Detection of Subclinical Cardiovascular Disease by Cardiovascular Magnetic Resonance in Lymphoma Survivors

Background Long-term survivors of Hodgkin lymphoma (HL) and mediastinal non-Hodgkin lymphoma experience late adverse effects of radiotherapy and/or anthracycline-containing chemotherapy, leading to premature cardiovascular morbidity and mortality. Objectives The aim of this study was to identify markers for subclinical cardiovascular disease using cardiovascular magnetic resonance (CMR) in survivors of HL and non-Hodgkin lymphoma. Methods CMR was performed in 80 lymphoma survivors treated with mediastinal radiotherapy with or without anthracyclines, and results were compared with those among 40 healthy control subjects matched for age and sex. Results Of the 80 lymphoma survivors, 98% had histories of HL, the mean age was 47 ± 11 years, and 54% were male. Median radiotherapy dose was 36 Gy (interquartile range: 36-40 Gy), and radiotherapy was combined with anthracyclines in 70 lymphoma survivors (88%). Mean time between diagnosis and CMR was 20 ± 8 years. Significantly lower left ventricular (LV) ejection fraction (53% ± 5% vs 60% ± 5%; P < 0.001) and LV mass (47 ± 10 g/m2 vs 56 ± 8 g/m2; P < 0.001) and higher LV end-systolic volume (37 ± 8 mL/m2 vs 33 ± 7 mL/m2; P = 0.013) were found in lymphoma survivors. LV global strain parameters were also significantly worse in lymphoma survivors (P < 0.02 for all). Native myocardial T1 was significantly higher in lymphoma survivors compared with healthy control subjects (980 ± 33 ms vs 964 ± 25 ms; P = 0.007), and late gadolinium enhancement was present in 11% of the survivors. Conclusions Long-term lymphoma survivors have detectable changes in LV function and native myocardial T1 on CMR. Further longitudinal studies are needed to assess the implication of these changes in relation to treatment and clinical outcome.

I n recent decades, improvements in the treatment of patients with Hodgkin lymphoma (HL) and non-Hodgkin lymphoma (NHL) have resulted in an increased survival rate of approximately 80% to 90%.
Nevertheless, long-term survivors may experience several late adverse effects that may lead to premature morbidity and mortality (1)(2)(3). Cardiovascular diseases (CVDs), such as myocardial dysfunction, heart failure, myocardial infarction, and valvular disease, are among the most important adverse effects of mediastinal radiotherapy and/or anthracycline-containing chemotherapy, which can develop even decades after treatment initiation (1,(4)(5)(6). However, lymphoma survivors may have CVD without having clinical symptoms, and normal left ventricular (LV) ejection fraction does not preclude the presence of myocardial dysfunction (7,8). To detect early stages of late cardiotoxicity, screening for subclinical CVD through appropriate monitoring, including noninvasive cardiac imaging on transthoracic echocardiography or cardiovascular magnetic resonance (CMR), should be considered (1,9). Both transthoracic echocardiography and CMR are safe, nonionizing imaging modalities that are able to assess myocardial and valvular function. In addition, CMR has the unique ability to characterize tissue. Late gadolinium enhancement (LGE) imaging and mapping techniques enable the detection of (diffuse) myocardial fibrosis and/or edema (1,9,10). Limited data exist on CMR abnormalities associated with subclinical CVD in survivors of HL and NHL (5). Therefore, in this study we aimed to identify markers of subclinical CVD using CMR in survivors of HL and NHL.

RESULTS
A total of 80 consecutive HL (n ¼ 78) and mediastinal NHL (n ¼ 2) survivors and 40 healthy control subjects matched for age and sex on a group level were included in this study ( Figure 1). Baseline characteristics of both groups are summarized in Table 1.
Overall, the mean age was 47 AE 11 years, and 53% were male.
Most lymphoma survivors had a history of stage II disease (74%), and mean age at diagnosis was 27 AE  Table 1.
Electrocardiographic and echocardiographic parameters are shown in Table 2. There was a low prevalence of electrocardiographic abnormalities. Grade II LV diastolic dysfunction and any hemodynamically significant (moderate or greater) valvular disease were found in 4 (5%) and 10 (13%) of the survivors, respectively.
Differences in CMR characteristics between lymphoma survivors and healthy control subjects are shown in Table 3   LGE quantification 3.0% [interquartile range: In a sensitivity analysis in which anthracycline dose and time between diagnosis and CMR were also included in the model (multivariable model 2), the

(4)
Continued on the next page  Tables 4 to 7). As is known from published research, LV mass may decline following anthracycline administration, with an inverse relationship between dose and LV mass (10,18). The exact pathophysiology behind this process is not fully elucidated, but it has been hypothesized that anthracyclines causes injury to the cardiomyocytes, resulting in extracellular remodeling, atrophy, or apoptosis (19). We also observed reduced LV mass and reduced cell volume with normal extracellular volume in lymphoma survivors, which might suggest cardiomyocyte atrophy. Of note, although LV mass in our lymphoma survivors was lower than in control subjects, it was still within the  normal range (20). Analysis showed no significant relationship between LV mass and high-dose anthracyclines. Although it cannot be excluded that this lack of relationship was due to power, our multivariable analysis showed a significant inverse relationship between LV mass and total mediastinal radiotherapy dose. This assumption is supported by the findings of Adams et al (21), which showed significantly lower LV mass in HL survivors who had been treated with mantle-field radiotherapy. However, it is difficult to  (26,27). Although it is also difficult to determine the influence of both therapies separately here, total mediastinal radiotherapy dose was not found to be an independent predictor of GLS, and neither radiotherapy nor anthracyclines were associated with GCS. Nevertheless, we can conclude that LV function and strain parameters are markers of the detection of subclinical CVD and should be incorporated as standard parameters in clinical protocol before, during, and after cancer therapy (8).

DISCUSSION
Native myocardial T1 was significantly higher in lymphoma survivors compared with healthy control subjects. This may suggest that diffuse myocardial fibrosis is actually present in lymphoma survivors, but it may still be explained by partial volume effect due to decrease of the amount of myocardial tissue. Extracellular volume fraction, which is associated with native T1, was not increased in lymphoma survivors, although our control subjects also demonstrated higher than expected myocardial extracellular volume. However, research has shown that extracellular volume fraction has a lower discriminatory performance compared with native T1 (28). Previous studies that demonstrated elevated native myocardial T1 and/or extracellular volume fractions in patients treated with anthracyclines were either conducted at the time of anthracycline treatment or did not concern only lymphoma survivors (24,29).
Subendocardial and transmural LGE was present in 5% of lymphoma survivors, which was most likely consistent with myocardial infarction. This can be a consequence of coronary artery disease due to treatment with radiotherapy. In 6% of the survivors, the LGE pattern was midmyocardial or epicardial. No differences were found in the presence and localization of LGE between higher and lower doses of radiotherapy and/or anthracyclines. LGE at the hinge points was also prevalent in our cohort, although this is not specific for late cardiotoxicity. Hinge-point LGE is rather common in the (older) general population but also in specific populations such as athletes, making it difficult to identify the cause of this nonischemic LGE pattern.
Cardio-oncology is a relatively new field and still growing because of increased survival rates. More  have been performed to detect subclinical CVD using CMR in this study population (24,29). Therefore, LGE quantification (% of left ventricle