSTRACT
Patients with cancer are often at elevated risk for cardiovascular disease due to overlapping risk factors and cardiotoxic
anticancer treatments. Their cancer diagnoses may be the predominant focus of clinical care, with less of an emphasis on
concurrent cardiovascular risk management. Widely adopted technology platforms, including electronic health records
and mobile devices, can be leveraged to improve the cardiovascular outcomes of these patients. These technologies
alone may be insufficient to change behavior and may have greater impact if combined with behavior change strategies.
Behavioral economics is a scientific field that uses insights from economics and psychology to help explain why indi-
viduals are often predictably irrational. Combining insights from behavioral economics with these scalable technology
platforms can positively impact medical decision-making and sustained healthy behaviors. This review focuses on the
principles of behavioral economics and how “nudges” and scalable technology can be used to positively impact clinician
and patient behaviors. (J Am Coll Cardiol CardioOnc 2020;2:84–96) Published by Elsevier on behalf of the
A dvances
neoplastic
in early
therapy
detection
have
reduced cancer mortality, resulting in a
growing population of cancer survivors (1,2). Over
and anti-
substantially
treatments, such as anthracyclines
epidermal growth factor receptor 2 targeted agents,
are augmented by overlapping risk factors for cardio-
vascular disease and cancer, such as obesity, tobacco
and human
the next decade, there are expected to be more than use, and age. Childhood cancer survivors who live
20 million cancer survivors in the United States, and beyond 35 years of age have a 5-fold increased risk
half will be over the age of 70 years (3,4). Cancer of stroke or myocardial infarction compared with
therapy–related cardiotoxicity is an increasingly healthy siblings (11). This risk is escalated in survivors
recognized cause of morbidity and mortality for these with dyslipidemia, obesity, high blood pressure, or
individuals (5–10). Known cardiac side effects of diabetes mellitus (11,12). Patients with cancer may
From the aCrescenz Veterans Affairs Medical Center, Philadelphia, Pennsylvania; bPenn Medicine Nudge Unit, University of
Pennsylvania, Philadelphia, Pennsylvania; cPerelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania;
and dThe Wharton School, University of Pennsylvania, Philadelphia, Pennsylvania. Dr. Waddell is supported by the Department of
Veterans Affairs Advanced Fellowship Program in Health Services Research & Development (HSR&D). Dr. Patel is supported by a
career development award from the Department of Veterans Affairs HSR&D; is founder of Catalyst Health, a technology and
behavior change consulting firm; is an advisory board member for Healthmine Services Inc, LifeVest Health, and Holistic In-
dustries; and has received research funding from Deloitte, which is not related to the work described in this manuscript. The
contents of this manuscript do not represent the views of the U.S. Department of Veterans Affairs or the United States Govern-
ment. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose. Larry
Allen, MD, served as Guest Editor for this paper.
Manuscript received December 30, 2019; accepted February 3, 2020.

JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020 Waddell et al. 85
MARCH 2020:84–96 Behavioral Economics, Technology, and Cardio-Oncology
suboptimal decisions that are not aligned ABBREVIATIONS

HIGHLIGHTS AND ACRONYMS
with longer-term goals (20). Insights from
Patients with cancer are often at behavioral economics can be combined with
ASCVD = atherosclerotic
increased risk of cardiovascular disease. these scalable technology platforms to cardiovascular disease
develop more effective interventions that
Electronic health records and mobile CV = cardiovascular
lead to improvements in medical decision-
devices are scalable technology CVD = cardiovascular disease
making and sustained healthy behaviors
platforms. EHR = electronic health record
(21).
DOAC = direct oral
Insights from behavioral economics In this article we describe how heuristics
anticoagulant
combined with scalable technology plat- and cognitive biases influence medical
LVEF = left ventricular
forms can nudge clinician decisions and decision-making and health behaviors. We ejection fraction
improve patient behaviors. highlight a growing body of evidence that

demonstrates how to implement behavioral economic
Combining technology with insights from
interventions through scalable technology platforms
behavioral economics can help address
to change clinician decisions and motivate patients to
unmet needs of cardio-oncology
engage in healthy behaviors. We provide recommen-
patients.
dations for the field of cardio-oncology to combine
technology with behavioral economics to improve
experience clinical care that focuses on their oncolog-
cardiovascular outcomes for patients with cancer.
ical diagnosis with less of an emphasis on modifiable
cardiovascular risk factors such as lack of physical ac- PRINCIPLES OF BEHAVIORAL ECONOMICS
tivity and dyslipidemia (1). As cancer-related survival
gains are made, greater attention is being directed to- Behavioral economics incorporates concepts from
ward understanding, preventing, and treating cardiac economics and psychology to explain human
sequelae of anti-neoplastic therapy in a burgeoning decision-making (20,22). Standard economic theory
population of cancer survivors. assumes that individuals consistently synthesize
A growing body of literature addresses surveil- available information and make rational choices that
lance and therapeutic strategies focused on the align with their longer-term goals (20). However, in-
early detection and treatment of cardiotoxicity to dividuals are often unable to consistently synthesize
help improve survivor outcomes (13–16). A signifi- complex information and make choices that maxi-
cant unmet need is the optimal implementation of mize their longer-term outcomes. As a result, in-
strategies that leverage principles of behavioral dividuals rely on heuristics, or mental shortcuts, to
economics and scalable technology platforms to help make decisions in their complex, daily environ-
change clinician decisions and patient behaviors in ments. These heuristics can lead to cognitive biases
a way that can positively impact cardiovascular that result in predictable decision errors (23). Insights
outcomes among patients with cancer. Widely from behavioral economics reveal patterns of irra-
adopted technology platforms including electronic tional, predictable deviations that result in subopti-
health records (EHR) and mobile devices could be mal decision-making. For example, individuals are
better used to monitor and change clinician motivated by immediate compared with delayed
decisions and patient behaviors to improve cardio- gratification (24,25), motivated more by avoiding
oncology outcomes. The EHR design can signifi- losses than receiving equivalent gains (26), and often
cantly impact clinician decision-making and this overestimate the probability of positive events while
creates an opportunity to implement interventions underestimating the probability of negative events
to improve medical decision-making (17,18). Smart- (27,28). Table 1 provides examples of common cogni-
phones and wearable devices can remotely monitor tive biases and their potential impact on clinician
daily behaviors and biometrics including physical decisions, patient behaviors, and how to operation-
activity, heart rate, and sleep patterns. However, alize interventions to account for these biases.
evidence indicates that these technologies alone are The availability bias, or availability heuristic, ex-
often not sufficient for changing behavior and could plains why people miscalculate the likelihood of an
be more impactful if they were combined with event happening based on how easily an example or
effective behavior change strategies (19). specific event comes to mind (23,29). For example,
Behavioral economics is a scientific field that le- prior research has shown that physicians who
verages principles from economics and psychology recently encountered bacteremia in a patient were
to help explain why individuals often make more likely to diagnose the condition in future
86 Waddell et al. JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020
Behavioral Economics, Technology, and Cardio-Oncology MARCH 2020:84–96
T A B L E 1 Cognitive Biases, Their Potential Impact on Clinician Decisions and Patient Behavior, and Intervention Strategies
Cognitive Bias Definition Impact on Clinician Decisions Impact on Patient Behavior Behavioral Intervention
Availability Relies on recent events when evaluating Not implementing a comprehensive Underestimating CVD risk due Provide more immediate feedback on
bias a decision or situation cardiotoxicity risk assessment prior to to limited information or patient outcome that is otherwise
cancer treatment prior exposure unknown
Optimism Overestimating probability of positive Underestimating patient risk for Underestimating risk of Frame information to accurately
bias outcomes and underestimating probability adverse CV event adverse CV event from convey risk
of negative outcomes cancer treatment
Status quo Preference for the current situation Low referral rates for diagnostic tests Not enrolling in medical Set optimal choice as the default
bias or things to stay the same due to opt-in default management programs due option
to opt-in default
Present Prefer immediate gratification compared Deprioritize CV surveillance to address Not achieving recommended Provide immediate feedback and
bias with delayed gratification immediate cancer attributable amount of daily physical reward preventative behaviors
morbidity and mortality activity
CV ¼ cardiovascular; CVD ¼ cardiovascular disease.
patients based on their previous exposure (30). In- Additionally, individuals tend to assign dispro-
terventions that provide feedback on outcomes that portionate weight to the immediate, present circum-
are otherwise unknown or reviewed can help address stance(s) and discount future events (24). Present
availability bias. For example, mailing a notification bias is the inherent tendency to show a stronger
letter to emergency medicine physicians when a pa- preference for immediate gratification relative to
tient for whom they had prescribed opioids died of an delayed gratification (24). Present bias helps explain
overdose resulted in a 10% reduction in opioid pre- why many individuals choose immediate behaviors
scribing (31). (snacking on unhealthy food) that may undermine
Optimism bias is defined as people’s tendency to their long-term health (obesity). Offering immediate
overestimate the probability of a positive event and performance feedback and/or rewarding preventative
underestimate the probability of a negative event behaviors can help leverage present bias to improve
happening in the future (27,28). Individuals regularly behavior. For example, providing daily performance
underestimate their probability of a negative health feedback regarding step goals coupled with loss-
outcome, such as obesity, cancer, heart disease, or framed financial incentives resulted in a significant
risk of cardiotoxicity related to cancer therapy. This increase in physical activity levels for adults with
may help explain sustained unhealthy behaviors; for ischemic heart disease (36).
example, individuals choose to smoke because they These decision biases can be leveraged to help
may grossly underestimate their odds of lung cancer facilitate guideline-consistent clinician behavior and
in the future. Framing information in such a way that increase healthy patient behaviors (37,38). The
offers comparison feedback or accurately conveys behavioral economics field proposes an approach,
risks can help account for optimism bias. termed asymmetric paternalism, to help improve
The status quo bias explains why individuals are human decision-making without restricting indi-
more likely to stick with the default option when vidual choice (20). The hallmark of asymmetric
presented with alternative choices. The default op- paternalism is helping individuals achieve their
tion is the selected condition if no alternative is goals while not restricting choice or interfering with
chosen (18,20,32). The default option, often consid- those who would otherwise make the optimal de-
ered the safe option, requires little energy and cision on their own (20,39,40). Subtle or simple
cognitive demand and strongly influences behavior changes to the choice environment that encourage
(33). An opt-in default requires a person to actively the optimal decision without restricting choice are
choose to participate, whereas an opt-out default as- called nudges. Nudges can be designed to remind,
sumes participation unless an individual elects to guide, or motivate different behaviors and vary in
opt-out of whatever intervention is offered. In- strength and complexity (17). The Central Illustration
terventions that place the optimal choice as the provides a ladder of nudge interventions with ex-
default choice can leverage status quo bias and have amples of how nudges can be applied to cardio-
an outsized effect on decision-making. In healthcare, oncology. The following section includes examples
changing the default from opting into generic medi- of how specific nudges can improve clinician
cations to opting out of them significantly increased decision-making and how previous research can
generic prescription rates from 75% to 98%, which led inform the design and delivery of cardio-oncology
to sustained effects (34,35). nudge interventions.
C ENTR AL I LL U STRA T I O N A Nudge Ladder With Clinician and Patient Examples
Waddell, K.J. et al. J Am Coll Cardiol CardioOnc. 2020;2(1):84–96.
Nudges at the bottom of the ladder focus on framing information. Nudges higher on the ladder are more likely to influence decision-making and, although they are
often more effective than nudges lower on the ladder, they are also considered more paternalistic. This figure is adapted from the Nuffield Council on Bioethics, Patel
et al. (38,93,94). EHR ¼ electronic health record; DOAC ¼ direct oral anticoagulant; ASCVD ¼ atherosclerotic cardiovascular disease; LVEF ¼ left ventricular ejection
fraction.
USING NUDGES TO IMPROVE peer comparison data can also be an impactful nudge
CLINICIAN DECISION-MAKING to improve decision-making and limit the effect of
availability and optimism bias (Table 1). Individuals
Over the last decade, EHR adoption among health often measure their behavior by how far they are from
systems and clinicians has increased from 20% to the norm and frequently use peer norms to guide
nearly 90% (41). As a result, decisions that were once their own decisions (47). An unintended consequence
made by paper or voice are now funneled through this of norming feedback, however, is that the descriptive
technology platform. The EHR is a choice environment norm (i.e., peer average) can cause regression toward
with clinical decision support embedded in the clini- the mean (“boomerang effect”) (47). To prevent this
cian’s workflow. For ongoing monitoring of complex scenario, a recent randomized trial focused on
conditions, such as cardiotoxicity or cardiovascular improving statin prescribing used a tiered feedback
disease, there are multiple opportunities to develop approach and demonstrated a tripling in statin pre-
scalable EHR interventions to facilitate guideline- scribing rates relative to usual care (43).
consistent decisions over time (41). Prior evidence In the field of cardio-oncology, a framing inter-
has demonstrated how nudges can be implemented vention might be providing peer comparison feed-
while reducing clinician workload (42–46). back to increase rates of serial left ventricular ejection
fraction monitoring for trastuzumab-related car-
FRAME INFORMATION. Framing of choices or infor- diotoxicity. Additionally, providing peer comparison
mation to highlight the positive or negative features feedback for prescribing appropriate medications for
(e.g., emphasize gains vs. emphasize losses) can be an patients with baseline hypertension may help in-
effective nudge that can alter behavior and the crease prescription rates among clinicians. For pa-
decision-making process (Central Illustration) (26). tients, showing their risk score for cardiotoxicity or
Additionally, framing performance feedback using likelihood of experiencing an adverse cardiovascular
T A B L E 2 Summary of Behavioral Economics and Physical Activity Studies
First Author
(Ref. #) N Duration Control Program Intervention Technology Outcome Result(s)
Chokshi et al. 105 24 weeks Passive Loss-framed financial incentives þ Wearable þ Change in mean daily Participants in financial
(36) monitoring of personalized step goal daily feedback steps during incentive arm had
daily step count messaging intervention period significantly greater increase
in mean steps per day
compared with control
(1,388 vs. 385)
Patel et al. (71) 200 24 weeks Daily feedback if Control program þ gamification Wearable þ Proportion of days Participants in gamification arm
step goal was daily feedback step goal was achieved step goals on a
achieved or not messaging achieved significantly greater
proportion of days
(0.53 vs. 0.32)
Patel et al. 602 36 weeks Daily feedback Control program þ gamification with Wearable þ Change in mean daily All gamification arms
(73) from wearable 3 arms: 1) competition; 2) daily feedback steps from baseline significantly increased their
support; 3) collaboration messaging to end of mean daily steps compared
intervention with control (920
(24 weeks) competition, 689 support,
637 collaboration)
Patel et al. 281 13 weeks Daily feedback if Control program þ 7,000 daily step Smartphone þ Proportion of days Loss-framed financial incentive
(69) step goal was goal þ financial incentives daily feedback step goal was group achieved step goal on
achieved or not where: Arm 1) gain-framed messaging achieved significantly greater number
incentive; Arm 2) lottery of days compared with
incentive; Arm 3) loss-framed control (0.16 adjusted
incentive difference)
Patel et al. 304 26 weeks Daily feedback if Control program þ financial Smartphone þ Proportion of days The combined incentive group
(70) step goal was incentives where: Arm 1) daily feedback step goal was achieved their step goal on a
achieved or not individual; Arm 2) team; Arm 3) messaging achieved significantly greater number
combined of days compared with
control (0.35 vs. 0.18)
Patel et al. 286 26 weeks N/A Social comparison feedback þ Smartphone þ Proportion of days The social comparison group
(72) financial incentives where: Arm 1) daily and step goal was compared with the 50th
weekly feedback compared with weekly achieved percentile with financial
50th percentile; Arm 2) weekly feedback incentive (arm 3) achieved
feedback compared with 75th messaging their step goal on a
percentile; Arm 3) weekly significantly greater number
performance feedback compared of days (0.18)
with 50th percentile þ regret
lottery; Arm 4) weekly
performance feedback compared
with 75th percentile þ regret
lottery
N/A ¼ not applicable.
event, contextually framed with normative data may help bridge the gap between intention and behavior
be an effective nudge to help increase motivation to change. Asking patients to sign a pre-commitment
improve healthy behaviors. An example could be pledge stating they will strive to monitor their blood
contextualizing patients’ risk score using a visual aide pressure at home may help improve motivation and
that has “very high risk, high risk, medium risk, low adherence rates for cardio-oncology patients. Given
risk, or no risk” levels may effectively communicate that the prevalence of hypertension is greater in
the importance of healthy decisions and vigilant oncology patients than the general population (50),
monitoring of their cardiovascular health. remote monitoring of hypertension may improve
PROMPT IMPLEMENTATION INTENTIONS. Prompt- disease management and potentially prevent adverse
ing clinicians and patients to state their imple- cardiovascular events (16,50).
mentation intentions helps increase motivation ENABLE CHOICE. Nudges that enable choice (i.e.,
(Central Illustration) (48,49). A commitment contract active choice) might require clinicians to make a de-
can help individuals bridge the gap between their cision in real time, rather than delaying the decision
current goals and future behaviors (48). Prompting until later (Central Illustration) (46,51). Most active
clinicians to sign a pre-commitment pledge to order a choice nudge interventions alert physicians (44,46)
cardiology referral for patients with malignancy or but this approach can cause alert fatigue. Such fatigue
high atherosclerotic cardiovascular disease risk may can be managed by shifting or redirecting the active
T A B L E 3 Summary of Behavioral Economics and Weight Loss Studies
First Author
John et al. (95) 66 32 weeks Dietician consult Control program þ financial Weight scale Weight loss at 32 w Intervention arm lost
and monthly incentives using deposit weeks significantly more weight
weigh-in contracts compared with control
(8.7 lb vs. 1.2 lb)
Kullgren et al. 105 24 weeks Monthly weigh-in Financial incentives where: Arm 1) Weight scale þ Weight loss at Group incentive participants lost
(96) individual financial reward automated text 24 weeks more weight compared with
($100) for meeting or exceeding messaging individual incentive and
weight loss goal; Arm 2) group control conditions (9.7 lb)
financial reward ($500) split
between group members who
met weight loss goal
Volpp et al. 57 16 weeks Monthly weigh-in Financial incentive where: Arm 1) Weight scale þ Weight loss at Lottery financial incentive arm
(97) deposit contract with matching; automated 16 weeks and deposit contract arm
Arm 2) lottery-based financial messaging lost significantly more
incentives weight compared with
control (13.1 lb and 14 lb vs.
3.9 lb)
choice nudge to a different medical team member priority for the cardio-oncology field (15). The medical
(e.g., medical assistant or nurse). For example, a assistant can receive an alert when a patient is at high
study evaluating cancer screening showed a 22% in- risk for a thromboembolic event and, in real time,
crease in breast cancer screening referral rates when accept or decline the option to template orders for the
medical assistants were trained to accept recom- physician to sign during the patient’s visit. Such an
mendations for cancer screening in patients who were instrument can save time during a clinical visit and
due and then template an order for a physician to reduce clinician EHR burden, both major limiting
sign (51). factors for implementing care interventions (53). This
An active choice nudge intervention could be used approach reduces friction along the decision-making
to improve cardiovascular care for patients with pathway. For appropriate patients, issuing a remote
cancer. Clinical practice guidelines recommend monitoring blood pressure cuff and assisting with
screening oncology patients for cardiotoxicity risk account or device set-up during their clinic may help
prior to treatment (14). Many documented risk factors improve remote blood pressure monitoring rates.
(14) are structured within the EHR but exist in scat- Indeed, issuing a blood pressure cuff and asking pa-
tered locations. Indeed, a patient’s age, body mass tients to complete device set-up in real time (in-
index, smoking status, lipid profile, and comorbid- clinic) can prevent delayed decision-making and help
ities can exist in demographic, medical history, increase in-home remote monitoring adherence rates.
problem list, laboratory, or other sections. Cancer SET DEFAULT OPTIONS. Higher on the nudge inter-
therapy drug and dose information are often in yet a vention ladder is the use of default options (Central
different section such as an infusion flowsheet. Illustration). This approach places the optimal deci-
Rather than creating a new screening assessment, a sion along the path of least resistance and is often an
tool that is embedded within the EHR that integrates effective nudge for facilitating evidence-based de-
these existing risk factors into a single view with an cisions when stakeholders have weakly held prefer-
incorporated risk stratification algorithm can identify ences and are unlikely to opt out (38). Currently,
patients at elevated risk for cardiotoxicity. If a patient clinicians are required to enter, or “opt-in,” to such
is identified as having a high risk for cardiotoxicity, interventions like cardiology referrals or echocardio-
the medical assistant can receive an alert and make a grams. Changing the default to automatically tem-
real-time decision to template an order for a diag- plate referral or diagnostic orders for a clinician to
nostic test or cardiology referral for the physician to review and sign versus filling in all the required in-
sign. formation for every order can save time and facilitate
A second cardio-oncology active choice interven- the desired behavior.
tion could be integrating the Khorana risk assessment For example, creating an EHR default pathway/
into routine practice using existing EHR data (52). order set with laboratory or imaging test orders at
This may help increase screening rates and prevent guideline-directed time intervals (e.g., echocardio-
adverse thromboembolic events, an important grams every 3 months with trastuzumab therapy) may
T A B L E 4 Summary of Behavioral Economics and Smoking Cessation Studies
First Author
Halpern et al. 2,538 6 months Issued local smoking cessation Financial incentives with: Arm 1) N/A Sustained smoking Rewards program resulted
(98) resources, cessation guides, individual reward; Arm 2) abstinence at 6 months in higher abstinence rates
behavioral modification individual deposit; Arm 3) after target quit date (15.7%) compared with
program, nicotine collaborative reward; Arm 4) deposit program (10.2%)
replacement therapy collaborative deposit and control (6%).
Volpp et al. 179 6 months 5 free smoking cessation classes Control program þ financial N/A Quit rate at 75 days Financial incentive group
(99) incentives ($20 for attending had higher quit rate at
class, $100 if not smoking at 75 days compared with
30 days post class completion) control (16.3% vs. 4.6%)
Volpp et al. 878 18 months Issued smoking cessation Control program þ financial N/A Smoking cessation at Financial incentive group
(100) information incentives ($100 for 9 or 12 months, had higher smoking
completing cessation program, depending on initial cessation rate compared
$250 for sustained cessation at cessation date with control (14.7% vs.
6 months, $400 for additional 5.0%)
6 months cessation)
Abbreviation as in Table 2.
help improve overall risk assessment rates and USING BEHAVIORAL ECONOMICS WITH
improve ongoing cardiovascular monitoring efforts. SCALABLE TECHNOLOGY TO INCREASE
Changing the default order to no daily imaging during HEALTHY PATIENT BEHAVIORS
palliative radiotherapy significantly reduced unnec-
essary imaging during palliative care from 68.2% to Patients are influenced by irrational, yet predictable,
32.4% (54). behaviors that contribute to lack of physical activity,
In many cases, it may not be possible to automat- obesity, and other unhealthy behaviors (20,21).
ically set a default, such as engaging in a longitudinal Helping patients to increase their physical activity,
surveillance program. In these cases, participation weight management, smoking cessation, and medi-
can be framed as opt-out. For example, a randomized cation adherence during cancer treatment and survi-
trial framing colorectal cancer screening as opt-out vorship may be one of the single most important
versus opt-in tripled colorectal cancer screening investments to improve individual outcomes. In-
rates over a 3-month period from 9.6% to 29.1% (55). terventions to improve these modifiable risk factors
Another randomized trial demonstrated significantly are an important part of post-chemotherapy car-
higher enrollment rates into a diabetes management diotoxicity prevention (5,14,16,58,59). The evolution
program for the opt-out group (38%) compared with of wearable devices now provides individuals with
the opt-in group (13%) (56). The cardio-oncology field immediate, constant access to an abundance of per-
could use an opt-out framing approach to help sonal health data. Wearable devices accurately
improve enrollment into a longitudinal left ventric- quantify daily physical activity levels (60) while also
ular ejection fraction surveillance program or the recording the quantity and quality of sleep, heart
recently proposed concept of cardio-oncology reha- rate, and other behaviors. Pairing this technology
bilitation (57). Changing the default from opt-in to with behavioral economic principles can help bridge
opt-out of cardiac rehabilitation for appropriate pa- the gap between quantifying and sharing information
tients resulted in significant increases in rehabilita- and sustained behavior change (19).
tion referrals (12% opt-in to 78% opt-out) (42). PHYSICAL ACTIVITY. Nearly 80% of US adults fail to
Together, nudges embedded with the EHR can achieve guideline-recommended physical activity
have a profound effect on clinician behavior without levels (61), and the majority of cancer survivors
added cost or burden. Indeed, many of the above remain inactive (62,63). Regular physical activity
nudge interventions required little time to implement before, during, and after cancer treatment has been
and had an outsized effect on rates of guideline- shown to reduce fatigue and improve overall physical
specific behavior. While nudges specifically target health (64,65). Indeed, low cardiorespiratory fitness
subtle or simple changes to choice architecture, there is associated with higher short- and long-term treat-
are additional principles of behavioral economics that ment-related cardiotoxicity, higher symptom burden,
can be leveraged to help facilitate healthy behavior and an increased risk of all-cause and cancer-specific
change in patients. mortality (66–68). Recently, the concept of cardio-
T A B L E 5 Summary of Behavioral Economics and Medication Adherence Studies
First Author
(Ref. #) N Medication Duration Control Program Intervention Technology Outcome Result(s)
Volpp et al. 1,509 Statin, aspirin, 12 months Standard of Issued electronic pill bottle þ Electronic pill Time to vascular No difference between
(101) b-blocker, or care daily lottery incentive þ bottles rehospitalization study arms in time
antiplatelet agent supportive partner þ or death to first
access to social work rehospitalization or
resources þ staff death (1.04 hazard
engagement advisor ratio)
Volpp et al. 20 Warfarin 3 months N/A Daily lottery incentives ($10 Computerized Proportion of Proportion of increased
(102) reward or $100 reward) pill box þ incorrect warfarin pills taken
daily doses significantly
reminders decreased from
22% to 2.3%
Linnemayr 155 Antiretroviral 9 months Educational Control program þ small Electronic pill Adherence rate Incentive group was
et al. (103) materials incentives (financial or bottles 23.7 percentage
prize) points more likely
to achieve 90%
adherence rate
compared with
control
Abbreviation as in Table 2.
oncology rehabilitation, analogous to cardiac reha- (36). All participants received a wrist-worn wearable
bilitation, has even been proposed to improve car- device to record daily steps. In the intervention arm
diovascular outcomes for cancer survivors (57). A of this trial, $14 was allocated to a virtual account of a
critical next step is designing and testing novel inpatient each week. Starting each week with a new
terventions paired with behavioral economics prin- deposit leveraged the fresh start effect, which is the
ciples to help improve physical activity for cardio- tendency for aspirational behavioral around temporal
oncology patients. landmarks (e.g., beginning of the year, month, or
Designing novel physical activity interventions that week) (74). If the patient did not meet a pre-
leverage the principles of behavioral economics can determined step goal each day, $2 were removed
help increase physical activity levels (36,69–73). from their account. Compared with the control arm,
Table 2 summarizes previous studies leveraging patients in the incentive arm had a significantly
behavioral economic principles and technology to greater increase in steps throughout the intervention
improve physical activity. These studies have (1,388 vs. 385 steps) and follow up period (1,066 vs. 92
deployed incentives, either financial or social, as a steps), even after removal of the financial incentive.
source of immediate gratification to leverage present Loss-framed financial incentives are effective at
bias (Table 1). Special consideration should be paid, motivating behavior change but a key limitation is the
however, to the framing of incentives when attempt- cost associated with these studies. Social incentives
ing to improve physical activity. Individuals will exert have profound influence on health behaviors and can
more effort to protect themselves against losses than be leveraged through gamification interventions to
gains, even when the two are equal (26,69). increase physical activity (71).
Previous work has shown that loss framing Gamification uses game components (i.e., points,
increased physical activity relative to a control con- levels) in nongame contexts to help improve physical
dition but gain framing did not (26,36,69,70,72). The activity and other health behaviors (71,73,75–77).
key difference between loss and gain framing is that Gamification interventions can leverage loss aversion
gain framing requires a patient to meet their goal through points and levels in addition to social in-
prior to receiving a monetary, point-based, or other centives through a variety of designs. For example,
reward. Conversely, loss-framing initially endows an families who were enrolled in the Framingham Heart
individual with a virtual incentive amount that the Study were randomized (as a family) into a 24-week
individual is at risk to lose if they do not achieve a study to increase physical activity (monitored with a
particular objective. wrist-worn wearable device) (71). Each family mem-
The ACTIVE-REWARD clinical trial randomized ber selected an individualized step goal, received
patients with a history of ischemic heart disease to daily performance feedback, and signed a pre-
control or loss-framed financial incentives with commitment pledge stating that they would try
personalized goal-setting to increase physical activity their best to achieve their daily step goal. Families in
the gamification intervention received 70 points in a automated hovering intervention in adults with heart
virtual account at the start of each week. Every day, failure (84). This study uses electronic, wireless
an individual family member was randomly selected weight scales to monitor daily weight and electronic
to represent their team. If the selected person did not pill bottles to monitor diuretic adherence.
meet their step goal, 10 points were deducted from Smoking is the leading preventable cause of cancer
the family account. If the selected person met their and cancer mortality (85) and is a key risk factor for
step goal, the family retained their points. At the end cardiovascular disease (86). Tobacco use negatively
of each week, families that had at least 50 points in impacts cancer treatment and the length and quality
their account advanced a level and families with of survival post-treatment (80). Table 4 summarizes
<50 points dropped a level. This gamification inter- studies that used behavioral economic principles to
vention leveraged accountability, peer support, and improve smoking cessation rates. These studies
collaboration to help increase physical activity. Par- leveraged varying financial incentives that resulted in
ticipants in the gamification arm met their daily step higher quit rates compared with standard educational
goal a significantly higher proportion of days interventions.
compared with the control arm (0.53 vs. 0.32) and Historically, anticoagulant and statin adherence
experienced a significantly higher change in average rates are alarmingly low in the general population
daily steps (1,661 vs. 636 steps) (71). (87–89). Nearly 60% of cardiovascular patients are
Gamification interventions that leverage social in- nonadherent with medications (90). Table 5 summa-
centives by providing peer support, accountability, rizes available studies that leveraged principles of
and even competition may be particularly beneficial behavioral economics with electronic pill bottles and
because they can provide social support that may automated messaging to improve medication adher-
improve emotional and psychological health, impor- ence rates. Results from these studies may benefit
tant priorities for cancer survivors (78). Additionally, 1 future research in the cardio-oncology field because
in every 4 cancer survivors use a support group, which the number of oncology patients and survivors
may serve as a unique space to design gamification receiving cardiovascular medications (e.g., statins)
interventions that leverage social support to increase will likely increase over time.
physical activity or other modifiable risk factors (79).
RECOMMENDATIONS AND
ADDITIONAL MODIFIABLE RISK FACTORS: WEIGHT FUTURE DIRECTIONS
MANAGEMENT, SMOKING CESSATION, AND MEDICATION
ADHERENCE. Combining technology with behavioral The principles of behavioral economics can explain
economic principles is efficacious for improving suboptimal decision-making and the underuse of
weight loss, smoking cessation rates, and medication guideline-based care while offering a framework for
adherence. These modifiable risk factors influence optimizing patient and clinician decisions. Selecting
cardiovascular outcomes for oncology patients and the right intervention requires thoughtful evaluation
survivors to varying degrees (80–82). Healthcare of the target behavior. The cardio-oncology field has
systems, clinicians, and researchers who are inter- recently identified several priorities, strategies, and
ested in examining the efficacy of combining tech- behaviors important to the field, such as aggressively
nology with behavioral economic principles to managing modifiable cardiovascular risk factors using
improve cardio-oncology outcomes can use previous a prevention framework (e.g., ABCDE framework),
research from cardiovascular and other clinical pop- developing, deploying, and using evidence-based
ulations as a guide. treatment protocols for cardiotoxicity, and acceler-
Obesity is a leading risk factor for cardiovascular ating the process of translating cardio-oncology
disease and cancer, (83) and increased adiposity can research to practice (14–16). Each strategy/behavior
result in higher mortality for nearly all types of can- will require a different behavioral economic inter-
cers (81,82). Designing novel interventions that vention because there is no one size fits all approach.
combine technology, such as weight scales with Furthermore, managing the care of oncology patients
wireless syncing capabilities, with principles of at high risk for cardiotoxicity is not a one-time event.
behavioral economics (e.g., financial incentives) can The need for ongoing surveillance is clear. Using the
result in a significantly greater weight loss compared EHR to embed nudge interventions can support both
with standard education interventions. Table 3 sum- simple and complex decision pathways that are likely
marizes previous research that leveraged the princi- sustainable over time.
ples of behavioral economics to improve weight loss. The field must extend considerable thought to the
Of note, an ongoing trial is currently evaluating an timing of patient interventions. Oncology patients at
the time of diagnosis and treatment are ill and man- factors could be embedded within these programs to
aging a range of medical issues. Clinicians may want help improve patient outcomes.
to consider testing a passive, remote monitoring Nudge interventions reach hospital systems, in-
program during the treatment phase using wearable stitutions, clinicians, and patients but require stake-
devices or smartphone data. For example, patients holder investment. EHR nudges can often fall into the
can be approached using opt-out framing into a gap between quality improvement and research.
remote monitoring program and issued a wearable Indeed, embedding experimentation into routine
device if they do not already own one. For clinical clinical practice is often a combined quality
purposes, approaching patient participation using improvement and research effort that requires novel
opt-out framing will likely increase patient uptake evaluation (92). As a result, embedding research
of remote monitoring strategies; for research pur- teams within hospital systems who can systemati-
poses, such opt-out defaults can contribute to a more cally test the effectiveness of nudge interventions is
diverse cohort (56,91). Behavioral and health data essential. Nudge units are behavioral design teams
from wearable devices can potentially be transmitted that can help researchers and clinicians identify
to the EHR to aid in the detection and management of appropriate opportunities to deliver a nudge and
clinical indicators of cardiotoxicity. For example, a when to choose a different path (17).
decrease in physical activity levels or an increase in
CONCLUSIONS
average heart rate could indicate worsening cardio-
vascular parameters or health behaviors in the pre-
There is a significant opportunity to improve cardio-
symptomatic stage. Detecting these changes before
vascular outcomes for patients with cancer by
the appearance of symptoms may be an indicator of
combining scalable technology platforms and behav-
subclinical cardiotoxicity and prompt expedited
ioral economic insights. Approaching the EHR as a
diagnostic testing, such as echocardiogram, and
choice environment, with thoughtful consideration to
potentially help to reduce potential hospitalizations
the design and presentation of choices, can promote
(84). Transmitted data could potentially alert a
optimal decision-making while reducing clinician
clinician prompting them to contact that patient and
burden. Mobile devices such as wearables can
take the appropriate action. Remote monitoring and
remotely monitor patient behaviors and be used to
integrating wearable device data with the EHR is a
deploy behavioral interventions that leverage in-
relatively unexplored space but may be particularly
centives and gamification. These approaches have
beneficial for the ongoing surveillance of cardio-
been used successfully across a wide range of clinical
oncology patients. Future research can help deter-
domains. They will need to be adapted and designed
mine optimal approaches for integrating health and
specifically for patients with cancer. Because these
behavioral data to the EHR, and what individual
approaches rely on widely used technology plat-
variables may be most informative in the remote
forms, effective interventions could be scaled more
monitoring of patients.
broadly at lower cost and this creates significant po-
Robust interventions targeting modifiable risk
tential to improve outcomes for patients with cancer.
factors that leverage the principles of behavioral
economics may be best delivered post-treatment.
Here, patients who successfully completed cancer ADDRESS FOR CORRESPONDENCE: Dr. Kimberly J.
treatment and are moving into survivorship may have Waddell, Crescenz Veterans Affairs Medical Center,
increased motivation to improve their health (i.e., 423 Guardian Drive, Blockley Hall 1005, Philadelphia,
fresh start effect). In light of the recently proposed Pennsylvania 19104. E-mail: Kimberly.Waddell@
concept of cardio-oncology rehabilitation programs pennmedicine.upenn.edu. Twitter: @kwaddell022,
(57), novel interventions targeting modifiable risk @miteshspatel, @sriadu.
REFERENCES
1. Herrmann J, Lerman A, Sandhu NP, Villarraga HR, 3. Siegel RL, Miller KD, Jemal A. Cancer statistics, in cancer treatment-related cardiotoxicity. Curr
Mulvagh SL, Kohli M. Evaluation and management 2018. CA Cancer J Clin 2018;68:7–30. Oncol Rep 2019;21:9.
of patients with heart disease and cancer: cardio-
6. Abdel-Qadir H, Austin PC, Lee DS, et al.
oncology. Mayo Clin Proc 2014;89:1287–306. 4. Miller KD, Siegel RL, Lin CC, et al. Cancer
treatment and survivorship statistics, 2016. CA A population-based study of cardiovascular mor-
2. Hayek SS, Ganatra S, Lenneman C, et al. Pre- tality following early-stage breast Cancer. JAMA
Cancer J Clin 2016;66:271–89.
paring the cardiovascular workforce to care for Cardiol 2017;2:88–93.
oncology patients: JACC review topic of the week. 5. Shelburne N, Simonds NI, Adhikari B, et al.
7. Mertens AC, Yasui Y, Neglia JP, et al. Late
J Am Coll Cardiol 2019;73:2226–35. Changing hearts and minds: improving outcomes
mortality experience in five-year survivors of
childhood and adolescent cancer: the Childhood 23. Tversky A, Kahneman D. Judgment under un- among U.S. non-federal acute care hospitals:
Cancer Survivor Study. J Clin Oncol 2001;19: certainty: heuristics and biases. Science 1974;185: 2008-2015. [Internet]. 2016. Available at: https://
3163–72. 1124–31. dashboard.healthit.gov/evaluations/data-briefs/
non-federal-acute-care-hospital-ehr-adoption-2
8. Patnaik JL, Byers T, DiGuiseppi C, Dabelea D, 24. O’Donoghue T, Rabin M. Doing it now or later.
008-2015.php. Accessed February 2020.
Denberg TD. Cardiovascular disease competes Am Econ Rev 1999;89:103–24.
with breast cancer as the leading cause of death 25. O’Donoghue T, Rabin M. The economics of 42. Adusumalli S, Jolly EG, Chokshi NP, et al.
for older females diagnosed with breast cancer: a immediate gratification. J Behav Decis Mak 2000; A change in cardiac rehabilitation referral defaults
retrospective cohort study. Breast Cancer Res 13:233–50. from opt-in to opt-out increases referral rates
2011;13:R64. among patients with ischemic heart disease. Cir-
26. Kahneman D, Tversky A. Prospect theory: an culation 2017;136 Suppl_1:A19699.
9. Armstrong GT, Chen Y, Yasui Y, et al. Reduction analysis of decision under risk. In Handbook of the
in late mortality among 5-year survivors of child- fundamentals of financial decision making: Part I. 43. Patel MS, Kurtzman GW, Kannan S, et al. Ef-
hood cancer. N Engl J Med 2016;374:833–42. Singapore: World Scientific, 2013:99–127. fect of an automated patient dashboard using
active choice and peer comparison performance
10. Chapman J-AW, Meng D, Shepherd L, et al. 27. Sharot T. The Optimism Bias: A Tour of the
feedback to physicians on statin prescribing: the
Competing causes of death from a randomized Irrationally Positive Brain. New York, NY:
PRESCRIBE cluster randomized clinical trial. JAMA
trial of extended adjuvant endocrine therapy for Pantheon, 2011.
Netw Open 2018;1:e180818.
breast cancer. JNCI J Natl Cancer Inst 2008;100:
28. Sharot T. The optimism bias. Curr Biol 2011;21:
252–60. 44. Patel MS, Volpp KG, Small DS, et al. Using
R941–5.
active choice within the electronic health record to
11. Armstrong GT, Oeffinger KC, Chen Y, et al.
29. Folkes VS. The availability heuristic and increase physician ordering and patient comple-
Modifiable risk factors and major cardiac events
perceived risk. J Consum Res 1988;15:13–23. tion of high-value cancer screening tests. Healthc
among adult survivors of childhood cancer. J Clin
30. Poses RM, Anthony M. Availability, wishful (Amst) 2016;4:340–5.
Oncol 2013;31:3673.
thinking, and physicians’ diagnostic judgments for 45. Patel MS, Day S, Small DS, et al. Using default
12. Armstrong GT, Kawashima T, Leisenring W, patients with suspected bacteremia. Med Decis options within the electronic health record to in-
et al. Aging and risk of severe, disabling, life- Making 1991;11:159–68. crease the prescribing of generic-equivalent
threatening, and fatal events in the childhood
31. Doctor JN, Nguyen A, Lev R, et al. Opioid medications: a quasi-experimental study. Ann
cancer survivor study. J Clin Oncol 2014;32:1218.
prescribing decreases after learning of a patient’s Intern Med 2014;161 10_Suppl:S44–52.
13. Cardinale D, Colombo A, Bacchiani G, et al. fatal overdose. Science 2018;361:588–90. 46. Patel MS, Volpp KG, Small DS, et al. Using
Early detection of anthracycline cardiotoxicity and
32. Samuelson W, Zeckhauser R. Status quo bias in active choice within the electronic health record to
improvement with heart failure therapy. Circula-
decision making. J Risk Uncertain 1988;1:7–59. increase influenza vaccination rates. J Gen Intern
tion 2015;131:1981–8. Med 2017;32:790–5.
33. Dean M, Kıbrıs Ö, Masatlioglu Y. Limited
14. Armenian SH, Lacchetti C, Barac A, et al. Pre- 47. Schultz PW, Nolan JM, Cialdini RB,
attention and status quo bias. J Econ Theory 2017;
vention and monitoring of cardiac dysfunction in Goldstein NJ, Griskevicius V. The constructive,
169:93–127.
survivors of adult cancers: American Society of destructive, and reconstructive power of social
Clinical Oncology clinical practice guideline. J Clin 34. Olshan D, Rareshide CAL, Patel MS. Longer-
norms. Psychol Sci 2007;18:429–34.
Oncol 2017;10:893–911. term durability of using default options in the
electronic health record to increase generic pre- 48. Rogers T, Milkman KL, Volpp KG. Commitment
15. Lenihan DJ, Fradley MG, Dent S, et al. Pro- scribing rates. J Gen Intern Med 2019;34:349–50. devices: using initiatives to change behavior.
ceedings from the Global Cardio-Oncology Sum- JAMA 2014;311:2065–6.
35. Patel MS, Day SC, Halpern SD, et al. Generic
mit. J Am Coll Cardiol CardioOnc 2019;1:256.
medication prescription rates after health system– 49. Ariely D, Wertenbroch K. Procrastination,
16. Marvel FA, Whelton SP, Blumenthal RS. wide redesign of default options within the elec- deadlines, and performance: self-control by pre-
A cardio-oncology cardiovascular prevention tronic health record. JAMA Intern Med 2016;176: commitment. Psychol Sci 2002;13:219–24.
framework. J Am Coll Cardiol CardioOnc 2019;1: 847–8.
50. Cohen JB, Geara AS, Hogan JJ, Townsend RR.
252–5.
36. Chokshi NP, Adusumalli S, Small DS, et al. Hypertension in cancer patients and survivors:
17. Patel MS, Volpp KG, Asch DA. Nudge units to Loss-framed financial incentives and personalized epidemiology, diagnosis, and management. J Am
improve the delivery of health care. N Engl J Med goal-setting to increase physical activity among Coll Cardiol CardioOnc 2019;1:238–51.
2018;378:214–6. ischemic heart disease patients using wearable
51. Hsiang EY, Mehta SJ, Small DS, et al. Associ-
devices: the ACTIVE REWARD randomized trial.
18. Halpern SD, Ubel PA, Asch DA. Harnessing the ation of an active choice intervention in the elec-
J Am Heart Assoc 2018;7:e009173.
power of default options to improve health care. tronic health record directed to medical assistants
N Engl J Med 2007;357:1340–4. 37. Loewenstein G, John LK, Volpp K. Using with clinician ordering and patient completion of
Decision Errors to Help People Help Themselves. breast and colorectal cancer screening tests.
19. Patel MS, Asch DA, Volpp KG. Wearable de-
In: The Behavioral Foundations of Public Policy, JAMA Netw Open 2019;2:e1915619.
vices as facilitators, not drivers, of health behavior
edited by Eldar Shafir. Princeton, NJ: Princeton
change. JAMA 2015;313:459–60. 52. Key NS, Khorana AA, Kuderer NM, Bohlke K,
University Press, 2012:361–79.
Lee AY, Arcelus JI, et al. Venous thromboembo-
20. Loewenstein G, Brennan T, Volpp KG. Asym- 38. Patel MS, Navathe AS, Liao JM. Using nudges
lism prophylaxis and treatment in patients with
metric Paternalism to Improve Health Behaviors. to improve value by increasing imaging-based
cancer: ASCO clinical practice guideline update.
JAMA 2007 Nov 28;298:2415–7. cancer screening. J Am Coll Radiol 2020;17:38–41.
J Clin Oncol 2019;19.
21. Loewenstein G, Asch DA, Volpp KG. Behavioral 39. Camerer C, Issacharoff S, Loewenstein G, 53. Geerligs L, Rankin NM, Shepherd HL, Butow P.
economics holds potential to deliver better results O’donoghue T, Rabin M. Regulation for conserva- Hospital-based interventions: a systematic review
for patients, insurers, and employers. Health Aff tives: behavioral economics and the case for“ of staff-reported barriers and facilitators to
(Millwood) 2013;32:1244–50. asymmetric paternalism”. Univ Pa Law Rev 2003; implementation processes. Implement Sci 2018;
151:1211–54. 13:36.
22. Camerer CF, Loeweknstein G. Behavioral
40. Thaler RH, Sunstein CR. Libertarian pater-
economics: past, present, future. [Internet] 2003. 54. Sharma S, Guttmann D, Small DS, et al. Effect
nalism. Am Econ Rev 2003;93:175–9.
p 1-61. Available at: https://www.its.caltech.edu/ of introducing a default order in the electronic
wcamerer/ribe239.pdf. Accessed November 30, 41. Henry J, Pylypchuk Y, Searcy T, Patel V. medical record on unnecessary daily imaging
2019. Adoption of electronic health record systems during palliative radiotherapy for adults with
cancer: a stepped-wedge cluster randomized of life in lymphoma patients. J Clin Oncol 2009; cancer in a prospectively studied cohort of US
clinical trial. JAMA Oncol 2019;5:1220–2. 27:4605–12. adults. N Engl J Med 2003;348:1625–38.
55. Mehta SJ, Khan T, Guerra C, et al. 68. Lakoski SG, Willis BL, Barlow CE, et al. Midlife 83. Koene RJ, Prizment AE, Blaes A, Konety SH.
A randomized controlled trial of opt-in versus opt- cardiorespiratory fitness, incident cancer, and Shared risk factors in cardiovascular disease and
out colorectal cancer screening outreach. Am J survival after cancer in men: the Cooper Center cancer. Circulation 2016;133:1104–14.
Gastroenterol 2018;113:1848–54. Longitudinal Study. JAMA Oncol 2015;1:231–7.
84. Mehta SJ, Volpp KG, Asch DA, et al. Rationale
56. Aysola J, Tahirovic E, Troxel AB, et al. 69. Patel MS, Asch DA, Rosin R, et al. Framing and design of EMPOWER, a pragmatic randomized
A randomized controlled trial of opt-in versus opt- financial incentives to increase physical activity trial of automated hovering in patients with
out enrollment into a diabetes behavioral inter- among overweight and obese adults: a random- congestive heart failure. Circ Cardiovasc Qual
vention. Am J Health Promot 2018;32:745–52. ized, controlled trial. Ann Intern Med 2016;164: Outcomes 2019;12:e005126.
385–94.
57. Gilchrist SC, Barac A, Ades PA, et al. Cardio- 85. US Department of Health and Human Ser-
oncology rehabilitation to manage cardiovascular 70. Patel MS, Asch DA, Rosin R, et al. Individual vices. The health consequences of smoking—50
outcomes in cancer patients and survivors: a sci- versus team-based financial incentives to increase years of progress: a report of the Surgeon Gen-
entific statement from the American Heart Asso- physical activity: a randomized, controlled trial. eral; 2014.
ciation. Circulation 2019;139:e997–1012. J Gen Intern Med 2016;31:746–54.
86. Fryar CD, Chen T-C, Li X. Prevalence of un-
58. Lakoski SG, Eves ND, Douglas PS, Jones LW. 71. Patel MS, Benjamin EJ, Volpp KG, et al. Effect
controlled risk factors for cardiovascular disease:
Exercise rehabilitation in patients with cancer. Nat of a game-based intervention designed to enhance
United States, 1999-2010. US Department of
Rev Clin Oncol 2012;9:288. social incentives to increase physical activity
Health and Human Services, Centers for Disease
among families: the BE FIT randomized clinical
59. Hunt SA, Abraham WT, Chin MH, et al. 2009 Control and Prevention, National Center for
trial. JAMA Intern Med 2017;177:1586–93.
focused update incorporated into the ACC/AHA Health Statistics, 2012.
2005 guidelines for the diagnosis and manage- 72. Patel MS, Volpp KG, Rosin R, et al.
87. Ellis JJ, Erickson SR, Stevenson JG,
ment of heart failure in adults: a report of the A randomized trial of social comparison feedback
Bernstein SJ, Stiles RA, Fendrick AM. Suboptimal
American College of Cardiology Foundation/ and financial incentives to increase physical ac-
statin adherence and discontinuation in primary
American Heart Association Task Force on Prac- tivity. Am J Health Promot 2016;30:416–24.
and secondary prevention populations. J Gen
tice Guidelines developed in collaboration with 73. Patel MS, Small DS, Harrison JD, et al. Effec- Intern Med 2004;19:638–45.
the International Society for Heart and Lung tiveness of behaviorally designed gamification in-
Transplantation. J Am Coll Cardiol 2009;53: 88. Jackevicius CA, Mamdani M, Tu JV. Adherence
terventions with social incentives for increasing
e1–90. with statin therapy in elderly patients with and
physical activity among overweight and obese
without acute coronary syndromes. JAMA 2002;
adults across the United States: the STEP UP
60. Case MA, Burwick HA, Volpp KG, Patel MS. 288:462–7.
randomized clinical trial. JAMA Intern Med 2019;
Accuracy of smartphone applications and wearable
179:1–9. 89. Kimmel SE, Chen Z, Price M, et al. The influ-
devices for tracking physical activity data. JAMA
2015;313:625–6. 74. Dai H, Milkman KL, Riis J. The fresh start ef- ence of patient adherence on anticoagulation
fect: temporal landmarks motivate aspirational control with warfarin: results from the Interna-
61. Piercy KL, Troiano RP, Ballard RM, et al. The tional Normalized Ratio Adherence and Genetics
behavior. Manag Sci 2014;60:2563–82.
physical activity guidelines for Americans. JAMA (IN-RANGE) Study. Arch Intern Med 2007;167:
2018;320:2020–8. 75. Edwards EA, Lumsden J, Rivas C, et al. Gami- 229–35.
fication for health promotion: systematic review of
62. Sabiston CM, Brunet J, Vallance JK, 90. Kravitz RL, Hays RD, Sherbourne CD, et al.
behaviour change techniques in smartphone apps.
Meterissian S. Prospective examination of objec- Recall of recommendations and adherence to
BMJ Open 2016;6:e012447.
tively assessed physical activity and sedentary advice among patients with chronic medical con-
time after breast cancer treatment: sitting on the 76. Kawachi I. It’s all in the game—the uses of
ditions. Arch Intern Med 1993;153:1869–78.
crest of the teachable moment. Cancer Epidemiol gamification to motivate behavior change. JAMA
Prev Biomark 2014;23:1324–30. Intern Med 2017;177:1593–4. 91. Mehta SJ, Troxel AB, Marcus N, et al.
Participation rates with opt-out enrollment in a
63. Lynch BM, Dunstan DW, Healy GN, Winkler E, 77. Kurtzman GW, Day SC, Small DS, et al. Social
remote monitoring intervention for patients with
Eakin E, Owen N. Objectively measured physical incentives and gamification to promote weight
myocardial infarction. JAMA Cardiol 2016;1:
activity and sedentary time of breast cancer sur- loss: the LOSE IT randomized, controlled trial.
847–8.
vivors, and associations with adiposity: findings J Gen Intern Med 2018;33:1669–75.
from NHANES (2003–2006). Cancer Causes Con- 92. Oakes AH, Patel MS. A nudge towards
78. Catt S, Sheward J, Sheward E, Harder H.
trol 2010;21:283–8. increased experimentation to more rapidly improve
Cancer survivors’ experiences of a community-
healthcare. BMJ Qual Saf 2020;29:179–81.
64. Winningham ML, MacVicar MG, Bondoc M, based cancer-specific exercise programme: re-
Anderson JI, Minton JP. Effect of aerobic exercise sults of an exploratory survey. Support Care Can- 93. Patel MS. Nudges for influenza vaccination.
on body weight and composition in patients with cer 2018;26:3209–16. Nat Hum Behav 2018;2:720.
breast cancer on adjuvant chemotherapy. Oncol 79. Owen JE, Goldstein MS, Lee JH, Breen N, 94. Policy process and practice. [Internet]. Nuf-
Nurs Forum 1989;16:683–9. Rowland JH. Use of health-related and cancer- field Council on Bioethics. 2007. p. 31–47. (Public
65. Velthuis MJ, Agasi-Idenburg SC, specific support groups among adult cancer sur- Health: Ethical Issues). Available at: https://
Aufdemkampe G, Wittink HM. The effect of vivors. Cancer 2007;109:2580–9. nuffieldbioethics.org/wp-content/uploads/2014/
physical exercise on cancer-related fatigue during 80. Shoemaker ML, White MC, Hawkins NA, 07/Public-health-Chapter-3-Policy-process-and-
cancer treatment: a meta-analysis of randomised Hayes NS. Prevalence of smoking and obesity practice.pdf. Accessed December 29, 2019.
controlled trials. Clin Oncol 2010;22:208–21. among US cancer survivors: estimates from the 95. John LK, Loewenstein G, Troxel AB, Norton L,
66. Jones LW, Courneya KS, Mackey JR, et al. National Health Interview Survey, 2008–2012. In Fassbender JE, Volpp KG. Financial incentives for
Cardiopulmonary function and age-related decline NIH Public Access 2016. p. 436. extended weight loss: a randomized, controlled
across the breast cancer survivorship continuum. 81. Calle EE, Kaaks R. Overweight, obesity and trial. J Gen Intern Med 2011;26:621–6.
J Clin Oncol 2012;30:2530. cancer: epidemiological evidence and proposed
96. Kullgren JT, Troxel AB, Loewenstein G, et al.
mechanisms. Nat Rev Cancer 2004;4:579.
67. Courneya KS, Sellar CM, Stevinson C, et al. Individual vs. group-based incentives for weight
Randomized controlled trial of the effects of aer- 82. Calle EE, Rodriguez C, Walker-Thurmond K, loss: a randomized, controlled trial. Ann Intern
obic exercise on physical functioning and quality Thun MJ. Overweight, obesity, and mortality from Med 2013;158:505.
97. Volpp KG, John LK, Troxel AB, Norton L, 100. Volpp KG, Troxel AB, Pauly MV, et al. warfarin adherence. BMC Health Serv Res 2008;
Fassbender J, Loewenstein G. Financial incentive– A randomized, controlled trial of financial in- 8:272.
based approaches for weight loss: a randomized centives for smoking cessation. N Engl J Med
103. Linnemayr S, Stecher C, Mukasa B. Behavioral
trial. JAMA 2008;300:2631–7. 2009;360:699–709.
economic incentives to improve adherence to an-
98. Halpern SD, French B, Small DS, et al. Random- 101. Volpp KG, Troxel AB, Mehta SJ, et al. Effect tiretroviral medication. AIDS Lond Engl 2017;31:
ized trial of four financial-incentive programs for of electronic reminders, financial incentives, and 719–26.
smoking cessation. N Engl J Med 2015;372:2108–17. social support on outcomes after myocardial
infarction: the HeartStrong randomized clinical
99. Volpp KG, Levy AG, Asch DA, et al.
trial. JAMA Intern Med 2017;177:1093–101.
A randomized controlled trial of financial in- KEY WORDS behavior change, behavioral
centives for smoking cessation. Cancer Epidemiol 102. Volpp KG, Loewenstein G, Troxel AB, et al. economics, cardiotoxicity, electronic health
Prev Biomark 2006;15:12–8. A test of financial incentives to improve records, nudges, wearables
PRIMERS IN CARDIOONCOLOGY
CAR T Cell Therapy–Related

Cardiovascular Outcomes
and Management
Systemic Disease or Direct Cardiotoxicity?
Arjun K. Ghosh, MBBS, MSC, PHD,a,b,* Daniel H. Chen, BMED,a,b,* Avirup Guha, MD,c Strachan Mackenzie, MBBS, BSC,d
J. Malcolm Walker, MD,b Claire Roddie, MD, PHDd
JACC: CARDIOONCOLOGY CME/MOC/ECME
This article has been selected as the month’s JACC: CardioOncology Method of Participation and Receipt of CME/MOC/ECME Certificate
CME/MOC/ECME activity, available online at http://www.acc.org/jacc-
To obtain credit for JACC CME/MOC/ECME, you must:
journals-cme by selecting the JACC Journals CME/MOC/ECME tab.
1. Be an ACC member or JACC: CardioOncology subscriber.
Accreditation and Designation Statement 2. Carefully read the CME/MOC/ECME-designated article available on-
line and in this issue of the Journal.
The American College of Cardiology Foundation (ACCF) is accredited by
3. Answer the post-test questions. A passing score of at least 70% must be
the Accreditation Council for Continuing Medical Education to provide
achieved to obtain credit.
continuing medical education for physicians.
4. Complete a brief evaluation.
The ACCF designates this Journal-based CME activity for a maximum
5. Claim your CME/MOC/ECME credit and receive your certificate
of 1 AMA PRA Category 1 Credit(s). Physicians should claim only the credit
electronically by following the instructions given at the conclusion of
commensurate with the extent of their participation in the activity.
the activity.
Successful completion of this CME activity, which includes participa-
CME/MOC/ECME Objective for This Article: Upon completion of this
tion in the evaluation component, enables the participant to earn up to
activity, the learner should be able to: 1) identify the mechanism of action
1 Medical Knowledge MOC point in the American Board of Internal
of CD19-specific CAR T cell therapy in relapsed and refractory B-cell he-
Medicine’s (ABIM) Maintenance of Certification (MOC) program. Par-
matological malignancies; 2) dDiscuss the common adverse effects
ticipants will earn MOC points equivalent to the amount of CME credits
associated with CAR T cell therapy, and the patients most at risk of
claimed for the activity. It is the CME activity provider’s responsibility
developing these adverse effects; and 3) discuss the treatment options for
to submit participant completion information to ACCME for the pur-
common adverse effects associated with CAR T cell therapy.
pose of granting ABIM MOC credit.
CME/MOC/ECME Editor Disclosure: JACC: CardioOncology CME/MOC/ECME
CAR T Cell Therapy–Related Cardiovascular Outcomes and Management: Editor Bonnie Ky, MD, MSCE, has reported that she has received research
Systemic Disease or Direct Cardiotoxicity? will be accredited by the support from the NIH; and consulting support from Bristol-Myers Squibb.
European Board for Accreditation in Cardiology (EBAC) for 1 hour of
Author Disclosures: Dr. Roddie has received speaker fees from Novartis,
External CME credits. Each participant should claim only those hours of
Gilead, and Celgen. All other authors have reported that they have no
credit that have actually been spent in the educational activity.
relationships relevant to the contents of this paper to disclose.
The Accreditation Council for Continuing Medical Education (ACCME)
and the European Board for Accreditation in Cardiology (EBAC) have Medium of Participation: Online (article and quiz).
recognized each other’s accreditation systems as substantially equiva-
CME/MOC/ECME Term of Approval
lent. Apply for credit through the post-course evaluation. While offering
the credits noted above, this program is not intended to provide exten- Issue Date: March 2020
sive training or certification in the field. Expiration Date: February 28, 2021
From the aCardio-Oncology Service, Bart’s Heart Centre, St Bartholomew’s Hospital West Smithfield, London, United Kingdom;
b
Hatter Cardiovascular Institute, Institute of Cardiovascular Science UCL, University College London Hospital, London, United
Kingdom; cDivision of Cardiology, Case Western Reserve University, Harrington Heart and Vascular Institute, Ashland, Ohio; and
the dDepartment of Haematology, University College London Hospital, London, United Kingdom. *Drs. Ghosh and Mr. Chen
shared first authorship. Dr. Roddie has received speaker fees from Novartis, Gilead, and Celgen. All other authors have reported
that they have no relationships relevant to the contents of this paper to disclose.

98 Ghosh et al. JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020
CAR T-Cell Therapy–Related Cardiovascular Outcomes and Management MARCH 2020:97–109
CAR T Cell Therapy–Related Cardiovascular

Outcomes and Management
Systemic Disease or Direct Cardiotoxicity?
Arjun K. Ghosh, MBBS, MSC, PHD,a,b,* Daniel H. Chen, BMED,a,b,* Avirup Guha, MD,c Strachan Mackenzie, MBBS, BSC,d
J. Malcolm Walker, MD,b Claire Roddie, MD, PHD,d
ABSTRACT
CD19-specific chimeric antigen receptor (CAR) T cell therapies have shown remarkable early success in highly refractory
and relapsing hematological malignancies. However, this potent therapy is accompanied by significant toxicity. Cytokine
release syndrome and neurotoxicity are the most widely reported, but the true extent and characteristics of cardiovas-
cular toxicity remain poorly understood. Thus far, adverse cardiovascular effects observed include sinus tachycardia and
other arrhythmias, left ventricular systolic dysfunction, profound hypotension, and shock requiring inotropic support. The
literature regarding cardiovascular toxicities remains sparse; prospective studies are needed to define the cardiac safety
of CAR T cell therapies to optimally harness their potential. This review summarizes the current understanding of the
potential cardiovascular toxicities of CD19-specific CAR T cell therapies, outlines a proposed cardiac surveillance protocol
for patients receiving this new treatment, and provides a roadmap of the future direction of cardio-oncology research in
this area. (J Am Coll Cardiol CardioOnc 2020;2:97–109) © 2020 The Authors. Published by Elsevier on behalf of
G enetically modified chimeric antigen recep-

tor (CAR) T cells specifically targeting CD19
have shown remarkable promise in the
treatment of highly refractory and relapsing hemato-
antigen, resulting in the killing of target cells. The
first CAR designs aimed at combating cancer emerged
in the late 1980s and early 1990s, but only recent
designs have seen significant clinical success. In
logical malignancies in pediatric and adult popula- particular, immense excitement has been generated
tions, including acute lymphoblastic leukemia (ALL) by the early success of CD19-specific CAR T cells in
and large B-cell lymphoma. However, this potent the treatment of highly refractory and relapsing he-
therapy is tempered by serious, potentially life- matological malignancies (2,3). CD19 is an effective
threatening complications (1). Cytokine release syn- target because it is expressed throughout B-cell line-
drome (CRS) and neurotoxicity are the most common age development and has frequent and high-level
and widely appreciated, but the true extent and char- expression on the surface of nearly all B-cell malig-
acteristics of cardiovascular toxicities remain poorly nancies. In addition, it is not found on other normal
defined. This review appraises the current literature tissues, including the heart, and is not shed as a sol-
on the latter and seeks to outline research priorities uble form (4).
in cardiotoxicity to maximize the cardiovascular The process of manufacturing CAR T cells requires
safety of CAR T cell therapies. 2 to 3 weeks (Figure 1). Autologous T cells are first
collected from the patient and then genetically
CAR T CELL THERAPY modified ex vivo with lentiviral or retroviral vectors
to re-program the T cells to recognize tumor cells
A CAR is a recombinant fusion protein that is capable expressing a tumor-associated antigen (in this
of activating T cells upon recognition of a specific context, CD19). The CAR T cells undergo rapid
Manuscript received November 1, 2019; revised manuscript received January 28, 2020, accepted February 5, 2020.
JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020 Ghosh et al. 99
MARCH 2020:97–109 CAR T-Cell Therapy–Related Cardiovascular Outcomes and Management
39% to 50% at 1 year. Retrospective analyses ABBREVIATIONS

HIGHLIGHTS AND ACRONYMS
of adults with relapsed/refractory large B-cell
CD19-specific CAR T cell therapies lymphoma suggest, in stark contrast to those
ALL = acute lymphoblastic
represent a new breakthrough in the receiving CAR T cell therapies, an objective leukemia
treatment of relapsing and refractory response of 26% and CR in only 7% with
CAR = chimeric antigen
hematological malignancies. conventional (non-CAR) therapies. The poor receptor
response to conventional therapies was CI = confidence interval
The adverse effects observed in early
associated with a median survival of CMR = cardiac magnetic
trials and recent retrospective studies
6.3 months, and low 1- and 2-year survival resonance imaging
have suggested that CAR T cell therapies
rates of 28% and 20%, respectively (7); this CR = complete response
are associated with cardiovascular
should be compared with the 76% and 50% CRS = cytokine release
toxicity.
survival rates reported in the statistical ana- syndrome
Prospective studies using multimodality lyses of the early pivotal CAR T cell therapy IL = interleukin
cardiac imaging and novel and estab- trial cohorts (8,9). LVSD = left ventricular systolic
dysfunction
lished biomarkers are needed to define
TISAGENLECLEUCEL. Tisagenlecleucel was
the cardiac safety of CAR T cell therapies TTE = transthoracic
developed through a collaboration between echocardiography
to fully harness their life-saving
the University of Pennsylvania and Novartis.
potential.
It achieved early success through its pilot study
within the Philadelphia Children’s Hospital with a
multiplication to generate therapeutic quantities
90% CR rate in a pediatric and young adult popula-
before being administered to the patient in a single
tion with relapsed/refractory B-cell ALL (10). There-
infusion (3,5).
after, it became the first CAR T cell product to receive
Before the infusion of CAR T cells, patients un-
U.S. Food and Drug Administration approval
dergo lymphodepleting chemotherapy, most
following the single-arm, global multicenter ELIANA
commonly with a combination of fludarabine and
(Determine Efficacy and Safety of CTL019 in Pediatric
cyclophosphamide. This approach is to suppress the
Patients With Relapsed and Refractory B-cell ALL and
patient’s endogenous T cell compartment, which in
High Risk B-cell ALL at First Relapse. Determine
turn promotes the in vivo expansion of the trans-
Feasibility and Safety of CTL019 Therapy in Pediatric
ferred CAR T cell product (6).
Patients With High Risk B-cell ALL That Relapsed < 6
Two CD19-specific CAR T cell therapies have now
Months Post All-HSCT) study. Sixty-one (81%) of its 75
been licensed for use with hematological cancers:
patients with follow-up out to 3 months after infusion
axicabtagene ciloleucel (Yescarta, Gilead Sciences,
were in ongoing CR. Relapse-free survival, in those
Inc., Foster City, California) and tisagenlecleucel
with a response to treatment, at 6 and 12 months was
(Kymriah, Novartis Pharmaceuticals Corporation,
80% and 59%, respectively (8).
East Hanover, New Jersey) have both received
Tisagenlecleucel has also been tested in adult
approval by the U.S. Food and Drug Administration
populations through the JULIET (Study of Efficacy
and the European Commission and European Medi-
and Safety of CTL019 in Adult DLBCL Patients) study,
cines Agency during 2017 and 2018. Axicabtagene
a Phase II, single-arm, global multicenter study. The
ciloleucel has been approved for the treatment of
analysis of 93 adults with relapsed/refractory diffuse
relapsed/refractory large B-cell lymphoma in adult
large B-cell lymphoma with follow-up out to 3 months
patients; tisagenlecleucel has been approved for the
demonstrated either complete or partial responses in
treatment of both the former and in relapsed/re-
52% of the 93 patients (11,12).
fractory B-cell precursor ALL in pediatric and young
adult patients up to 25 years of age. AXICABTAGENE CILOLEUCEL. Axicabtagene cil-
The unprecedented response to CAR T cell thera- oleucel is another CD19-targeting CAR T cell therapy
pies in these studies must be appreciated in the and has been explored in adult populations with large
context of a patient population in which curative B-cell lymphomas. The initial results of the single-
options are limited or even exhausted, and for whom arm, global and multicenter ZUMA-1 (Long-Term
palliative therapy would usually be the mainstay of Safety and Activity of Axicabtagene Ciloleucel in Re-
treatment. The early pivotal studies that established fractory Large B-Cell Lymphoma) study were pub-
the efficacy of CAR T cell therapies in this patient lished in 2017 and the 2-year follow-up analysis was
cohort achieved complete remission (CR) rates of 52% published in 2019. Of the 101 patients who were
to 90% with responses maintained in the range of assessed after infusion with axicabtagene ciloleucel,
F I G U R E 1 Manufacturing CAR T Cell Therapy
Autologous T cells are collected from the patient. A chimeric receptor antibody (CAR) that recognizes CD19 is inserted into the cell surface using lentiviral or retroviral
vectors. The CAR T cells undergo ex vivo rapid multiplication to generate therapeutic quantities. The patient undergoes lymphodepleting chemotherapy (usually a
combination of fludarabine and cyclophosphamide) before receiving an infusion of the CAR T cells to reduce the native population of T cells, which in turn promotes CAR
T-cell expansion. Reproduced from Makita et al. (45) with permission.
83% (n ¼ 84) reached the primary endpoint of an ELIANA study was 77% in its pediatric cohort with
objective response at 1 month (a composite of both CR relapsed/refractory ALL (8).
[n ¼ 59, 58%] and partial response). Median CYTOKINE RELEASE SYNDROME. CRS is a phenom-
progression-free survival in the whole cohort was enon that is associated with (but not limited to) CAR T
reported as 5.9 months with the duration of response cell therapy. It is also recognized in a number of
lasting 11.1 months (9,13). Subgroup analysis at different settings, including monoclonal antibody
24 months indicates that in patients who achieve CR, therapies such as with the anti-CD20 agent rituximab.
the progression-free survival is estimated to be 72.0% The clinical syndrome occurs in response to wide-
(95% confidence interval [CI]: 56.0% to 83.0%), and spread release of inflammatory cytokines and che-
although the median overall survival was not reached mokines when large numbers of lymphocytes (B cells,
during the study period, this was estimated to be T cells, or natural killer cells) or myeloid cells (mac-
50.5% (95% CI: 40.2% to 59.7%). These findings rophages, dendritic cells, or monocytes) are activated
illustrate the durability of responses that can be (15); in this setting, it is triggered by activation of T
achieved after a single infusion of CAR T cells. cells upon engagement of the CAR by CD19. This ac-
tion leads to a release, among others, of interleukin
ADVERSE EFFECTS OF CAR T CELL THERAPY (IL)-2, soluble IL-2R a , interferon gamma, IL-6, soluble
IL-6R, and granulocyte-macrophage colony-stimu-
The utilization of CAR T-cell therapies has, however, lating factor by the activated T cells as well as other
been accompanied by significant toxicities (Table 1, inflammatory cytokines and chemokines by
Figure 2). The most common and well-appreciated is bystander immune cells (16). In particular, IL-6 has
that of CRS. The pivotal ZUMA-1 and JULIET studies been shown to be an important mediator of toxicity in
of CAR T-cell therapies in adults with relapsed/re- CRS (15).
fractory B-cell lymphoma were complicated by CRS in Symptoms of CRS can vary significantly. Diag-
93% and 58% of their patient cohorts, respectively; nosing the grade of CRS is key to appropriate man-
severe grade 3 or 4 CRS was only noted in 13% and agement. Early CAR T cell clinical trials produced a
22%, however (12,13). Similarly, the rate of CRS in the number of different CRS grading systems, but there
T A B L E 1 CD19-Specific CAR T Cell Therapy Toxicities
Neelapu et al. Maude et al. Schuster et al.

Maude et al. (ZUMA-1) Fitzgerald et al. (ELIANA) Burstein et al. (JULIET) (11,12), Alvi et al.
(10), 2014 (13), 2017 (25), 2017 (8), 2018 (24), 2018 2018/2019 (26), 2019
(N ¼ 30) (N¼ 101) (N¼ 39) (N ¼ 75) (N ¼ 98) (N ¼ 93) (N ¼ 137)
Cardiovascular toxicities
Profound hypotension requiring 27% 14% 33% 17% 24% 9% 4% (6 patients
inotropic support or shock developed shock
leading to cardiac
deaths)
Left ventricular systolic dysfunction — — 2% — 10% — 6%
Pulmonary edema — — — 6% — — 4% (shortness of
Fluid overload — — 5% breath, hypoxia,
signs of volume
overload, and NT-
proBNP >3000
pg/ml)
ECG and rhythm abnormalities — — Sinus — ST- segment — New-onset
tachycardia changes (18%) arrhythmias
(median peak including
HR 170 beats/ supraventricular
min) tachycardia, atrial
fibrillation, or
flutter (4%)
Biomarker abnormalities — — — — NT-proBNP - Troponin elevation
(92%) (54%)
Lactate (79%)
Mixed venous
saturation (52%)
Other toxicities
Cytokine release syndrome 100% 93% 92% 77% — 58% 59%
Neurotoxicity 43% 64% 33% 40% — 21% —
Management
Treatment with tocilizumab 30% 43% 33% 37% 21% 14% 41%
Treatment with corticosteroids 20% 27% 21% — — 10% —
Treatment with cardioprotective — — — — Beta-blockers — —
medications (21%)
ACE inhibitors
(17%)
CD19-specific chimeric antigen receptor (CAR) T cell therapy is associated with toxicities, including cytokine release syndrome, neurotoxicity, and a number of cardiovascular complications as shown in the
early pivotal trials.
— ¼ data have not been studied or reported within the publication; ACE ¼ angiotensin-converting enzyme; ECG ¼ electrocardiography; ELIANA ¼ Determine Efficacy and Safety of CTL019 in Pediatric
Patients With Relapsed and Refractory B-cell ALL and High Risk B-cell ALL at First Relapse. Determine Feasibility and Safety of CTL019 Therapy in Pediatric Patients With High Risk B-cell ALL That Relapsed
<6 Months Post All-HSCT; HR ¼ heart rate; JULIET ¼ Study of Efficacy and Safety of CTL019 in Adult DLBCL Patients; NT-proBNP ¼ N-terminal pro–B-type natriuretic peptide; ZUMA-1 ¼ Long-Term Safety
and Activity of Axicabtagene Ciloleucel in Refractory Large B-Cell Lymphoma.
was a lack of consensus and significant variability High disease burden (and thus antigenic load) re-
among them. More recently, the American Society for mains the strongest risk factor for severity of CRS
Transplantation and Cellular Therapy published a (18,19), but the dose of cells infused, presence of
consensus grading for CRS (17), the details of which other comorbidities, and early onset of CRS (within
are described in Table 2. CRS is characterized mainly the first 3 days) are also important. Translational
by constitutional symptoms such as high fever, mal- research continues in an attempt to identify and
aise and fatigue, anorexia, myalgias, and nausea. At validate predictive biomarkers (16,19).
the other end of the spectrum, CRS can involve any Management of CRS depends on the grade (17) and
organ system in the body, including the cardiovas- is outlined in detail in an excellent review by Neelapu
cular, nervous, respiratory, gastrointestinal, hepatic, et al. (16); it was also explored in detail in the 2019
renal, and hematological systems. American Society of Clinical Oncology Educational
The symptoms typically manifest within days and Book (20). Low-grade CRS can often be managed with
occasionally weeks after CAR T cell infusion, in supportive care alone, but in more severe cases,
keeping with maximal in vivo T cell expansion in blockade of the IL-6 pathway is recommended either
some translational studies (15). Research has focused with tocilizumab, a monoclonal antibody targeting IL-
on attempts to identify which groups of patients are 6 receptors, or with siltuximab, a monoclonal anti-
at the highest risk for severe CRS to target prevention. body that binds to soluble IL-6. The experience with
F I G U R E 2 Management of CRS
Low-grade cytokine release syndrome (CRS) can be managed with supportive care alone. Tocilizumab, an interleukin (IL)-6 receptor blocker, is
considered as a first line of treatment in patients with grade 2 or higher CRS. When this approach is ineffective, a dose of corticosteroids is the
second-line treatment. ASTCT ¼ American Society for Transplantation and Cellular Therapy.
tocilizumab in CRS is more extensive; it has been CRS is represented in a treatment algorithm in
approved by the U.S. Food and Drug Administration Figure 3.
and has become standard of care for the treatment of NEUROTOXICITY. Neurotoxicity or immune effector
CRS complicating CAR T cell therapy (5). However, cell–associated neurotoxicity syndrome (17) is
not all cases of severe CRS will respond to IL-6 another recognized complication of CAR T cell ther-
pathway interruption; in these cases in which CRS is apies that can occur in up to 50% of treated patients
refractory to tocilizumab, corticosteroid therapy is (18). It has a wide spectrum of manifestations. Early
recommended. Thus far, the use of IL-6 receptor signs include diminished attention, language distur-
blockade and/or corticosteroid therapy does not bance, and dysgraphia before progressing to confu-
result in a higher rate of cancer relapse (21). This has sion, disorientation, agitation, aphasia, somnolence,
led to the administration of tocilizumab and cortico- and tremors. At the severe end of the spectrum of this
steroid earlier in the course of CRS. The usual practice syndrome, patients may experience seizures, motor
in many centers is use of tocilizumab as a first-line and sensory deficits, transverse myelitis, and
therapy in patients who develop grade 2 or higher decreased levels of consciousness accompanied by
CRS, with the addition of steroids as a second-line evidence of increased intracranial pressure (16,22).
therapy when tocilizumab has been ineffective IL-6 has been implicated as a potential underlying
(16,20). This stepwise approach to the management of mechanism for neurotoxicity, much as it has been for
T A B L E 2 ASTCT Consensus Grading for Cytokine Release Syndrome
Grade 1 Grade 2 Grade 3 Grade 4
Fever þ þ þ þ
Temperature $38 C
With or without constitutional
symptoms (which include myalgia,
arthralgia, and malaise)
With
Hypotension – þ þ þ
No vasopressors required Requiring one vasopressor without Requiring multiple vasopressors
vasopressin (excluding vasopressin)
And/or
Hypoxia – þ þ þ
Requiring low-flow nasal cannula or Requiring high-flow nasal cannula, Requiring positive pressure
blow-by facemask, nonrebreather mask, or (i.e., CPAP, BiPAP, intubation,
Venturi mask mechanical ventilation)
The assessment and grading of cytokine release syndrome has not been uniform, particularly early on in the experience with CAR T-cell therapy. Subsequently, this has prompted a consensus approach to
grading the severity of cytokine release syndrome, which was released by the American Society for Transplantation and Cellular Therapy (ASTCT) in 2019.
BiPAP ¼ bilevel positive airway pressure; CPAP ¼ continuous positive airway pressure.
CRS. However, neurotoxicity does not always occur a-agonists and, in 1 patient, by LVSD managed with
concurrently with CRS and may even be detected in milrinone. Hypotension always occurred in the setting
the absence of CRS, suggesting additional or alter- of fever. It was often resistant to treatment with a
native mechanisms (18,23). single inotropic agent; in 10 of the 17 patients requiring
CARDIOVASCULAR ADVERSE EFFECTS. C l i n i c a l inotropic support, more than one agent was required.
m a n i f e s t a t i o n s . Cardiovascular complications and Ninety-six percent of the patients who developed
toxicity are less well defined due to the paucity of cardiovascular dysfunction also had CRS; 18% had CRS
data in the literature. Two retrospective studies grade 3 and 28% had CRS grade 4 according to the Penn
(24,25) reported by the investigators at the Children’s Grading Scale for CRS (28).
Hospital of Philadelphia assessed the cardiovascular A second retrospective analysis of the potential
adverse effects of CAR T cell therapies specifically in cardiovascular toxicity associated with CD19 CAR
pediatric and young adult populations. Until recently, T cell therapy was published by Burstein et al. (24) in
data within the adult population were lacking. How- 2018. They described a cohort of 98 patients, of whom
ever, the retrospective study by Alvi et al. (26) in 24% (n ¼ 24) developed cardiovascular dysfunction;
December 2019 has provided significant new insights. this dysfunction was defined as significant hypoten-
Table 1 provides a summary of the cardiovascular sion requiring inotropic support, which included the
complications that have been observed. use of milrinone, dopamine, epinephrine, norepi-
The most prominent cardiovascular complication nephrine, or vasopressin. All of the patients who
thus far is profound hypotension requiring vaso- developed cardiovascular dysfunction had high-
pressor support and admission to the intensive care grade CRS; 12% had CRS grade 3 and 88% had CRS
unit (24,25). The full spectrum of clinical manifesta- grade 4 according to the Penn Grading Scale for CRS
tions, however, is broad. It can include sinus tachy- (28). There were no cardiovascular deaths in this
cardia (which is often attributed to the CRS fever), cohort; there was, however, a successfully resusci-
elevations in serum troponin levels, left ventricular tated cardiac arrest in one patient 2 months after the
systolic function (LVSD), decompensated heart fail- CAR T cell infusion. As previously identified in CRS, a
ure, new-onset arrhythmias, and cardiovascular heavier burden of disease, defined by a pre-treatment
death (26,27). blast percentage >25% on bone marrow biopsy spec-
The initial analysis by Fitzgerald et al. (25) in 2017 imen, was the most significant predictor of hypoten-
retrospectively assessed cardiovascular toxicity sion requiring inotropic support in this patient cohort
within 39 pediatric and young adult patients receiving (odds ratio: 15.5; 95% CI: 5.1 to 47.1; p < 0.001). Other
CAR T cell therapy for relapsed/refractory ALL. significant predictors included a lower pre-treatment
Tachycardia was observed in the setting of prolonged ejection fraction (p ¼ 0.019) or diastolic dysfunction
high fevers, with a median peak heart rate of 170 beats/ (p ¼ 0.021) prior to infusion of CAR T cells. In both
min. Cardiovascular dysfunction was reported in 36% studies, most of the patients with profound hypo-
(n ¼ 14) of the cohort. This was defined in 13 patients by tension (93% and 88%, respectively) received tocili-
fluid-refractory vasoplegic shock managed with zumab (24,25).
3.05 mmol/l), and mixed venous saturations (52% of

F I G U R E 3 CAR T Cell Toxicities
patients; median 68.5%) were observed (24).
The retrospective review of 137 adult patients by
Alvi et al. (26) in December 2019 provided the first
description of cardiovascular complications associ-
ated with CAR T cell therapy in the adult population
outside of case reports and isolated events within the
earlier clinical trials. Thirty-eight percent of the pa-
tients (n ¼ 53) had pre- and post-CAR T cell infusion
troponin levels measured. In 54% (n ¼ 29), the
troponin result was elevated after CAR T-cell infu-
sion. Cardiovascular events occurred in 12% (n ¼ 17).
There were 6 cardiovascular deaths due to new-onset
heart failure, arrhythmias, hypotension, shock, or
cardiac arrest; an additional 6 patients developed
decompensated cardiac failure requiring intravenous
diuretic agents; and 5 patients had new-onset ar-
rhythmias, including supraventricular tachycardias,
atrial fibrillation, and atrial flutter.
Patients at higher risk of cardiovascular complica-
tions were identified in this study. All cardiovascular
events occurred in patients who also had concomitant
CRS. Furthermore, troponin elevation after CAR T cell
infusion (cardiovascular event rate among patients
with positive troponin vs. negative troponin after
CAR T cell infusion: 55% vs. 4%; p < 0.001) and the
time from onset of CRS to administration of tocilizu-
mab (adjusted odds ratio: 1.22; 95% CI: 1.01 to 1.53;
p ¼ 0.022) were both associated with an increased risk
of developing a cardiovascular event.
Significant multiorgan toxicities can be associated with chimeric receptor antibody (CAR)
It would seem from the currently available data
T-cell therapy.
that cardiovascular toxicities observed in the CD19
CAR T cell space are predominantly noted early and
are associated with CRS. Potential long-term and
Of the patients in the analysis by Burstein et al. (24) late-onset cardiovascular and systemic adverse ef-
who had hypotension requiring inotropic support, at fects from CAR T cell therapy remain incompletely
least 41% (n ¼ 10) had new LVSD during hospitaliza- described. A retrospective review by Cordeiro et al.
tion but not all patients underwent echocardiogra- (29) published in 2020 examined an 86-patient
phy. A deterioration in left ventricular systolic cohort who had all survived beyond 1 year
function in this study was defined as a decrease in following CAR T cell infusion. The authors identified
ejection fraction >10% or fractional shortening >5% the following late adverse events: significant cyto-
compared with baseline or an ejection fraction <55% penias, hypogammaglobulinemia, infections, subse-
or fractional shortening <28% when previously quent malignancies, immune-related events, graft-
normal at baseline. Longer term follow-up also versus-host disease in previous allogeneic stem cell
exhibited cardiac systolic and diastolic dysfunction in transplant recipients, and neurologic psychiatric
7% (n ¼ 7) at the time of discharge, and persistent events. Importantly, there were no cardiovascular
dysfunction in just 1% (n ¼ 1) at 6 months. New complications producing late morbidity or mortality
electrocardiography abnormalities related to ST- in the analysis by Cordeiro et al. Therefore, although
segment changes were present in 18% (n ¼ 6) of pa- we still do not have optimal long-term follow-up
tients with hypotension, requiring inotropic support. data in the CAR T cell field regarding long-term
In the small group of patients who had serum cardiac toxicity and complications, current evidence sug-
biomarkers obtained, frequent abnormalities in gests that most of the cardiovascular complications
B-type natriuretic peptide (92% of patients; median associated with this novel therapy are likely to be
580 pg/mL), lactate (79% of patients; median early post-infusion and transient.
C ENTR AL I LL U STRA T I O N A Single Institution Proposed Cardiac Screening and Monitoring in

Patients Undergoing Chimeric Antigen Receptor T Cell Therapy
Ghosh, A.K. et al. J Am Coll Cardiol CardioOnc. 2020;2(1):97–109.
Patient journey through cardio-oncology screening and monitoring as part of the chimeric receptor antibody (CAR) T cell therapy program at
the University College of London Hospital. CMR ¼ cardiac magnetic resonance; ECG ¼ electrocardiography; NT-proBNP ¼ N-terminal pro–B-
type natriuretic peptide; TnT ¼ troponin T; TTE ¼ transthoracic echocardiography.
P a t h o p h y s i o l o g y . The underlying mechanisms of FUTURE DIRECTIONS IN CARDIO-ONCOLOGY

cardiovascular complications, such as refractory hy- RESEARCH OF CAR T CELL THERAPIES
potension and myocardial toxicity in CAR T cell
therapy, remain poorly understood but are likely to There is a limited body of literature that informs the
be multifactorial. Vascular leak contributes to this field of cardio-oncology with respect to the impact of
clinical phenomenon; significantly positive fluid bal- CAR T cell therapies in hematological malignancies,
ances have been reported by Fitzgerald et al. (25). Hu and additional research is needed to fully define the
et al. (30) observed a frequent diagnosis of capillary extent of the problem. It is important to understand
leak syndrome, defined as a triad of hypotension, whether CAR T cell–associated cardiotoxicity is simply
edema, and hemoconcentration with hypo- an early phenomenon associated with CRS or whether
proteinemia. However, data collection exploring car- there are more directly cardiotoxic effects from CAR T
diovascular complications has thus far been cells, both early and late, that have yet to be defined.
retrospective and ad hoc. Not all of these patients Indeed, once the scale and chronology of the associa-
underwent echocardiography at the time of their hy- tion are clearer, further exploration of the possible
potensive episodes. It is therefore not yet possible to pathophysiology and mechanisms will then be desir-
completely exclude a contribution from LVSD. able such that evidence-based interventions may be
It has also been proposed that LVSD in this setting developed.
represents a nonspecific stress-induced or Takotsubo In our view, the most immediate emphasis should
cardiomyopathy (15), and not direct toxicity from CAR be to clearly define whether cardiotoxicity exists as a
T cells. Stress-induced or Takotsubo cardiomyopathy distinct, separate identity underpinned by direct
is described as acute and transient LVSD associated toxicity from CAR T cells, or if observed, presumed
with catecholamine surges classically precipitated by transient hemodynamic abnormalities are epiphe-
emotionally or physically stressful events (31). The nomena associated with CRS and systemic shock.
LVSD associated with CAR T cell therapy occurs Prospective studies will need to use systematic
within the context of physiological stress from the cardiac surveillance strategies to define the true inci-
CRS and has thus far been shown to be reversible by dence and extent of cardiac injury. These must include
6-month follow-up (24). pre-treatment assessments and close monitoring
CAR T cell therapies have also been investigated in while on treatment. Imaging will play a crucial role in
the treatment of other cancers, including stage III/IV these strategies, including detailed, protocol-driven
melanoma and high-risk or relapsed myeloma, in echocardiography, using up-to-date parameters as
which melanoma-associated antigen 3 (MAGE-A3)– well as cardiac magnetic resonance imaging (CMR)
specified CARs are used. In a detailed case report, (33,34). The utility of biomarkers, including troponins
Linette et al. (32) closely examined the histopathology and natriuretic peptides, as part of the cardiac sur-
in 2 cardiac deaths in response to MAGE-A3–specified veillance strategy in CAR T cell therapy in some centers
CAR T cell therapy for melanoma and myeloma in 2 will also need to be evaluated (35). Many novel bio-
separate patients. In both patients, there was robust markers have generated significant interest in other
in vivo expansion of the T cell population with signif- cancer treatment–related cardiotoxicities, such as ni-
icantly elevated levels of IL-6. The CAR T cell bio- tric oxide metabolites and microRNAs (miRNAs),
distribution was analyzed, and some of the highest among many others; the role of these novel biomarkers
concentrations were found in the myocardium and within the context of cardiovascular toxicity due to
pericardial fluid. The histopathological analysis found CAR T cell therapy should also be considered (36–38).
extensive cardiac myonecrosis and lymphoid infiltra- In addition, the underlying mechanism for cardiac
tion with CD3þ T cells. The CD3 þ T cell infiltration injury needs to be elucidated to mitigate off-tumor
appeared to be specific to the myocardium, as this was toxicities and inform the design of next-generation
not noted in other organs and tissues examined (32). CAR T cells (39). Future CAR validation must incor-
This finding suggests that there may be direct T cell– porate in vitro studies, including peptide scanning
mediated cardiac injury and toxicity due to off- and tissue screening to help identify CARs which
tumor, off-target cross-reactivity of the MAGE-A3– have the potential to be directly myocardiotoxic (40).
specific CAR T cells with the myocardium. However, Experience with MAGE-A3–specific CAR T cells in-
similar histopathological analyses in CD19-specific dicates that off-tumor, off-target toxicity poses a
CAR T cell therapy deaths have not been reported, significant risk to patients. Conversely, toxicity may
and there is no evidence of direct myocardial toxicity also be due to off-tumor, on-target effects; this has
of CD19-specific CAR T cells. generated interest in multiple-antigen targeting
through the use of logic-gated T cells developed may also have other cardiovascular comorbidities.
within the AND, OR, and NOT concepts of Boolean Although, the hemodynamic instability and cardio-
logic (41). Adding AND gated circuits will require both vascular complications associated with CAR T cell
antigens to be present for CAR activation to occur, therapy have thus far been reversible and transient,
whereas NOT gated circuits will allow CAR activation they can be exacerbated and more challenging to
in the presence of one antigen and only if the other is manage in the setting of pre-existing cardiovascular
not present. This strategy has the potential to in- disease (43). In this young field, cardiac evaluation
crease specificity of CARs and reduce toxicity. before receiving CAR T cell therapy varies signifi-
Beyond these initial research objectives, the rele- cantly between centers due to paucity of data in the
vance of cardio-oncology within CAR T cell therapy literature and CAR T cell–specific guidelines and
will be to prevent and/or minimize cardiovascular recommendations. The American Society of Clinical
complications and toxicity from occurring and pro- Oncology has, however, provided a Clinical Practice
gressing (33,34). Subsequently, the next phase of Guideline from 2017 for the prevention and moni-
research will also need to examine the utility of pri- toring of cardiac dysfunction in the setting of either
mary and secondary prevention strategies and to anthracycline exposure or radiotherapy, or both. It
provide longer term follow-up data to assess the advocates for a careful history taking and physical
reversibility of cardiovascular toxicity and what examination, as well as further risk stratification and
impact CAR T cell therapies will have on late cardio- assessment through the appropriate use of trans-
vascular outcomes. thoracic echocardiography (TTE), cardiac magnetic
The success of CAR T cell therapy in cancer treat- resonance imaging (CMR) or multigated acquisition
ment has led to interest in whether engineered T cells scan, and serum cardiac biomarkers (39).
could also be used to target noncancer cells, and To advance our understanding of CAR T car-
preclinical studies have sought to translate this suc- diotoxicity, the cardio-oncology service at the Uni-
cess into the treatment of myocardial disease. versity College of London Hospital has established a
Excessive cardiac fibrosis, mediated by cardiac fibro- cardiac screening and monitoring program in line
blasts, is a key pathophysiological component of with recommendations from the American Society of
many cardiac diseases and heart failure. Fibroblast Clinical Oncology guidelines (39) and in collaboration
activation protein is highly expressed by cardiac fi- with the CAR T cell program team (Central
broblasts in human hearts with dilated cardiomyop- Illustration). Phased implementation of this program
athy and hypertrophic cardiomyopathy; it is not commenced in late 2019.
expressed by cardiomyocytes, is weakly expressed in As part of treatment planning for CAR T cell ther-
normal human hearts, and is therefore a suitable po- apy, all patients undergo risk assessment of cardiac
tential target for immunotherapy. Subsequently, the function. In the first instance, all patients complete a
administration of fibroblast activation protein- questionnaire to identify those with a high cardiac
specific CAR T cells into a cohort of mice, in which risk profile. This includes patients who have had a
cardiac injury had been induced, has been shown to history of chemotherapy-induced or other cardio-
significantly reduce, and in some cases completely myopathies, pre-existing arrhythmias or coronary
eliminate, cardiac fibrosis compared with control artery disease, or a high risk of coronary artery dis-
mice (42). In addition, there was partial rescue of both ease based on their calculated QRISK3-2018 score
systolic and diastolic cardiac function. Future (44). All patients have blood samples drawn for
research will therefore focus not just on cardiac injury troponin and N-terminal pro–B-type natriuretic pep-
associated with CAR T cell therapy, but indeed tide levels, a 12-lead electrocardiogram, cardiac im-
whether this exciting preclinical study can be trans- aging with a TTE (including 2-, 3-, and 4-dimensional
lated to humans for CAR T cell therapy to successfully left ventricular ejection fraction, diastolic, and global
treat and prevent myocardial disease. longitudinal strain assessments) and a CMR using a
fast scan protocol. Significant abnormalities of the
HOW TO SCREEN AND MONITOR PATIENTS patient’s biomarkers or cardiac diagnostic tests, or a
UNDERGOING CAR T CELL THERAPY high cardiac risk profile identified by the question-
naire, will subsequently prompt cardio-oncology re-
Adults receiving CAR T cell therapy are likely to view before commencing CAR T cell therapy.
represent a particularly at-risk patient population for Initiation of cardioprotective therapies with beta-
cardiac injury. They have often already had previous blockers and either angiotensin-converting enzyme
exposure to cardiotoxic cancer treatments, including inhibitors or angiotensin receptor blockers is
anthracyclines and thoracic radiotherapy, and often considered.
After administration of CAR T cell therapy, the bio- population most at risk through the use of serial
markers, 12-lead electrocardiogram, and TTE are echocardiography, cardiac biomarkers, clinical data,
repeated as part of cardiac surveillance. If there are and quality of life questionnaires.
abnormalities on these cardiac investigations or an
elevation in the cardiac biomarkers, or in the context of
CONCLUSIONS
clinical deterioration (including profound hypoten-
sion, arrhythmias, clinical signs and symptoms of CAR T cell therapies have revolutionized the thera-
cardiac failure, and grade III or IV CRS), these patients peutic approach to highly refractory and relapsing
receive inpatient cardio-oncology review and further hematological malignancies in which management
imaging with CMR. Initiation of a beta-blocker, options had previously been limited or altogether
angiotensin-converting enzyme inhibitor, or angio- exhausted. The potency of this treatment is accom-
tensin receptor blocker is considered if LVSD is noted panied by potentially significant toxicity that
on TTE or CMR imaging. includes a potential risk of cardiovascular complica-
All patients are subsequently followed up from a tions, with the early evidence suggesting the latter is
cardio-oncology perspective at 3 months after CAR T likely to be transient. The evidence base surrounding
cell infusion. In addition to clinical assessment by the cardiovascular complications and toxicity in partic-
cardio-oncology team, the biomarkers, 12-lead elec- ular remains limited. Multimodality cardiac imaging
trocardiogram, and TTE are repeated. and established and novel biomarkers will need
There is, as yet, no evidence in the literature that to drive future research to more fully define the car-
examines the utility of increased cardiac surveillance diac safety profile of CAR T cell therapies, to identify
in the prevention, management, or improvement of the population who may be at the highest risk and
outcomes in patients receiving CAR T cell therapy. enable early detection of cardiovascular complica-
Subsequently, the data from this cardiac screening and tions and toxicity, and guide their management
monitoring program are being prospectively collated along evidence-based interventions applicable to this
with the aim of providing further information to the unique class of therapies.
CAR T cell and cardio-oncology community in the
future.
In addition, there is at least one prospective ADDRESS FOR CORRESPONDENCE: Mr. Daniel H.
observational study, led by the Abramson Chen, Hatter Cardiovascular Institute, Institute of
Cancer Centre of the University of Pennsylvania Cardiovascular Science UCL, University College Lon-
(NCT04026737), that is underway. Similarly, they will don Hospital, 67 Chenies Mews, London WC1E 6HX,
seek to define the incidence, natural history, and United Kingdom. E-mail: daniel.chen1@nhs.net.
progression of cardiac dysfunction and determine the Twitter: @DanielH_Chen, @arjunkg.
REFERENCES
1. Park JH, Geyer MB, Brentjens RJ. CD19-targeted 7. Crump M, Neelapu SS, Farooq U, et al. Out- 12. Schuster SJ, Bishop MR, Tam CS, et al. Tisa-
CAR T-cell therapeutics for hematologic malig- comes in refractory diffuse large B-cell lymphoma: genlecleucel in adult relapsed or refractory diffuse
nancies: interpreting clinical outcomes to date. results from the international SCHOLAR-1 study. large B-cell lymphoma. N Engl J Med 2019;380:
Blood 2016;127:3312. Blood 2017;130:1800–8. 45–56.
8. Maude SL, Laetsch TW, Buechner J, et al. 13. Neelapu SS, Locke FL, Bartlett NL, et al. Axi-
2. Brown CE, Mackall CL. CAR T cell therapy: in-
Tisagenlecleucel in children and young adults with cabtagene ciloleucel CAR T-cell therapy in re-
roads to response and resistance. Nature Rev
B-cell lymphoblastic leukemia. N Engl J Med 2018; fractory large B-cell lymphoma. N Engl J Med
Immunol 2019;19:73–4.
378:439–48. 2017;377:2531–44.
3. Salter AI, Pont MJ, Riddell SR. Chimeric antigen 14. Schuster SJ, Svoboda J, Chong EA, et al.
receptor–modified T cells: CD19 and the road 9. Locke FL, Ghobadi A, Jacobson CA, et al. Long-
Chimeric antigen receptor T cells in refractory B-
beyond. Blood 2018;131:2621. term safety and activity of axicabtagene ciloleucel
cell lymphomas. N Engl J Med 2017;377:2545–54.
in refractory large B-cell lymphoma (ZUMA-1): a
4. Maude S, Barrett DM. Current status of chimeric single-arm, multicentre, phase 1-2 trial. Lancet 15. Lee DW, Gardner R, Porter DL, et al. Current
antigen receptor therapy for haematological ma- Oncol 2019;20:31–42. concepts in the diagnosis and management of cyto-
lignancies. Br J Haematol 2016;172:11–22. kine release syndrome. Blood 2014;124:188–95.
10. Maude SL, Frey N, Shaw PA, et al. Chimeric
5. Subklewe M, von Bergwelt-Baildon M, antigen receptor T cells for sustained remissions in 16. Neelapu SS, Tummala S, Kebriaei P, et al.
Humpe A. Chimeric antigen receptor T cells: a race leukemia. N Engl J Med 2014;371:1507–17. Chimeric antigen receptor T-cell therapy—assess-
to revolutionize cancer therapy. Transfus Med ment and management of toxicities. Nat Rev Clin
11. Schuster SJ, Bishop MR, Tam C, et al. Sustained
Hemother 2019;46:15–24. Oncol 2018;15:47–62.
disease control for adult patients with relapsed or
6. Brudno JN, Kochenderfer JN. Chimeric antigen refractory diffuse large B-cell lymphoma: an 17. Lee DW, Santomasso BD, Locke FL, et al.
receptor T-cell therapies for lymphoma. Nat Rev updated analysis of JULIET, a global pivotal phase ASTCT consensus grading for cytokine release
Clin Oncol 2017;15:31. 2 trial of tisagenlecleucel. Blood 2018;132:1684. syndrome and neurologic toxicity associated with
immune effector cells. Biol Blood Marrow Trans- 28. Porter DL, Hwang WT, Frey NV, et al. Chimeric with anthracycline chemotherapy. J Am Heart
plant 2019;25:625–38. antigen receptor T cells persist and induce sus- Assoc 2017;6:e004653.
tained remissions in relapsed refractory chronic
18. Brudno JN, Kochenderfer JN. Toxicities of 39. Armenian SH, Lacchetti C, Barac A, et al. Pre-
lymphocytic leukemia. Sci Transl Med 2015;7:
chimeric antigen receptor T cells: recognition and vention and monitoring of cardiac dysfunction in
303ra139.
management. Blood 2016;127:3321. survivors of adult cancers: American Society of
29. Cordeiro A, Bezerra ED, Hirayama AV, et al. Clinical Oncology Clinical Practice Guideline. J Clin
19. Frey N. Cytokine release syndrome: who is at
Late events after treatment with CD19-targeted Oncol 2017;35:893–911.
risk and how to treat. Best Prac Res Clin Oncol
chimeric antigen receptor modified T cells. Biol
2017;30:336–40. 40. Cameron BJ, Gerry AB, Dukes J, et al. Identi-
Blood Marrow Transplant 2020;26:26–33.
fication of a titin-derived HLA-A1–presented pep-
20. Santomasso B, Bachier C, Westin J, Rezvani K, 30. Hu Y, Feng J, Shao M, Huang H. Profile of tide as a cross-reactive target for engineered
Shpall EJ. The Other Side of CAR T-Cell Therapy: capillary-leak syndrome in patients received MAGE A3–directed T cells. Sci Transl Med 2013;5:
Cytokine Release Syndrome, Neurologic Toxicity, chimeric antigen receptor T cell therapy. Blood 197ra103.
and Financial Burden. American Society of Clinical 2018;132:5204.
Oncology Educational Book 2019:433–44. 41. Charrot S, Hallam S. CAR-T cells: future per-
31. Medina de Chazal H, Del Buono MG, Keyser- spectives. HemaSphere 2019;3:e188.
21. June CH, Sadelain M. Chimeric antigen recep- Marcus L, et al. Stress cardiomyopathy diagnosis
42. Aghajanian H, Kimura T, Rurik JG, et al. Tar-
tor therapy. N Engl J Med 2018;379:64–73. and treatment: JACC State-of-the-Art Review.
geting cardiac fibrosis with engineered T cells.
22. Belin C, Simard C, Dos Santos A, et al. JS1.4 J Am Coll Cardiol 2018;72:1955–71.
Nature 2019;573:430–3.
Neurological follow-up of neurotoxicity after CAR 32. Linette GP, Stadtmauer EA, Maus MV, et al.
43. Ganatra S, Carver JR, Hayek SS, et al. Chimeric
T cell therapy in lymphoma patients: a French Cardiovascular toxicity and titin cross-reactivity of
antigen receptor T-cell therapy for cancer and
neurological multi-center survey. J Neurooncol affinity-enhanced T cells in myeloma and mela-
heart: JACC Council Perspectives. J Am Coll Car-
2019;21. iii5–iii5. noma. Blood 2013;122:863.
diol 2019;74:3153–63.
23. Gust J, Finney OC, Li D, et al. Glial injury in 33. Plana JC, Thavendiranathan P, Bucciarelli-
44. Hippisley-Cox J, Coupland C, Brindle P.
neurotoxicity after pediatric CD19-directed Ducci C, Lancellotti P. Multi-modality imaging in
Development and validation of QRISK3 risk pre-
chimeric antigen receptor T cell therapy. Ann the assessment of cardiovascular toxicity in the
diction algorithms to estimate future risk of car-
Neurol 2019;86:42–54. cancer patient. J Am Coll Cardiol Img 2018;11:1173.
diovascular disease: prospective cohort study.
24. Burstein DS, Maude S, Grupp S, Griffis H, 34. Seraphim A, Westwood M, Bhuva AN, et al. BMJ 2017;357:j2099.
Rossano J, Lin K. Cardiac profile of chimeric anti- Advanced imaging modalities to monitor for 45. Makita S, Imaizumi K, Kurosawa S, Tobinai K.
gen receptor T cell therapy in children: a single- cardiotoxicity. Curr Treat Options Oncol 2019; Chimeric antigen receptor T-cell therapy for B-cell
institution experience. Biol Blood Marrow Trans- 20:73. non-Hodgkin lymphoma: opportunities and chal-
plant 2018;24:1590–5. 35. Yu AF, Ky B. Roadmap for biomarkers of cancer lenges. Drugs Context 2019;8:212567.
25. Fitzgerald JCMDP, Weiss SLMDM, therapy cardiotoxicity. Heart 2016;102:425.
Maude SLMDP, et al. Cytokine release syndrome 36. Michel L, Rassaf T, Totzeck M. Biomarkers for
after chimeric antigen receptor T cell therapy for the detection of apparent and subclinical cancer KEY WORDS cardiomyopathy, cytokine
acute lymphoblastic leukemia. Crit Care Med 2017; therapy-related cardiotoxicity. J Thorac Dis 2018: release syndrome, immunotherapy, leukemia,
45:e124–31. S4282–95. lymphoma
26. Alvi RM, Frigault MJ, Fradley MG, et al. Car- 37. Tan LL, Lyon AR. Role of biomarkers in pre-
diovascular events among adults treated with diction of cardiotoxicity during cancer treatment.
Go to http://www.acc.org/
chimeric antigen receptor T-cells (CAR-T). J Am Curr Treat Options Cardiovasc Med 2018;20:55.
jacc-journals-cme to take
Coll Cardiol 2019;74:3099–108.
38. Leger Kasey J, Leonard D, Nielson D, de Lemos the CME/MOC/ECME quiz
27. Asnani A. Cardiotoxicity of immunotherapy: James A, Mammen Pradeep PA, Winick Naomi J. for this article.
incidence, diagnosis, and management. Curr Oncol Circulating microRNAs: potential markers of car-
Rep 2018;20:44. diotoxicity in children and young adults treated
ª 2020 THE AUTHOR. PUBLISHED BY ELSEVIER ON BEHALF OF THE AMERICAN
EDITORIAL COMMENT
CARTing Away Cardiac Fibrosis*

Jonathan A. Epstein, MD
P athological cardiac fibrosis

nearly all forms of heart failure and contrib-
utes to mechanical dysfunction
diseased heart. After myocardial infarction, the
accompanies
of the
scar, teach that scar is not immutable. Fibroblasts
themselves, as well as myocytes, macrophages,
and other cells, make enzymes that
matrix including metalloproteases, collagenases, and
degrade
fibrotic response is a critical defense against myocar- endopeptidases. Thus, mouse models in which
dial rupture, but adverse remodeling subsequently cardiac fibroblasts are destroyed using genetic
contributes to progression of heart failure. Fibrosis approaches (including the expression of diphtheria
results, at least in part, from inflammation and im- toxin in activated fibroblasts) after scar has already
mune activation that accompanies cardiac stress, formed reveal the possibility in at least some
ischemia, infarction, or exposure to a host of toxic circumstances of rapid resorption of existing scar and
agents. Therapeutic approaches to intervene in the
improved cardiac function as a result (2).
inflammation-fibrosis axis have been the focus of
SEE PAGE 97
intense research and clinical investigation, albeit
with generally disappointing results to date (1). At Based on these observations, we recently tested
present, there are essentially no effective therapies the efficacy of engineered T cells targeted to seek and
for established cardiac fibrosis. destroy activated fibroblasts in the diseased heart (3).
Emerging results from animal studies suggest that One of the most exciting advances in cancer therapy
targeting cardiac fibroblasts, even after fibrosis and in recent times has been the advent of cell-based
heart failure are established, is an approach worthy of therapies to target leukemia and other malignant
further investigation. At first glance, this is not an cell types (4). Patients with acute lymphoblastic leu-
intuitive conclusion because large portions of an kemia who did not respond to all conventional ther-
established cardiac scar are composed of extracellular apies and have little hope of survival have been
matrix, which may in some cases be “fixed” and untreated with a single dose of engineered T cells and
responsive to destruction of the fibroblasts that have achieved lasting remission. In these cases, the
originally produced the matrix. However, we are patients’ own T cells are removed from the body,
learning that scar tissue is often actively maintained exposed ex vivo to a lentivirus encoding a chimeric
by ongoing production and concomitant resorption antigen receptor (CAR) that recognizes the CD19
processes. Diseases such as scurvy, in which old molecule expressed on all B cells (including leukemic
wounds can reopen due to breakdown of existing B cells), and reinfused with functional expression of
the engineered CAR. Rapid tumor lysis ensues,
frequently resulting in release of potent cytokines
*Editorials published in JACC: CardioOncology reflect the views of the that can induce hyperactivation of the immune
authors and do not necessarily represent the views of JACC: system, a life-threatening complication known as
cytokine release syndrome (CRS), which can include
From the Perelman School of Medicine at the University of Pennsylvania, cardiac toxicity, as reviewed by Ghosh et al. (5) in this
Philadelphia, Pennsylvania. Dr. Epstein reports a U.S. Provisional Patent
issue of JACC: CardioOncology. With appropriate
Application No. 62/563,323, filed September 26, 2017 entitled: “Treatment
of Heart Disease via CAR-T Cell Immunotherapy.” supportive therapy, patients often emerge disease-
The author attests they are in compliance with human studies commit- free (although requiring ongoing immunoglobulin
therapy due to the lack of B cells) and long-lasting
appropriate. For more information, visit the JACC: CardioOncology author clones of engineered T cells can provide ongoing
instructions page. surveillance against tumor recurrence.

JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020 Epstein 111
MARCH 2020:110–3 CARTing Away Cardiac Fibrosis
The opportunity to extend this approach to treat recognize FAP (Figure 1) (3). Infused FAP CAR T cells
fibrotic diseases is attractive, although it is necessary accumulated in the heart within 1 day of delivery.
to identify antigens expressed by pathological, acti- Treated animals did not exhibit signs of CRS, such as
vated fibroblasts that can allow them to be distin- fever, lethargy, or other untoward reactions, and only
guished from quiescent, physiological fibroblasts that mild changes in serum cytokines or gene expression
are necessary for tissue integrity and homeostasis. abnormalities indicative of cardiotoxicity were
A growing body of work using powerful tools of single recorded. This is likely due to the fact that the overall
cell and single nucleus analysis coupled with genetic burden of cell lysis in this model is far lower than in
lineage tracing in mouse models suggests that there leukemia, where CRS has been observed, and perhaps
are many types of fibroblasts in the heart and in other because fibroblasts contain fewer cytokines released
tissues, and that fibroblasts induced to proliferate, to upon cell lysis than lymphocytes.
secrete extracellular matrix, and to modulate the Interestingly, animals treated with FAP CAR T
immune response can be identified through gene cells were able to heal skin wounds at an identical
expression characteristics and cell surface markers rate to control animals, despite the fact that some
(6,7). activated fibroblasts in the skin express FAP. This is
To identify an appropriate antigenic target for likely due to the fact that many activated fibroblasts
CAR T therapy in human fibrotic hearts, we analyzed in the skin do not express FAP and are able to
RNA sequencing data from diseased explanted hu- compensate for those that are ablated by FAP CAR T
man hearts obtained at the time of transplantation cells. This result is consistent with the notion that
and compared these with healthy controls. We there are many types of fibroblasts in the body, with
identified fibroblast-specific transcripts that were populations characterized by different markers of
expressed in hearts of patients with hypertrophic and activation. Indeed, in the hearts treated with
dilated cardiomyopathies, but not in healthy control angiotensin and phenylephrine we noted a popula-
hearts. Among these candidates, Fibroblast Activa- tion of perivascular fibroblasts that form a “cuff”
tion Protein (FAP) showed the greatest difference around small coronary arteries that do not express
between control and diseased specimens (3). FAP and are not removed by FAP CAR T therapy. The
FAP is a transmembrane and secreted glycoprotein clinical implications of this finding are yet to be
with prolyl endopeptidase activity that has been determined, but the existence of multiple pop-
extensively studied. It is expressed at very low levels ulations of activated fibroblasts raises the prospect
or not at all in most healthy tissues, but expression is of developing an array of engineered T cells target-
up-regulated in many activated fibroblasts associated ing different populations relevant to specific dis-
with injury and tumor stroma. FAP is expressed by eases or tissues.
fibroblasts in hearts of patients with cardiomyopathy These proof-of-concept results in mice are
and in scar associated with infarction (8). It is also encouraging, although multiple hurdles exist before
expressed by fibroblasts in the cirrhotic liver, arthritic first-in-human clinical trials can be initiated. Unlike
joints, and other diseased tissues. CAR T cells tar- cancer, where immune surveillance and long-term
geting FAP have been developed for use in cancer, persistence of targeted T cells is an advantage and
including pancreatic cancer where tumor stroma is a where every last cancer cell should be destroyed,
dominant feature of disease, and in mesothelioma (9). fibrotic disorders demand only an overall reduction in
Several patients in Europe have received FAP CAR T disease burden to achieve clinical benefit. Persistence
cells without reported complications (NCT01722149), of anti-fibroblast T cells is likely to be associated with
although the results of these trials have not been untoward effects and complications. A treated patient
reported in the literature. who has a subsequent myocardial infarction, for
A mouse model of hypertensive heart disease that example, might be prone to myocardial rupture. A
has been well validated in the literature involves short-lived, transient CAR T cell is far more attractive,
chronic infusion of angiotensin II and phenylephrine or one in which activity could be regulated with a
via implanted osmotic pumps over a 2-week period. “switch” or small-molecule regulator, such as those
These animals develop significant cardiac fibrosis and being developed in multiple laboratories. A short-
signs of cardiac dysfunction. Treatment with FAP lived CAR T cell can be produced by engineering the
CAR T cells, delivered after the appearance of cardiac cells with messenger RNA instead of virally encoded
fibrosis, dramatically attenuated fibrosis at 4 weeks DNA, thus avoiding the need for a viral vector and the
and improved function when compared with animals possibility of genomic integration (10). A messenger
treated with T cells that had not been engineered to RNA–generated CAR T cell could be delivered in doses
112 Epstein JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020
CARTing Away Cardiac Fibrosis MARCH 2020:110–3
F I G U R E 1 FAP CAR T Cells Can Target Cardiac Fibrosis
(A) Schematic of experiments for FAP CAR T cell targeting of cardiac fibroblasts. (B) Picro-Sirius red staining of heart coronal sections in mice treated with saline (left),
AngII/PE (center), or AngII/PE and FAP CAR T cells (right) to evaluate fibrosis (red). (Bottom) Magnification of left ventricular fibrosis. Scale bar, 100 mm.
(C) Quantification of cardiac fibrosis. ****p < 0.0001; one-way ANOVA between groups, p < 0.0001; Tukey test was used for post-hoc multiple comparisons; n ¼ 10, 9,
and 7 biologically independent mice, from left to right. Reproduced with permission from Aghajanian et al. (3). ANGII/PE ¼ angiotensin II and phenylephrine;
ANOVA ¼ analysis of variance; CAR ¼ chimeric antigen receptor; FAP ¼ fibroblast activation protein.
as needed to produce clinical benefit, perhaps by also be important tools to facilitate clinical trans-
infusion into a coronary artery to produce maximal lation. FAP–positron emission tomography (PET)
“first pass” effect. Such “second-generation” CAR T probes are already in development and identify
cells will have to be developed against FAP or other robust signal in the hearts of animals after myocar-
fibroblast-specific antigens and validated in animal dial infarction but not in healthy controls (11) and
models before moving to the clinic. these probes have also been used in humans. FAP-
The ability to assess cardiac fibrotic burden and PET imaging may be useful to screen for potential
the expression of the target antigen (e.g., FAP) will non-cardiac effects of FAP CAR T therapy due to
JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020 Epstein 113
MARCH 2020:110–3 CARTing Away Cardiac Fibrosis
FAP-expressing scar in non-cardiac tissues, and for

monitoring the effectiveness of FAP CAR T therapy. ADDRESS FOR CORRESPONDENCE: Dr. Jonathan A.
Although it is early days, the prospect of using Epstein, Perelman School of Medicine at the Univer-
immunotherapy and engineered immune cells to sity of Pennsylvania, 602 South Tower, PCAM, 3400
treat cardiac disease, and other forms of fibrotic Civic Center Boulevard, Philadelphia, Pennsylva-
illness, represents a new and exciting frontier in nia 19104. E-mail: epsteinj@upenn.edu. Twitter:
medicine. @JonEpsteinLab.
REFERENCES
1. Mann DL, McMurray JJ, Packer M, et al. Tar- 5. Ghosh AK, Chen DH, Guha A, Mackenzie S, 9. Petrausch U, Schuberth PC, Hagedorn C, et al.
geted anticytokine therapy in patients with Walker JM, Roddie C, et al. CAR T-cell therapy– Re-directed T cells for the treatment of fibroblast
chronic heart failure: results of the Randomized related cardiovascular outcomes and manage- activation protein (FAP)-positive malignant
Etanercept Worldwide Evaluation (RENEWAL). ment: systemic disease or direct cardiotoxicity? pleural mesothelioma (FAPME-1). BMC Cancer
Circulation 2004;109:1594–602. J Am Coll Cardiol CardioOnc 2020;2:97–109. 2012;12:615.
2. Kaur H, Takefuji M, Ngai CY, et al. Targeted 6. Kanisicak O, Khalil H, Ivey MJ, et al. Genetic 10. Foster JB, Barrett DM, Kariko K. The emerging
ablation of periostin-expressing activated fibro- lineage tracing defines myofibroblast origin and role of in vitro-transcribed mRNA in adoptive T
blasts prevents adverse cardiac remodeling in function in the injured heart. Nat Commun 2016;7: cell immunotherapy. Mol Ther 2019;27:747–56.
mice. Circ Res 2016;118:1906–17. 12260.
11. Varasteh Z, Mohanta S, Robu S, et al. Molecular
3. Aghajanian H, Kimura T, Rurik JG, et al. Tar- 7. Tallquist MD, Molkentin JD. Redefining the imaging of fibroblast activity after myocardial
geting cardiac fibrosis with engineered T cells. identity of cardiac fibroblasts. Nat Rev Cardiol infarction using a (68)Ga-labeled fibroblast acti-
Nature 2019;573:430–3. 2017;14:484–91. vation protein inhibitor, FAPI-04. J Nucl Med
2019;60:1743–9.
4. June CH, O’Connor RS, Kawalekar OU, 8. Tillmanns J, Hoffmann D, Habbaba Y, et al.
Ghassemi S, Milone MC. CAR T cell immuno- Fibroblast activation protein alpha expression
therapy for human cancer. Science 2018;359: identifies activated fibroblasts after myocardial KEY WORDS cardiac fibrosis, chimeric antigen
1361–5. infarction. J Mol Cell Cardiol 2015;87:194–203. receptor, fibroblast activation protein
ISSN 2666-0873
RESEARCH LETTER
Fatigue, Cardiovascular Participants receive a laminated schedule card of

visits and monthly motivational text messages to
Decline, and Events promote retention as well as the following incentives:
after Breast Cancer an UPBEAT tote bag, honoree stickers for study visits,
and gift cards at each study visit. This plan has
Treatment resulted in 94% retention (in both groups) at the 3-
Rationale and Design of UPBEAT Study month visit. UPBEAT is a highly collaborative study,
and welcomes those interested in participating as a
study site or submitting an ancillary study proposal to
Although survival rates for stage I to III breast cancer contact the corresponding author.
have greatly increased, in part due to improved All patients meeting eligibility criteria at partici-
treatment options, this progress is threatened by pating sites are invited to enroll and may identify
increased cardiovascular (CV) events for survivors. family or friends as healthy control subjects; partici-
More than 35% of women experience CV injury, left pating sites may also locally recruit through flyers.
ventricular (LV) dysfunction, exercise intolerance, Baseline for control subjects is considered to be time
and/or fatigue after receipt of neoadjuvant or adju- of consent and study enrollment.
vant chemotherapy for stage I to III breast cancer Inclusion criteria common between the groups
(1,2). There are, however, major gaps in scientific include: age 18 years or older; ability to hold breath
knowledge regarding the origins and causes of for 10 s; ability to walk at least 2 blocks without chest
increased CV dysfunction, exercise intolerance, and pain, dyspnea, or syncope; and ability to exercise on a
fatigue in this population. treadmill or stationary cycle. For participation in the
The UPBEAT (Understanding and Predicting Fa- chemotherapy group, women must have stage I to III
tigue, Cardiovascular Decline, and Events After breast cancer (including inflammatory, newly diag-
Breast Cancer) study will prospectively determine the nosed, or locally recurrent being treated with curative
influence of neoadjuvant and adjuvant chemotherapy intent but not metastatic breast cancer), be scheduled
on the following: 1) CV and physical function; 2) to receive chemotherapy, and have an ECOG perfor-
submaximal and maximal exercise capacity; 3) fa- mance status of 0 to 2. Patients may also participate
tigue; and 4) future CV events in a cohort of 1,000 in other ongoing clinical trials. Women in the control
women (840 with stage I to III breast cancer and 160 group must be healthy; have no history of cancer or
healthy control participants). This study ascertains breast surgery; have no previous exposure to
pre-existing and dynamic changes during and after chemotherapy, radiation therapy, or immunotherapy;
chemotherapy receipt with active surveillance of and have an ECOG performance status of 0 to 1.
subsequent CV events for 7 to 10 years following Exclusion criteria include inability to provide
treatment. UPBEAT is designed to address these informed consent, contraindications to cardiovascu-
knowledge gaps by prospectively identifying which lar magnetic resonance (CMR) imaging, a known LV
baseline and early dynamic changes may precede CV ejection fraction (LVEF) <50%, pregnant or breast-
dysfunction and exercise intolerance (Figure 1). feeding at baseline, severe pulmonary hypertension,
The UPBEAT study is currently recruiting partici- history of either acute pulmonary embolus or deep
pants through academic and community-based med- vein thrombosis within 6 months, or any history of
ical centers who are affiliated with either the Wake the following conditions within the past 1 month:
Forest NCI Community Oncology Research Program myocardial infarction, unstable or stable angina, left
(NCORP) or the Eastern Cooperative Oncology Group– main coronary artery disease, symptomatic heart
American College of Radiology Imaging Network failure, uncontrolled hypertension, severe valvular
(ECOG-ACRIN)/ECOG-ACRIN NCORP. Currently, 239 heart disease, uncontrolled metabolic disease, aortic
participants have been enrolled (127 in the breast aneurysm or dissection, uncontrolled arrhythmia
cancer group; 112 control subjects) across 12 centers. causing symptoms or hemodynamic compromise, or
JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020 Research Letter 115
MARCH 2020:114–8
F I G U R E 1 Conceptual Model for the UPBEAT Study
The study will address the relations between cancer therapy, cardiovascular (CV) injury, exercise intolerance, fatigue, and CV events as primary
outcomes. Relative contributions of age, menopausal status, race/ethnicity, and radiation therapy, as well as psychosocial and behavioral risk
factors, will be analyzed as covariates. CRP ¼ C-reactive protein; LVEF ¼ left ventricular ejection fraction; NT-proBNP ¼ N-terminal pro–B-
type natriuretic peptide; PWV ¼ pulse wave velocity; TnI ¼ troponin I; UPBEAT ¼ Understanding and Predicting Fatigue, Cardiovascular
Decline, and Events After Breast Cancer.
hypertrophic obstructive cardiomyopathy. Additional measured at a subset of sites, is the secondary exer-
control group exclusion criteria include inflammatory cise capacity outcome because it is an objective
conditions and known coronary artery disease or measure of maximal exercise capacity, integrates the
heart failure. Because the healthy control group may physiological response to exercise, and provides ex-
lack representation of common CV risk factors, this ercise and exertional fatigue data (6).
selection bias may limit generalizability. To address Fatigue and health-related quality of life (HRQOL)
this limitation, UPBEAT has a data-sharing agreement questionnaires and both cognitive and physical
with the MESA (Multi-Ethnic Study of Atheroscle- function measures are collected. Fatigue is measured
rosis) to explore data in a more generalizable control by using the Functional Assessment of Cancer
group. Therapy–Fatigue scale that assesses fatigue and
The primary endpoints include a 6% LVEF decline anemia-related concerns in people with cancer.
or LVEF <50% (2) and aortic pulse wave velocity. HRQOL and health status are measured by using the
These primary outcomes are acquired with a 10- to RAND 36-item Short Form Health Survey to assess
15-min rapid CMR examination at baseline, 3 months, general HRQOL and the shortened Kansas City Car-
and 24 months (Figure 2). CMR is the gold standard diomyopathy Questionnaire to assess physical limi-
for volumetric quantification; it measures cardiac and tations, symptoms, HRQOL, and social limitations.
vascular structure and function, and does not use Cognitive function is assessed with a Controlled Oral
ionizing radiation. Word Association test, the Trail Making A & B cogni-
The secondary CMR outcomes include LV end- tive test, and self-reported cognitive function ques-
diastolic and end-systolic volumes and mass (3), tionnaire. Physical function is assessed with the Short
myocardial strain and strain rate, mid-cavity short- Physical Performance Battery protocol, grip strength,
axis LV myocardial T1 and T2 (4), and abdominal body and upper body range of motion.
composition measures. The primary exercise perfor- In addition, the following psychosocial and
mance measure is the 6-min walk distance to measure behavioral covariates (which also affect fatigue and
submaximal exercise capacity, an independent pre- exercise intolerance in breast cancer survivors) are
dictor of CV mortality (5). Peak oxygen consumption, assessed: depressive symptoms, assessed by using
116 Research Letter JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020
MARCH 2020:114–8
F I G U R E 2 UPBEAT Study Measures
Serial UPBEAT study measures through 24 months of observation encompass a range of cardiovascular and psychosocial measures of fatigue
and exercise intolerance in breast cancer patients and healthy control subjects with up to 10 years of active annual surveillance for CV events.
Abbreviations as in Figure 1.
the Center for Epidemiological Studies Depression Age, race/ethnicity, performance status, preg-
Scale; perceived stress as measured by using the nancy status, menopausal status, and hostility are
Perceived Stress Scale; sleep disturbance, assessed by collected for each participant at study enrollment.
using the PROMIS (Patient-Reported Outcomes Mea- Serial assessments of weight, height, body mass
surement Information System) scale; social support, index, blood pressure, heart rate, health and family
measured by using the RAND Social Support Scale to history, CV comorbidities (hypertension, diabetes,
assess social, emotional, tangible information and hypercholesterolemia, and coronary artery disease),
appraisal support; walking self-efficacy as measured smoking and tobacco use, and medications occur at
according to a self-efficacy scale; the PACE (Patient- study visits. Serum and plasma biomarkers (i.e.,
Centered Assessment and Counseling for Exercise) hematocrit, glucose, creatinine, low- and high-
questionnaire to assess sedentary behaviors; and the density lipoprotein cholesterol, triglycerides, total
Godin Leisure-Time Exercise Questionnaire to assess cholesterol, C-reactive protein, troponin I, N-termi-
self-reported physical activity. nal pro–B-type natriuretic peptide) are ascertained
Finally, adjudicated CV events include myocardial in control subjects and participants with cancer at
infarction or cardiac death (death in the presence of the study visits.
acute myocardial infarction, significant cardiac Cancer diagnosis and treatment information are
arrhythmia, or refractory heart failure) (7). UPBEAT collected in participants with breast cancer; this in-
also identifies symptomatic heart failure warranting formation includes cancer diagnosis/staging, hor-
hospitalization, all-cause mortality, and coronary ar- mone receptor and Her2 status, and current cancer
tery revascularization (7). treatment (e.g., chemotherapy, surgery, radiation
JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020 Research Letter 117
MARCH 2020:114–8
therapy planning records/images, radiation dose). cardio-oncology research will reduce the knowledge
UPBEAT also collects information regarding heredi- gaps surrounding CV health of breast cancer survivors
tary gene mutations, LVEF at diagnosis, and prior through a well-characterized cohort study. In a study
primary cancer diagnosis/treatment, if available. of this magnitude, it is important to assess a wide
The sample size of 840 breast cancer patients and range of measures to understand pre-existing risk
160 healthy participants was chosen to address the factors and dynamic changes associated with
specific aims for between-group (healthy participants chemotherapy for breast cancer that increase risk for
vs. healthy participants) and within-cancer-group CV damage. The UPBEAT study may also influence
(anthracycline vs. non-anthracycline) comparisons future monitoring of subclinical CV disease in cancer
(assuming 85% retention rate) to identify a detectable treatment trials, identify potential mechanisms, and
difference for a measure assuming a correlation be- inform cardio-oncology study development to reduce
tween repeated measures of 0.50, 90% power, and CV disease and improve overall survival and quality
alpha ¼ 0.025 (given 2 primary outcomes). For binary of life in women with breast cancer.
outcomes, there is $90% power to detect a 9% dif-
ference between groups and an 8% difference in Jennifer H. Jordan, PhD, MS
within-group comparisons assuming that 10% in the Ralph B. D’Agostino, Jr., PhD
Nancy E. Avis, PhD
non-anthracycline group have this outcome
Shannon L. Mihalko, PhD
and $18% of anthracycline-treated patients have this
Peter H. Brubaker, PhD
outcome.
Dalane Kitzman, MD
Longitudinal mixed models will be used to
Karen M. Winkfield, MD, PhD
compare both continuous and binary outcomes,
Cynthia K. Suerkin, MS
which accounts for the repeated assessments of the
Kerryn W. Reding, PhD, MPH, RN
patients in modeling and will include fixed effects for
Heidi D. Klepin, MD, MS
group (e.g., anthracycline-treated breast cancer pa- Glenn J. Lesser, MD
tients, non-anthracycline–treated breast cancer pa- *W. Gregory Hundley, MD
tients, healthy participants), baseline assessment of
*Virginia Commonwealth University
the outcome of interest (e.g., LVEF, pulse wave ve-
1200 East Broad Street
locity, fatigue), and the time point at which the
P.O. Box 980335
measurements are made relative to the baseline pre–
Richmond, Virginia 23298
CMR imaging assessment. Additional patient-level
E-mail: greg.hundley@vcuhealth.org
risk factors will also be considered as fixed effects in
Twitter: @VCUHealthHeart, @JenJordanPhD
this model. Of primary interest in many of our hy-
https://dx.doi.org/10.1016/j.jaccao.2020.02.013
potheses will be the main effect terms for group and
Ó 2020 The Author. Published by Elsevier on behalf of the American College of
the group time interaction terms. In addition, we
Cardiology Foundation. This is an open access article under the CC BY-NC-ND
will examine in sequential models whether cardiac or license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
vascular measures mediate the association between

Please note: Funding for this study was provided by a National Institutes of
groups and outcomes. Some women may not com- Health, National Cancer Institute grant (1R01CA199167). The Wake Forest NCI
Community Oncology Research Program and Eastern Cooperative Oncology
plete the study, leading to missing data. We will
Group–American College of Radiology Imaging Network support is provided by
compare the baseline characteristics of participants the National Institutes of Health, National Cancer Institute, program grants
5UG1CA189824 and 5UG1CA189828. Dr. Lesser has served as a consultant for
who drop out versus those who do not to determine if
Cancer Expert Now, Agios, and Stemline Therapeutics; and has received
missingness is noninformative, and analyses can be research funding from Vascular Biogenics, Incyte, and NewLink Genetics. All
other authors have reported that they have no relationships relevant to the
performed by using the proposed repeated measures
contents of this paper to disclose. (UPBEAT [Understanding and Predicting
mixed models adjusted with these covariates. How- Fatigue, Cardiovascular Decline, and Events After Breast Cancer]; NCT0279158).
The authors attest they are in compliance with human studies committees and
ever, if evidence exists that the missing data are
animal welfare regulations of the authors’ institutions and Food and Drug
informative, an analysis using more sophisticated Administration guidelines, including patient consent where appropriate. For
more information, visit the JACC: CardioOncology author instructions page.
statistical methods such as multiple imputation will
be applied.
To the best of our knowledge, this is the first pro- REFERENCES
spective cohort study designed to fully characterize 1. Bowles EJ, Wellman R, Feigelson HS, et al. Risk of heart failure in breast
the time course and correlates of subclinical CV cancer patients after anthracycline and trastuzumab treatment: a retrospec-
tive cohort study. J Natl Cancer Inst 2012;104:1293–305.
dysfunction, exercise intolerance, fatigue, and CV
2. Pinder MC, Duan Z, Goodwin JS, Hortobagyi GN, Giordano SH. Congestive
events in women treated with chemotherapy for heart failure in older women treated with adjuvant anthracycline chemo-
breast cancer. The immediate impact of this therapy for breast cancer. J Clin Oncol 2007;25:3808–15.
118 Research Letter JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020
MARCH 2020:114–8
3. Hundley WG, Bluemke DA, Finn JP, et al. ACCF/ACR/AHA/NASCI/SCMR 5. Forman DE, Fleg JL, Kitzman DW, et al. 6-Min walk test provides
2010 expert consensus document on cardiovascular magnetic resonance: a prognostic utility comparable to cardiopulmonary exercise testing in
report of the American College of Cardiology Foundation Task Force on Expert ambulatory outpatients with systolic heart failure. J Am Coll Cardiol 2012;
Consensus Documents. J Am Coll Cardiol 2010;55:2614–62. 60:2653–61.
4. Messroghli DR, Moon JC, Ferreira VM, et al. Clinical recommendations for 6. Scott JM, Haykowsky MJ, Eggebeen J, Morgan TM, Brubaker PH,
cardiovascular magnetic resonance mapping of T1, T2, T2* and extracellular Kitzman DW. Reliability of peak exercise testing in patients with heart failure
volume: a consensus statement by the Society for Cardiovascular Magnetic with preserved ejection fraction. Am J Cardiol 2012;110:1809–13.
Resonance (SCMR) endorsed by the European Association for Cardiovascular 7. Hicks KA, Hung J, Mahaffey KW, et al. Standardized definitions for end point
Imaging (EACVI). J Cardiovasc Magn Reson 2017;19:75. events in cardiovascular trials; U.S. Food & Drug Administration, 2010.
CLINICAL CASE CHALLENGES
Heart Failure Associated With the

Epidermal Growth Factor Receptor
Inhibitor Osimertinib
Andrew J. Piper-Vallillo, MD,a Daniel B. Costa, MD, PHD, MMSC,a Marwa A. Sabe, MD, MPH,b Aarti Asnani, MDc
I ncorporation of molecular targeted therapies into the care of patients with sensitizing mutations has
driven significant improvements in outcomes in non-small cell lung cancer (NSCLC). Epidermal growth
factor receptor (EGFR)-mutant lung cancer includes the greatest percentage of these cases, and accounts
for approximately 15% of lung adenocarcinoma cases in the United States (1). Following the FLAURA (AZD9291
Versus Gefitinib or Erlotinib in Patients With Locally Advanced or Metastatic Non-small Cell Lung Cancer) trial,
osimertinib, a third-generation oral irreversible EGFR tyrosine kinase inhibitor (TKI), became part of first-line
therapy for EGFR-mutant lung cancer. Although generally well-tolerated, emerging evidence suggests that
treatment with osimertinib can result in significant cardiac toxicity (2–4). We present here a case of heart fail-
ure in a patient treated with osimertinib who improved following cessation of therapy (Figure 1).
CASE DESCRIPTION
A 67-year-old woman with history of hypertension, hypothyroidism, pulmonary embolism treated with
enoxaparin, and metastatic EGFR-mutant lung adenocarcinoma was treated with 4 cycles of platinum-based
chemoradiation for stage III disease in 2010. In 2012, the cancer recurred, and she was started on the first-
generation EGFR TKI erlotinib, which she tolerated well with disease control from May 2012 until September
2018, when her disease progressed in the form of a painful new vertebral lesion. Although circulating tumor
DNA did not show evidence of a resistant clone (i.e., T790M mutation), she was transitioned to osimertinib in
September 2018 because of the likelihood that disease progression was due to this common mechanism of
resistance. An electrocardiogram (ECG) obtained shortly before osimertinib initiation revealed normal sinus
rhythm with a normal QRS duration and normal corrected QT interval. There was no evidence of baseline
conduction disease. A transthoracic echocardiogram (TTE) previously obtained in March 2016 demonstrated a
left ventricular ejection fraction (LVEF) of 70% with left ventricular end-diastolic dimension normal at 35 mm.
Global longitudinal strain was not measured on this previous study. Following initiation of osimertinib, the
patient improved clinically and radiographically, with her first monitoring imaging studies revealing stable to
improved burden of metastatic disease. ECGs obtained in October and November 2018 demonstrated no
significant changes attributable to osimertinib.
In April 2019, 7 months after osimertinib initiation, she presented with shortness of breath and was diag-
nosed with heart failure. Her ECG demonstrated sinus rhythm and a left bundle branch block that was new
From the aDepartment of Medicine, Division of Medical Oncology, Beth Israel Deaconess Medical Center, Harvard Medical School,
Boston, Massachusetts; bAdvanced Heart Failure and Mechanical Circulatory Support Program, Beth Israel Deaconess Medical
Center, Harvard Medical School, Boston, Massachusetts; and the cCardio-Oncology Program and CardioVascular Institute, Beth
Israel Deaconess Medical Institute, Beth Israel Deaconess Medical Center/Harvard Medical School, Boston, Massachusetts. Dr.
Costa has received consulting fees, honoraria, and nonfinancial support (institutional research support) from Takeda/
Millennium Pharmaceuticals; has received consulting fees and nonfinancial support (institutional research support) from
AstraZeneca; has received honoraria and nonfinancial support (institutional research support) from Pfizer; and has received
nonfinancial support (institutional research support) from Merck Sharp & Dohme Corporation, Merrimack Pharmaceuticals,
Bristol-Myers Squibb, Clovis Oncology, and Tesaro, all outside the submitted work. All other authors have reported that they
have no relationships relevant to the contents of this paper to disclose.
Manuscript received October 19, 2019; revised manuscript received January 7, 2020, accepted January 13, 2020.

120 Piper-Vallillo et al. JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020
HF Associated With Osimertinib MARCH 2020:119–22
ABBREVIATIONS compared with a previous tracing obtained in November 2018. There was no significant prolongation
AND ACRONYMS of the corrected QT interval. Troponin was repeatedly negative, although brain natriuretic peptide was
elevated at 1,104 pg/ml (normal: 0 to 353 pg/ml). A TTE demonstrated decreased LVEF (42% by
ECG = electrocardiogram
quantitative biplane measurement) with global hypokinesis, abnormal global longitudinal strain
EGFR = epidermal growth
factor receptor
(12%), septal dyssynchrony, and mild biventricular dilation, findings that were all new compared
with the previous TTE obtained in March 2016. A pharmacological nuclear stress test at that time
fraction demonstrated a LVEF of 35% with normal perfusion. Lisinopril and metoprolol succinate were initi-
NRG = neuregulin
ated, although metoprolol was subsequently discontinued due to fatigue, and lisinopril was replaced
NSCLC = non-small cell lung

by valsartan/sacubitril.
cancer A follow-up TTE in June 2019, while still on osimertinib, demonstrated that her LVEF had decreased
TKI = tyrosine kinase inhibitor to 24%. Cardiac magnetic resonance imaging subsequently confirmed severe left ventricular systolic
TTE = transthoracic dysfunction with no late gadolinium enhancement to suggest focal fibrosis or scar and no evidence of
echocardiography myocarditis. In the absence of alternative explanations, osimertinib-associated cardiotoxicity was
considered to be the most likely etiology. Osimertinib was discontinued, with subsequent positron
emission tomographycomputed tomography demonstrating several new and worsening sites of osseous
progression without evidence of visceral progression. A repeat TTE performed 12 weeks after osimertinib
cessation showed some improvement in LVEF to 34%, at which time erlotinib was restarted.
Her LVEF was stable 4 weeks after resuming erlotinib, and since then, she has been asymptomatic from a
heart failure perspective. Further optimization of guideline-directed medical therapy has been limited by
hypotension. A post-progression iliac biopsy did not reveal evidence of either erlotinib resistance (e.g., a
T790M mutation) or a targetable mechanism of resistance. She currently remains clinically stable on erlotinib,
having also undergone palliative radiation to sites of painful osseous progression. If further symptomatic
progression occurs, the plan is to recommend a transition to cytotoxic chemotherapy.
DISCUSSION
The first reported case of heart failure with reduced LVEF with osimertinib developed within 3 weeks of
starting therapy (5), whereas our patient presented with symptoms of heart failure 7 months after initiation.
ECGs and TTE obtained before osimertinib initiation were not suggestive of baseline conduction disease or
subclinical cardiomyopathy. At the time of her diagnosis of heart failure, a comprehensive evaluation con-
sisting of ECG, TTE, pharmacological nuclear stress testing, and cardiac magnetic resonance imaging confirmed
that there was no evidence of ischemic or inflammatory injury to the myocardium, which suggested osi-
mertinib as the only identifiable offending agent. Some recovery in LVEF was observed following osimertinib
cessation, albeit while simultaneously initiating guideline-directed therapy for nonischemic cardiomyopathy.
Her LVEF has subsequently remained stable following resumption of erlotinib as post-progression therapy.
Although generally well-tolerated, there is increasing evidence that osimertinib may be uniquely cardiotoxic
compared with earlier generation EGFR TKIs. Osimertinib gained approval by the Food and Drug Adminis-
tration for use in the first-line setting for EGFR-mutant NSCLC following the phase III FLAURA study, which
demonstrated superior progression-free survival and improved tolerability versus first-generation EGFR TKIs
(erlotinib and gefitinib). In this study, 10% of patients developed QTc prolongation on osimertinib compared
with 5% as previously reported with first-generation TKIs (2). The FLAURA study reported a single case of
cardiac failure as a serious adverse effect of osimertinib (<1% of 279 patients). In contrast, 5% of patients in the
osimertinib arm of the AURA3 (AZD9291 [Osimertinib] Versus Platinum-Based Doublet-Chemotherapy in
Locally Advanced or Metastatic Non-Small Cell Lung Cancer) trial experienced an LVEF decline of >10% and a
decline to <50% (14 of 258 patients) (3). Of these, 3% (6 of 258 patients) developed clinically significant heart
failure. More recently, a retrospective review of data from the Food and Drug Administration Adverse Event
Reporting System database described a number of cardiovascular events associated with osimertinib use,
including heart failure, atrial fibrillation, and QT prolongation (4). In this review, heart failure was twice as
likely to develop in patients treated with osimertinib versus other EGFR TKIs (reporting odds ratio: 2.2; 95%
confidence interval: 1.5 to 3.2). Because of increasing evidence that osimertinib affects the conduction system,
it remains unclear in our case whether osimertinib treatment could have caused heart failure indirectly
through the development of left bundle branch block and mechanical dyssynchrony or through direct effects
on the myocardium. Although left bundle branch block has not been reported in published studies of
JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020 Piper-Vallillo et al. 121
MARCH 2020:119–22 HF Associated With Osimertinib
F I G U R E 1 Case Timeline
Key clinical events are presented in chronological order outlining oncological diagnosis, timing of therapy, and development of cardiac toxicity. CMR ¼ cardiac magnetic
resonance imaging; EGFR ¼ epidermal growth factor receptor; GDMT ¼ guideline directed medical therapy; gen ¼ generation; LVEF ¼ left ventricular ejection fraction;
NSCLC ¼ non-small cell lung cancer; SOB ¼ shortness of breath; TKI ¼ tyrosine kinase inhibitors.
osimertinib, it is possible that QT prolongation in some patients occurred in the setting of QRS prolongation, an
area worthy of further investigation.
Although a mechanistic explanation for the cardiotoxicity of osimertinib is currently lacking, perturbation of
neuregulin (NRG)-1/ErbB signaling with agents such as trastuzumab is well-associated with cardiomyopathy.
The NRG-1/ErbB axis mediates the cardiac myocyte response to stress via Akt-dependent anti-apoptotic effects
and other mechanisms (6). ErbB family receptors are activated by ligands such as NRG-1, leading to phos-
phorylation of homo- and/or heterodimerization partners and ultimately downstream signaling to Ras—ERK
and PI3-kinase—Akt pathways (Figure 2). Trastuzumab disrupts this axis by preventing ErbB4 (activated by
NRG-1) from heterodimerizing with ErbB2 (HER2), a preferred signaling partner (7). Although osimertinib is
highly specific for ErbB1 (EGFR), mouse models have demonstrated greater inhibition of wild-type ErbB2
(HER2) than that observed with erlotinib or afatinib (8). Whether this effect is responsible for increased car-
diotoxicity requires further exploration. If the mechanism of myocardial injury is similar to that observed with
F I G U R E 2 ErbB Family Receptor Signaling
ErbB family receptors are stimulated by several ligands including neuregulin (NRG)-1, which leads to homo- and/or hetero-dimerization of
ErbB receptors and downstream signaling to PI3-kinase—Akt and Ras—ERK. Disruption of ErbB4 heterodimerization with ErbB2 by trastu-
zumab impairs downstream signaling and normal cardiac myocyte stress response. Osimertinib may interfere with cardiovascular function
through its affinity for the wild-type ErbB2 receptor. ATP ¼ adenosine triphosphate.
122 Piper-Vallillo et al. JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020
HF Associated With Osimertinib MARCH 2020:119–22
other HER2-directed therapies, cessation of therapy alone or initiation of neurohormonal therapy may result in
recovery of systolic function (9).
CONCLUSIONS
As demonstrated by this case, we currently lack diagnostic findings specific for osimertinib-associated car-
diotoxicity, which remains a diagnosis of exclusion. The Food and Drug Administration label for osimertinib
suggests obtaining baseline and on-treatment LVEF assessment only in patients with known cardiac risk
factors. The label also recommends initiating a cardiac evaluation in patients on osimertinib who present with
new symptoms potentially attributable to heart failure. Further investigation is needed to understand the risk
factors for and mechanism by which osimertinib induces cardiomyopathy in a small number of patients. Until
then, clinicians must be aware of the potentially serious but uncommon risk of cardiotoxicity from an other-
wise well-tolerated and highly effective therapy.
ADDRESS FOR CORRESPONDENCE: Dr. Aarti Asnani, Center for Life Sciences, Room 911, 3 Blackfan Circle,
Boston, Massachusetts 02115. E-mail: aasnani@bidmc.harvard.edu. Twitter: @AartiAsnaniMD, @apipervallillo.
REFERENCES
1. Shi Y, Au JS-K, Thongprasert S, et al. review of the FDA Adverse Events Reporting Sys- cardiotoxic drugs in cancer patients. Card Fail Rev
A prospective, molecular epidemiology study of tem (FAERS). J Am Coll Cardiol CardioOnc 2019;1: 2019;5:112–8.
EGFR mutations in Asian patients with advanced 172–8.
8. Liu S, Li S, Hai J, et al. Targeting HER2
non–small-cell lung cancer of adenocarcinoma
5. Watanabe H, Ichihara E, Kano H, Ninomiya K, aberrations in non–small cell lung cancer with
histology (PIONEER). J Thorac Oncol 2014;9:
Tanimoto M, Kiura K. Congestive heart failure osimertinib. Clin Cancer Res 2018;24:
154–62.
during osimertinib treatment for epidermal 2594–604.
2. Soria J-C, Ohe Y, Vansteenkiste J, et al. FLAURA growth factor receptor (EGFR)-mutant non-small
9. Nowsheen S, Aziz K, Park JY, et al. Trastuzumab
osimertinib in untreated EGFR-mutated advanced cell lung cancer (NSCLC). Intern Med 2017;56:
in female breast cancer patients with reduced left
non–small-cell lung cancer. N Engl J Med 2017; 2195–7.
ventricular ejection fraction. J Am Heart Assoc
378:113–25.
6. Lemmens K, Doggen K, De Keulenaer GW. 2018;7.
3. Mok TS, Wu Y-L, Ahn M-J, et al. Osimertinib Role of neuregulin-1/ErbB signaling in cardio-
or platinum–pemetrexed in EGFR T790M–posi- vascular physiology and disease: implications
tive lung cancer. N Engl J Med 2016;376: for therapy of heart failure. Circulation 2007;
629–40. 116:954–60. KEY WORDS cardiotoxicity, epidermal
4. Kartik A, Ensor J, Trachtenberg B, Bernicker E. 7. Cuomo A, Rodolico A, Galdieri A, et al. Heart growth factor receptor, heart failure, lung
Osimertinib induced cardiotoxicity: a retrospective failure and cancer: mechanisms of old and new cancer, osimertinib
Aortic Stenosis and the

Tyrosine Kinase Inhibitor Nilotinib
in Chronic Myeloid Leukemia
Miguel Carracedo, PHD,a Leif Stenke, MD, PHD,a,b Anders Franco-Cereceda, MD, PHD,c,d Magnus Bäck, MD, PHDa,b
A 61-year-old woman diagnosed with chronic myeloid leukemia (CML) (Sokal intermediate risk) was
treated with the tyrosine kinase inhibitor (TKI) nilotinib 600 mg/day. The patient was a nonsmoker
and had no cardiovascular history, diabetes, or other relevant diseases. Blood tests revealed hypercho-
lesterolemia with a total cholesterol of 263 mg/dl and a low-density lipoprotein of 186 mg/dl. Renal function
was normal. Echocardiography performed at diagnosis showed moderate septal hypertrophy (18 mm) with
mild outflow obstruction. The aortic valve was tricuspid with some sclerosis but had a normal opening. The
maximum velocity (V max ) across the aortic valve was 2.8 m/s (Figure 1), and the V max in the left ventricular
outflow tract was 2.1 m/s, which did not indicate any significant aortic stenosis, with a calculated aortic valve
area of 2.0 cm 2 by the continuity equation (indexed aortic valve area of 1.2 cm 2/cm 2).
The patient responded favorably and rapidly to nilotinib, achieving a deep/complete molecular response
within 6 months of TKI treatment. Follow-up echocardiography demonstrated stable aortic valve sclerosis at
6 months (Figure 1). However, 12 months after nilotinib therapy initiation, there was a marked increase in aortic
valve velocity, with a V max of 3.5 m/s (Figure 1), which corresponded to an increase in Vmax of 0.7 m/s/year.
During the third treatment year, the patient developed dyspnea and chest pain on exertion.
Echocardiography showed progression of aortic valve disease, fulfilling the criteria of severe aortic stenosis,
with a V max of 4.2 m/s (Figure 1) without any increases in the left ventricular outflow tract parameters. Because
of the uncertainty of the continuity equation in the setting of left ventricular outflow tract obstruction (1),
severe aortic stenosis was confirmed by transesophageal echocardiography, with an aortic valve area by
planimetry of 0.8 cm2 (0.5 cm 2/m2 ). The patient underwent surgical aortic valve replacement with a bio-
prothesis and concomitant subvalvular myectomy, after which symptom relief was achieved.
The unexpectedly rapid progression into symptomatic severe valvular aortic stenosis that required valve
replacement surgery within 3 years after initiation of nilotinib in a relatively young patient with CML with a
tricuspid aortic valve prompted us to investigate mechanisms of possible TKI-associated valvular disease.
CLINICAL PERSPECTIVE
In addition to myocardial damage and heart failure, direct vascular effects of both chemotherapy and radio-
therapy are also important considerations in cardio-oncology (2). Specifically in patients with CML, recent
evidence has raised concerns of cardiovascular adverse events associated with the use of second-generation
TKIs (3). Imatinib is a first-generation TKI used in CML that resulted in dramatic increases in patient
a b
From the Department of Medicine, Karolinska Institutet, Stockholm, Sweden; Theme Cancer, Division of Hematology,
Karolinska University Hospital, Stockholm, Sweden; cDepartment of Molecular Medicine and Surgery, Karolinska Institutet,
Stockholm, Sweden; and the dTheme Heart and Vessels, Division of Valvular and Coronary Disease, Karolinska University
Hospital, Stockholm, Sweden. Dr. Bäck was supported by grants from the Swedish Heart and Lung Foundation (grant 20180571),
the King Gustaf V and Queen Victoria Freemason Foundation, the Stockholm County Council (grant 20170365), and the
Marianne and Marcus Wallenberg Foundation (grant 2015.0104). Dr. Franco-Cereceda was supported by a donation from Fredrik
Lundberg. The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
Manuscript received August 12, 2019; revised manuscript received December 15, 2019, accepted December 17, 2019.

124 Carracedo et al. JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020
Aortic Stenosis and Nilotinib in Leukemia MARCH 2020:123–6
ABBREVIATIONS survival. However, to overcome imatinib resistance or intolerance, more potent second-generation
AND ACRONYMS inhibitors, such as nilotinib, were developed. Cardiovascular concerns for nilotinib were initially
raised based on observations of peripheral arterial disease (4). However, recent observational data also
BMP = bone morphogenic
protein
pointed to an increased incidence of myocardial infarction and other vascular events in patients
CML = chronic myeloid

treated with TKIs (5). Importantly, this might not be a general drug class effect but might specifically
leukemia apply to the second-generation TKIs (3,6). Differential cardiovascular effects between first- and
TKI = tyrosine kinase inhibitor second-generation TKIs were supported by a previous study in patients with CML in which the inci-
VIC = valvular interstitial cell
dence of myocardial infarction was approximately 3.6-fold higher among patients treated with nilo-
tinib compared with those treated with imatinib (5). This led to the introduction of cardiovascular risk
monitoring and prevention in hematological treatment recommendations for TKI (6).
In our case, the annual progression of the aortic valve V max was 0.7 m/s. Rapid aortic stenosis progression is
defined by an increase in Vmax of $0.3 m/s/year (1), which suggested accelerated progression of aortic stenosis
after initiation of nilotinib in this patient. Close monitoring of the patient allowed the appropriate diagnosis of
valvular aortic stenosis. Aortic valve replacement surgery was performed in accordance with guideline rec-
ommendations for severe, symptomatic aortic stenosis. Because of concern that the accelerated valvular
disease was a result of nilotinib treatment, the patient was switched to imatinib treatment before cardiac
surgery.
TRANSLATIONAL PERSPECTIVE
Calcification and thickening of the aortic valve, accompanied by inflammation, extracellular matrix remod-
eling, and intra-leaflet hemorrhages, are part of the underlying pathological processes of aortic valve stenosis.
Several contributors to aortic stenosis incidence and progression have been identified, including age, tradi-
tional cardiovascular risk factors, genetic susceptibility, and certain comorbidities. In addition, direct exposure
to ionizing radiation during radiotherapy increases aortic valve calcification and stenosis (7). However, the
potential effects of targeted therapies (e.g., TKI) on valvular heart disease have received less attention in
cardio-oncology. Understanding the direct valvular effects of cancer therapies and determining their
F I G U R E 1 Echocardiography Demonstrating Rapid Progression of Aortic Valve Stenosis in a Patient With CML Treated With Nilotinib
Aortic valve maximum velocity (Vmax) increased from 2.8 m/s at treatment initiation to 4.2 m/s after 3 years of nilotinib treatment. CML ¼
chronic myeloid leukemia.
JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020 Carracedo et al. 125
MARCH 2020:123–6 Aortic Stenosis and Nilotinib in Leukemia
F I G U R E 2 Nilotinib Enhanced Valvular Interstitial Cell Calcification and Osteogenic Signaling
Treatment of human valvular interstitial cells with nilotinib increased phosphate-induced calcification after 9 days and mRNA expression of
the osteogenic activator bone morphogenic protein (BMP)-2 after 24 h. In contrast, treatment with imatinib was not significantly different
from vehicle control (dimethyl sulfoxide). Data represented as mean SD. **p < 0.01; ***p < 0.001.
mechanistic implications in valve calcification may be fundamental for understanding valvular heart disease in
various cancer populations.
With the aim of understanding if there is a potential role of first- and second-generation TKIs in aortic valve
stenosis, we applied an experimental approach. Our protocol was approved by the local ethical committee, and
informed consent was obtained from all subjects. Human aortic valves were obtained from 7 patients who
underwent aortic valve replacement surgery. Valvular interstitial cells (VICs) were isolated as previously
described (8). Each aortic valve was used to generate an individual cell culture, which was exposed to either
nilotinib (10 m M) or the first-generation TKI imatinib (10 m M). A corresponding volume of dimethyl sulfoxide
vehicle was administered to control cultures. In vitro calcification was assessed after 9 days of culture in the
presence of 2.6-m M inorganic phosphate according to previously described protocols (9). After 24 h of culture,
mRNA was isolated for the determination of bone morphogenic protein(BMP)-2 expression levels with Ther-
moFisher 7900HT fast real-time polymerase chain reaction using primers and probes from Thermo Fischer
Scientific (Waltham, Massachusetts (Hs00154192_m1) and normalized to glyceraldehyde 3-phosphate dehy-
drogenase (Hs99999905_m1) as a housekeeping gene. Statistically significant differences were determined
using a 1-way analysis of variance followed by a recommended post hoc test. Statistical significance was
assigned at p < 0.05. Statistical analysis was performed using GraphPad Prism 8 (GraphPad Software, San
Diego, California).
These experiments revealed that nilotinib (10 m M) significantly increased phosphate-induced VIC calcifi-
cation compared with that in control cultures (Figure 2). Likewise, nilotinib induced a 24.0 5.1-fold increase
in BMP-2 expression (Figure 2). Because BMP-2, through downstream pSMAD 1/5/8 signaling, was shown to
drive valvular calcification (10), these findings supported a direct pro-calcifying effect of nilotinib in the aortic
valve. In contrast, the first-generation TKI imatinib did not significantly alter either calcification or BMP-2
expression in human VICs (Figure 2).
DISCUSSION
In summary, a possible causal relation behind the reported case of rapid aortic valve stenosis development
during nilotinib treatment was supported by the capacity of nilotinib to increase calcification and induce
126 Carracedo et al. JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020
Aortic Stenosis and Nilotinib in Leukemia MARCH 2020:123–6
osteogenic activation in human VICs. The present case report and the associated experimental data therefore
suggest the direct valvular toxic effects of nilotinib. Similar effects were not observed for imatinib. The effects
of other TKIs remain to be established.
However, it might be premature to make recommendations for routine echocardiographic screening in TKI-
treated patients with CML. It should also be emphasized that the main purpose of TKI treatment in CML is the
antileukemic effect, which should not be compromised by concerns for adverse effects without cause, as
indicated in hematological treatment recommendations (6). However, an awareness of a possible link between
TKI and aortic stenosis could potentially lead to earlier diagnosis and treatment of valve disease in patients
with CML.
Studies are ongoing to decipher the mechanisms involved in nilotinib-induced valvular effects. Further-
more, we are currently exploring the incidence of aortic valve stenosis in larger populations with CML. Patient
registries, such as that previously used to monitor myocardial infarction in CML and nilotinib treatment (5),
might provide further insights into a possible increased risk of aortic stenosis associated with nilotinib
treatment.
CONCLUSIONS
These observations provide the first indication of a possible link between TKI and valvular heart disease in
CML.
ADDRESS FOR CORRESPONDENCE: Dr. Magnus Bäck, Karolinska University Hospital, Theme Heart and Vessels,
Division of Valvular and Coronary Disease, M85, 141 86 Stockholm, Sweden. E-mail: Magnus.Back@ki.se. Twitter:
@TransCardio.
REFERENCES
1. Baumgartner H, Hung J, Bermejo J, et al. Rec- 5. Dahlen T, Edgren G, Lambe M, et al. 8. Laguna-Fernandez A, Carracedo M, Jeanson G,
ommendations on the echocardiographic assess- Cardiovascular events associated with et al. Iron alters valvular interstitial cell function
ment of aortic valve stenosis: a focused update use of tyrosine kinase inhibitors in chronic and is associated with calcification in aortic ste-
from the European Association of Cardiovascular myeloid leukemia: a population-based nosis. Eur Heart J 2016;37:3532–5.
Imaging and the American Society of Echocardi- cohort study. Ann Intern Med 2016;165:
9. Carracedo M, Artiach G, Witasp A, et al. The G-
ography. J Am Soc Echocardiogr 2017;30:372–92. 161–6.
protein coupled receptor ChemR23 determines
2. Herrmann J, Yang EH, Iliescu CA, et al. 6. Steegmann JL, Baccarani M, Breccia M, et al. smooth muscle cell phenotypic switching to
Vascular toxicities of cancer therapies: the old European LeukemiaNet recommendations for the enhance high phosphate-induced vascular calcifi-
and the new–an evolving avenue. Circulation management and avoidance of adverse events of cation. Cardiovasc Res 2019;115:1557–66.
2016;133:1272–89. treatment in chronic myeloid leukaemia. Leukemia 10. Gomez-Stallons MV, Wirrig-Schwendeman EE,
3. Herrmann J. Tyrosine kinase inhibitors and 2016;30:1648–71. Hassel KR, et al. Bone morphogenetic protein
vascular toxicity: impetus for a classification sys- signaling is required for aortic valve calcification.
7. Donnellan E, Masri A, Johnston DR, et al.
tem? Current Oncol Rep 2016;18:33. Arterioscler Thromb Vascul Biol 2016;36:1398–405.
Long-term outcomes of patients with medias-
4. Kim TD, Rea D, Schwarz M, et al. Peripheral tinal radiation-associated severe aortic stenosis
artery occlusive disease in chronic phase chronic and subsequent surgical aortic valve replace- KEY WORDS aortic stenosis, leukemia,
myeloid leukemia patients treated with nilotinib or ment: a matched cohort study. J Am Heart nilotinib, tyrosine kinase inhibitor, valvular
imatinib. Leukemia 2013;27:1316–21. Assoc 2017;6. disease
Evidence of Concurrent Light Chain

and Transthyretin Cardiac Amyloidosis
in 2 Patients
Joseph P. Donnelly, MD,a Andrej Gabrovsek, MD,a Lidiya Sul, BS,a Claudiu Cotta, MD, PHD,b E. Rene Rodriguez, MD,b
Carmela D. Tan, MD,b Mazen Hanna, MDa
T wo distinct types of systemic amyloidosis, light chain (AL) and transthyretin (ATTR), account for >95%
of diagnosed cardiac amyloidosis (CA) (1). AL arises from a clonal population of plasma cells that pro-
duce misfolded immunoglobulin light chains that aggregate in organs such as the heart, kidneys, pe-
ripheral and autonomic nerves, liver, and gastrointestinal tract (1). ATTR arises from the liver-derived protein
transthyretin (TTR) that becomes kinetically unstable, misfolds, and aggregates into amyloid fibrils. This can
occur due to a hereditary point mutation (ATTRv for “variant”) or is acquired as a form known as wild-type
(ATTRwt) (1), which is most commonly associated with aging. ATTRv affects the peripheral and autonomic
nerves and/or the heart, with the phenotype varying depending on the mutation, whereas ATTRwt exhibits
a more cardiac-dominant phenotype (1).
The traditional method used to diagnose amyloidosis is histological confirmation of the affected organ using
Congo red or thioflavin S staining (2). Further characterization of the amyloid fibril protein content is made
using immunohistochemistry (IHC) or liquid chromatography tandem mass spectrometry (LC-MS/MS) (2).
Determination of the type of amyloidosis is crucial because the prognosis and treatment significantly differ
between AL and ATTR amyloidosis (1). Current analytical techniques are very good at distinguishing the 2
types; however, occasionally, there can be diagnostic confusion by identifying the presence of multiple pre-
cursor proteins in an amyloid deposit, including reports of co-deposition of AL and TTR (3–5).
Herein, we present 2 cases with clinical and histological evidence of both ATTR and AL-CA occurring
concurrently.
CASE 1
An 89-year-old Eastern European woman with a history of hypertension and stage 4 chronic kidney disease
presented with heart failure. Initial echocardiography revealed a preserved ejection fraction of 60% with
thickened left ventricular walls and a mitral inflow pattern suggestive of restrictive cardiomyopathy. Elec-
trocardiography (ECG) showed sinus rhythm with occasional premature atrial complexes. Due to a high sus-
picion for cardiac amyloidosis, work-up with nuclear 99m-technetium pyrophosphate scintigraphy
( 99m Tc-PYP) was pursued. This showed a semi quantitative visual score of grade 3 with diffuse cardiac uptake
on single-photon emission computed tomography-computed tomography that was highly suggestive of ATTR-
From the aDepartment of Cardiovascular Medicine, Heart and Vascular Institute, Cleveland Clinic Foundation, Cleveland, Ohio;
and the bDepartment of Pathology, Pathology and Laboratory Medicine Institute, Cleveland Clinic Foundation, Cleveland,
Ohio. This work was supported by the Cleveland Clinic Term Chair in Amyloid Heart Disease. Dr. Hanna has served on
advisory boards for Pfizer, Alynam, Eidos, and Akcea. All other authors have reported that they have no relationships relevant
to the contents of this paper to disclose.
Manuscript received November 13, 2019; revised manuscript received December 20, 2019, accepted December 21, 2019.

128 Donnelly et al. JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020
Concurrent AL and ATTR Cardiac Amyloidosis MARCH 2020:127–30
ABBREVIATIONS CA. However, simultaneous laboratory work-up for AL amyloidosis showed a lambda restricted M
AND ACRONYMS protein present on both serum and urine immunofixation, along with serum free kappa light chains of
99m
20.6 mg/l and serum free lambda of 531.3 mg/l with a highly abnormal ratio of 0.04, which indicated a
Tc-PYP = 99m-technetium
pyrophosphate scintigraphy
lambda monoclonal gammopathy.
AL-CA = light chain cardiac

Because of the potential diagnosis of ATTR-CA or AL-CA, endomyocardial biopsy was obtained. IHC
amyloidosis stains displayed a striking positivity for TTR in a nodular pattern within the interstitium and weak
ATTR-CA = transthyretin positivity in a peri-myocytic linear pattern around the myocytes for a lambda AL. Laser capture
cardiac amyloidosis microdissection and LC-MS/MS identified both TTR and lambda AL as the main amyloidogenic pro-
ATTRv = hereditary variant teins in the sample. Subsequent bone marrow biopsy revealed a monoclonal lambda plasma cell
transthyretin amyloidosis
neoplasm that represented approximately 20% to 30% of the bone marrow cellularity without evi-
ATTRwt = wild-type dence of amyloid deposits. Concurrent diagnosis of lambda AL-CA and ATTR-CA was established. TTR
transthyretin amyloidosis
genetic testing revealed no mutation, consistent with ATTRwt-CA.
IHC = immunohistochemistry
Antiplasma cell therapy was recommended to treat the AL component of CA. At the time of diag-
ECG = electrocardiography
nosis, no treatments were available for ATTR-CA. The patient opted for palliative care, was discharged
LC-MS/MS = liquid to hospice, and died 5 months after initial onset of symptoms.
chromatography tandem mass
spectrometry CASE 2
PSM = protein spectrum match
TTR = transthyretin An 85-year-old Russian man presented with heart failure with acute volume overload and an
NT-pro-B-type natriuretic peptide level of 7,913 pg/ml, which required admission for intravenous
diuresis. ECG showed normal sinus rhythm, but chest radiographs revealed a lytic lesion in the left
scapula, prompting further investigation. Echocardiography revealed increased left ventricular septal and
posterior wall thickness and grade III diastolic dysfunction. Cardiac magnetic resonance displayed diffuse
delayed myocardial enhancement of the left ventricle with overall features suggestive of amyloidosis, and the
patient was referred to an amyloidosis specialist.
99m
Upon initial work-up for CA, Tc-PYP showed grade 3 cardiac uptake suggestive of ATTR-CA. However,
monoclonal protein testing, which was ordered at the same time, showed a lambda M protein present on serum
and urine immunofixation, along with serum free kappa of 26.3 mg/l and free lambda of 365.5 mg/l with a
highly abnormal ratio of 0.07, which indicated a lambda monoclonal gammopathy. Bone marrow biopsy
revealed a monoclonal lambda plasma cell neoplasm representing approximately 10% of the bone marrow
cellularity with trace amyloid deposition. An endomyocardial biopsy was performed to reconcile the findings,
and IHC stains displayed nodular deposits positive for TTR and distinct perimyocytic staining positive for
lambda AL. The IHC results were confirmed by LC-MS/MS, establishing a diagnosis of concurrent lambda AL-CA
and ATTR-CA (Figure 1). TTR genetic testing showed no mutation, consistent with ATTRwt-CA.
Antiplasma cell therapy with bortezomib was initiated to treat the AL component of the CA. Like the pre-
vious patient, no treatments were available for ATTR-CA at the time of diagnosis. One day after receiving his
fourth bortezomib infusion, the patient decompensated and was admitted for heart failure. Following a
palliative care consultation, the patient was discharged to hospice and died.
DISCUSSION
These 2 cases represent the rare concurrence of clinical and histological evidence of AL-CA and ATTR-CA (3–5).
Co-deposition of 2 types of amyloid protein has been previously reported but reports of cardiac ATTR and AL
co-deposition are exceedingly rare (3–5). Concomitant ATTR and AL cardiac amyloidosis have been described
in 2 patients by Siddiqi et al. (5) and in a case report by Mahmood et al. (4). Liepnieks and Benson reported an
autopsy finding of ATTR and AL co-deposition in a patient diagnosed with ATTR-CA (3).
99m
In our report, the work-up of both patients entailed simultaneous Tc-PYP scan and laboratory testing for
AL amyloidosis. However, most guidelines for the diagnostic work-up of suspected CA recommend testing for
monoclonal protein first and, if abnormal, pursuing the pathway of bone marrow and fat pad or endomyo-
cardial biopsy. In this pathway, there is typically no need for 99mTc-PYP in the setting of abnormal monoclonal
99m
protein. However, in some clinical scenarios, a clinician may opt to pursue both Tc-PYP and monoclonal
laboratory testing simultaneously, particularly if the patient is an older adult with a higher pre-test likelihood
of ATTR-CA.
Initially, both cases were presumed to be ATTR-CA because of the high-grade cardiac uptake on nuclear
scintigraphy and the patients’ ages. However, abnormal monoclonal protein testing necessitated an
JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020 Donnelly et al. 129
MARCH 2020:127–30 Concurrent AL and ATTR Cardiac Amyloidosis
F I G U R E 1 PSM Matches From LC-MS/MS Analysis of Endomyocardial Biopsy From Case 2
Protein spectrum matches (PSM) convey information on the relative abundance of particular proteins. In most cases of amyloid, a few proteins are always identified and
in relatively abundant quantities: serum P component, apolipoprotein E, apolipoprotein A1, apolipoprotein AIV, and vitronectin (yellow). Further examination identifies
multiple proteins with amyloidogenic properties, including transthyretin (green) and the immunoglobulin light chains (red). The identification of transthyretin in the
samples, at a PSM comparable to that of the other amyloid-associated proteins, is evidence that the amyloid has a significant transthyretin component. Immuno-
globulin heavy and light chains are present in virtually all patient samples, a consequence of serum contamination. However, in this case, the light chains are present
at concentrations similar to those of other amyloid-associated proteins. Although kappa immunoglobulin light chains are significantly more abundant than lambda in
most patients, in this case, lambda light chains dominated, providing another piece of evidence that amyloid has a significant lambda light chain component.
PSM duplicates 1 and 2 represent duplicate testing results from the same patient. LC-MS/MS ¼ liquid chromatography with tandem mass spectrometry.
endomyocardial biopsy to reconcile the findings. These cases highlight the importance of completeness of the
diagnostic work-up for CA, because the existence of one disease process does not preclude the existence of the
other. In addition, it is known that AL-CA can occasionally lead to significant cardiac uptake on nuclear
scintigraphy (6).
The observation that >95% of CA cases are due to either ATTR or AL was pivotal in the development of the
nonbiopsy diagnostic method for ATTR-CA, which is a less invasive strategy compared with traditional his-
tological confirmation (6). Nonbiopsy diagnosis of ATTR-CA can be reliably made when there is a semi
quantitative visual score of grade 2 or 3 cardiac uptake on nuclear scintigraphy coupled with a normal
serum free light chain ratio and no M protein on serum and urine immunofixation to rule out AL amyloidosis
(6). In cases in which AL amyloidosis cannot be ruled out due to an abnormal serum free light chain ratio and/or
an M protein on immunofixation, endomyocardial biopsy is recommended for confirmation of the amyloid
diagnosis and type. As these cases demonstrated, a presumptive diagnosis of ATTR-CA with positive nuclear
scintigraphy alone risks missing the diagnosis of AL-CA in the setting of an underlying abnormal free light
chain ratio or presence of an M protein on serum or urine immunofixation.
LC-MS/MS relies on the purification of proteins from histological sections, followed by their digestion with
trypsin and analysis by LC-MS/MS. It is important to emphasize that analysis by LC-MS/MS is semi quantitative
and does not allow a precise quantification of each protein, but the protein spectrum matches (PSMs) convey
information on the relative abundance of particular proteins. In most cases of amyloidosis, a few proteins are
130 Donnelly et al. JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020
Concurrent AL and ATTR Cardiac Amyloidosis MARCH 2020:127–30
always identified and present in relatively abundant quantities: serum P component, apolipoprotein E,
apolipoprotein A1, apolipoprotein AIV, and vitronectin. The identification of these proteins in the cases pre-
sented was not only a confirmation of the initial diagnosis but also indicative of significant amounts of amyloid
in the sample obtained by microdissection.
In our cases, further examination of the mass spectrometry data identified multiple proteins with amyloi-
dogenic properties, including TTR and immunoglobulin lambda. The identification of TTR in the samples, at a
PSM comparable to that of the other amyloid-associated proteins, was evidence that the amyloid had a sig-
nificant TTR component. Immunoglobulin heavy chains and light chains are present in virtually all patient
samples, which is a consequence of serum contamination. However, in this case, the immunoglobulin light
chains were present at concentrations similar to those of other amyloid-associated proteins. Although kappa
immunoglobulin light chains are significantly more abundant than lambda light chains in most patients, in the
cases illustrated, lambda light chains dominated, providing another piece of evidence that amyloid has a
significant lambda AL component.
CONCLUSIONS
These cases represented the rare presentation of clinical and histological ATTR- and AL-CA occurring
concurrently. Although a dominant amyloidosis type can often be identified, AL and ATTR arise from
independent pathological mechanisms, and the existence of one does not preclude the existence of the other.
As awareness for CA increases with the emergence of new therapies (7–10), it is important to be vigilant in the
work-up and complete the diagnostic process.
ADDRESS FOR CORRESPONDENCE: Dr. Mazen Hanna, Cleveland Clinic, 9500 Euclid Avenue, J3-4, Cleveland,
Ohio 44195. E-mail: hannam@ccf.org. Twitter: @JoeDonnellyMD, @maz_hanna.
REFERENCES
1. Donnelly JP, Hanna M. Cardiac amyloidosis: an 4. Mahmood S, Gilbertson JA, Rendell N, et al. 8. Khouri J, Kin A, Thapa B, et al. Daratumumab
update on diagnosis and treatment. Cleve Clin J Two types of amyloid in a single heart. Blood proves safe and highly effective in AL amyloidosis.
Med 2017;84:12 Suppl 3:12–26. 2014;124:3025–7. Br J Haematol 2019;185:342–4.
2. Tan CD, Rodriguez ER. Cardiac amyloidosis. In: 5. Sidiqi MH, McPhail ED, Theis JD, et al. Two 9. Maurer MS, Schwartz JH, Gundapaneni B, et al.
Picken M, Herrera G, Dogan A, editors. Amyloid types of amyloidosis presenting in a single patient: Tafamidis treatment for patients with trans-
and Related Disorders. Current Clinical Pathology. a case series. Blood Cancer J 2019;9:30. thyretin amyloid cardiomyopathy. N Engl J Med
Totowa, NJ: Humana Press, 2015;391–411. 2018;379:1007–16.
6. Gillmore JD, Maurer MS, Falk RH, et al. Non-
3. Liepnieks JJ, Benson MD. Codeposition of biopsy diagnosis of cardiac transthyretin 10. Buxbaum JN. Oligonucleotide drugs for
transthyretin and immunoglobulin lambda light amyloidosis. Circulation 2016;133:2404–12. transthyretin amyloidosis. N Engl J Med 2018;379:
chain in senile cardiac (ATTR) amyloidosis. In: 82–5.
7. Sperry BW, Ikram A, Hachamovitch R, et al. Ef-
Grateu G, Kyle RA, Skinner M, editors. Proceedings
ficacy of chemotherapy for light-chain amyloidosis
of Xth International Symposium on Amyloidosis.
in patients presenting with symptomatic heart KEY WORDS amyloidosis, cardiomyopathy,
Washington, DC: CRC Press, 2004;332–3.
failure. J Am Coll Cardiol 2016;67:2941–8. diagnosis, heart failure
Metastatic Merkel Cell Carcinoma

Resulting in Complete Heart Block
Brenden S. Ingraham, MD, Kyle W. Klarich, MD, Nandan S. Anavekar, MB BCH, Martha Grogan, MD
M erkel cell carcinoma (MCC) is an aggressive, rare cutaneous neuroendocrine carcinoma that can
metastasize to the heart in exceedingly rare cases. The incidence of MCC in the United States is
0.7 per 100,000 person-years with a median age of 75 to 80 years at diagnosis and a male predom-
inance (1,2). MCC is seen in older, fair-skinned individuals with immunocompromise, particularly those with
leukemia, lymphoma (especially B-cell malignancies), and HIV. Merkel cells are found in the stratum basale
of the epidermis and are associated with mechanoreceptors. Merkel cell polyoma virus is a double-stranded
deoxyribonucleic acid virus causally linked to MCC. Carcinogenesis in the absence of Merkel cell polyoma virus
is related to ultraviolet light exposure. Tumors grow rapidly as firm, painless, fleshy, or bluish-red nodules
within the skin, usually on the head or neck in 40% to 50% of cases. Regional metastases and local recurrence
are more common than distant metastases. When distant metastases occur, they tend to be found in skin,
lymph nodes, adrenal glands, lung, liver, brain, and bone (1). Cardiac metastases of MCC are extraordinarily
rare, with only about a dozen cases reported in the literature (3–6). We herein present a unique case of MCC
with complete heart block related to an aortic root mass with diagnostic uncertainty given this patient’s history
of both MCC and lymphoma.
CASE REPORT
A 73-year-old woman with a history of marginal zone (B-cell) lymphoma of the spleen, hypertension, and
hypothyroidism presented for evaluation of left axillary mass. Her lymphoma was diagnosed 6 years prior
and was initially treated with splenectomy and rituximab. She had a recurrence 5 years after her initial lym-
phoma diagnosis and was treated with rituximab, which controlled her disease. The new left axillary mass
relevant to her current presentation was mildly painful with a burning sensation down her left arm. She had no
fever, chills, night sweats, or weight loss. There were no other new skin lesions. Her vital signs and cardio-
pulmonary examination were normal. The left axillary mass was 6 cm with mild pain on palpation but no
erythema or fluctuance, and there was no other cervical, axillary, or inguinal adenopathy.
Chest computed tomography (CT) imaging revealed a 6-cm left axillary mass involving the cephalic vein and
brachial plexus, concerning for lymphoma. She underwent incisional biopsy, the results of which were inde-
terminate. Excisional biopsy was then pursued for definitive diagnosis. Immunohistochemical staining
revealed MCC, and the specimen tested positive for Merkel cell polyomavirus. She did not have a primary
culprit cutaneous lesion. The mass was deemed surgically unresectable given the neurovascular involvement.
Pembrolizumab with concomitant radiation therapy (60 Gy/6,000 rad in 30 fractions) was initiated. Surveil-
lance imaging at 3 months after 4 cycles of radiation and 4 doses of pembrolizumab showed a marked interval
treatment response of the axillary mass without residual mass or nodularity. A positron emission tomography
(PET)-CT scan did not show any further fluorodeoxyglucose (FDG) uptake. Pembrolizumab was stopped, and
she was followed up expectantly.
From the Department of Cardiovascular Medicine, Mayo Clinic, Rochester, Minnesota. This work was supported by the
Department of Cardiovascular Medicine, Mayo Clinic. The authors have reported that they have no relationships relevant to
the contents of this paper to disclose.
Manuscript received November 13, 2019; revised manuscript received January 28, 2020, accepted January 28, 2020.

132 Ingraham et al. JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020
Merkel Cell Carcinoma With Cardiac Metastases MARCH 2020:131–4
ABBREVIATIONS Eighteen months later, the patient presented with weakness. An electrocardiogram revealed Mobitz
AND ACRONYMS type II with a ventricular rate of 69 beats/min. Transthoracic echocardiogram revealed a circumfer-
ential aortic root mass. CT angiography demonstrated involvement of the aortic root, aortic valve
CT = computed tomography
annulus, tricuspid valve annulus, and the proximal right coronary artery, as well as narrowing of the
FDG = fluorodeoxyglucose
right ventricular inflow tract. A surgical approach was thus not a treatment option. Cardiac magnetic
MCC = Merkel cell carcinoma
resonance imaging (CMR) was subsequently obtained for better characterization of the mass. CMR
CMR = cardiac magnetic
revealed an ill-defined infiltrative mass involving the interatrial septum with expansion across the
resonance imaging
aortic root (Figure 1). Double inversion recovery sequences revealed infiltration of the mass into the
PET = positron emission
tomography
atrial septum with lobulations protruding into the right atrium and tricuspid valve. Diffusion-
weighted imaging showed increased signaling, while myocardial delayed enhancement revealed
SUV = standardized uptake
value gradual low-level enhancement of the mass. A repeat fludeoxyglucose F 18 PET-CT (14.91 mCi; stan-
dard fasting protocol before the examination; CT fusion imaging performed) demonstrated moderate
FDG avidity throughout the aortic root mass (7.2 maximum standardized uptake value) (Figure 2). The area of
uptake was concerning for transformed lymphoma versus MCC metastasis. There were no pleural effusions or
other sites suspicious for malignancy, including the left axillary region. Of note, C-reactive protein levels were
within normal limits, and results of blood cultures were negative. Telemetry revealed intermittent third-
degree atrioventricular block.
The patient underwent a right atrial endocardial biopsy with concomitant temporary pacemaker insertion,
and pathology confirmed metastatic MCC. Immunotherapy with avelumab and radiation (40 Gy/4,000 rad in 5
fractions) were resumed. Repeat PET-CT imaging after these therapies revealed a decrease in right atrial and
septal uptake (Figure 2). A single endocardial lead permanent pacemaker was placed to the right ventricle given
the symptomatic complete heart block and to provide protection from varying degrees of heart block that could
occur with changes to the mass during treatment. A right atrial lead was avoided due to the location of the
mass, and the pulse generator was placed in the right chest to avoid the irradiated left chest further compli-
cated by recurrent seroma. There was 99.8% pacing at the time of implantation. Repeat cardiac CT angiogram
after 2 months of avelumab and radiation showed considerable shrinkage of the mass and 1.3% pacing. She
initially responded well but had to discontinue subsequent immunotherapy due to severe avelumab-related
pneumonitis and esophagitis.
F I G U R E 1 Cardiac Magnetic Resonance Imaging in Patient With Merkel Cell Carcinoma
Cardiac magnetic resonance imaging/fast imaging employing steady-state acquisition image revealed an ill-defined infiltrative mass involving
the interatrial septum with expansion across the aortic root. The involvement of the aortic root and the nearby conduction system resulted in
the development of complete atrioventricular block.
JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020 Ingraham et al. 133
MARCH 2020:131–4 Merkel Cell Carcinoma With Cardiac Metastases
F I G U R E 2 PET-CT Imaging in a Patient With Merkel Cell Carcinoma
Positron emission tomography/computed tomography (PET-CT) imaging shows moderate (7.2 standardized uptake value) fludeoxyglucose F
18 tracer uptake throughout the aortic root mass before (left) and after (right) avelumab and radiation therapies directed against the
metastatic Merkel cell lesion.
DISCUSSION
This case shows that: 1) MCC can metastasize to the heart and cause complete heart block; 2) multiple imaging
modalities helped in management but did not ultimately negate the need for tissue diagnosis; and 3) immune
checkpoint inhibitor therapy was effective in controlling the MCC.
We present a patient with advanced (stage IV) MCC with aortic root cardiac metastasis disrupting the
conduction system and leading to third-degree atrioventricular block. The patient’s history of marginal zone
lymphoma predisposed her to MCC and also provided another possible etiology for the intracardiac mass,
obfuscating the diagnosis. Multiple imaging modalities were used to guide the diagnostic process. The initial
echocardiogram was grossly abnormal with an apparent mass infiltrating the atrial septum and the aortic root.
The mass’s lack of respect for tissue planes on the initial echocardiographic study increased the likelihood of
malignancy. The additional imaging modalities further characterized the tissue and supported the suspicion
for malignant infiltration. PET-CT imaging (Figure 2) showed the FDG avidity of the mass, which provided
additional evidence for the malignant nature of this tumor. CMR with double inversion recovery sequence
revealed infiltration of the mass into the atrial septum, with lobulations protruding into the right atrium and
tricuspid valve. Repeat CT angiogram after 2 months of avelumab treatment revealed a substantial response to
therapy. Despite having excellent characterization of the lesion, however, tissue biopsy was ultimately pur-
sued given the rarity of cardiac MCC, the possibility of lymphoma as an alternative etiology, and the need for a
definitive tissue diagnosis to guide disease-directed therapy.
Immune checkpoint inhibitor therapy should be considered for stage IV MCC involving the myocardium, and
placement of a permanent pacemaker may be necessary when patients have heart block or lesions threatening
the conduction system. It is important to note that heart block, among other cardiac toxicities, resulting from
immune checkpoint inhibitors has been reported and should be considered when patients taking these agents
present with this condition (7). However, this patient had been off of pembrolizumab for 18 months before
developing heart block, and the intracardiac lesion was positioned to cause disruption of the conduction
system. Another possible cause for the atrioventricular block is radiation fibrosis of the myocardium following
radiation to the left axillary mass. Although these were all considered, the aortic root lesion affecting the
conduction system is the most likely cause of her complete heart block. The percent pacing dramatically
decreased from nearly 100% to 1.3% with decrease in tumor size while being treated with avelumab, suggesting
that MCC was the culprit.
134 Ingraham et al. JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020
Merkel Cell Carcinoma With Cardiac Metastases MARCH 2020:131–4
Cardiac masses can be benign or malignant and primary or metastatic. The differential for cardiac masses
also includes thrombus, vegetations, and/or foreign bodies. These masses can present with symptoms related
to heart failure (especially if the mass is obstructive), arrhythmias, atrioventricular block, embolization, and
pericardial effusions [particularly if the mass is malignant]). Primary cardiac tumors are rare, with a frequency
of 0.02%, and three-quarters are benign (8). The most common cardiac neoplasms are benign myxomas, with
most located in the left atrium. Other benign tumors include fibromas, lipomas, and hamartomas. Primary
malignant cardiac tumors are exceedingly rare; the most common is angiosarcoma. Secondary cardiac tumors
originating from a distant primary are much more common and w100 times more common than a primary
cardiac neoplasm. Melanoma is the most likely cancer to cause cardiac metastases. However, cardiac metas-
tases from other more prevalent cancers, such as lung, breast, esophageal, leukemia, and lymphoma, are also
frequently encountered.
In the detection and evaluation of cardiac masses, symptoms and signs may be present on history and
physical examination. An electrocardiogram may reveal nonspecific changes or other findings such as electrical
alternans in the event of an associated pericardial effusion. Echocardiography has become the gold standard
for initial detection of masses and effusions. CT, CMR, and PET imaging modalities are also used to assess
tissue characteristics that favor certain pathologies. Furthermore, advanced cross-sectional imaging provides
accurate measurements in different imaging planes and can identify structures affected by the mass. Advanced
imaging can guide surgical candidacy and planning as well as assess for response to therapies (9). FDG uptake
on PET-CT imaging has been used more frequently in recent years to help differentiate benign and malignant
tumors. Malignant tumors generally have high FDG uptake in the 8.0 to 10.8 standardized uptake value range,
whereas benign tumors have only slight FDG uptake (mean standardized uptake value for benign tumors 2.8
0.6). However, there are reports of benign myxomas exhibiting high FDG uptake; thus, although PET is helpful,
it cannot conclusively predict whether a tumor is malignant (10). Ultimately, in cases in which the diagnosis is
unclear, endomyocardial biopsy may be used to obtain tissue diagnosis. Exploratory thoracotomy may be
necessary in some cases depending on the location of the tumor (8). Management of MCC includes wide local
excision of the primary tumor with consideration of adjuvant radiation and immune checkpoint inhibitors;
chemotherapy has limited evidence and has not been shown to be particularly effective.
Given the aging population and the close association of MCC with age, the incidence of MCC is expected to
rise in the coming years (2). Survival depends largely on staging; 5-year survival for stage IV is 13.5% (1). Our
unique case of metastatic MCC with myocardial infiltration around the aortic root resulting in complete heart
block supports the use of multimodality imaging combined with tissue biopsy to establish the correct diagnosis
and guide management strategy.
ADDRESS FOR CORRESPONDENCE: Dr. Martha Grogan, Department of Cardiovascular Diseases, Mayo Clinic,
200 First Street SW, Rochester, Minnesota 55905. E-mail: Grogan.Martha@Mayo.edu. Twitter: @MarthaGrogan1.
REFERENCES
1. Becker JC, Stang A, DeCaprio JA, et al. Merkel 5. Suttie CF, Hruby G, Horvath L, Thompson J. 9. Motwani M, Kidambi A, Herzog BA, Uddin A,
cell carcinoma. Nat Rev Dis Primers 2017;3:17077. Cardiac metastasis in Merkel cell carcinoma. J Clin Greenwood JP, Plein S. MR imaging of cardiac
Oncol 2014;32:e52–3. tumors and masses: a review of methods and
2. Paulson KG, Park SY, Vandeven NA, et al. Mer-
clinical applications. Radiology 2013;268:26–43.
kel cell carcinoma: current US incidence and pro- 6. Wang LW, Walker BD, Omari A, Tay AE,
jected increases based on changing demographics. Subbiah RN. Metastatic Merkel cell carcinoma of 10. Rahbar K, Seifarth H, Schafers M, et al. Dif-
J Am Acad Dermatol 2018;78:457–63.e2. the heart. Eur Heart J 2014;35:2927. ferentiation of malignant and benign cardiac tu-
mors using 18F-FDG PET/CT. J Nucl Med 2012;53:
3. Di Loreto M, Francis R. Merkel cell carcinoma 7. Ball S, Ghosh RK, Wongsaengsak S, et al. Car-
856–63.
cardiac metastasis causing cardiac tamponade. diovascular toxicities of immune checkpoint in-
BMJ Case Rep 2017;2017. hibitors: JACC Review Topic of the Week. J Am
Coll Cardiol 2019;74:1714–27.
4. Keeling AN, Johnson NP, Farrelly C, et al. Right KEY WORDS cardiac metastases, complete
atrial Merkel cell carcinoma metastasis. J Am Coll 8. Reynen K, Kockeritz U, Strasser RH. Metastases heart block, Merkel cell carcinoma,
Cardiol 2010;55:496. to the heart. Ann Oncol 2004;15:375–81. pacemaker
Infiltrative Lymphoma-Associated
Bradycardia and Cardiac Conduction
Abnormalities
Pritha Subramanyam, MD,a Syed S. Mahmood, MD, MPH,a William Dinsfriend, MD,c Raymond D. Pastore, MD,b
Peter Martin, MD,b Angel T. Chan, MD, PHD,c Jonathan W. Weinsaft, MD,a Jim W. Cheung, MDa
CASE 1
In 2015, a 31-year-old woman presented with cough and lymphadenopathy. On positron emission tomography-
computed tomography (PET-CT), she was found to have cervical and mediastinal lymph node involvement and
superior vena cava encasement. Open left neck lymph node biopsy confirmed the diagnosis of nodular
sclerosing Hodgkin’s lymphoma, with immunohistochemistry positive for CD30 and CD15. She underwent
transthoracic echocardiography (TTE), which revealed a large pericardial effusion with evidence of tampo-
nade. She was admitted and underwent pericardiocentesis, with drainage of 500 ml of serosanginous fluid.
Repeat TTE revealed a small pericardial effusion. The patient developed sinus pauses of up to 9 s in duration,
in addition to episodes of Mobitz I second-degree atrioventricular block.
Cardiac magnetic resonance (CMR) imaging revealed a normal left ventricular ejection fraction (EF) of 67%
and an irregularly contoured tissue prominence on the posterior aspect of left and right atria that measured
1.4 0.9 cm with perivascular thickening adjacent to the aorta and superior vena cava. Tissue characterization
with contrast enhancement was consistent with neoplastic invasion. The patient continued to have frequent
sinus pauses consistent with direct sinoatrial node or carotid sinus involvement. A temporary transvenous
pacemaker (TVP) was placed. Cycle 1 of adriamycin, bleomycin, vinblastine, dicarbazine (ABVD) chemotherapy
was initiated in two 14-day infusions. The TVP was removed due to reduced sinus pauses. However, a repeat
CMR showed an increase in the size of the right atrial mass (1.8 1.0 cm), with a septal thickness of 1.1 cm
(Figure 1). She also developed nonocclusive thrombi in both the proximal left internal jugular and right
subclavian veins.
Despite receiving chemotherapy, the patient continued to have prolonged sinus pauses. After a discussion
between the cardiology and oncology services, the decision was made to implant a permanent pacemaker. A
compassionate use exemption for a leadless pacemaker being studied at our institution as part of a clinical trial
was denied. Therefore, the patient underwent placement of a single-chamber right ventricular lead pacemaker
(Biotronik Eluna, Portland, Oregon) due to presence of continued sinus node dysfunction and atrioventricular
block. The patient completed 6 cycles of ABVD and has remained in full remission. She has had minimal pacing
requirements and has been planned for device removal and lead extraction.
From the aDepartment of Medicine, Division of Cardiology, Weill Cornell Medicine, New York-Presbyterian Hospital, New York,
New York; bDepartment of Medicine, Division of Hematology and Oncology, Weill-Cornell Medicine, New York-Presbyterian
Hospital, New York, New York; and cDepartment of Medicine, Memorial Sloan Kettering Cancer Center, New York, New York.
Dr. Mahmood is supported by the New York Academy of Medicine’s Glorney-Raisbeck Fellowship Award; and has received
consulting fees from Medicure, OMR Globus, Alpha Detail, and Opinion Research Team. Dr. Pastore has received consulting
fees from Sanofi Genzyme, Dova, and Rigel; has been a member of the Medical Advisory Boards of Rigel, Dova, and
BioVerativ; and has been a speaker for Sanofi Genzyme. Dr. Martine has been a consultant for Bayer, Celgene, Cellectar,
Janssen, Kit, Morphosys, Regeneron, and Teneobio. Dr. Cheung has received consulting fees from Abbott and Biotronik; and
has received fellowship grant support from Abbott, Biotronik, Boston Scientific, and Medtronic. All other authors have
reported that they have no relationships relevant to the contents of this paper to disclose.
Manuscript received November 19, 2019; revised manuscript received January 5, 2020, accepted January 7, 2020.

136 Subramanyam et al. JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020
Lymphoma-Associated Bradycardia and Conduction Abnormalities MARCH 2020:135–8
ABBREVIATIONS CASE 2
AND ACRONYMS
ABVD = adriamycin, In 2019, a 57-year-old man presented with abdominal pain and jaundice and was found to have biliary
bleomycin, vinblastine, obstruction. He underwent CT imaging, which was notable for lung and mediastinal masses. Following
dacarbazine
an endoscopic retrograde cholangio-pancreatography to relieve the biliary obstruction, a duodenal
CMR = cardiac magnetic
biopsy was performed. Pathology revealed diffuse large B-cell lymphoma. At subsequent follow-up, he
resonance
complained of dizziness, and an electrocardiogram (ECG) was notable for complete heart block with a
ECG = electrocardiography
junctional escape rhythm at a ventricular rate of 40 beats/min.
EF = ejection fraction
The patient was admitted, and he remained in a stable junctional rhythm for several days. CMR
FDG = fluorodeoxyglucose
revealed an irregularly contoured intrapericardial mass involving the interatrial septum, pulmonary
LGE = late gadolinium veins, aortic root, and membranous interventricular septum, measuring up to 1.9 cm in diameter
enhancement
(Figure 1). Post-contrast enhancement demonstrated tissue characterization consistent with neoplasm,
PET-CT = positron emission
with a preserved EF of 62%, and no evidence of myocardial scar or infarction. PET-CT showed fluo-
tomography-computed
tomography rodeoxyglucose (FDG) avid areas in the mediastinum that extended into the pericardium and right
R-EPOCH = ritixubimab, lung, and encased the left mainstem bronchus, left main pulmonary artery, proximal biliary duct, and
etoposide, prednisone, proximal gastric wall. Before biopsy, a TVP was placed in anticipation of worsening conduction dis-
vincristine, cyclophosphamide,
ease. Right lung core biopsy was performed and confirmed the diagnosis of germinal center-like
doxorubicin
diffuse large B-cell lymphoma. Immunohistochemistry was positive for BCL-2 and BCL-6 and
TTE = transthoracic
echocardiography
negative for MYC, and FISH did not demonstrate translocations of BCL-2 and MYC.
Cycle 1 of ritixubimab, etoposide, prednisone, vincristine, cyclophosphamide, doxorubicin (R-
TVP = transvenous pacemaker
EPOCH) was completed. He remained in complete heart block. Because of the extent of the patient’s
tumor burden, full recovery of conduction disease was believed to be unlikely to occur within a short time
frame. The decision was made to implant a permanent pacemaker and the options of a dual-chamber pace-
maker and leadless pacemaker were considered. To minimize the risk of infection and hematoma with
anticipated pancytopenia following completion of chemotherapy, a leadless pacemaker (Medtronic Micra,
Minneapolis, Minnesota) was implanted after cycle 1 of R-EPOCH. The patient has been stable and is
continuing chemotherapy. Pacemaker interrogation 1-month post-implantation revealed 1% ventricular pacing
burden.
DISCUSSION
Although primary cardiac lymphoma is rare, lymphomas are one of the most common cancers with cardiac
metastases (1,2). In an autopsy study of 150 patients with lymphoma, 13 (8.7%) had histological evidence of
cardiac involvement, one-half of whom had abnormal ECGs but did not have specific conduction abnormalities
(3). Despite the histological presence of cardiac involvement in a substantial percentage of patients with
lymphoma, its presentation is often subclinical, and its prevalence is poorly understood. Other infiltrative
masses can cause similar conduction disturbances, including neck masses such as thymoma, catecholamine-
secreting tumors, cardiac metastasis, and amyloidosis (4). The overall prevalence of the association of car-
diac conduction disease with these malignancies is unknown.
Generally, T-cell lymphomas invade the myocardium more frequently than B-cell lymphomas (3,5). There
are 3 mechanisms for cardiac involvement by lymphoma: 1) extension of tumor from the mediastinum; 2)
hematogeneous spread; and 3) retrograde flow via cardiac lymphatics. Direct extension is the most destructive,
with the highest frequency of cardiac dysfunction (5).
WHO IS AT RISK? It is important to maintain a high index of suspicion for the development of cardiac
conduction disease if lymphoma rapidly spreads to different anatomic structures, especially in patients with
previous conduction disease. It is essential to understand the anatomic locations of tumor involvement
and its proximity to the primary cardiac conduction system or structures that mediate cardiac
autonomic input.
HOW TO DIAGNOSE? The optimal tools for visualizing cardiac involvement in lymphoma include
echocardiography, CMR, and PET-CT. Tissue characterization by CMR or PET provides added usefulness for
identification of neoplastic involvement in the heart. Long inversion time late gadolinium enhancement CMR
(LGE-CMR) identifies neoplasms based on tissue vascularity, whereas PET identifies neoplasms based on
metabolic activity. Studies have shown that CMR tissue characterization can stratify prognosis among patients
JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020 Subramanyam et al. 137
MARCH 2020:135–8 Lymphoma-Associated Bradycardia and Conduction Abnormalities
F I G U R E 1 Summary of Managing Patients With Lymphoma
Summary of approach to managing patients with lymphoma at risk for bradycardia and cardiac conduction disease. Timing of permanent pacemaker
implantation (asterisk) requires consideration of timing of chemotherapy and its impact on tumor burden and on blood cell counts. (Top)
Representative cine cardiac magnetic resonance (CMR) imaging and (Bottom) late gadolinium enhancement (LGE) CMR images of Case 1 and Case 2
patients are also shown to the upper right. Cine-CMR was acquired at different phases of the cardiac cycle. LGE CMR was acquired using long
inversion recovery pulse sequence for which inversion times (TI) were tailored to null thrombus (TI 600), whereas contrast enhancement was
observed in tissue with vascularity. Neoplastic invasion of the interatrial septum (yellow circles) was present in both patients.
with cardiac neoplasms (6,7). PET imaging with FDG-avid uptake can allow early detection of
cardiac involvement by neoplasm (8). However, tumors with low metabolic activity may be challenging to
discern on PET.
The diagnostic sensitivity of FDG-PET for detection of cardiac neoplasms may be dependent upon the extent
of tissue necrosis. In a study of 121 patients with cardiac neoplasms who underwent CMR and PET within a
3-month interval, PET sensitivity for tumor detection varied in relation to the extent of CMR-evidenced
avascularity (9). PET detection of diffusely enhanced or mixed lesions was higher than for predominantly
avascular neoplasms (87% vs. 63%). PET-FDG uptake decreased stepwise when neoplasms were stratified
based on the extent of avascularity on LGE-CMR (p # 0.001). Therefore, in cases with a high suspicion of
cardiac neoplasm, CMR with tissue mass characterization will have a higher yield (10).
WHAT IS THE TREATMENT STRATEGY? The management of lymphoma-associated bradycardia and cardiac
conduction disease requires close coordination between cardiology and oncology teams because decisions on
both cardiac and cancer treatment should be made in parallel. On the basis of our clinical experience and
expert consensus, we propose a flowchart for management of this clinical problem in Figure 1. In the acute
presentation, if advanced atrioventricular block or sinus node dysfunction with symptoms or signs of hemo-
dynamic instability are encountered, then temporary pacing is indicated. If the patient is stable, CMR imaging
should be performed before placement of a TVP to determine the extent of cardiac involvement by the tumor.
After placement of a TVP, there are 2 decision points: 1) timing of transition from a temporary to permanent
pacemaker if necessary; and 2) implantation of a TVP versus a leadless pacemaker. Transition to a permanent
pacemaker is recommended if there is a persistent need for pacing and if the extent of lymphoma burden and
its impact on cardiac conduction are not expected to be resolved within several weeks. The decision to implant
138 Subramanyam et al. JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020
Lymphoma-Associated Bradycardia and Conduction Abnormalities MARCH 2020:135–8
a TVP versus a leadless pacemaker should be based on weighing the need for dual-chamber pacing against the
risks associated with device system infection, pocket hematoma, and lead-related vascular issues. Currently,
there are no published reports comparing outcomes of infection, bleeding, and complication rates associated
with TVPs versus leadless pacemakers in patients with cancer.
Proper timing of device implantation with respect to chemotherapy is critical because chemotherapy-
induced pancytopenia can significantly increase the risks of infection and device-related hematoma forma-
tion. In addition, tumor involvement or thrombotic occlusion of venous structures may preclude the use of
transvenous pacing. Overall, we recommend the timing and type of implantable pacemaker be decided
collectively and within clinical context (9).
IS THERE A CHANCE OF RECOVERY? The extent of cardiac involvement by tumor invasion, the chance of
complete cancer remission with treatment, and the presence of pre-existing conduction disease will determine
the chances of cardiac recovery. Patients with extensive tumor burden or significant baseline conduction
disease before cancer diagnosis are more likely to require long-term permanent pacing.
CONCLUSIONS
These cases highlight the need for recognition of bradycardia and cardiac conduction abnormalities caused by
lymphoma infiltration, the usefulness of imaging (e.g., CMR and PET-CT), and the importance of a multidis-
ciplinary and patient-centered shared decision-making approach to management.
ADDRESS FOR CORRESPONDENCE: Dr. Jim W. Cheung, Weill Cornell Medicine, New York-Presbyterian Hospital,
520 East 70th Street, Starr 4th Floor, New York, New York 10021. E-mail: jac9029@med.cornell.edu. Twitter:
@DrJCheungEP.
REFERENCES
1. Allen DC, JA, Morton P, Mollan PA, Morris TC. 25 autopsy cases based on the WHO classification. 9. Nagano M, Uike N, Suzumiya J, et al. Successful
Pathology of the heart and conduction system in Ann Hematol 2005;84:498–505. treatment of a patient with cardiac lymphoma
lymphoma and leukemia. J Clin Pathol 1987;40: who presented with a complete
6. Chan AT, Plodkowski AJ, Pun SC, et al.
746–50. atrioventricular block. Am J Hematol 1998;59:
Prognostic utility of differential tissue
171–4.
2. Palaskas NTK, Gladish G, Agha AM, et al. Eval- characterization of cardiac neoplasm and
uation and management of cardiac tumors. Curr thrombus via late gadolinium enhancement 10. Chan AT, Fox J, Perez Johnston R, et al. Late
Treat Options Cardiovasc Med 2018;20:29. cardiovascular magnetic resonance among pa- gadolinium enhancement cardiac magnetic
tients with advanced systemic cancer. J Cardiovasc resonance tissue characterization for
3. McDonnell PJMR, Bulkley BH. Involvement of
Magn Reson 2017;19:76. cancer-associated cardiac masses: metabolic and
the heart by malignant lymphoma: a clinicopath-
prognostic manifestations in relation to
ologic study. Cancer 1982;49:944–51. 7. Pun SC, Plodkowski A, Matasar MJ, et al.
whole-body positron emission tomography. J Am
4. Chang HM, Okwuosa TM, Scarabelli T, Pattern and prognostic implications of cardiac
Heart Assoc 2019;8:e011709.
Moudgil R, Yeh ETH. Cardiovascular complications metastases among patients with advanced sys-
of cancer therapy: best practices in diagnosis, temic cancer assessed with cardiac magnetic
prevention, and management: part 2. J Am Coll resonance imaging. J Am Heart Assoc 2016;5.
Cardiol 2017;70:2552–65.
8. Cui J, Ren Z, Wang X, et al. Evaluation of 18F- KEY WORDS bradycardia, cardiac magnetic
5. Chinen K, Izumo T. Cardiac involvement by FDG PET/CT in the diagnosis of cardiac lymphoma. resonance, cardiac masses, conduction
malignant lymphoma: a clinicopathologic study of J Nucl Med 2018;59:S1574. disease, lymphoma, pacemaker
INTERNATIONAL PERSPECTIVES
Cardio-Oncology in China
We Are on the Go!
Yuhui Zhang, MD,a Zhiren Zhang, MD, PHD,b Ying Liu, MD,c Jian Zhang, MD, PHDa
A s the most populous country in the world, in

the year 2020, there will be approximately
4.5 million new cases of cancer in China (1).
Cancer and cardiovascular disease are the 2 most
oncology, hematology, and radiation oncology. Dur-
ing that conference, we collectively agreed to develop
a formalized cardio-oncology program in China.
Since then, cardio-oncology clinics and multidis-
common causes of mortality in China. However, ciplinary teams have been established across the
because of the development of newer therapies, can- entire country, including: 1) cardio-oncology clinics
cer patients are living longer, and as a consequence, and multidisciplinary teams in general hospitals,
also suffering from an increased risk of cardiovascular such as The First Affiliated Hospital of Dalian Medical
disease. This Perspective aims to present the recent University and Zhongshan Hospital of Fudan Uni-
development of cardio-oncology in China. versity; 2) cardio-oncology clinics in cancer-specific
hospitals, such as the Harbin Medical University
THE DEVELOPMENT AND GROWTH OF
Cancer Hospital; Chongqing Cancer Hospital; Cancer
CARDIO-ONCOLOGY IN CHINA
Hospital, Chinese Academy of Medical Science
(CAMS); and the Peking Union Medical College
In 2011, Dr. Bonnie Ky delivered a keynote lecture at
(PUMC); 3) cardio-oncology clinics in cardiac-specific
the China Heart Congress, and first introduced the
hospitals, such as the Fuwai Hospital, CAMS, and
concept of cardio-oncology to medical specialists in
PUMC; and 4) multidisciplinary teams across cardio-
China. At that time, most Chinese cardiovascular and
vascular and oncology hospitals, including collabo-
oncology physicians never heard of cardio-oncology
rations between the cardiac-specific centers Fuwai
and lacked awareness of the treatment and preven-
Hospital, CAMS, and PUMC with cancer-specific cen-
tion of cardiotoxicity in cancer patients. The first
ters in CAMS, PUMC, and the Beijing Cancer Hospital.
cardio-oncology–specific forum was then held at the
To date, there are more than 10 large cardio-
China International Heart Failure Congress in March
oncology clinics or multidisciplinary groups in China
2016, followed by the first cardio-oncology workshop
(Figure 1). We anticipate these numbers will continue
in June 2016 in Dalian, China. This meeting was
to grow to meet the demands of an increasing number
attended by 17 experts from different provinces and
of cancer patients. Alongside this, many professional
cities, and brought together experts from cardiology,
cardio-oncology societies have also been established
in China (Figure 2), providing professional and orga-
nizational platforms to promote the development of
From the aHeart Failure Center, State Key Laboratory of Cardiovascular cardio-oncology programs across China.
Disease, Fuwai Hospital, National Center for Cardiovascular Diseases,
Chinese Academy of Medical Sciences and Peking Union Medical College, CARDIO-ONCOLOGY EDUCATION AND
Beijing, China; bDepartment of Cardiology of Harbin Medical University TRAINING IN CHINA
Cancer Hospital; and the cDepartment of Cardiology, The First Affiliated
Hospital of Dalian Medical University, Dalian, China. The authors have
reported that they have no relationships relevant to the contents of this
With the leading efforts by major professional soci-
paper to report. eties, many cardio-oncology congresses and specific
The authors attest they are in compliance with human studies forums have been held over the past 4 years. Cardio-
committees and animal welfare regulations of the authors’ institutions
oncology congresses include meetings such as the
and Food and Drug Administration guidelines, including patient consent
where appropriate. For more information, visit the JACC: CardioOncology China CardioOncology Conference, China Cancer
author instructions page. Management Symposium, and the International

Cardio-Oncology in China MARCH 2020:139–43
F I G U R E 1 Distribution of Major Cardio-Oncology Clinics in China
Conference for Metabolic Disorders and Cancer research findings to clinicians and researchers. The
Related Cardiovascular Diseases. iCardio-oncology account is one example of a specific
Cardio-oncology specific sessions have been held WeChat network. The Journal of Heart Failure and
during the scientific meetings of cardiovascular and Cardiomyopathy has also published a special column
oncology societies, including the Integrated Cardio- on JACC: CardioOncology highlighting each issue with
Oncology Society of China Anti-Cancer Association, translated synopses. Efforts are ongoing to also pro-
Chinese Society of Clinical Oncology, Great Wall In- vide translated summaries of cardio-oncology papers
ternational Congress of Cardiology, China Heart published internationally as we believe this will
Congress, China International Heart Failure Congress, enable physicians in China to access the leading pa-
Chinese Society of Cardiology, Oriental Congress of pers and facilitate the practice of cardio-oncology
Cardiology, and South China International Congress according to the latest guidelines, recommenda-
of Cardiology. tions, and research. Several textbooks have been
In 2018, the Integrated Cardio-Oncology Society translated, written, and published in China. Popular
launched the “CardioOncology Academy” and texts have also been written specifically for patients
“Speaking Tour” program in several provinces. A total to empower them and increase awareness. Together,
of 58 specific conferences on cardio-oncology have all of these efforts have produced great impact on the
been held, and more than 10,000 physicians have education and training of cardio-oncologists in China.
attended these activities to help them exchange ideas
and promote collaborations. CARDIO-ONCOLOGY RESEARCH IN CHINA
Academic websites, journals, and WeChat official
accounts on nationwide social media platforms have There are many ongoing basic, clinical, and trans-
been established to foster communications and lational research programs in cardio-oncology,
dissemination of cardio-oncology clinical and investigating the mechanisms, pathophysiology,
MARCH 2020:139–43 Cardio-Oncology in China
F I G U R E 2 Timeline Depicting the Development of Cardio-Oncology Societies in China
epidemiology, evaluation, prevention, and manage- fibrillation in breast cancer patients treated with
ment of cardio-oncology. For example, over the past 2 trastuzumab was approximately 1.2%, with the high-
years in China, approximately 100 peer-reviewed est incidence of atrial fibrillation occurring within
publications have been published in Chinese and 90 days after cancer diagnosis. In the Kailuan study,
English journals. which included 68,759 Chinese male adults with
In the clinical research field, Liu et al. (2) found approximately 8 years of follow-up, total cholesterol,
that hypertension was the most prevalent cardiovas- low-density lipoprotein (LDL)–cholesterol, and non-
cular risk factor (26.0%) among 22,500 newly diag- LDL-cholesterol not only increased the risk of car-
nosed cancer patients; Liu et al. (3) reported a study diovascular disease, but also negatively impacted
which included 710,170 cancer patients, and showed cancer risk. Lipophilic statins have been shown to
that the presence of hypertension was associated improve the prognosis of breast cancer patients (8).
with a higher all-cause mortality in patients with Xu et al. (9) found that echocardiographic 3-dimen-
heart failure, myocardial infarction, atrial fibrillation, sional speckle tracking imaging with strain was useful
stroke, and diabetes. Cao et al. (4) found that in the early detection of myocardial injury caused by
increased radiation dose was associated with a higher chemotherapeutic drugs. Anqi et al. (10) showed that
risk of heart failure in breast cancer patients; Hu et al. segmental left ventricular strain, indicative of regions
(5) showed that thoracic cancer survivors who un- supplied by the left anterior descending coronary
derwent radiotherapy have higher SYNTAX scores artery, was a more sensitive indicator of the car-
and were at a higher risk of developing anatomically diotoxicity of anthracyclines.
severe coronary artery disease, independent of In basic research, An et al. (11) found that
chemotherapy exposure. Meta-analyses conducted by neuregulin-1 could attenuate doxorubicin-induced
Yuan et al. (6,7) showed the incidence of atrial autophagy and apoptosis in cardiomyocytes. Xie
Cardio-Oncology in China MARCH 2020:139–43
et al. (12) showed mitochondrial ONOO - represents an cardioprotection prophylaxis in Chinese cancer pa-
early biomarker to predict subclinical anthracycline tients, using not only Western but also traditional
cardiotoxicity. Li et al. (13) reported that L-arginine Chinese medicines; 6) Issuing guidelines and
was effective in suppressing doxorubicin-induced consensus statements on the diagnosis, prevention
vascular dysfunction by attenuating vascular NO and treatment of the cardiotoxicity of cancer thera-
release and apoptosis. Huang et al. (14) found that pies tailored to Chinese cancer patients; 7) Devel-
NH4 CL significantly improved doxorubicin-induced oping educational platforms aimed to improve the
contractile dysfunction, inflammation, apoptosis, quality of education for cardio-oncologists and phy-
and autophagy in mice. Ma et al. (15) found that rit- sicians at all levels; and 8) Promoting international
uximab can prevent or reverse ventricular remodeling exchange and collaboration in cardio-oncology
by interfering with B cells. Liu et al. (16) showed that globally.
autophagy-related single nucleotide polymorphisms In the future, cardio-oncology will continue to
were associated with chemotherapy-induced rapidly evolve in China, and together we will face the
cardiotoxicity. challenges to maintain and develop the standards to
Some original studies and clinical trials are best serve our patients. We are on the go!
ongoing: Dr. Zhiren Zhang received funding from the ACKNOWLEDGMENTS The authors thank Xia Yun-
National Natural Science Foundation of China to long, MD, Director of the First Affiliated Hospital of
study the mechanisms of cardiotoxicity caused by Dalian Medical University; Ma Fei, MD, Director of
anti-cancer medications; Dr. Yuhui Zhang launched a Oncology of Cancer Hospital, CAMS & PUMC; David H.
randomized clinical study on the prevention of Hsi, MD, Chief of Cardiology and Co-Director, Heart &
cardiovascular toxicity with a traditional Chinese Vascular Institute, Professor of Clinical Medicine,
medicine; Dr. Yun Zhang is leading a multicenter Columbia University College of Physicians & Surgeons
study on the use of echocardiography to evaluate Stamford Hospital, Stamford, Connecticut; Gary Tse,
cardiotoxicity with chemotherapy and the role of MA, Department of Medicine and Therapeutics, Fac-
targeted cardioprotective therapy in patients with ulty of Medicine, Chinese University of Hong Kong,
breast cancer; Dr. Yunlong Xia initiated a multicenter Hong Kong, China; Cheng Leilei, MD, Department of
clinical registry study on the preventive treatment of Echocardiography, Zhongshan Hospital, Fudan Uni-
bevacizumab-related hypertension with olmesartan. versity; Shao Qun, MD, Vice Director of Cardiology of
THE FUTURE OF CARDIO-ONCOLOGY Harbin Medical University Cancer Hospital, Sandra
IN CHINA Wang, China iCardio-Oncology Network, for their
shared vision and invaluable support in the estab-
Over the next few years, we hope to focus on lishment in China cardio-oncology service.
accomplishing the following goals: 1) Improving the
awareness of cardio-oncology amongst both cardio- ADDRESS FOR CORRESPONDENCE: Dr. Jian Zhang,
vascular and oncology physicians; 2) Growing our Heart Failure Center, State Key Laboratory of Cardio-
cardio-oncology clinics and multidisciplinary teams vascular Disease, Fuwai Hospital, National Center for
to serve more cancer patients; 3) Establishing the Cardiovascular Diseases, Chinese Academy of Medical
standards for the diagnosis, evaluation, monitoring, Sciences and Peking Union Medical College, Beijing
and risk stratification for Chinese cardio-oncology 100037, China. E-mail: fwzhangjian62@126.com. Yuhui
patients; 4) Initiating multicenter registries to study Zhang wechat ID: 15901314243; Zhiren Zhang wechat
cardiotoxicity in Chinese cancer patients; 5) Planning ID: wxid_oog1uon88rag22; Ying Liu wechat ID:
multicenter randomized clinical trials for 18098875801; Jian Zhang wechat ID: 13911102015.
REFERENCES
1. Global Cancer Observatory. Cancer Today. Lyon, 3. Liu D, Ma Z, Yang J, et al. Prevalence and coronary artery disease: a propensity score
France: International Agency for Research on prognosis significance of cardiovascular disease in matching comparison among adult-onset
Cancer, 2018. Available at: https://gco.iarc.fr/ cancer patients: a population-based study. Aging thoracic cancer survivors. Cardiology 2018;140:
today. Accessed February 20, 2019. (Albany NY) 2019;11:7948–60. 239–46.
2. Liu F, Hidru TH, Gao R, et al. Cancer pa- 4. Cao L, Cai G, Chang C, et al. Early cardiac 6. Yuan M, Zhang Z, Tse G, et al. Association of
tients with potential eligibility for vascular toxicity following adjuvant radiotherapy of left- cancer and the risk of developing atrial fibrillation:
endothelial growth factor antagonists use sided breast cancer with or without concurrent a systematic review and meta-analysis. Cardiol Res
have an increased risk for cardiovascular dis- trastuzumab [J]. Oncotarget 2016;7:1042–54. Pract 2019;2019:8985273.
eases comorbidities. J Hypertens 2020;38:
5. Hu S, Gao H, Zhang J, et al. Association be- 7. Yuan M, Tse G, Zhang Z, et al. The incidence of
426–33.
tween radiotherapy and anatomic severity of atrial fibrillation with trastuzumab treatment: a
MARCH 2020:139–43 Cardio-Oncology in China
systematic review and meta-analysis. Cardiovasc 11. An T, Huang Y, Zhou Q, et al. Neuregulin-1 at- 15. Ma XL, Lin QY, Wang L, Xie X, Zhang YL,
Ther 2018;36:e12475. tenuates doxorubicin-induced autophagy in neonatal Li HH. Rituximab prevents and reverses car-
rat cardiomyocytes. J Cardiovasc Pharmacol 2013;62: diac remodeling by depressing B cell function
8. Guan XM, Wu SL, Yang XL, et al. Associa-
130–7. in mice. Biomed Pharmacother 2019;114:
tion of total cholesterol, low-density lipopro-
12. Xie X, Tang F, Liu G, et al. Mitochondrial per- 108804.
tein cholesterol, and non-high-density
lipoprotein cholesterol with atherosclerotic oxynitrite mediation of anthracycline-induced 16. Liu B, An T, Li M, et al. The association
cardiovascular disease and cancer in a Chinese cardiotoxicity as visualized by a two-photon between early-onset cardiac events caused by
male population. Int J Cancer 2018;142: near-infrared fluorescent probe. Anal Chem neoadjuvant or adjuvant chemotherapy
1209–17. 2018;90:11629–35. in triple-negative breast cancer patients
13. Li X, Gu J, Zhang Y, et al. l-arginine alleviates and some novel autophagy-related poly-
9. Xu Y, Shi J, Zhao R, et al. Anthracycline induced
doxorubicin-induced endothelium-dependent morphisms in their genomic DNA: a real-
inconsistent left ventricular segmental systolic func-
dysfunction by promoting nitric oxide generation world study. Cancer Commun (London) 2018;
tion variation in patients with lymphoma detected by
and inhibiting apoptosis. Toxicology 2019;423: 38:71.
three-dimensional speckle tracking imaging. Int J
Cardiovasc Imaging 2019;35:771–9. 105–11.
10. Anqi Y, Yu Z, Mingjun X, et al. Use of echocar- 14. Huang X, Liu Y, Yang X, et al. NH4 Cl
diography to monitor myocardial damage during treatment prevents doxorubicin-induced KEY WORDS cardiotoxicity, cardio-oncology,
anthracycline chemotherapy. Echocardiography myocardial dysfunction in vivo. Life Sci 2019; China, evolvement of cardio-oncology,
2019;36:495–502. 227:94–100. multidisciplinary team
ª 2020 PUBLISHED BY ELSEVIER ON BEHALF OF THE AMERICAN COLLEGE OF
LEADERSHIP PAGE
Working Together
to Advance
Cardio-Oncology in China
Yunlong Xia, MD, PHD, President, Cardio-Oncology Committee,
China Society of Cardiology
Jiwei Liu, MD, PHD, President, Cardio-Oncology Committee,
Chinese Society of Clinical Oncology
WHY WE PROMOTE cancer patients as CVD and cancer have many shared
CARDIO-ONCOLOGY IN CHINA traditional risk factors, such as age, tobacco, and
obesity; and 3) primary or metastatic cardiac tumors,
The number of cancer patients and survivors is which are comparatively less common.
increasing in China. However, many cancer therapies The responsibilities of cardio-oncology care pro-
have short- and long-term cardiovascular (CV) toxic- viders include the diagnosis and management of
ities. The risk of cardiovascular disease (CVD) in can- multiple potential CV conditions in cancer patients.
cer patients is substantially higher than the incidence In our practice, we typically recommend a strategy of
observed in general population (1–4). CV events comprehensive cardiovascular management, which
may occur during anti-cancer therapy or decades involves a general CV evaluation before anti-cancer
after treatment. Once patients receive chemotherapy, therapy, individualized monitoring and manage-
radiotherapy, targeted therapy, or immunotherapy, ment of CV risk factors and CVD during anti-cancer
the adverse CV effects may persist over their therapy, and life-long CV follow-up after anti-cancer
lifetime. As a result, CVD has become a leading cause therapy is completed. We have tried to adapt this
of morbidity or mortality in cancer survivors. strategy in the oncology units of several medical
CV risk factors and CVD in cancer patients tend to centers in China. As a result, we hope to detect the
be complex, and neither the cardiologist nor the potential CV effects of anti-cancer therapy earlier so
oncologist can fully manage them alone. In the past, that we can more intensively manage these condi-
we have paid inadequate attention to this patient tions and ultimately improve the lives of our patients.
population. The clinical needs of CV care of cancer
CURRENT STATUS OF
patients are increasing rapidly, and experts from
CARDIO-ONCOLOGY IN CHINA
cardiology and oncology must collaborate closely to
overcome these complexities and challenges, with
The incidence rates of CVD and cancer are both
the goal of improving CV prognosis in the growing
increasing in China. As a result, these diseases have
cancer population.
become the 2 leading causes of death in our country.
THE MULTIDISCIPLINARY NATURE OF Approximately 290 million Chinese patients suffer
CARDIO-ONCOLOGY from CV risk factors or CVD. According to the 2018
Report on Cardiovascular Diseases in China (6), the
Cardio-oncology lays at the intersection of oncology number of patients with stroke, coronary heart dis-
and cardiology, and is largely focused on the man- ease, pulmonary heart disease, heart failure, rheu-
agement of CV risk factors and CVD in cancer pa- matic heart disease, and congenital heart disease are
tients (5). In China, cardio-oncology practice is 13 million, 11 million, 5 million, 4.5 million, 2.5 million,
typically focused on the following: 1) CV toxicities and 2 million, respectively. Hypertension is prevalent
related to anti-cancer therapy; 2) diagnosis and in 245 million. CVD-related deaths account for more
management of co-existent CV risk factors or CVD in than 40% of deaths. Meanwhile, it is estimated that

JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020 Xia and Liu 145
MARCH 2020:144–5 Working Together to Advance Cardio-Oncology in China
more than 4 million new cases of cancer are diagnosed must strengthen awareness and education, to not only
each year in China. Lung, hepatocellular, gastrointes- the physicians, but also to cancer patients. In China,
tinal, and breast cancers are the most common. there are more than 10,000 patients being newly
The field of cardio-oncology started relatively diagnosed with cancer every day. Therefore, the pop-
recently in China (7). In 2016, the first conference ulation of cancer survivors is also growing rapidly. We
dedicated entirely to cardio-oncology was held in the hope to implement a nationwide standardized CV
city of Dalian, which marked an important point in our screening program, using available modalities
history, when Chinese physicians began to focus on the including electrocardiograms, echocardiography,
CV care in the cancer population. Since then, the Chi- Holter monitors, or cardiac biomarkers to recognize
nese Cardio-Oncology Conference has been held patients at increased risk for CVD. Furthermore, we
annually. Efforts from Chinese cardiologists, oncolo- also look forward to learning, communicating, and
gists, hematologists, imaging specialists, and other collaborating with international cardio-oncologists
scientists are ongoing to promote the new discipline. more closely to serve our patients with the highest
As the number of epidemiologic studies, clinical trials, quality care, and provide the best training for our cli-
and basic science investigations has increased, the nicians. Our hope is to provide state-of-the-art cardio-
pace of scientific development in this new field has oncology care for our patients, to improve not just the
accelerated rapidly. Similarly, there are a growing quantity of years, but also quality of life.
number of clinical units. Currently, there are 31 cardio- JACC: CARDIOONCOLOGY IS INVALUABLE,
oncology medical units, including 22 in general hos- NECESSARY, AND IMPORTANT TO CHINA
pitals, 8 in cancer hospitals, and 1 in a cardiovascular
hospital, which have been built across 13 provinces To advance this growing field, we need more evi-
nationwide. From our view, a single spark of cardio- dence and greater experience. Experts in the Chinese
oncology has boosted a fire throughout our country. Society of Cardiology and the Chinese Society of
WE STAND TOGETHER TO ADVANCE Clinical Oncology are now collaborating in both sci-
CARDIO-ONCOLOGY entific research and clinical practice. We look to JACC:
CardioOncology to promote new knowledge—both
With the increasing recognition and the support of scientific and clinical—and for this to be platform of
numerous cardiologists, oncologists, hematologists, shared experiences for us to learn. We look to JACC:
cardiac imaging specialists, and basic research scien- CardioOncology for inspiration and to help stimulate
tists, cardio-oncology has been steadily gaining deeper discussions in the field and in China. It has
attention in China (8). In 2019, both the Chinese So- already been of great value to us, and it will help us
ciety of Cardiology and Chinese Society of Clinical improve the clinical care we provide to our patients.
Oncology set up a committee dedicated to cardio- We believe it will also serve to increase collaboration
oncology. There are now 15 cardio-oncology working across cardio-oncologists and expand the dialogue
groups throughout the country. These groups focus globally, ultimately benefiting a large, growing pop-
on promoting the CV health of cancer patients and ulation of patients.
cancer survivors, as well as promoting clinical trials,
developing basic mechanistic investigations, and ADDRESS FOR CORRESPONDENCE: Dr. Yunlong Xia,
establishing multidisciplinary training programs for Department of Cardiology, The First Affiliated Hos-
cardiologists specializing in cardio-oncology. pital of Dalian Medical University, No. 193, Lian He
All the members of our specialty society must work Road, Dalian, Liaoning, 116011, P. R. China. E-mail:
together to advance this new field in our country. We yunlong_xia@126.com.
REFERENCES
1. Abe J, Martin JF, Yeh ET. The future of onco- adolescent, and young adult survivors of Hodg- 7. Zhang Y, Zhang Z, Liu Y, Zhang J. Cardio-
cardiology: we are not just “side effect hunters.” kin’s lymphoma: an analysis from the St. Jude oncology in China: we are on the go! J Am Coll
Circ Res 2016;119:896–9. Lifetime Cohort Study. Lancet Oncol 2016;17: Cardiol CardioOnc 2020;2:139–43.
2. Cardinale DM, Barac A, Torbicki A, et al. Cardio- 1325–34.
8. Liu Y, Zhang YL, Liu JW, et al. Emergence,
oncological management of patients. Semin Oncol 5. Herrmann J. From trends to transformation: development, and future of cardio-oncology in
2019;46:408–13. where cardio-oncology is to make a difference. Eur China. Chin Med J (Engl) 2018;131:2640–4.
3. Howard E, Steingart RM, Armstrong GT, et al. Heart J 2019;40:3898–900.
Cardiovascular events in cancer survivors. Semin
Oncol 2019;46:426–32. 6. Shengshou H, Runlin G, Lisheng L, et al. Sum-
4. Bhakta N, Liu Q, Yeo F, et al. Cumulative burden mary of the 2018 report on Cardiovascular Dis- KEY WORDS cancer survivor, cardio-oncology,
of cardiovascular morbidity in paediatric, eases in China. Chinese Circ J 2019;3:209–20. cardiovascular disease (CVD), China
EDITOR’S PAGE
Cardio-Oncology and the

Patient–Physician Relationship
Bonnie Ky, MD, MSCE, FACC, Editor-in-Chief, JACC: CardioOncology
“I am a survivor. Bring it on.” distraction, to what our patients are trying to tell us
—From my patient, a survivor of 4 cancers, (2). We need our health care system to allow us the
cardiac arrest, and heart failure in response to time establish a patient-centered approach to med-
the question, “What first comes to mind when ical care. In cardio-oncology, in particular, the
hearing cancer and heart disease?” medical problems our patients face can be extraor-
T
dinarily complex, requiring detailed discussions and
he resiliency and strength of my patients
a careful navigation of concerns across multiple
inspire and motivate me every day. They suf-
disciplines.
fer from the burden of both cancer and car-
diovascular disease, but they are able to effectively
WE NEED SUPPORT
manage their health and navigate the multiple com-
plexities of their diseases. They work full time and
To practice at the “highest level of certification,” we
strive to live life each day to its fullest, as mothers, fa-
need greater support to provide coordinated care to
thers, sisters, brothers, daughters, sons, wives, hus-
our patients (2). The multidisciplinary nature of
bands, friends, and colleagues. But they also feel
cardio-oncology necessitates a team-based approach
anxious, overwhelmed, and sad. They cry, and I cry
to care, where silos of practice cannot exist. We
with them.
need institutional support to build an effective
OUR PATIENTS infrastructure that meets the needs of individual
providers and patients, to collaboratively develop
Caring for the needs of our patients is at the core of innovative solutions to overcome the challenges
what we—as medical professionals—do every day. But that we face in everyday practice, which include
how can we do this in the best possible way: How can lack of access to coordinated care and excessive
we remain firmly patient-centered, and partner with administrative burdens (1). We need an environ-
our patients to provide them with evidence-based, ment that educates, communicates, and values
personalized care? team-based care, so that an individual can use his
or her strengths and skill sets to most effectively
WE NEED TO LISTEN
contribute to the common goal of providing the best
possible care to our patients (2).
Fundamentally, we need to take the time, and be
afforded the time, to empathically listen to our WE NEED EVIDENCE
patients. As medical professionals, we are acutely
aware of the many needs and demands on our time We need to practice evidence-based and inter-
and the pressures of meeting productivity and per- personal medicine. Cardio-oncology is a growing
formance metrics within our practice. We are field, and we need science to inform our best
acutely aware of the many strengths, but also the practices. We need to develop guidelines based on
many burdens of the electronic health record sys- rigorously executed studies, the available data,
tem, which can provide ready access to data, but and graded evidence. We need to practice based
also impose overwhelming stress and intrusion on on scientific evidence, rather than intuition (3,4).
the human connection (1). We need to give our- Moreover, we need to understand how to best
selves the time to listen intently, without apply guidelines, using interpersonal medicine

JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020 Ky 147
MARCH 2020:146–8 Editor’s Page
F I G U R E 1 Artist’s Sketch Depicting the Discussions at the ACC Patient Forum “At a Crossroads: Surviving Cancer, Taking Care of Your Heart”
that addresses the behavioral and social factors for themselves. To help address these needs, the
that are barriers to our patients’ acceptance of National Cancer Institute–funded Eastern Coopera-
our recommendations (3). tive Group–American College of Radiology Imaging
Network and the American College of Cardiology
have collaborated to develop patient education
WE NEED TO EDUCATE AND ADVOCATE
modules specific to cardio-oncology (6).
The American College of Cardiology convened a To close, a recent mixed-methods study pub-
patient forum in September 2019, “At a Crossroads: lished in JAMA identified 5 practices with the po-
Surviving Cancer, Taking Your Care to Heart” tential to positively affect the patient-physician
(Figure 1), where 20 patients were invited to share relationship (7). Although there is a need for a
their experiences, in an effort to better understand greater understanding of the impact these measures
the unique challenges confronting cancer patients can have on clinical outcomes, they provide specific
with cardiovascular disease. The overarching mes- guiding principles for us to live by, as we rein-
sage was that patients wanted a physician who vigorate the patient-physician relationship:
would listen to them and advocate for them (5). Prepare with intention: Review the patient’s his-
Patients wanted more knowledge and to be better tory. Focus your attention before you walk into the
educated and informed, so they could also advocate room.
148 Ky JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020
Editor’s Page MARCH 2020:146–8
Listen intently and completely: Thoughtfully po-

sition your body. Listen without interruption. ADDRESS FOR CORRESPONDENCE: Dr. Bonnie Ky,
Department of Cardiovascular Medicine, Perelman
Agree on what matters most: Determine your pa-
School of Medicine at the University of Pennsyl-
tient’s concerns and priorities.
vania, Smilow Center for Translational Research,
Connect with your patient’s story: Empathize.
3400 Civic Center Boulevard, Philadelphia, Penn-
Acknowledge your patient’s efforts.
sylvania 19104. E-mail: bonnie.ky@pennmedicine.
Explore emotional cues: Be attentive, elicit, reflect
upenn.edu. Twitter: @pennmedicine.
and validate your patient’s cues.
REFERENCES
1. Lieu TA, Altschuler A, Weiner JZ, et al. Primary 4. Minasian LM, Dimond E, Davis M, et al. The 6. American College of Cardiology. CardioSmart.
care physicians’ experiences with and strategies evolving design of NIH-funded cardio-oncology Available at: www.cardiosmart.org. Accessed
for managing electronic messages. JAMA Netw studies to address cancer treatment-related car- February 5, 2020.
Open 2019;2:e1918287. diovascular toxicity. J Am Coll Cardiol CardioOnc
2019;1:105–13. 7. Zullman DM, Haverfield MC, Shaw JG,
2. Noseworthy J. The future of care—preserving
et al. Practices to foster physician
the patient-physician relationship. N Engl J Med
5. Strong SR. Decades after diagnosis: the unrec- presence and connection with patients
2019;381:2265–9.
ognized trauma of surviving. J Am Coll Cardiol in the clinical encounter. JAMA 2020;323:
3. Cheng S, Lee TH. Beyond evidence-based
CardioOnc 2020;3:149–51. 70–81.
medicine. N Engl J Med 2018;379:1983–5.
VIEWPOINT
Decades After Diagnosis

The Unrecognized Trauma of Surviving
Susan R. Strong, MA
W hen I met Darryl in 2012, I was in the best

shape of my life. Our first year together
was joyful and active: we enjoyed moun-
tain hikes and long talks, with the crunch of graveled
your opportunity to move on, because things are
about to get real.”
How could cancer be impacting me now, decades
after diagnosis? After all, at age 22, I had reached the
trail underfoot and an expanse of azure sky overhead; 5-year survival milestone. I was released from my
summer concerts under a canopy of stars at the oncologist’s care and told to go live my life.
Red Rocks amphitheater; and shared laughter and I was determined to live (and pushed down the
conversation over wine and gourmet dinners with simmering fear that I wouldn’t). Afraid of not having
friends. At 46, I felt certain that the second half of a future, I rushed into adulthood. Completing college
my life held a joy that my younger self had only in 7 semesters, I married at 21, had a baby at 24,
hoped for. started grad school at 25, continuing on this trajectory
One short year later, at a routine appointment, my until my life unraveled in my early 40s and my
primary care physician said, “Your murmur is 24-year marriage ended.
changing. Let’s send you to a cardiologist.” The life- No medical provider suggested the need for any
altering chain of events that followed left me blind- special follow-up. I intuitively sensed that a healthy
sided. lifestyle was important — I exercised, rarely drank,
A few weeks later, an echocardiogram revealed never smoked, and maintained a normal body weight.
severe aortic stenosis. The cardiologist said I would The only late effect I experienced happened within
need valve replacement within the next 6 months. He the first year of treatment: hypothyroidism. That
explained that the stenosis was a result of high-dose seemed like more of an inconvenience that I
(36Gy) mantle radiation I had received in 1983— accepted. Along with secondary infertility, premature
when at 17 years old, I had been diagnosed with menopause, and constant neck pain caused by
Hodgkin lymphoma. degenerative thoracic disc disease (within the
What?? My experience of cancer as a teenager had
radiation field) and muscular pain that I now under-
become a distant memory.. The staging laparotomy
stand to be radiation fibrosis. None were life-
and splenectomy, lymphangiogram, bone marrow
threatening. Overall, I had experienced relatively
biopsy, MOPP chemotherapy, radiation.the impact
good health.
of that brutal year seemed distant and unreal, like a
On November 12, 2014, just after my 49th birthday,
nightmare that was difficult to recall after the morn-
I had an aortic valve replacement via transcatheter
ing light and a cup of coffee.
aortic valve replacement (TAVR). The 6 months to a
Stunned, I sat on a metal bench in front of the
new valve my cardiologist had predicted in 2013 had
hospital, waiting for the valet to retrieve my car. I
stretched into 18 months; my procedure was delayed
called Darryl to tell him the shocking news about my
while waiting for FDA approval of a second-
heart valve. I braced myself and said stoically, “Now’s
generation valve. The night before the TAVR pro-
cedure, I lay awake wondering if I would wake up
after surgery, a fear creeping over me like a shadow in
the darkness. In retrospect, my fears seemed un-
From Heart Valve Voice US, Washington, DC. Ms. Strong has reported
that she has no relationships relevant to the contents of this paper to founded, but my aortic stenosis diagnosis had been
report. accompanied by an increasingly paralyzing anxiety.

150 Strong JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020
The Unrecognized Trauma of Surviving MARCH 2020:149–51
I realized I had spent my life thinking I would be Even though the events had occurred decades before,
lucky to live to be 50. the steps leading up to that cardio-oncology
My TAVR was a great success. Darryl had been by appointment were reminiscent of those 90-min,
my side the entire lengthy process. Ten days after the early morning drives from my home to the cancer
procedure, as we reached the top of a hill on my fa- center of another large university hospital in another
vorite trail near our home, Darryl said, “Stop. Do you state where I had lived as a child. Every infusion then
realize what you just did? You walked uphill, talking was given through a peripheral intravenous line, as
the whole time, and weren’t short of breath!” I had there were no chemo ports in the early 1980s. And no
my life back. medication was given to prevent the violent nausea
While in the hospital for my valve replacement, I from the MOPP regimen of nitrogen mustard,
met a cardio-oncologist for the first time. On rounds, vincristine, procarbazine and prednisone. I recalled
she introduced herself and told me about the hospi- how even after treatment ended, every time I
tal’s survivorship clinic. returned for follow-up, as soon as the familiar sur-
I made an appointment with her, feeling incredibly roundings indicated that the hospital was 15 min
fortunate to have access to this specialist. Our first away, I began to feel anxious and nauseated.
appointment was in the hospital’s cardiovascular The full impact of my cancer experience remained
center. I was comfortable in the familiar surroundings dormant for decades. I did not fully recognize the
where I had been successfully treated via TAVR, unresolved trauma I carried emotionally, nor was I
content to wait in the impressive lobby featuring a prepared for the extent of the collateral physical
contemporary, wall-mounted fireplace and colorful, damage initiated by high-dose mantle radiation.
floor-to-ceiling glass created from microscopic images This new, first-hand experience of late effects led
of cardiac tissue. The appointment was uneventful. me to search for information and support online.
The following year, my return appointment took Disappointingly, I found no patient organization
place in the hospital’s cancer center, as my doctor whose primary focus is to provide support and in-
met with patients in both clinics. formation to long term survivors of childhood and
I wasn’t nervous about this appointment. I felt adolescent and young adult cancers. I became
great and I fully expected my 2-year-old TAVR to be involved as a patient advocate in hopes of easing
fine. What I didn’t expect was the physical and the journey for others, creating a website for sur-
emotional impact of returning to the scene of the vivors who were facing cardiac late effects, and I
crime, so to speak. I had not stepped foot into an became a Heart Valve Ambassador with the Amer-
oncology clinic for more than 30 years. I walked into ican Heart Association. On May 7, 2016, I discovered
the cancer center, and as I checked in at the front the Hodgkin’s Lymphoma Survival & Late Effects
desk, I noticed people waiting who were obviously 1960-early 2000s support group on Facebook. Today
patients receiving chemotherapy. Standing there I this private, patients-only group includes 654
felt my muscles tense, my jaw clench, and a wave of members, with more added every week. These
nausea come over me, seemingly out of nowhere. The women and men, most of whom I’ve never met,
waiting room chairs were arranged in lines, like a have become my Tribe, my friends, my family. They
classroom. I took a seat and suddenly felt afraid and are a source of support, information, and inspira-
alone. When my name was called, I felt shaky, and my tion. And a source of heartbreak as well, as we have
disobedient eyes had welled with tears. I felt lost more friends than I care to count in a few short
confused and embarrassed by my reaction. I hadn’t years. I experience heartbreak over the dispropor-
had anxious thoughts. I wasn’t worried. And yet, my tionate, collective burden of physical and emotional
body, specifically my central nervous system, had suffering represented here among very real, very
been hijacked by this physical experience. By the human individuals.
time the nurse came into the exam room, I was My experience with my new-found peer group
shaking with tears freely flowing, despite my at- helped me grasp the need for high-risk surveillance
tempts to wipe them away and my strong desire for for breast cancer. During my experience with valvular
composure. When I apologetically explained that I disease, I neglected mammograms. When I went for
didn’t know why I was suddenly falling apart, she my first high-risk screening protocol of mammogram
kindly said it was common for cancer survivors to and MRI, it had been 3 years since my last screening.
have this response. Fortunately, the tests were clear, as well as the next
Reflecting on the drive home, I realized how the round 6 months later. Thirteen months later, in July
events of the morning had unconsciously elicited 2017, I received my first-ever call-back after a
physical memories of my own prior cancer treatment. mammogram. The radiologist who read the diagnostic
JACC: CARDIOONCOLOGY, VOL. 2, NO. 1, 2020 Strong 151
MARCH 2020:149–51 The Unrecognized Trauma of Surviving
images sent a letter with the opinion that the calcifi- a trauma psychologist who uses eye movement
cations were likely benign, recommending that I desensitization and reprocessing to treat the post-
follow-up with repeat tests in 6 months. My primary traumatic stress disorder (PTSD) symptoms that
care provider and I both agreed on biopsy. Two weeks originated with childhood cancer.
later, ironically on the same day, I received that letter Two important mentors from the online Hodgkin
from the radiologist as well as the pathology report group have given me hope in spite of an uncertain
from the biopsy: Cancer. Again. future. Dave (from Australia) and Dolly (from Texas)
Life seemed to be moving in slow motion. I was taught me how to face death: to speak of it openly,
teaching middle school full- time, but folding under and to love with an open heart. They each wrote of
the pressure of decisions about surgery and recon- savoring life moment by moment, and each lived
struction, multiple medical appointments, and un- fully and courageously to the end of their lives. From
certainty about the future. Nine weeks after the them I learned that someday we will die, but on all
biopsy, I took a leave-of-absence and underwent a the other days, we will not. The key is to keep that
double mastectomy. I met my oncologist for the first balance in perspective. They showed me that among
time a full 3 months after I learned the breast biopsy the collateral damage there is a certain collateral
was malignant. At our first meeting, she had to break beauty: although we may not be cured, we can be
the news that the surgical pathology indicated I had healed.
invasive Her2þ, hormone receptor negative, breast Through this process, I am discovering that it is OK
cancer. I would need 12 weeks of chemo and a full to feel vulnerable and afraid and to ask for help, even
year of trastuzumab (Herceptin). And I learned that when others affirm me for being strong. (With my last
Herceptin carried a risk of heart failure. name, I hear that one a lot!)
After experiencing valve disease and breast cancer When treatment ends, cancer is not won and done.
as direct results from radiation treatment, the risk of The psychosocial burden and impact on quality of life
heart failure felt all too real. The anxiety I had often goes unaddressed. I’m learning that it takes
managed to mostly avoid now had a strangle-hold. I courage to face and grieve our losses. I grieve that for
told my primary care physician, “I know that late ef- 7 of our 8 years together, my identity and my time
fects are here now. What I don’t know is whether I’m have been consumed by illness. Yet I am infinitely
standing at the edge of a glacier or in the direct path grateful to share present moments with Darryl as my
of an avalanche.” partner in life. I grieve also for the tens of thousands
A few months ago in 2019, I was granted disability of fellow survivors living with the physical and
retirement from teaching. To look at me, no one emotional burdens of their cures, whose needs are
would ever guess I have serious health issues. My not adequately addressed by our current system. My
breast oncologist once said, “Susan, you’re in pretty greatest hope is that within my most likely shortened
good shape for the shape that you’re in!” This week I lifetime, significant change takes place to meet these
have appointments with a cardiac electrophysiologist needs.
for follow-up on arrhythmias (including afib) from a This piece is dedicated to the z16,900,000 cancer
heart monitor test in October. Two days later, I’ll survivors, to include the z500,000 childhood cancer
drive back to see a gastroenterologist who will review survivors living in the United States today, to survi-
with me the multiple markedly abnormal results from vors around the globe, and to the memory of those
upper endoscopy, motility, and pH tests. Several whose lives were cut short by cancer and its ruthless
months ago, an ENT specialist diagnosed partial pa- collateral damage. Our deep gratitude goes to the
ralysis in my left vocal cord. The treatments that dedicated professionals whose life’s work has given
saved my life now seem to be slowly stealing it, and us decades of life that others before us never had the
on difficult days, I feel as if survivorship is a pro- chance to live.
gressive, terminal illness.
Cancer survivors with late effects, if they are
fortunate enough to have no barriers to care, see not ADDRESS FOR CORRESPONDENCE: Susan R. Strong,
only cardio-oncology specialists, but multiple others. MA, Heart Valve Voice US, 1700 K Street, NW, Suite
Keeping up with my health care is time-consuming, 740, Washington, DC 200006. E-mail: susan@
costly, and emotionally draining. I also regularly see susanstrong.com.

Cardiooncology

Uploaded by

Copyright:

Available Formats

Cardiooncology

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Cardiooncology

Uploaded by

Copyright:

Available Formats

CardioOncology

EDITOR-IN-CHIEF STATISTICAL EDITOR DIRECTOR, PRODUCT MANAGEMENT,

ASSOCIATE EDITORS GUEST EDITOR EDITORIAL CONSULTANTS

Paul W. Burridge, PhD, Carol Fabian, MD, L. Cheng, MD, PhD,

Carlo Gabriele Tocchetti, MD, PhD, P.S. Douglas, MD,

Richard J. Kovacs, MD, FACC,

Secretary and Board of Governors Chair

Chief Executive Ofﬁcer

2019-2020 PUBLICATIONS AND EDITORIAL COORDINATION COMMITTEE

Viviany R. Taqueti, MD, MPH, FACC, Chair

ª 2020 THE AUTHORS. PUBLISHED BY ELSEVIER ON BEHALF OF THE AMERICAN

COLLEGE OF CARDIOLOGY FOUNDATION. THIS IS AN OPEN ACCESS ARTICLE UNDER

THE CC BY-NC-ND LICENSE (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Cardiac Adverse Events in EGFR-Mutated

ISSN 2666-0873 https://doi.org/10.1016/j.jaccao.2020.02.003

Osimertinib-Associated Cardiac Adverse Events in NSCLC Patients MARCH 2020:1–10

O simertinib (Tagrisso, AstraZeneca)

geﬁtinib or erlotinib (3,5). With the success of these

medical records of all NSCLC patients with an EGFR

sensitizing, activating EGFR mutation; and treatment Adenocarcinoma 122 (99.2)

Within this cohort, there were 72 patients who 1st 33 (26.8)

Echocardiography/European Association of Cardio- Hyperuricemia 1 (0.8)

LVIDs) were obtained by 2-dimensional (2D) imaging,

Osimertinib-Associated Cardiac Adverse Events in NSCLC Patients MARCH 2020:1–10

T A B L E 2 Cases of Osimertinib-Induced Cardiac Adverse Events

Case 1 Case 2 Case 3 Case 4 Case 5 Case 6

F I G U R E 2 Histology Assessment of Myocardial Biopsy Specimens

Osimertinib-Associated Cardiac Adverse Events in NSCLC Patients MARCH 2020:1–10

QTc, and NT-proBNP and troponin I concentrations

(between 450 and 480 ms) in 18 patients. No episodes

T A B L E 4 Comparison of Echocardiographic Parameters Before and After Osimertinib Administration

NSCLC (N ¼ 36) Non-CTRCD (n ¼ 32) CTRCD (n ¼ 4)

Before After p Value* Before After p Value† Before After p Value‡

Osimertinib-Associated Cardiac Adverse Events in NSCLC Patients MARCH 2020:1–10

C E NT R AL IL L U STR AT IO N Osimertinib and Adverse Cardiac Events

Kunimasa, K. et al. J Am Coll Cardiol CardioOnc. 2020;2(1):1–10.

DISCUSSION higher CTCAE version 5.0 cardiac AEs. Five of the 6

Osimertinib-Associated Cardiac Adverse Events in NSCLC Patients MARCH 2020:1–10

In this retrospective cohort study, the incidence of

ª 2020 THE AUTHORS. PUBLISHED BY ELSEVIER ON BEHALF OF THE AMERICAN

COLLEGE OF CARDIOLOGY FOUNDATION. THIS IS AN OPEN ACCESS ARTICLE UNDER

THE CC BY-NC-ND LICENSE (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Targeting the Cardiotoxicity of

T argeted therapies have revolutionized the

ISSN 2666-0873 https://doi.org/10.1016/j.jaccao.2020.02.005

Targeting the Cardiotoxicity of EGFR Inhibitors MARCH 2020:11–2

ª 2020 THE AUTHORS. PUBLISHED BY ELSEVIER ON BEHALF OF THE AMERICAN

COLLEGE OF CARDIOLOGY FOUNDATION. THIS IS AN OPEN ACCESS ARTICLE UNDER

THE CC BY-NC-ND LICENSE (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Early Detection and Prediction

ISSN 2666-0873 https://doi.org/10.1016/j.jaccao.2020.01.007

RV Cardiotoxicity After Anthracycline Exposure MARCH 2020:13–22

A nthracyclines are widely used in reproducible (11,12). Therefore, a prospective cohort

LVEF = left ventricular ejection

Manuscript received August 28, 2019; accepted January 28, 2020.

RV Cardiotoxicity After Anthracycline Exposure MARCH 2020:13–22

and coronary artery disease. Four patients were

CHANGES IN 3DE PARAMETERS WITH ANTHRACYCLINE

and strain parameters are shown in Table 3. Compared T0 T1 T2 T3