Cardiomyopathies

Occult Cardiotoxicity in Childhood Cancer Survivors

Exposed to Anthracycline Therapy
Olga H. Toro-Salazar, MD; Eileen Gillan, MD; Michael T. O’Loughlin, MD;
Georgine S. Burke, PhD; Joanna Ferranti, MPH; Jeffrey Stainsby, MSc; Bruce Liang, MD;
Wojciech Mazur, MD; Subha V. Raman, MD, MSEE; Kan N. Hor, MD

Background—More than 50% of >270 000 childhood cancer survivors in the United States have been treated with
anthracyclines and are therefore at risk of developing cardiotoxicity. Cardiac magnetic resonance (CMR) has
demonstrated utility to detect diffuse interstitial fibrosis and changes in regional myocardial function. We hypothesized
that CMR would identify occult cardiotoxicity characterized by structural and functional myocardial abnormalities in a
cohort of asymptomatic pediatric cancer survivors with normal global systolic function.
Methods and Results—Forty-six long-term childhood cancer survivors with a cumulative anthracycline dose ≥200 mg/m2
and normal systolic function were studied 2.5 to 26.9 years after anthracycline exposure. Subjects underwent transthoracic
echocardiography, CMR with routine cine acquisition, tissue characterization, and left ventricular strain analysis using
a modified 16-segment model. Extracellular volume was measured in 27 subjects, all of whom were late gadolinium
enhancement negative. End-systolic fiber stress was elevated in 45 of 46 subjects. Low average circumferential
strain magnitude (εcc) −14.9±1.4; P<0.001, longitudinal strain magnitude (εll) −13.5±1.9; P<0.001, and regional peak
circumferential strain were seen in multiple myocardial segments, despite normal global systolic function by transthoracic
echocardiography and CMR. The mean T1 values of the myocardium were significantly lower than that of control subjects
at 20 minutes (458±69 versus 487±44 milliseconds; P=0.01). Higher mean extracellular volume was observed in female
subjects (0.34 versus 0.22; P=0.01).
Conclusions—Asymptomatic postchemotherapy pediatric patients have abnormal myocardial characteristics and strain
parameters by CMR despite normal global cardiac function by standard transthoracic echocardiography and CMR
measures.  (Circ Cardiovasc Imaging. 2013;6:873-880.)
Downloaded from http://ahajournals.org by on April 30, 2024

Key Words: anthracyclines ◼ cardiotoxicity ◼ magnetic resonance imaging ◼ myocardial strain

A nthracycline-induced cardiotoxicity (AIC) has an insidi-

ous onset with early features of dilated cardiomyopathy,
followed by a period of latency with clinical recovery of sys-
been damaged.7 Abnormalities in blood flow and tissue
Doppler indices of systolic and diastolic functions have
been proposed as early indicators of cardiotoxicity but lack
tolic function, before the development of overt heart failure.1,2 specificity to guide preventive therapy.8 Changes in regional
AIC has been attributed to a variety of factors, including the myocardial function can be present after anthracycline
generation of free oxygen radicals that induce myocyte mem- therapy in children and adults9–11 and in women undergoing
brane and DNA damage, as well as induction of apoptosis.3,4 breast cancer treatment.12
Previous studies have also demonstrated that AIC is mediated, Cardiac magnetic resonance (CMR) is considered the gold
in part, by an effect on the collagen matrix with loss of extra- standard for quantification of ventricular volumes and global
cellular matrix and increase in fibrosis.5,6 systolic function.13 Measures of myocardial strain by tagged
cine MRI are increasingly being used to detect subclini-
Clinical Perspective on p 880 cal myocardial dysfunction.14,15 CMR T1 mapping has been
Transthoracic echocardiographic (TTE) measurements of proposed as a noninvasive marker of interstitial myocardial
global systolic function, currently the standard of care to fibrosis,16–19 a fundamental change associated with early path-
diagnose AIC, allow the detection of the disease only when ological remodeling in the course of AIC.20 Measurement of
a significant percentage of cardiomyocytes have already the extracellular volume (ECV) fraction by CMR has been

Received May 7, 2013; accepted September 26, 2013.

From the Department of Pediatrics, Division of Pediatric Cardiology and Radiology (O.H.T.-S.), Department of Pediatrics, Division of Hematology-
Oncology (E.G.), and Department of Research (G.S.B., J.F.), Connecticut Children’s Medical Center, Hartford, CT; The Heart Center at Nationwide
Children’s Hospital and Ohio State University, Columbus, OH (K.N.H.); Hartford Hospital, Hartford, CT (M.T.O.); Applied Science Laboratory, GE
Healthcare, Toronto, ON, Canada (J.S.); Pat and Jim Calhoun Cardiology Center, University of Connecticut Health Center, Farmington, CT (B.T.L.); Ohio
Heart and Vascular Center, The Christ Hospital, Cincinnati (W.M.); and Ohio State University, Columbus, OH (S.V.R.).
Guest Editor for this article was David A. Bluemke, MD, PhD.
The online-only Data Supplement is available at http://circimaging.ahajournals.org/lookup/suppl/doi:10.1161/CIRCIMAGING.113.000798/-/DC1.
Correspondence to Olga H. Toro-Salazar, MD, Connecticut Children’s Medical Center, 282 Washington St, Hartford, CT 06106. E-mail [email protected]
© 2013 American Heart Association, Inc.
Circ Cardiovasc Imaging is available at http://circimaging.ahajournals.org DOI: 10.1161/CIRCIMAGING.113.000798

873
 874  Circ Cardiovasc Imaging  November 2013

correlated with the collagen volume fraction in human myo- CMR Techniques
cardium and has been validated as a measurement of extracel- Subjects were imaged on a 1.5-Tesla scanner (GE CV software
lular matrix volume expansion in several disease processes.17 version 16.0/M4, Milwaukee, WI) with the following protocol:
The purpose of this study was to use CMR T1 mapping tech- (1) standard multislice, multiphase cine imaging using a steady-
state free-precession acquisition technique (fast imaging employing
niques and measures of myocardial strain by tagged cine MRI steady-state acquisition [FIESTA]) in the 2-chamber, 4-chamber, and
to detect interstitial disease and changes in regional myocardial contiguous short-axis planes. FIESTA parameters include the follow-
function in childhood cancer survivors. We tested the hypothesis ing: echo time (TE), min full; pulse repetition time (TR), 3.4 mil-
that individuals with a history of high-dose anthracycline ther- liseconds; flip angle (FA), 45; bandwidth (BW), 125; i­mage matrix,
224×224 (VPS=20). (2) Tagged cine CMR acquired in the 4-cham-
apy may have occult cardiotoxicity, manifested by a decrease
ber and 3 short-axis planes (basal, midcavity, and apical) with an
in circumferential εcc and longitudinal εll strain magnitude and ECG-triggered segmented k-space fast gradient echo sequence with
changes in myocardial T1 and ECV, despite normal standard spatial modulation of magnetization in orthogonal planes. Grid tag
measures of global left ventricular (LV) systolic function. spacing, 7 to 8 mm; field of view, (30–32)×(25–26) cm2; slice thick-
ness, 7 to 8 mm; flip angle, 12°; TE/TR, 3/4.2 milliseconds; views/
segment, 6. (3) T1 mapping using modified Look-Locker with satura-
Methods tion ­recovery sequence (MLLSR)23: scan parameters: 2-2-4 pattern;
time to inversion (TI)0, 100 milliseconds; TIinc, 100 milliseconds+RR
Study Population interval resulting in TIs of 100, 1100; 200,1200; 300,1300; and 2300,
Eligible subjects with a cumulative anthracycline dose ≥200 mg/m2 3300 milliseconds; echo time/pulse repetition time, 1.7/3.9 milli-
and normal LV systolic function defined as TTE-based shortening seconds; flip angle, 45°; 256×160 matrix; phase field of view, 0.75;
fraction ≥29% were identified and prospectively enrolled through a slice thickness, 8 mm; number of excitations, 0.5; 38 views/segment
registry of pediatric cancer survivors treated with anthracyclines be- (148-millisecond readout window). Gadopentetate dimeglumine
tween 1985 and January 2011. Inclusion criteria for this study are (Gd-DTPA; Magnevist; Schering, Berlin, Germany) was adminis-
shown in Figure 1. Subjects with high-dose radiation exposure to the tered at a dose of 0.1 mmol per 1 kg body weight through a 20-gauge
chest (>3000 cGy) were excluded from the study to minimize the cannula placed in the antecubital vein. MLLSR was performed for all
known synergistic effect of therapeutic radiation on cardiotoxicity.5,21 3 short-axis levels (midcavity, basal, and apical) before and at 4, 10,
This study was approved by the Institutional Review Board at the and 20 minutes after the contrast injection (Figure 2). (4) Late gado-
Connecticut Children’s Medical Center. The medical records of all en- linium enhancement was acquired 10 to 15 minutes after intravenous
rolled subjects were reviewed to identify known risk factors associated injection of contrast for assessment of signal intensity changes in the
with cardiotoxicity. Conversions to isotoxic equivalents of anthracy- 4-chamber and contiguous short-axis planes. TI was calculated to
cline antibiotics were performed to calculate the total ­cumulative dose. null the myocardium (TI, 150–250 milliseconds; TR, 5.5 millisec-
onds; TE, min full; 6-mm slice thickness; no spacing; flip angle, 20°;
Echocardiography bandwidth, 31; matrix, 256×128). The delayed-­enhancement images
were independently reviewed by 2 cardiologists (O.T.S., K.N.H.) in
Downloaded from http://ahajournals.org by on April 30, 2024

Echocardiographic assessments were obtained using a Philips iE33

a blinded fashion. Hematocrit was measured at the time of intrave-
echocardiography system. A TTE was performed for evaluation of
nous placement for CMR examination. The CMR and TTE were per-
systolic and diastolic functions, including 2-dimensional imaging,
formed on the same day.
M mode, spectral, and tissue Doppler parameters of diastolic func-
tion; measurements of ventricular afterload (end-systolic wall stress);
and contractility (stress/velocity index), using American Society of Image Analysis
Echocardiography guidelines22 (echocardiographic image analysis is Ventricular volumes, ventricular mass, and global function were
given in the Appendix in the online-only Data Supplement). obtained via standard planimetry techniques with semiautomated

Figure 1. Identification of eligible patients for

study group.
 Toro-Salazar et al   Novel CMR Parameters and Anthracycline Cardiotoxicity   875

coefficients were used to assess the relationship between 2 continuous

variables. All tests are 2 tailed, and P<0.05 is considered to indicate
a statistically significant difference. SPSS software (IBM, Armonk,
NY) was used for statistical analysis.

Results
Study Population
Five hundred twenty-one patients were identified through
the cancer registry, of whom 428 patients had been off che-
motherapy for ≥2 years (Figure 1). Of these, 193 patients
had received a cumulative anthracycline dose ≥200 mg/m2.
Twenty-eight patients died, 50 patients were lost to follow-
up, and 43 patients did not meet the inclusion criteria. Car-
diomyopathy contributed to 6 of the 28 deaths (Table I in the
online-only Data Supplement). Of the remaining 72 eligible
patients, 46 patients agreed to participate in the study. The
most frequent type of cancer diagnosis in these subjects was
acute myelogenous leukemia (21.7%), osteosarcoma (13%),
Hodgkin lymphoma (10.9%), and Ewing sarcoma (10.9%;
Figure 2. Modified look-locker with saturation recovery sequence Table II in the online-only Data Supplement). Patient charac-
(MLLSR pulse sequence). The MLLSR pulse sequence uses a teristics are summarized in Table 1. There were no statistically
flexible number (typically 3) Look-Locker (LL) imaging blocks,
significant differences in patient characteristics between the
consisting of 2, 2, and 4 heartbeats each. To sample T1 recovery,
8 serial single-shot diastolic images are acquired every heartbeat study participants and the 26 subjects who were not included
after a non–slice-selective saturation pulse. For generating the (data not shown). Ten of the 46 subjects enrolled had a prior
relaxation curve, the signal intensity for each pixel at each time to history of systolic dysfunction (indicated by a shortening
inversion (TI) is fit to a 3-parameter model for longitudinal relax-
ation: S(t)=A×exp(−TI/T1)+B, where A, B, and T1* are the fitting fraction <29%), which had normalized before participating
parameters. The equation describes the relaxation curve in terms in this study. None of these subjects received previous cardiac
of T1*, often referred to as the apparent T1, rather than the true medications, and all are currently asymptomatic.
tissue T1. SR indicates saturation recovery.
Downloaded from http://ahajournals.org by on April 30, 2024

Table 1. Demographic and Anthropometric Characteristics of

computer software (QMASS version 6.1.5; Medis Medical Imaging Study Subjects*
Systems, Leiden, the Netherlands) by an experienced reader (P.I.).
Ventricular analysis calculations including ejection fraction, stroke n=46
volume, end-diastolic volume, LV mass-volume, and end-systolic fi- Years postchemotherapy 9.6 (2.5–26.9)
ber stress (ESFS) were performed. ESFS was calculated using the
following formula: ESFS=1.35×P×bm×2h (expressed in g/cm2), Mean age at diagnosis, y 11 (±5.1)
where P is the end-systolic pressure,24 D is the end-systolic LV in- Mean current age, y 22 (12–42)
ternal diameter (in cm), h is the end-systolic posterior wall thickness, Male/female, n 33/13
and bm is the midwall minor semiaxis expressed as h/ln(0.5D+h)–
ln(0.5D).25 To assess the interobserver variability of volumetric analy- Total cumulative dose anthracyclines, mg/m2 328 (200–600)
ses, a second expert (K.H.) performed the same analysis on subsets Radiation to chest, n (%)† 4 (8.7)
of 10 subjects. Tagged images were analyzed with the HARmonic
Vinca alkaloids, n (%) 30 (65.2)
Phase (Diagnosoft, Palo Alto, CA) technique26 for calculation of peak
average circumferential and longitudinal strain magnitude. Regional Previous SF <29%, n (%) 11 (23.9)
changes in peak circumferential strain (εcc) magnitude were per- Previous bone marrow transplantation, n (%) 4 (8.7)
formed in the midventricular level only given the limited reproduc-
Previous heart disease, n (%) 3 (6.5)
ibility of the basal and apical slice. Midwall strain data were exported
to a spreadsheet file for offline analysis. To assess intraobserver and Height, cm 170.7 (±9.7)
interobserver variability of HARmonic Phase strain analyses, a sec- Weight, kg 74.8 (±16.7)
ond expert reader (O.T.S., K.N.H.) performed the same analysis in
a subset of 10 randomly selected subjects in a blinded fashion. Epi- BSA, m2 1.9 (±0.23)
and endocardial contours were traced for T1 analysis using custom BMI, kg/m 2
26 (20–45)
software (Vizpack, GE Healthcare, Waukesha, WI).23 ECV values Heart rate, bpm 67.5 (±12.8)
were calculated as previously reported27 using the following formu-
las: ΔR1myo=1/T1myo-post–1/T1myo-pre; ΔR1blood=1/T1blood-post–1/T1blood-pre; Systolic blood pressure, mm Hg 119 (98–148)
and ECV=ΔR1myo/ΔR1blood×(100–HCT), where ECV and Hematocrit Diastolic blood pressure, mm Hg 67 (46–86)
(HCT) are given as percentages.
NYHA functional class, n (%)
 I 39 (84.8)
Statistical Analysis
 II 4 (8.7)
Subject values were combined in a single study database. t Tests were
used to assess differences between groups (subjects versus control BMI indicates body mass index; BSA, body surface area; NYHA, New York
subjects, TTE versus CMR, men versus women). The primary com- Heart Association; and SF, shortening fraction.
parisons of interest were between subjects and control subjects, as *Shown as frequency (percent), mean (±SD), or mean (range) as appropriate.
well as between the 2 imaging methods: CMR and TTE. Correlation †All radiation to the chest is low dose.
 876  Circ Cardiovasc Imaging  November 2013

Echocardiographic Parameters Table 2. CMR Parameters

All 46 subjects had normal standard echocardiographic CMR Parameters (n=46) Mean±SD P Value Z Score
parameters of global systolic function, including stress-
velocity index and stress-shortening index. Myocardial per- LVEF, %
formance index was elevated in 44 (96%) subjects (Table III  Male (n=33) 55.9±3.8 0.648 −1.8±0.83
in the online-only Data Supplement). One subject (age, 42  Female (n=13) 56.5±5.5 −1.6±1.1
years) was diagnosed with diastolic dysfunction on the basis LVEDV, mL/m2
of American Society of Echocardiography guidelines.28 Indi-  Male (n=33) 84.9±15.4 0.001 0.19±1.1
vidual parameters of diastolic function are presented in Table  Female (n=13) 69.3±8.3 −0.77±0.76
IV in the online-only Data Supplement.
LVESV, mL
 Male (n=33) 72.9±18.9 0.001
Basic MRI Parameters
Global left and right ventricular systolic function by standard  Female (n=13) 52.4±11.4
MRI parameters was normal in all subjects (Table 2). LV mass LV SV, mL
index was normal in both men (65.6±10.9; Z score, 0.27±1.4)  Male (n=33) 91.0±17.6 <0.0001
and women (49.8±5.9; Z score, −0.37±0.76). Men had larger  Female (n=13) 67.1±9.9
right ventricular and LV end-diastolic and end-systolic vol- LV mass, g/m2
umes and a higher LV mass than women. ESFS was elevated  Male (n=33) 65.6±10.9 <0.0001 0.27±1.4
in 45 of 46 subjects. A low mass-volume ratio was inversely
 Female (n=13) 49.8±5.9 −0.37±0.76
correlated with an increase in ESFS (r=−0.397; P=0.006).
LV mass volume, g/mL
Late gadolinium enhancement was not present in any of the
study subjects, with 100% interobserver agreement.  Male (n=33) 0.81±0.22 0.137 0.02±1.4
 Female (n=13) 0.72±0.08 0.45±0.81
Regional Myocardial Function RV EDV, mL/m2
Compared with normal subjects, 45 of 46 patients studied  Male (n=33) 84.5±16.0 <0.0001 −0.04±1.2
showed lower average midpeak circumferential strain mag-  Female (n=13) 67.9±7.3 −0.55±0.55
nitude (εcc) −14.9±1.4 versus −19.5±2.1 (P<0.001) and peak RV ESV, mL
longitudinal strain magnitude (ειι) −13.5±1.9 versus 17.3±1.4
 Male (n=33) 73.6±22.5 <0.0001
(P<0.001) by CMR, despite having normal values of global
Downloaded from http://ahajournals.org by on April 30, 2024

 Female (n=13) 50.4±9.9

systolic function by TTE and CMR (Figure 3). Furthermore,
segmental εcc magnitude was significantly lower in all seg- RV SV, mL
ments, but in particular, the septal and inferior segments were  Male (n=33) 89.9±16.7 <0.0001
most affected (Table 3). Intraobserver (K.N.H.; mean differ-  Female (n=13) 65.8±9.4
ence=−0.25; limits of agreement=−3.05 to 3.55) and interob- RV EF, %
server (O.T.S. and K.N.H.; mean difference=−0.45; limits of  Male (n=33) 55.7±4.4 0.506 0.03±1.3
agreement=−2.61 to 1.71) variability of global peak circum-  Female (n=13) 56.7±4.9 −0.61±0.99
ferential strain was low on the basis of separate quantitative
End-systolic fiber stress, g/cm2
analysis in a subset of 9 subjects (Figure I in the online-only
Data Supplement). Average circumferential (P=0.005) and  Male (n=33) 125.6±17.1 0.192 4.5±1.5
longitudinal (P<0.001) strain magnitude by CMR was lower  Female (n=13) 133.9±23.0 5.3±2.0
among male subjects. Lower average εcc magnitude corre- CMR indicates cardiac magnetic resonance; LVEDV, left ventricular end-
lated with higher cumulative dose (P=0.049), lower ejection diastolic volume; LVEF, left ventricular ejection fraction; LVESV, left ventricular
fraction (r=−0.389; P=0.008), radiation therapy to the chest end-systolic volume; LV SV, left ventricular stroke volume; RV EDV, right ven­
tricular end-diastolic volume; RV EF, right ventricular ejection fraction; RV ESV,
(P=0.007), and increased myocardial performance index
right ventricular end-systolic volume; and RV SV, right ventricular stroke volume.
(r=0.309; P=0.037). A significant reduction in peak longitu-
dinal strain magnitude was seen by CMR (−13.5±1.9) but not
by echocardiography (−18.5±2.4; P<0.001). Compared with parameters, or regional changes in peak circumferential (εcc) or
CMR, speckle-tracking techniques by TTE demonstrated sig- longitudinal strain magnitude (ειι). Postgadolinium T1 values
nificant reductions in peak circumferential strain (εcc) limited were 373±60 versus 389±36 milliseconds at 4 minutes (P=0.12),
to the anteroseptal and inferoseptal segments (Figure 3). 442±129 versus 435±36 milliseconds at 10 minutes (P=0.75),
and 458±69 versus 487±44 milliseconds at 20 minutes (P=0.01).
T1 Mapping of the Myocardium Low T1 relaxation myocardial values at 20 minutes correlated
Segmental T1 maps at baseline and at 4, 10, and 20 minutes with low LV mass index (r=0.521; P<0.001), increase in ESFS
are depicted in Figure 4. Normative T1 data on 24 volunteers (r=−0.501; P<0.001), and female sex (P=0.01).
(46.6±11 years; 15 men) were used for comparison.23 Precon-
trast T1 values of the myocardium were higher in study subjects, Quality Scores
but the difference did not reach statistical significance (974±312 The T1 maps were scored for overall quality based on a
versus 910±93 milliseconds; P=0.26). Precontrast T1 values 3-point scale in all subjects, similar to previous reports.16 The
did not correlate with risk factors for AIC, global function MRI 2 most common sources of artifact observed were related to
 Toro-Salazar et al   Novel CMR Parameters and Anthracycline Cardiotoxicity   877

Figure 3. Strain by echocardiography compared with strain by MRI. A, Strain by echocardiograpy in subjects compared with control sub-
jects. B, Strain by MRI in subjects compared with control subjects. C, Strain by echocardiography compared with strain by MRI in sub-
jects. eAnt indicates anterior strain; eAS, anteroseptal strain; ecc, global circumferential strain; eIL, inferolateral strain; eIS, inferior strain;
eLat, lateral strain; ell, longitudinal; and eMS, inferoseptal strain.

physiological motion and off-resonance effects. Only images tagging techniques are able to detect significant decreases in
with quality scores of 1 and 2 were chosen for analysis. global and segmental LV myocardial peak circumferential
Image quality was given a grade of 1 (good) for 207 (40.6%) (εcc) and longitudinal (ειι) strain magnitude in subjects with
of the myocardial slices imaged, a grade of 2 (satisfactory) normal indices of global systolic function. Furthermore, T1
Downloaded from http://ahajournals.org by on April 30, 2024

for 224 (43.9%) of the slices, and a grade of 3 (nonevalu- techniques and ECV calculations quantify diffuse interstitial
able) for 79 (15.5%) of the slices (Table V in the online-only myocardial changes and expansion in the extracellular matrix
Data Supplement). in these patients.
Changes in global systolic function have been reported by
Extracellular Volume CMR in a subset of long-term survivors of childhood cancer.30
ECV values were calculated in 27 of 46 subjects. ECV values
were normal with an average of 0.25±0.1, similar to previ- Table 3. Global and Regional Myocardial Function by CMR
ous studies.27 Abnormally elevated ECV was identified in 5
of 27 subjects (0.38±0.07). Higher mean ECV was observed Control
Subjects Anthracycline
in female subjects (0.34 versus 0.22; P=0.01; Figure 5) and in
(n=18) (n=46)
association with increase in ESFS (r=0.369; P=0.049). ECV
was not associated with diastolic parameters or global and Age, y 20±8.6 22±6.0
regional myocardial systolic function by TTE or CMR. Height, cm 164.3±8.1 170.7±9.7
Weight, kg 67.1±19.9 74.8±16.7
Discussion Left ventricular ejection fraction, % 66.3±5.6 56.1±4.3**
Delayed anthracycline-related cardiotoxicity is manifested Left ventricular end-systolic fiber 74±1.0 128±19.0**
by chronic dilated cardiomyopathy that develops late in the stress, g/cm2
course of therapy.1 Cardiotoxicity is dose related, and myo- Left ventricular mass, g/m2 50.3±12.6 61.1±12.0**
cyte damage on biopsy has a linear relationship with cumu- Global (εcc), % −19.5±2.1 −14.9±1.4**
lative dose, whereas myocardial function is preserved until Longitudinal (εll), % −17.3±1.4 −13.5±1.5**
a critical dose or degree of damage is reached.29 Changes Anterior (εcc), % −18.8±3.2 16.0±2.4**
identified with light microscopy include an increase in inter-
Anteroseptal (εcc), % −18.2±2.6 −13.9±2.7**
stitial fibrous tissue and 2 main types of myocyte injury: par-
Inferoseptal (εcc), % −18.1±3.4 −13.9±2.1**
tial or complete myofibrillar loss and vacuolar degeneration
caused by dilatation of elements of the sarcoplasmic reticu- Inferior (εcc), % −20.7±2.5 −14.7±2.4**
lum.20,29 These lesions eventually progress until the death of Inferolateral (εcc), % −19.2±3.4 −16.9±3.1**
the myocyte. Echocardiography, the current standard of care, Lateral (εcc), % −19.7±3.3 −15.8±2.5**
does not detect myocardial damage until late in the disease. CMR indicates cardiac magnetic resonance.
In this study, we demonstrate for the first time that CMR *P<0.05; **P<0.01.
 878  Circ Cardiovasc Imaging  November 2013

Figure 4. Segmental T1 maps

for study sample and in a rep-
resentative subject. 1 indicates
basal anterior; 2, basal antero-
septal; 3, basal inferoseptal; 4,
basal inferior; 5, basal infero-
lateral; 6, basal anterolateral;
7, midanterior; 8, midantero-
septal; 9, midinferoseptal; 10,
midinferior; 11, midinfero-
lateral; 12, midanterolateral;
13, apical anterior; 14, apical
septal; 15, apical inferior; 16,
apical lateral; and 17, apex.

In contrast, in this study, we focused on subjects with normal Postcontrast myocardial longitudinal relaxation time (T1)
indices of global systolic function. Similar to previous investi- mapping constitutes a novel technique for quantifying intersti-
gations, adverse cardiac remodeling with inadequate compen- tial myocardial fibrosis,16,36 a fundamental change associated
satory hypertrophy is present, resulting in increased afterload, with early pathological remodeling in the course of AIC.5 An
as indicated by an increase in ESFS, preceding contractile increasing number of studies have identified diffuse myocar-
deficits in this patient population.1,31 dial fibrosis in several cardiac disease states36–38 with the use of
Measures of regional function have emerged as accurate various protocols for T1 mapping16,23,36 and in pediatric patients
tools for the detection of subclinical myocardial dysfunction.32 exposed to anthracyclines.39 In this study, we demonstrated
The presence of myocardial strain abnormalities, as well as lower myocardial T1 values 20 minutes after the injection of
Downloaded from http://ahajournals.org by on April 30, 2024

a serial decline in myocardial strain magnitude without a gadolinium. Lower postcontrast T1 values at 20 minutes corre-
change in ejection fraction, has been previously reported in lated with low LV mass, an increase in ESFS, and higher cumu-
other forms of nonischemic cardiomyopathy.14,33–35 Our data lative anthracycline dose. Tham et al39 recently demonstrated
provide evidence of similar findings in AIC subjects with similar associations with precontrast T1 values in this patient
reductions in regional performance in these subjects as dem- population. All subjects in our study were negative for late gad-
onstrated by low peak longitudinal (ειι) and circumferential olinium enhancement, confirming a low prevalence of replace-
strain (εcc) magnitude in all segments of the myocardium. In ment fibrosis in subjects with preserved ejection fraction.39,40
contrast to the widespread segmental involvement demon- Measurement of myocardial ECV has been shown to corre-
strated by CMR, speckle-tracking techniques by TTE demon- late with collagen volume fraction.17,27 Previous studies show
strated reductions limited to the anteroseptal and inferoseptal that myocardial ECV quantification correlates with diffuse
segments. This difference likely reflects the enhanced sensi- myocardial fibrosis17 and is elevated in anthracycline-treated
tivity of CMR for the detection of subtle changes in intramural adult subjects.39,41 Average ECV value in our patient popu-
contractile function, as previously reported.33 lation was similar to previous studies.27 In agreement with

Figure 5. Sex differences in extracellular volume.

1 indicates basal anterior; 2, basal anteroseptal; 3,
basal inferoseptal; 4, basal inferior; 5, basal infero-
lateral; 6, basal anterolateral; 7, midanterior; 8,
midanteroseptal; 9, midinferoseptal; 10, midinferior;
11, midinferolateral; 12, midanterolateral; 13, apical
anterior; 14, apical septal; 15, apical inferior; 16,
apical lateral; and 17, apex.
 Toro-Salazar et al   Novel CMR Parameters and Anthracycline Cardiotoxicity   879

previous reports,39,41 higher mean ECV and lower postcontrast References

myocardial T1 values were observed in female subjects, sug- 1. Lipshultz SE, Colan SD, Gelber RD, Perez-Atayde AR, Sallan SE,
gesting sex differences in extracellular matrix remodeling. No Sanders SP. Late cardiac effects of doxorubicin therapy for acute lympho-
blastic leukemia in childhood. N Engl J Med. 1991;324:808–815.
association was seen with measurements of diastolic function 2. Ferrans VJ. Overview of cardiac pathology in relation to anthracycline
by echocardiography and ECV in our study population. This cardiotoxicity. Cancer Treat Rep. 1978;62:955–961.
is most likely related to younger age in our study population 3. Myers CE, McGuire WP, Liss RH, Ifrim I, Grotzinger K, Young RC.
(22±6 years) compared with the older cohort (55±17 years) in Adriamycin: the role of lipid peroxidation in cardiac toxicity and tumor
response. Science. 1977;197:165–167.
the study by Neilan et al41 and the relatively low frequency of 4. De Angelis A, Piegari E, Cappetta D, Marino L, Filippelli A, Berrino L,
abnormal diastolic parameters. Ferreira-Martins J, Zheng H, Hosoda T, Rota M, Urbanek K, Kajstura J,
Leri A, Rossi F, Anversa P. Anthracycline cardiomyopathy is mediated by
Study Limitations depletion of the cardiac stem cell pool and is rescued by restoration of
progenitor cell function. Circulation. 2010;121:276–292.
Our study has limitations resulting in part from its cross-sectional 5. Caulfield JB, Bittner V. Cardiac matrix alterations induced by adriamycin.
design. Further longitudinal evaluation in individual patients is Am J Pathol. 1988;133:298–305.
required to determine whether early markers of cardiotoxicity 6. Bernaba BN, Chan JB, Lai CK, Fishbein MC. Pathology of late-onset an-
serve as predictors of eventual heart failure in this patient popu- thracycline cardiomyopathy. Cardiovasc Pathol. 2010;19:308–311.
7. Ewer MS, Ali MK, Mackay B, Wallace S, Valdivieso M, Legha SS, Benjamin
lation. Postcontrast T1 relaxation times may be affected by con- RS, Haynie TP. A comparison of cardiac biopsy grades and ejection fraction
trast type and dose of administration, in addition to the recently estimations in patients receiving Adriamycin. J Clin Oncol. 1984;2:112–117.
recognized sex and age dependency of T1 indices and ECV.17 T1 8. Stapleton GE, Stapleton SL, Martinez A, Ayres NA, Kovalchin JP, Bezold
LI, Pignatelli R, Eidem BW. Evaluation of longitudinal ventricular func-
normative data on MLLSR were derived in patients 46±11 years tion with tissue Doppler echocardiography in children treated with anthra-
of age. Given the younger age of our study population, com- cyclines. J Am Soc Echocardiogr. 2007;20:492–497.
parison of T1 values and ECV based on normative data from a 9. Ganame J, Claus P, Eyskens B, Uyttebroeck A, Renard M, D’hooge J,
relatively older cohort may lead to underestimation of the degree Gewillig M, Bijnens B, Sutherland GR, Mertens L. Acute cardiac func-
tional and morphological changes after anthracycline infusions in chil-
of interstitial fibrosis, given that normal younger individuals dren. Am J Cardiol. 2007;99:974–977.
would be expected to have lower precontrast T1 and higher T1 10. Poterucha JT, Kutty S, Lindquist RK, Li L, Eidem BW. Changes in left
postmyocardial values and lower ECV. Finally, in this study, we ventricular longitudinal strain with anthracycline chemotherapy in adoles-
cents precede subsequent decreased left ventricular ejection fraction. J Am
did not perform end myocardial biopsy for further validation of Soc Echocardiogr. 2012;25:733–740.
ECV and collagen volume fraction of the myocardium. 11. Cheung YF, Hong WJ, Chan GC, Wong SJ, Ha SY. Left ventricular myo-
cardial deformation and mechanical dyssynchrony in children with nor-
Conclusions mal ventricular shortening fraction after anthracycline therapy. Heart.
2010;96:1137–1141.
Downloaded from http://ahajournals.org by on April 30, 2024

This work demonstrates for the first time abnormal myocardial

12. Sawaya H, Sebag IA, Plana JC, Januzzi JL, Ky B, Tan TC, Cohen V,
characteristics and abnormal regional myocardial function in Banchs J, Carver JR, Wiegers SE, Martin RP, Picard MH, Gerszten RE,
a cohort of pediatric cancer survivors treated with high-dose Halpern EF, Passeri J, Kuter I, Scherrer-Crosbie M. Assessment of echo-
anthracycline chemotherapy in the face of normal values for cardiography and biomarkers for the extended prediction of cardiotoxic-
ity in patients treated with anthracyclines, taxanes, and trastuzumab. Circ
clinically routine measures of LV systolic function. Micro- Cardiovasc Imaging. 2012;5:596–603.
scopic fibrosis, as indicated by changes in T1 mapping–derived 13. Grothues F, Smith GC, Moon JC, Bellenger NG, Collins P, Klein HU,
relaxation time, ECV, and decreased LV myocardial peak cir- Pennell DJ. Comparison of interstudy reproducibility of cardiovascular
cumferential and longitudinal strain magnitude, is present in magnetic resonance with two-dimensional echocardiography in normal
subjects and in patients with heart failure or left ventricular hypertrophy.
asymptomatic postchemotherapy pediatric patients who have Am J Cardiol. 2002;90:29–34.
normal standard CMR parameters. Thus, LV myocardial peak 14. Hor KN, Wansapura J, Markham LW, Mazur W, Cripe LH, Fleck R,
circumferential strain, T1 mapping–derived myocardial relax- Benson DW, Gottliebson WM. Circumferential strain analysis identifies
strata of cardiomyopathy in Duchenne muscular dystrophy: a cardiac mag-
ation time, and ECV may serve as novel CMR parameters in the netic resonance tagging study. J Am Coll Cardiol. 2009;53:1204–1210.
evaluation of therapies aimed at reducing adverse cardiac out- 15. Yoneyama K, Gjesdal O, Choi EY, Wu CO, Hundley WG, Gomes AS, Liu
come in childhood cancer survivors exposed to anthracyclines. CY, McClelland RL, Bluemke DA, Lima JA. Age, sex, and hypertension-
related remodeling influences left ventricular torsion assessed by tagged
cardiac magnetic resonance in asymptomatic individuals: the Multi-Ethnic
Acknowledgments Study of Atherosclerosis. Circulation. 2012;126:2481–2490.
We acknowledge Sam Gizelis, RT, for postprocessing of T1 maps; 16. Messroghli DR, Radjenovic A, Kozerke S, Higgins DM, Sivananthan MU,
Steve Walling, RDCS, for advanced echocardiography postprocess- Ridgway JP. Modified Look-Locker inversion recovery (MOLLI) for high-
ing; Natalya Lim, MD, for medical record review, data abstraction, resolution T1 mapping of the heart. Magn Reson Med. 2004;52:141–146.
and data entry; Nancy Cronin, RN, MSN, for assisting in subject re- 17. Miller CA, Naish JH, Bishop P, Coutts G, Clark D, Zhao S, Ray SG,
cruitment and facilitating each subject’s study visit; and Amy Tipton, Yonan N, Williams SG, Flett AS, Moon JC, Greiser A, Parker GJ, Schmitt
BFA, for organizing the strain images. We also thank the echocardiog- M. Comprehensive validation of cardiovascular magnetic resonance
raphy lab at Connecticut Children’s Medical Center and the MRI lab at techniques for the assessment of myocardial extracellular volume. Circ
Hartford Hospital, as well as the echocardiography laboratory at Boston Cardiovasc Imaging. 2013;6:373–383.
Children’s Hospital for access to their database and normative echo data. 18. Sibley CT, Noureldin RA, Gai N, Nacif MS, Liu S, Turkbey EB, Mudd JO,
van der Geest RJ, Lima JA, Halushka MK, Bluemke DA. T1 Mapping in
cardiomyopathy at cardiac MR: comparison with endomyocardial biopsy.
Sources of Funding Radiology. 2012;265:724–732.
This study was supported in part by the St. Baldrick’s Foundation 19. Wouters KA, Kremer LC, Miller TL, Herman EH, Lipshultz SE. Protecting
Supportive Care Research grant to Dr Toro-Salazar. against anthracycline-induced myocardial damage: a review of the most
promising strategies. Br J Haematol. 2005;131:561–578.
20. Billingham ME, Mason JW, Bristow MR, Daniels JR. Anthracycline
Disclosures cardiomyopathy monitored by morphologic changes. Cancer Treat Rep.
None. 1978;62:865–872.
 880  Circ Cardiovasc Imaging  November 2013

21. Emami B, Lyman J, Brown A, Coia L, Goitein M, Munzenrider JE, Shank dystrophin-deficient children revealed by cardiac magnetic resonance
B, Solin LJ, Wesson M. Tolerance of normal tissue to therapeutic irradia- strain imaging. Circulation. 2005;112:2462–2467.
tion. Int J Radiat Oncol Biol Phys. 1991;21:109–122. 34. Carasso S, Yang H, Woo A, Vannan MA, Jamorski M, Wigle ED, Rakowski
22. López B, González A, Díez J. Circulating biomarkers of collagen metabo- H. Systolic myocardial mechanics in hypertrophic cardiomyopathy: novel
lism in cardiac diseases. Circulation. 2010;121:1645–1654. concepts and implications for clinical status. J Am Soc Echocardiogr.
23. Song T, Stainsby JA, Ho VB, Hood MN, Slavin GS. Flexible cardiac T1 2008;21:675–683.
mapping using a modified Look-Locker acquisition with saturation recov- 35. Dogan V, Öcal B, Orun UA, Ozgur S, Yılmaz O, Keskin M, Ceylan O,
ery. Magn Reson Med. 2012;67:622–627. Karademir S, Şenocak F. Strain and strain rate echocardiography findings
24. Iwahara M, Benson LN, Freedom RM. Noninvasive estimation of end- in children with asymptomatic congenital aortic stenosis. Pediatr Cardiol.
systolic aortic pressure in children. Echocardiography. 1991;8:559–562. 2013;34:1152–1158.
25. Regen DM. Calculation of left ventricular wall stress. Circ Res. 1990;67: 36. Iles L, Pfluger H, Phrommintikul A, Cherayath J, Aksit P, Gupta SN, Kaye
245–252. DM, Taylor AJ. Evaluation of diffuse myocardial fibrosis in heart failure
26. Osman NF, Kerwin WS, McVeigh ER, Prince JL. Cardiac motion tracking with cardiac magnetic resonance contrast-enhanced T1 mapping. J Am
using CINE harmonic phase (HARP) magnetic resonance imaging. Magn Coll Cardiol. 2008;52:1574–1580.
Reson Med. 1999;42:1048–1060. 37. Ng AC, Auger D, Delgado V, van Elderen SG, Bertini M, Siebelink HM,
27. Kellman P, Wilson JR, Xue H, Bandettini WP, Shanbhag SM, Druey KM, van der Geest RJ, Bonetti C, van der Velde ET, de Roos A, Smit JW, Leung
Ugander M, Arai AE. Extracellular volume fraction mapping in the myocardi- DY, Bax JJ, Lamb HJ. Association between diffuse myocardial fibrosis by
um, part 2: initial clinical experience. J Cardiovasc Magn Reson. 2012;14:64. cardiac magnetic resonance contrast-enhanced T1 mapping and subclinical
28. Nagueh SF, Appleton CP, Gillebert TC, Marino PN, Oh JK, Smiseth myocardial dysfunction in diabetic patients: a pilot study. Circ Cardiovasc
OA, Waggoner AD, Flachskampf FA, Pellikka PA, Evangelisa A. Imaging. 2012;5:51–59.
Recommendations for the evaluation of left ventricular diastolic function 38. Amano Y, Takayama M, Kumita S. Contrast-enhanced myocardial T1-
by echocardiography. Eur J Echocardiogr. 2009;10:165–193. weighted scout (Look-Locker) imaging for the detection of myocar-
29. Bristow MR, Mason JW, Billingham ME, Daniels JR. Dose-effect and struc- dial damages in hypertrophic cardiomyopathy. J Magn Reson Imaging.
ture-function relationships in doxorubicin cardiomyopathy. Am Heart J. 2009;30:778–784.
1981;102:709–718. 39. Tham EB, Haykowsky MJ, Chow K, Spavor M, Kaneko S, Khoo NS,
30. Ylänen K, Poutanen T, Savikurki-Heikkilä P, Rinta-Kiikka I, Eerola A, Pagano JJ, Mackie AS, Thompson RB. Diffuse myocardial fibrosis by
Vettenranta K. Cardiac magnetic resonance imaging in the evaluation of T1-mapping in children with subclinical anthracycline cardiotoxic-
the late effects of anthracyclines among long-term survivors of childhood ity: relationship to exercise capacity, cumulative dose and remodeling.
cancer. J Am Coll Cardiol. 2013;61:1539–1547. J Cardiovasc Magn Reson. 2013;15:48.
31. Lipshultz SE, Lipsitz SR, Sallan SE, Dalton VM, Mone SM, Gelber RD, 40. Neilan TG, Coelho-Filho OR, Pena-Herrera D, Shah RV, Jerosch-Herold
Colan SD. Chronic progressive cardiac dysfunction years after doxoru- M, Francis SA, Moslehi J, Kwong RY. Left ventricular mass in patients
bicin therapy for childhood acute lymphoblastic leukemia. J Clin Oncol. with a cardiomyopathy after treatment with anthracyclines. Am J Cardiol.
2005;23:2629–2636. 2012;110:1679–1686.
32. Shehata ML, Cheng S, Osman NF, Bluemke DA, Lima JA. Myocardial tis- 41. Neilan TG, Coelho-Filho OR, Shah RV, Feng JH, Pena-Herrera D,
sue tagging with cardiovascular magnetic resonance. J Cardiovasc Magn Mandry D, Pierre-Mongeon F, Heydari B, Francis SA, Moslehi J, Kwong
Reson. 2009;11:55. RY, Jerosch-Herold M. Myocardial extracellular volume by cardiac mag-
33. Ashford MW Jr, Liu W, Lin SJ, Abraszewski P, Caruthers SD, Connolly netic resonance imaging in patients treated with anthracycline-based che-
Downloaded from http://ahajournals.org by on April 30, 2024

AM, Yu X, Wickline SA. Occult cardiac contractile dysfunction in motherapy. Am J Cardiol. 2013;111:717–722.

CLINICAL PERSPECTIVE
The lack of suitable imaging biomarkers has been a barrier to clinical trials in this patient population. Microscopic fibrosis as
determined by T1 mapping techniques and increased extracellular volume attributable to interstitial expansion due, in part,
to deposition of collagen is present in asymptomatic subjects exposed to high-dose anthracyclines with preserved global sys-
tolic function. These structural abnormalities, coupled with a decrease in longitudinal and circumferential strain magnitude
as demonstrated by cardiac magnetic resonance, may represent potential therapeutic targets for modern heart failure therapy
aimed at preventing overt cardiac dysfunction and end-stage cardiomyopathy in this patient population.