Author + information
- Received November 23, 2012
- Revision received December 21, 2012
- Accepted January 8, 2013
- Published online April 9, 2013.
- Kaisa Ylänen, MD⁎,⁎ (, )
- Tuija Poutanen, MD, PhD⁎,
- Päivi Savikurki-Heikkilä, MD†,
- Irina Rinta-Kiikka, MD, PhD†,
- Anneli Eerola, MD, PhD⁎ and
- Kim Vettenranta, MD, PhD⁎
- ↵⁎Reprint requests and correspondence:
Dr. Kaisa Ylänen, Tampere University Hospital, Department of Pediatrics, P.O. Box 2000, FI-33521 Tampere, Finland
Objectives This study sought to examine the left ventricular (LV) and right ventricular (RV) function and signs of focal fibrosis among long-term survivors of childhood cancer with the use of cardiac magnetic resonance (CMR) imaging.
Background Increased myocardial fibrosis has been detected in the endomyocardial biopsies of survivors. CMR has established its role in the assessment of both cardiac function and structure, and focal fibrosis of the myocardium is detectable with late gadolinium enhancement (LGE).
Methods Sixty-two anthracycline-exposed long-term survivors of childhood cancer were studied at a mean age of 14.6 years. The LV and RV ejection fractions (EFs) and volumes were measured, and LGE was assessed using CMR.
Results An abnormal LV function (EF <45%) was detected in 18% (11 of 62) of the survivors, and an abnormal RV function was detected in 27% (17 of 62) of the survivors. Subnormal (45% ≤ EF <55%) LV function were demonstrated in 61% (38 of 62) and subnormal RV function in 53% (33 of 62) of the survivors, respectively. Both the LV and RV end-systolic and LV end-diastolic volumes were increased compared with reference values. None of the study patients showed LGE.
Conclusions A considerable proportion of the long-term survivors of childhood cancer with anthracycline exposure demonstrate signs of cardiac dysfunction detectable by CMR, with the RV also being involved. Yet, myocardial fibrosis does not seem to be detectable at a median of 7.8 years after anthracycline therapy.
Anthracyclines are widely used in the treatment of many childhood malignancies, but a dose-dependent cardiotoxicity is limiting their use. Late toxicity is defined as cardiac abnormalities developing no earlier than 1 year after anthracycline treatment, but in some cases it is not discernible until years or decades later. Late cardiotoxicity can manifest as subclinical abnormalities in the cardiac function, dilated cardiomyopathy, and/or heart failure and is usually irreversible. The cumulative drug dose, mediastinal irradiation, young age at exposure, female sex, and time post-exposure are known risk factors for anthracycline cardiomyopathy also modified by genetic factors (1–3). The prevalence of subclinical cardiotoxicity is between 0% and 57% (4), and the prevalence of anthracycline-induced clinical heart failure is between 0% and 16% (5). Mulrooney et al. (2) reported a cumulative incidence of congestive heart failure exceeding 7.5% at 30 years among those with a cumulative anthracycline dose ≥250 mg/m2 and 2.5% among those with a lower dose. Published data also indicate that a left ventricular (LV) dysfunction, frank heart failure (6), and terminal cardiomyopathy (7) may occur even at lower doses. Childhood cancer survivors thus seem at an increased risk for anthracycline cardiotoxicity with their young age at exposure and longer life expectancy.
Among the cancer survivors, echocardiography has been widely used in the follow-up of cardiac function, but it has limitations in sensitivity and problems caused by a suboptimal acoustic window (8,9). Cardiac magnetic resonance (CMR) imaging is considered a standard of reference in the assessment of cardiac function and structure against which other modalities are validated (10,11). With anthracycline cardiomyopathy having been thought to involve mainly the LV, the follow-up protocols of cancer survivors have, until recently, largely neglected the right ventricle (RV) with its complex anatomy and scarcity of optimal imaging methods. CMR is considered the best method for visualizing the RV (12) and thus used in the follow-up of heart diseases affecting the RV (13).
Many cardiac diseases lead to fibrosis of the myocardium classifiable for clinical purposes as local or diffuse (14). Late gadolinium enhancement (LGE) in CMR after the administration of a gadolinium-chelated contrast agent (Gd-CA) is a well-established method for assessing focal fibrosis (15). LGE for fibrosis or deposition of abnormal substrates in the myocardium can be used in patients with myocardial infarction, hypertrophic and dilated cardiomyopathies, infiltrative diseases, and myocarditis (16).
The pathogenesis of late-onset cardiac toxicity remains incompletely understood, but evidence exists on myocyte apoptosis after anthracycline exposure (17), resulting in myocardial fibrosis detectable on LGE. Steinherz et al. (18) have reported on increased fibrosis in the endomyocardial biopsies taken from the RV and at autopsy among the long-term cancer survivors with anthracycline cardiomyopathy.
The aim of our study was to evaluate the prevalence of LV and RV dysfunction and signs of focal fibrosis with CMR among anthracycline-exposed long-term childhood cancer survivors in a single-center setting.
Characteristics of the study population
The study population consisted of 62 long-term survivors of childhood cancer (34 female and 28 male) attending the population-based pediatric hematology-oncology service of the Tampere University Hospital (Tampere, Finland) between February 2010 and June 2011. The survivors enrolled had received anthracyclines (doxorubicin, daunorubicin, idarubicin, or mitoxantrone) as a part of their therapy, had a minimum follow-up of 5 years, had no congenital heart disease, and were in remission. Of the 86 survivors initially recruited, 62 agreed to participate in the study. The body surface area (BSA) at diagnosis was used to calculate the cumulative anthracycline dose. Dose conversion to doxorubicin isotoxic equivalents was performed according to the Children's Oncology Group recommendations. CMR was performed only for study purposes with only 1 child requiring anesthesia to achieve optimal images. The institutional review board of the Tampere University Hospital approved the study protocol. All survivors and their legal guardian(s) gave their written informed consent.
The key characteristics of the study patients are presented in Tables 1 and 2.⇓⇓ The age of the survivors was 14.6 ± 3.2 years. The median (range) age at the time of the malignancy diagnosis was 3.8 years (0.0 to 13.8 years), and the follow-up time was 7.8 years (4.9 to 18.0 years). Their median cumulative anthracycline dose was 222 mg/m2 (80 to 419 mg/m2). Seven survivors (11%) had received radiotherapy involving the heart with a median average cumulative cardiac dose of 10.0 Gy (3.6 to 12.0 Gy). The average cardiac dose of those without total body irradiation was derived from the radiotherapy planning charts. The dose among those with total body irradiation was estimated to equal the total radiation dose.
Three survivors (5%) had experienced a relapse but at the time of inclusion were in remission. Three survivors were taking enalaprile for their cardiomyopathy, and 1 survivor was taking bisoprolol for a long QT-syndrome. One survivor had an asymptomatic patent ductus arteriosus. One survivor had had a thrombus resected from the right atrium during primary therapy but remained asymptomatic.
Magnetic resonance imaging technique and analysis
CMR was performed on a 1.5-T scanner (Siemens Magnetom Avanto; Siemens Healthcare, Erlangen, Germany) using a 6-channel body array coil with a spine coil and electrocardiogram gating. Cine TrueFISP slices of 8 mm without any gap from the heart apex to valves were obtained in the short-axis plane to analyze the LV and in the axial plane to analyze the RV function. After injection of gadoterate meglumine (0.1 mmol/kg, Dotarem [Guerbet, Roissy, France]), a 5-min delay was used before obtaining a segmented inversion recovery cine TrueFISP pulse sequence at the midventricular short-axis level for the determination of the inversion time (TI) value for the nullification of the impact of the normal myocardium. Within the next 5 min, the LGE images were obtained using a TrueFISP gradient echo sequence with a determined TI (range: 280 to 330 ms) with a slice thickness of 8 mm in the short axis, 2-chamber view (vertical long axis through the left atrium and ventricle), 4-chamber view (horizontal long axis), and axial stacks covering the heart from apex to valves. Thereafter, the corresponding views were recovered using a phase-sensitive inversion recovery (PSIR) technique with a constant TI value of 300 ms. The late-enhancement imaging was performed within 15 min from the beginning of the gadoterate injection.
In the analysis, ARGUS software (Siemens AG, Munich, Germany) was used. All measurements were done manually. The end-systolic and diastolic frames were identified by determining the ventricular blood-pool areas. The LV volumes were calculated from the short axis, and the RV volumes were calculated from the axial cine views. The aortic outflow tract below the valve and the LV portion of the slice in the basal region near the left atrium were included in the LV volume measurements. The same principles were used for the RV. The free papillary muscles were included for both the ventricular volumes (Fig. 1). The LGE in the TrueFISP and PSIR sequences was independently analyzed by 2 radiologists (P.S.H, I.R.K.), who also independently analyzed 15 randomly selected data sets to evaluate the interobserver variability. One investigator (PSH) measured 15 data sets twice to assess the intraobserver variability.
Descriptive statistics and analyses were performed using PASW Statistics 18.0 software (SPSS Inc., Chicago, Illinois). Descriptive statistics are presented as the frequencies and percentages for categoric data and the mean ± SD or median and (range) for continuous data. The CMR parameters were compared with published data (19,20) using the 1-sample Student t test. The categoric variables were compared using the chi-square test (the Fisher exact test when appropriate). For continuous data, the independent samples t test and the Mann-Whitney U test were used. All tests were 2-sided, and p values <0.05 were considered statistically significant. The intraobserver and interobserver variations were assessed using the Bland-Altman analysis (21).
The reference values for the CMR were deduced from published data. The LV parameters were compared with the data of Robbers-Visser et al. (19) describing the short-axis slices from 60 healthy children age 8 to 17 years. For the RV analysis, the data of Sarikouch et al. (20) on 99 healthy children age 8 to 20 years obtained by the axial slices were used. Abnormally large ventricular volumes were defined as those exceeding the published mean volumes by at least 2 SD.
We classified the ejection fraction (EF) as normal (EF ≥55%), subnormal (EF 45% ≤ EF <55%), or abnormal (EF <45%). Of the survivors, 21% (13 of 62) had a normal LV EF, 61% (38 of 62) had a subnormal LVEF, and 18% (11 of 62) had an abnormal LVEF. There was a trend toward a male predominance among those with an abnormal LVEF (n = 11); 29% (8 of 28) were male and 9% (3 of 34) were female (p = 0.053). Survivors with an LVEF <45% were older than those with an LVEF ≥45% (16.3 vs. 14.2 years, p = 0.047). Eight percent of the survivors (5 of 62) thought their exercise tolerance was compromised when compared with peers and had an LVEF <55% (range: 44.3% to 54.7%). The cumulative anthracycline dose, age at diagnosis, and follow-up time seemed not to be associated with the LV EF (data not shown).
The LV parameters were compared in 3 age groups (8 to 11, 12 to 14, and 15 to 17 years) with the data of Robbers-Visser et al. (19) (Table 3). No reference data are available for the short-axis LV parameters taking both age group and sex simultaneously into account. The LV end-diastolic volume (EDV) and end-systolic volume (ESV) were significantly larger in the study group. A total of 82% (28 of 34) of the female subjects and 100% (28 of 28) of the male subjects had their LV ESV above the +2 SD of the reference values. The upper normal limits of the LV EDV were exceeded by 18% (6 of 34) of the female subjects and 43% (12 of 28) of the male subjects. The LVEF was significantly lower in all age groups compared with the reference values. The LV mass did not differ significantly from the normal values.
A normal RVEF was found in 19% (12 of 62) of the survivors, a subnormal RVEF was found in 53% (33 of 62) of the survivors, and an abnormal RVEF was found in 27% (17 of 62) of the survivors. More male subjects (39% [11 of 28]) than female subjects (18% [6 of 34]) had an abnormal RVEF (p = 0.057). The RVEF was not associated with the cumulative anthracycline dose, age at diagnosis, or follow-up time (data not shown).
For the RV parameters, a comparison with the reference values was done in 2 age groups (8 to 15 years and 16 to 20 years) and separately for female and male subjects (20) (Table 4). The RV ESV and stroke volume were significantly larger in the study group. some 41% (14 of 34) of the female and 64% (18 of 28) of the male subjects had their RV ESV above the +2 SD of the reference values. An abnormally large RV EDV was found in 15% (5 of 34) of the female subjects and 21% (6 of 28) of the male subjects. Only the male subjects aged 16 to 20 years had a significantly larger RV EDV compared with the reference values. The RVEF was significantly lower in all age groups compared with reference values. The RV mass was not measured in our study.
Those with a history of cardiac irradiation did not have larger ventricular volumes or an altered LV mass compared with the other survivors (Table 5). Among the irradiation-exposed subjects, 29% (2 of 7) had their LV EDV, 86% (6 of 7) had their LV ESV, and 14% (1 of 7) had their RV ESV above the +2 SD of the reference values.
None of the study patients showed LGE. The intraobserver and interobserver data are presented in Figure 2 and Table 6. The limits of agreement were defined as ±1.96 SD from the mean between 2 measurements. Intraobserver variability was lower than interobserver variability. Interobserver variability was lower for the LV than for the RV.
The current study is the first to report on the LGE and LV and RV function with CMR in a pediatric population after anthracycline treatment. The data unequivocally document a marked LV and RV dysfunction without signs of focal myocardial fibrosis among the anthracycline-exposed long-term survivors of childhood cancer. However, the myocardial dysfunction remains mostly asymptomatic at approximately 1 decade post-therapy. Only 6% (4 of 62) were taking medication, and 92% of the total considered their exercise tolerance normal. Armstrong et al. (9) found an abnormal LVEF with CMR in 32% of their adult survivors of childhood cancer previously undiagnosed with cardiotoxicity. Yet, although asymptomatic, the survivors may be at risk of a symptomatic heart failure on exposure to a stress such as pregnancy (22) or hypertension (23), warranting a regular cardiologic follow-up and lifestyle counseling.
CMR volumetry and function
CMR imaging is a well-established, accurate, reproducible, and noninvasive method for assessing both the LV and RV function (10,24). Advantages over echocardiography include independence of the acoustic window and geometric assumptions, as well as freely selectable imaging planes. Among the survivors of childhood cancer, the impact of factors such as obesity, postoperative abnormalities of the thorax, or a marked dilatation of the ventricles emphasizes the benefits of CMR. Furthermore, CMR remains optimal for the imaging of cardiomyopathies with an ability to visualize the myocardium (25).
Until recently, the CMR reference values in children have been based on the gradient-echo sequences from a limited number of small studies (26–28). Recent papers (19,20) report pediatric data using steady-state, free-precession, gradient-echo sequences, currently the standard method in a functional and volumetric CMR. The availability of age group–specific data (19,20) made it possible for us to compare the ventricular volumes with the pediatric reference values.
The screening for an anthracycline-induced cardiotoxicity has previously focused on the LV. The normal values for the fractional shortening and EF are age-independent, facilitating their use in a pediatric population. However, the BSA-indexed ventricular volumes offer a valuable tool for an earlier detection of overload changes in the heart chambers. A sizeable proportion of our study patients had abnormally large LV and RV ESVs, with the EDVs being less affected. This may represent an early stage in the progression of dysfunction. The fractional shortening and EF may remain within the normal limits despite LV dilatation. In many heart diseases, the BSA-indexed ventricular volumes are used at the clinic (13). Therefore, abnormally large ventricular volumes with a normal EF emphasize the need for a regular cardiac follow-up among pediatric cancer survivors.
Our results are in agreement with the data of Oberholzer et al. (29) on the biventricular cardiac function of anthracycline-exposed pediatric patients. One-half of their patients had an LVEF <55% but remained asymptomatic, and the RVEF after chemotherapy was 48.2 ± 7.1%, reflecting an additional impairment in the RV function (29).
The cardiotoxic effects of irradiation are more diverse than those of anthracyclines and are reported to be predominantly restrictive (30). Adams et al. (31) reported reduced LV dimensions by echocardiography among mediastinal irradiation–exposed long-term survivors of Hodgkin disease. In our study, there was a trend toward smaller ventricular volumes among the cardiac irradiation–exposed survivors. Yet, because of the small sample size and heterogeneity of the cohort, reliable conclusions could not be drawn.
Late gadolinium enhancement
A delayed enhancement could not be detected even though 2 different methods were used: one using the changing TI time and one using a sensitive PSIR method with a constant TI time. Gd-CA doses in the LGE imaging of various heart diseases have varied between 0.1 and 0.2 mmol/kg (32–34). It remains unsettled whether a larger Gd-CA dose could have produced more LGE among our study patients, but for safety reasons the lower recommended dose was used (35).
Data on the myocardial LGE among anthracycline-exposed long-term survivors of childhood cancer have not been reported. Wassmuth et al. (36) studied 22 adult patients and reported on a transient increase in the myocardial enhancement early after the administration of anthracycline. Perel et al. (37) described 2 adults with an anthracycline cardiomyopathy, LV dysfunction, and subendocardial LGE years after chemotherapy, with one having LGE in the RV. Furthermore, LGE also was detected in 29% of 36 adult patients with mediastinal radiotherapy for Hodgkin disease more than 2 decades earlier (38).
Data on fibrosis in anthracycline-induced cardiomyopathy remain limited but indicate a prolonged process (18,37). The incidence of frank cardiac pathology increases with time since exposure. The estimated risk of developing cardiomyopathy after anthracycline therapy has been reported to be 4.5% at 10 years and increases to 9.8% at 20 years after therapy among patients with cumulative doses of ≥300 mg/m2 (39). The median follow-up of 7.8 years in our study may not have been long enough for myocardial fibrosis to develop. This may be partly due to age-related cardiovascular events promoting cardiac damage and myocardial fibrosis, emphasizing the importance of lifestyle counseling among the long-term survivors. Furthermore, endomyocardial biopsies remain infrequently performed during pediatric follow-up. This may negatively affect the possibility of documenting myocardial fibrosis among children.
A diffuse myocardial fibrosis cannot be visualized on LGE with diffuse fibrosis being nulled out to highlight focal fibrosis and information on a possible background diffuse fibrosis thus being obliterated (14). Flett et al. (40) recently reported on a novel, noninvasive equilibrium-contrast CMR approach to detect the diffuse fibrosis validated using surgical myocardial biopsies in patients with aortic stenosis and hypertrophic cardiomyopathy. Bernaba et al. (41) found both interstitial and diffuse fibrosis in the myocardial tissue of 10 adult patients with an anthracycline-induced cardiomyopathy. It is conceivable that some of our study patients may have diffuse fibrosis not detectable with the technique used.
Published data (18,37,41) clearly indicate that some patients with anthracycline exposure will have focal or diffuse fibrosis detectable on LGE or through an equilibrium-contrast CMR. The limited number of reports on an established myocardial fibrosis with anthracycline cardiomyopathy invariably document a concomitant symptomatic heart failure, and many of these patients have undergone heart transplantation (18,37,41).
Because of ethical reasons and resource limitations, we had no own healthy controls analyzed with CMR.
A markedly high proportion of the long-term survivors of childhood cancer with anthracycline exposure appear to have a cardiac dysfunction detectable by CMR with the RV also being involved, but without focal myocardial fibrosis. In the follow-up of these patients, the CMR is highly usable, particularly among those with poor acoustic windows or the RV function at focus. Whether a longer follow-up will render putative anthracycline-induced myocardial fibrosis detectable (e.g., LGE) among the long-term survivors of childhood cancer remains to be established.
The authors thank Tuija Wigren, MD, for help with the radiation therapy case histories; Kirsi-Maria Lauerma, MD, for expert advice with CMR techniques; and Satu Ranta, RN, for practical assistance during the project.
This study was financially supported by the Competitive Research Funding of the Tampere University Hospital (Grant 9L114), the EVO funds of the Tampere University Hospital, the Emil Aaltonen Foundation, the Finnish Cultural Foundation, the Finnish Cultural Foundation Pirkanmaa Regional Fund, the Finnish Association of Hematology, the National Graduate School of Clinical Investigations, the Päivikki and Sakari Sohlberg Foundation, the Foundation for Pediatric Research, the Scientific Foundation of the City of Tampere, the Finnish Cancer Foundation, and the Blood Disease Research Foundation. All authors have reported they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- body surface area
- cardiac magnetic resonance
- end-diastolic volume
- ejection fraction
- end-systolic volume
- gadolinium-chelated contrast agent
- late gadolinium enhancement
- left ventricle/ventricular
- phase-sensitive inversion recovery
- right ventricle/ventricular
- inversion time
- Received November 23, 2012.
- Revision received December 21, 2012.
- Accepted January 8, 2013.
- American College of Cardiology Foundation
- Mulrooney D.A.,
- Yeazel M.W.,
- Kawashima T.,
- et al.
- Blanco J.G.,
- Sun C.L.,
- Landier W.,
- et al.
- Kremer L.C.,
- van der Pal H.J.,
- Offringa M.,
- van Dalen E.C.,
- Voute P.A.
- Kremer L.C.,
- van Dalen E.C.,
- Offringa M.,
- Voute P.A.
- Armstrong G.T.,
- Plana J.C.,
- Zhang N.,
- et al.
- American College of Cardiology Foundation Task Force on Expert Consensus, Documents,
- Hundley W.G.,
- Bluemke D.A.,
- et al.
- Haddad F.,
- Hunt S.A.,
- Rosenthal D.N.,
- Murphy D.J.
- Sarikouch S.,
- Peters B.,
- Gutberlet M.,
- et al.
- Ryberg M.,
- Nielsen D.,
- Cortese G.,
- Nielsen G.,
- Skovsgaard T.,
- Andersen P.K.
- Higgins C.B.
- Chen M.H.,
- Colan S.D.,
- Diller L.
- Adams M.J.,
- Lipsitz S.R.,
- Colan S.D.,
- et al.
- Moon J.C.,
- McKenna W.J.,
- McCrohon J.A.,
- Elliott P.M.,
- Smith G.C.,
- Pennell D.J.
- Assomull R.G.,
- Prasad S.K.,
- Lyne J.,
- et al.
- Assomull R.G.,
- Pennell D.J.,
- Prasad S.K.
- Flett A.S.,
- Hayward M.P.,
- Ashworth M.T.,
- et al.