Effect of anthracycline therapy on myocardial function and markers of fibrotic remodelling in childhood cancer survivors

Abstract

Aims

Anthracyclines are a cornerstone of paediatric cancer treatment. We aimed to quantify myocardial cardiac magnetic resonance (CMR) native T1 (NT1) and extracellular volume fraction (ECV) as markers of fibrosis in a cohort of childhood cancer survivors (CCS).

Methods and results

A cohort of CCS in remission underwent CMR T1 mapping. Diastolic function was assessed by echocardiography. Results were compared to a cohort of normal controls of similar age and gender. Fifty-five CCS and 46 controls were included. Both groups had similar mean left ventricular (LV) NT1 values (999 ± 36 vs. 1007 ± 32 ms, P = 0.27); ECV was higher (25.6 ± 6.9 vs. 20.7 ± 2.4%, P = 0.003) and intracellular mass was lower (37.5 ± 8.4 vs. 43.3 ± 9.9g/m², P = 0.02) in CCS. The CCS group had lower LV ejection fraction (EF) and LV mass index with otherwise normal diastolic function in all but one patient. The proportion of subjects with elevated ECV compared to controls did not differ between subgroups with normal or reduced LV EF (22% vs. 28%; P = 0.13) and no correlations were found between LVEF and ECV. While average values remained within normal range, mitral E/E′ (6.6 ± 1.6 vs. 5.9 ± 0.9, P = 0.02) was higher in CCS. Neither NT1 nor ECV correlated with diastolic function indices or cumulative anthracycline dose.

Conclusions

There is evidence for mild diffuse extracellular volume expansion in some asymptomatic CCS; myocyte loss could be part of the mechanism, accompanied by subtle changes in systolic and diastolic function. These findings suggest mild myocardial damage and remodelling after anthracycline treatment in some CCS which requires continued monitoring.

Introduction

Despite a dramatic improvement in survival for children diagnosed with cancer and more than 80% surviving long-term,^1,2 important morbidity and mortality prevail. Anthracycline chemotherapy agents are one of the cornerstones of cancer therapy, but cause dose-related myocardial injury and harbour a risk of progressive heart failure,^3–5 even at relatively low cumulative anthracycline doses.^6–9 Surveillance of cardiotoxicity in childhood cancer survivors (CCS) currently rests predominantly on echocardiographic parameters of global left ventricular (LV) function.¹⁰ However, LV shortening fraction (LV SF) and LV ejection fraction (LV EF) are often normal until the later stages of anthracycline cardiomyopathy.¹¹ Metrics of myocardial deformation, such as strain imaging echocardiography and, more recently, cardiac magnetic resonance (CMR) have been proposed as more sensitive for the detection of ventricular dysfunction, albeit with limited data in CCS.^12,13 In addition to indices of cardiac and myocardial function, CMR yields insights into myocardial tissue composition: myocardial native T1 times (NT1) and extracellular volume fraction (ECV) by CMR T1 relaxometry, often referred to as ‘T1 mapping’, correlate with histological diffuse fibrosis.^14,15 Early reports on NT1 and ECV in CCS are conflicting and paediatric control groups are universally missing. The aims of the current study were (i) to characterize NT1 and ECV, as candidate markers of myocardial fibrosis, in CCS in comparison to paediatric controls; and (ii) to elucidate whether these markers are associated with myocardial functional changes.

Methods

Study design

This single-centre, cross-sectional nested cohort study was approved by the authors’ institutional research ethics board and by Health Canada. Eligible subjects were in cancer remission, at least 3 years from their last dose of anthracycline therapy and were participants in a multicentre prospective study assessing cardiac outcome in children with cancer (Preventing Cardiac Sequelae in Pediatric Cancer Survivors, PCS²). Details of the study design have been previously published.¹⁶ Patients were recruited for the current sub-study during an outpatient visit between October 2014 and January 2018. Following informed written consent, participants underwent CMR. Subjects with congenital heart disease or contraindications to CMR were excluded. Clinical details were retrieved from the patient charts. Anthracycline doses were converted into a doxorubicin-equivalent dose.¹⁷ Healthy paediatric controls who were referred to CMR for a family history of cardiomyopathy or for coronary artery imaging in non-specific chest pain were included, as long as their entire work-up was negative.

Cardiac magnetic resonance

All CMR examinations were performed on a 1.5 T scanner (‘Avanto’, Siemens Medical Systems, Erlangen, Germany). The protocol included balanced steady-state free precession (b-SSFP) short-axis cine imaging for ventricular volumetry, mass and EF. T1 mapping was performed before and 15 min after 0.2 mmol/kg of gadopentetate dimeglumine (‘MultiHance’, Bracco Diagnostics, Princeton, NJ, USA), using a modified Look-Locker Inversion Recovery (MOLLI) sequence at a single mid short-axis level. The sequence consists of two inversion pulses followed by five and three single-shot image b-SSFP images, respectively, acquired during diastole and separated by a number of heart beats. This number was adjusted to the patient’s heart rate to allow for recovery of longitudinal magnetization before the next inversion experiment (Figure 1).¹⁸ The other sequence parameters included repetition and echo times of 2.68 and 1.13 ms, respectively, 8 mm slice thickness, 1.4 × 1.4 mm in-plane resolution, and flip angle of 35°. An inline motion correction algorithm was employed for co-registration of the eight individual images. A phase-sensitive inversion recovery sequence was used for late gadolinium enhancement (LGE) imaging ten minutes after contrast injection. Haematocrit for ECV computation¹⁹ was obtained at the time of intravenous line insertion for the CMR. Myocardial intracellular volume fraction (ICV) was calculated as 1-ECV²⁰ and indexed intracellular LV mass (ICMi) as the product of ICV and indexed LV mass (Mi).

Figure 1.

Example of modified Look-Locker Inversion Recovery (MOLLI) sequence and mid-ventricular segmentation for T1 mapping. TD, trigger delay; TI, inversion time.

Ventricular volumes, LV mass, LV strain, and T1 times were quantified using commercially available software (QMass 8.1, Qstrain 2.0 and Qmap 2.2.36; Medis Medical Imaging Systems, Leiden, The Netherlands). Z-scores were calculated for ventricular volumes, EFs, and LV mass (M) using published normal values based on comparable sequences and post-processing methods.²¹ T1 was measured in the entire mid short-axis circumference, in the LV free wall (FW) and in the interventricular septum (IVS), avoiding the blood pool, epicardial fat, and areas of LGE.²² Myocardial T1 values were obtained using a region-of-interest-based quantification method. Blood pool T1 values were obtained by defining a region of interest within the blood pool. LV myocardial strain analysis was obtained by applying CMR feature tracking to the acquired LV four-chamber and short-axis cines at the basal, mid and apical ventricular levels. The endocardial border was set semi-automatically in end-systole and end-diastole and manually readjusted as required. Segments with poor border tracking were excluded.

Echocardiography

A comprehensive functional echocardiogram was performed prospectively in all patients as part of the PCS² study and included in the current analysis if it occurred within 12 months of the CMR as significant changes in cardiovascular status are not expected to occur in that period in clinically stable chronic cancer survivors. Diastolic function was evaluated based on mitral valve inflow (E- and A-wave velocities, E-wave deceleration time, and isovolumic relaxation time) and pulmonary venous flow (systolic-to-diastolic peak velocity ratio, A-wave reversal amplitude and duration) pulsed-wave Doppler and tissue Doppler (mitral lateral and septal peak early diastolic tissue velocities, lateral E′, septal E′). As widely accepted paediatric guidelines for diastolic function assessment are lacking, patient measurements were converted to Z-scores based on published normative values from our institution,²³ and compared to a large group of normal children of similar age and gender distribution, using similar echocardiographic methodology. Given that no single parameter reliably identifies diastolic dysfunction, the latter was deemed present when the Z-scores of three or more of the above-listed parameters of diastolic function were outside of −2 to 2.

Statistical analysis

Statistical analysis was performed using GraphPad Prism 7 (GraphPad Software, La Jolla, CA, USA). Continuous variables are presented as means ± standard deviations if normally distributed, otherwise as medians and ranges. Categorical variables are expressed as counts and percentages of the total. Cases were compared to controls. When data were normally distributed, two-tailed t-tests were used to compare continuous variables and Pearson’s correlation coefficient for correlations. For non-normally distributed data, the Mann–Whitney U test was used for comparisons and the Spearman rank-order test for correlations. P-values <0.05 was considered statistically significant.

Results

Patient characteristics are summarized in Table 1. Fifty-five patients were included and compared to 46 controls. Of these, 48 patients and 25 controls consented to receive gadolinium injection for ECV estimation and LGE. The CCS and the control groups did not differ with regards to age, gender distribution, or body surface area. The median age at cancer diagnosis in the CCS group was 3.7 years (range 0.4–13.0 years). CMR was obtained after a median time of 9.8 years (3.6–17.0 years) from the last anthracycline dose. The median cumulative doxorubicin-equivalent dose was 212 mg/m² (60–550 mg/m²). There were seven subjects (12%) who received chest radiation therapy. At the time of the CMR, all patients were asymptomatic and were not receiving any cardiac medications.

Table 1.

Patient characteristics and cardiac magnetic resonance results

	Controls	Cancer survivors	P-value
	(n = 46)	(n = 55)
Age (years)	14.2 ± 2.4	15.1 ± 2.8	0.23
Male (%)	54 (n = 25)	49 (n = 27)	0.43
Heart rate (bpm)	67 ± 13	65 ± 20	0.27
BSA (m2)	1.65 ± 0.35	1.61 ± 0.32	0.55
Median age at cancer diagnosis (years)		3.7 (0.4–13.0)
Median time since last anthracycline therapy (years)		9.8 (3.6–17.0)
Cumulative anthracycline dose (mg/m2)		212 (60–550)
Chest radiation therapy (%)		12 (n = 7)
LV EDVi (mL/m2)	91 ± 13	89 ± 12	0.63
LV EDVi (Z-score)	0.3 ± 1.6	0.2 ± 1.2	0.69
LV ESVi (mL/m2)	38 ± 7	41 ± 8	0.12
LV ESVi (Z-score)	0.8 ± 1.1	1.2 ± 1.2	0.09
LV Mi (g/m2)	55 ± 12	49 ± 9	0.04
LV M (Z-score)	−0.6 ± 1.9	−1.2 ± 1.5	0.02
LV Mass/Vol (g/mL)	0.60 ± 0.10	0.56 ± 0.10	0.04
LV EF (%)	58 ± 5	55 ± 5	0.005
LV EF (Z-score)	−0.8 ± 1.1	−1.6 ± 1.2	0.001
RV EDVi (mL/m2)	98 ± 18	93 ± 14	0.6
RV EDVi (Z-score)	−0.1 ± 1.6	−0.5 ±1.3	0.5
RV ESVi (mL/m2)	46 ± 10	45 ±9	0.9
RV ESVi (Z-score)	0.4 ± 1.2	0.2 ± 1.1	0.9
RV EF (%)	52 ± 4	52 ± 5	0.8
RV EF (Z-score)	−0.8 ± 1.0	−0.9 ± 1.1	0.44
LV NT1 (ms)	1007 ± 32	998 ± 36	0.11
IVS NT1(ms)	1005 ± 40	1023 ± 53	0.08
LV FW NT1(ms)	985 ± 47	1012 ± 57	0.08
LV ECV (%)	20.7 ± 2.4 (n = 25)	25.5 ± 7.0 (n = 48)	0.003
IVS ECV (%)	21.7 ± 2.6 (n = 25)	26.6 ± 7.3 (n = 48)	0.04
LV FW ECV (%)	20.6 ± 2.6 (n = 25)	25.4 ± 7.7 (n = 48)	0.02
LV ICMi (g/m2)	43.3 ± 9.9 (n = 25)	37.5 ± 8.4 (n = 48)	0.02

BSA, body surface area; ECV, extracellular volume fraction; EDVi, end-diastolic volume indexed; EF, ejection fraction; ESVi, end-systolic volume indexed; FW, free wall; ICMi, intracellular mass indexed; IVS, interventricular septum; LV, left ventricle; Mass/Vol, mass to volume ratio; Mi, mass indexed; NT1, native T1 time; RV, right ventricle.

Ventricular volumes and function

Table 1 summarizes the CMR findings. Right and LV indexed end-diastolic volume (EDVi) and the LV indexed end-systolic volume (ESVi) did not differ between controls and patients. The CCS group had slightly lower LV EF and LV EF Z-score compared to controls (P = 0.005 and P = 0.001, respectively); 25 patients in the CCS group (45%) had a subnormal LVEF (LVEF < 55%, Figure 2). LV mass indexed (LVMi), LV mass Z-score, and LV mass/volume ratio were lower in the CCS group. Right ventricular (RV) volumes and EF were similar to controls. Global longitudinal strain (GLS) and basal circumferential strain (CS) did not differ between the 54 CCS subjects and 29 controls in whom it was obtained (Table 2). Mid-ventricular and apical CS were lower (P = 0.02 and P = 0.008, respectively) in CCS as was global CS (P = 0.001). CCS with LVEF ≥ 55% did not differ from those with LVEF <55% with regards to ventricular volumes, LVMi or LV strain. The subgroup of CCS who had received chest radiation therapy did not differ from the rest of the group or from controls with regards to any of the aforementioned parameters. None of the LV systolic and diastolic functional parameters correlated with systolic blood pressure.

Figure 2.

Distribution of left ventricular ejection fraction in the childhood cancer survivor group. LVEF, left ventricular ejection fraction.

Table 2.

LV strain by CMR

	Controls (n = 29)	Cancer survivors (n = 54)	P-value
GLS (%)	−20.9 ± 2.9	−19.3 ± 2.8 (n = 51)	0.99
Basal CS (%)	−31.6 ± 4.0	−28.7 ± 4.5	0.34
Mid CS (%)	−30.1 ± 4.8	−25.8 ± 4.4	0.03
Apical CS (%)	−44.2 ± 9.8	−32.8 ± 7.7	0.001
GCS (%)	−35.6 ± 4.6	−28.0 ± 7.2	0.001

CS, circumferential strain; GCS, global circumferential strain; GLS, global longitudinal strain; LVEF, left ventricular ejection fraction.

Diastolic function by echocardiography

Fifty-three CCS subjects had received an echocardiogram within 12 months of CMR. Echocardiography was obtained within a median of 1 month of CMR (0–9 months). In the CCS group, parameters of diastolic function did not differ from the control group (Table 3), although septal and lateral E′ were lower in CCS (P = 0.001 and P = 0.003, respectively) and septal E/E′ (P = 0.04) and average E/E′ (P = 0.02) were slightly higher. No patient had diastolic dysfunction according to adult guidelines²⁴ and only one patient in the CCS group had abnormal diastolic function on the basis of the criteria outlined in the methods. Despite a normal LVEF, this patient’s GLS, basal CS and mid CS were below the 10th percentile (−15%, −21%, and −19%, respectively) while ECV was above the 90th and NT1 above the 75th percentile (ECV 37%, NT1 1028 ms). This patient also had received a high cumulative anthracycline dose of 360 mg/m² (85th percentile in the group) in addition to chest radiation therapy. Diastolic function parameters did not differ amongst CCS with LVEF ≥55% vs. those with LVEF <55% or between patients who did and those who had not received chest radiation therapy.

Table 3.

Diastolic parameters by echocardiography

	Controls (n = 115)	Cancer survivors (n = 53)	P-value
Age (years)	15.0 ± 1.7	15.1 ± 2.8	0.24
Male (%)	53	49	0.55
BSA (m2)	1.62 ± 0.20	1.61 ± 0.32	0.77
Heart rate (bpm)	67 ± 12	69 ±14	0.33
Systolic blood pressure (mmHg)	109 ± 10	101 ± 23	0.2
Diastolic blood pressure (mmHg)	57 ± 7	57 ± 14	0.9
Mitral E velocity (cm/s)	99 ± 16	98 ± 17	0.85
Mitral A velocity (cm/s)	42 ± 11	44 ± 10	0.07
Mitral E/A	2.5 ± 0.7	2.4 ± 0.7	0.15
Mitral E DT (ms)	149 ± 18	174 ± 34	<0.001
Mitral A duration (ms)	118 ± 20	125 ± 30	0.18
PV S/D	0.73 ± 0.23	0.74 ± 0.28	0.82
PVa velocity (cm/s)	18 ± 10	14 ± 11	0.08
PVa duration (ms)	103 ± 24	79 ± 55	0.09
IVRT (ms)	75 ± 7	70 ± 14	0.07
Septal E′ (cm/s)	15 ± 2	14 ± 2	0.001
Lateral E′ (cm/s)	19 ± 2	17 ± 3	0.003
Septal E/E′	6.5 ± 1.2	7.4 ± 1.9	0.04
Lateral E/E′	5.3 ± 0.9	5.9 ± 1.7	0.12
Average E/E′	5.9 ± 0.9	6.6 ± 1.6	0.02
LVMPI	0.31 ± 0.09	0.34 ± 0.07	0.06

BSA, body surface area; DT, deceleration time; IVRT, isovolumetric relaxation time; LVMPI, left ventricular myocardial performance index; PV, pulmonary venous; S/D, systolic to diastolic ratio.

Native T1 and extracellular volume

Forty-eight subjects in the CCS group and 25 controls received gadolinium. None of the patients or the controls had LGE. Native T1 did not differ between CCS and controls (Table 1, Figure 3). Extracellular volume was higher in CCS compared to controls (25.5 ± 7.0 vs. 20.7 ± 2.4%; P = 0.003, Figure 3) although most were within normal range: Twelve CCS (22%) had ECV equal or higher than 2 SDs above the mean ECV value for controls. Of the 25 CCS who had reduced LV EF, only 7 (28%) had elevated ECV compared to controls (31.2 ± 5.8 vs. 20.7 ± 2.4%; P = 0.001). In the rest of the CCS group with normal LV EF, 5 of these 23 subjects (22%) had elevated ECV compared to controls (37.7 ± 6.3 vs. 20.7 ± 2.4%; P = 0.001). These two subgroups with elevated ECV in the CCS group did not differ from each other (P = 0.13) and no significant correlations were found between LV EF and ECV. Male and female patients did not differ with regards to NT1 (999 ± 40 ms vs. 998 ± 32 ms, P = 0.5) or ECV (25.9 ± 8.3% vs. 25.0 ± 5.7%, P = 0.9). Within the CCS group, subjects with LVEF < 55% or LVEF ≥ 55% did not differ with regards to NT1 or ECV, neither did those who had received chest radiation therapy compared to those who had not. Native T1 and ECV did not correlate with RV or LV volumes and EFs (including their z-scores), cumulative anthracycline dose (Figure 4), age at cancer diagnosis, or time from last anthracycline exposure. Both NT1 and ECV correlated inversely with LV mass to volume ratio (r = −0.31, P = 0.02; r = −0.30, P = 0.04, respectively) as well as LVMi (r = −0.29, P = 0.03; r = −0.36, P = 0.01, respectively) (Figure 4). Intracellular mass indexed was lower in CCS compared to controls (Figure 5), but was not different between subjects with LVEF < 55% or LVEF ≥ 55%. It did not correlate with ventricular volumes, function, anthracycline dose, or time since chemotherapy. No parameter of diastolic function correlated with NT1, ECV, or ICMi.

Figure 3.

Left ventricular native T1 (A) and extracellular volume fraction (B) in childhood cancer survivors vs. controls. CCS, childhood cancer survivors; CTL, control; ECV, extracellular volume fraction; LVEF, left ventricular ejection fraction; NT1, native T1.

Figure 4.

Correlation of left ventricular mass indexed with native T1 (A) and ECV (B), left ventricular mass to volume ratio with native T1 (C) and ECV (D) and cumulative anthracycline dose with native T1 (E) and ECV (F) in childhood cancer survivors. ECV, extracellular volume fraction; LVMi, left ventricular mass indexed; LV Mass/Vol, left ventricular mass/volume ratio; NT1, native T1.

Figure 5.

Left ventricular intracellular mass indexed in childhood cancer survivors compared to controls. CCS, childhood cancer survivors; CTL, control; ECV, extracellular volume fraction; LV ICMi, indexed left ventricular intracellular mass.

Segmental strain and T1 mapping

A trend towards higher septal vs. LV FW NT1 (1023 ± 53 ms vs. 1012 ± 57 ms; P = 0.08) and higher ECV (26.6 ± 7.3% vs. 25.4 ± 7.7%; P = 0.04) were observed in CCS. In controls, NT1 was higher in the septum (1005 ms ± 40 vs. 985 ± 47 ms; P = 0.03). There were no differences in septal vs. LV FW circumferential or longitudinal strain in the CCS group. Neither regional CS nor longitudinal strain correlated with NT1 or ECV in the same segments.

Discussion

Paediatric cancer survivors who are exposed to anthracycline chemotherapy are at risk for developing ventricular dysfunction and some of progress to clinical heart failure.^7,8 The current study adds the following to our understanding of cardiac health after exposure to anthracyclines during childhood and adolescence:

1. In comparison to their peers, more than one-fifth of CCS treated with anthracycline chemotherapy show an expansion of their myocardial extracellular space relative to cellular space. It is possible these changes reflect the loss of myocytes with replacement fibrosis, although this needs further confirmation.

2. On average, children and adolescents exposed to anthracyclines have preserved systolic function, although LV EF and myocardial strain are lower than in controls.

3. Diastolic function is mostly comparable to controls, although some subtle differences with no known clinical relevance exist.

Native T1 and ECV by CMR are increasingly used as non-invasive metrics of myocardial injury and diffuse fibrosis.^18,25 Two previous T1 relaxometry studies in CCS did not demonstrate elevated NT1 or ECV.^26,27 However, both studies lacked a paediatric control group which had been examined using the same methodology which is a significant limitation. A current expert consensus statement recommends using the same methodology when patients and controls are compared. As a consequence, these two studies may not be optimally positioned to elicit differences in T1 and ECV between CCS and controls. In fact, we might have failed to demonstrate a difference in ECV between CCS and controls if we had used adult control data, generated on a different CMR system.²⁸ Findings from this study are in keeping with the experience in adult cancer survivors.²⁹ In a study by Jordan et al., ECV was elevated in anthracycline-treated adults compared with both pre-treatment cancer patients as well as controls. Interestingly, cancer patients who had not received anthracyclines did not differ in T1 and ECV from patients prior to chemotherapy. This and other studies suggest that anthracycline plays a key role in myocardial changes leading to a rise in ECV and NT1. Tham et al.²⁷ demonstrated an association between anthracycline dose and ECV in children treated with anthracyclines which was not found in our study. The lack of such a correlation in the present cohort of CCS could be related to the lack in statistical power or may suggest that other factors beyond dose modulate myocardial response to anthracyclines. In chronic processes, an increase in ECV relative to total myocardial space can reflect an expansion of the extracellular space, as seen in reactive fibrosis, or it can be due to cellular atrophy and myocyte death, leading to replacement fibrosis. Our findings of lower LVMi and ICMi in CCS suggest myocyte loss as the prevailing mechanism. In keeping with published work,^26,30 a trend towards higher septal vs. LV FW NT1 and ECV was observed in both groups, reaching statistical significance for NT1 in controls and for ECV in CCS, with no effect on regional myocardial deformation.

Significant overlap between disease and health exists with regards to candidate imaging fibrosis markers, namely NT1 and ECV, in most conditions, including in cancer patients. This overlap limits the applicability of NT1 and ECV as one-time tests for the identification of pathological myocardial remodelling, unless levels are very high. It is tempting to speculate, however, that a rise of these markers within the same individual during post-chemotherapy follow-up or in comparison with their baseline could indicate unfavourable remodelling. In fact, in anthracycline treated rabbits that were studied serially, a gradual increase in ECV preceded a decline in LV EF, indicating both the value of serial T1 mapping during follow-up and its potential as a predictor of functional decline.^31,32 The present cohort will be followed longitudinally to monitor these parameters. Currently, however, the clinical and prognostic significance of the elevated ECV remains to be investigated.

On average, systolic function was relatively preserved in CCS and any differences in comparison with the control group are likely clinically insignificant in our cohort. Nonetheless, the subtle reduction in LV EF and lower CS measurements could indicate the presence of subclinical changes. The predictive value of these findings is presently unclear, but it is possible that imaging evidence of dysfunction precedes clinical heart failure in certain patients. RV volumes and EF were similar to controls, in keeping with clinical observation and previous reports that identified the LV as more vulnerable to adverse remodelling following chemotherapy.²⁶

Diastolic function deteriorates before systolic function in many conditions.^33,34 Up to 60% of anthracycline-exposed patients develop echocardiographic changes in diastolic function.⁸ In the present cohort of CCS, we found a slightly increased mitral valve inflow deceleration time and higher E/E′ compared to controls, but, on average, these values are within normal range. We could only identify one patient with conclusive evidence of diastolic dysfunction. This patient also had the lowest GLS and global CS values and the highest ECV and had received one of the highest anthracycline doses. However, beyond this anecdotal observation, there was no association between metrics of cardiac function and either T1 or ECV, neither on a global nor on a regional level.

In a paediatric population, Tham and colleagues found a modest correlation between ECV and maximum oxygen consumption during cardiopulmonary exercise testing.²⁷ In adult cancer survivors, Neilan et al.³⁵ demonstrated an association between ECV and diastolic function parameters. However, in contrast to non-anthracycline cardiomyopathies,^36,37 a connection of CMR fibrosis markers with outcomes has not yet been established in cancer survivors.

Several limitations pertaining to the current study warrant discussion. First, although all patients with clinical outpatient encounters were invited to participate, a selection bias cannot be excluded. Second, despite this being the largest T1 mapping study in paediatric CCS, the relatively modest sample size may have obscured further differences and associations. Third, ECV and NT1 were obtained only at the mid-ventricular level. Although it is not unreasonable to assume that interstitial fibrosis, if present, would be homogeneously distributed throughout the LV, it is possible that certain areas are more affected than others. Furthermore, the lack of tissue specimens prevented a correlation of ECV and NT1 with the histological gold standard for myocardial fibrosis, although the association has been demonstrated in prior studies.^14,15 As a consequence, we can only hypothesize that myocyte loss is the prevailing mechanism explaining the elevation of ECV in CCS.

Conclusions

Our data suggest that in some CCS treated with anthracycline chemotherapy, mild changes in myocardial tissue properties characterized by increased myocardial extracellular volume and decreased intracellular mass, can be detected in the presence of overall preserved systolic and diastolic function. Further long-term follow-up of this cohort will clarify the possible clinical significance of these findings.