Extracellular Volume Fraction of Cardiac MRI in the Detection of Cardiac Toxicity from Radiotherapy

Abstract

Extracellular volume (ECV) fraction, a cardiac MRI (CMR) marker of myocardial tissue remodeling and fibrosis, is analyzed in this proof of concept study to detect myocardial fibrosis in patients with previous thoracic radiotherapy. Twenty-eight patients with no cardiac history and chest tumors treated with radiotherapy, with a heart dose V5Gy ≥ 10%, underwent CMRs (1.5T, Siemens Aera, Erlangen, Germany) on a cross sectional study protocol. Estimation of the ECV (extracellular matrix) required measurement of T1 mapping values before and after administration of gadolinium-based contrast agents, in addition to the patients’ hematocrit value at the time of scan. Cardiac dosimetric variables were gathered from cumulative dose-volume histograms of delivered radiotherapy plans. Mann-Whitney-U tests were performed between ECV fractions and various clinical variables. Spearman’s correlation coefficients were calculated between dosimetric variables and ECV fractions at the left ventricle (LV) base, mid, and apex. Mean time from radiotherapy to CMR was 46.4 months (range 1.7-344.5). Mean V5Gy heart volume was 43.1% (range 10-97%). The cohort consisted of 14 (50%) patients with lung, 9 (32%) with breast, and 5 (18%) with other primary tumors. The median ECV at LV base, mid, and apex was 28% (25-30%), 27% (25-29%), and 30% (27-32%), respectively. Abnormal LV ECV fractions above reported reference range (25.3 ± 3.5%)1 included 8 patients (28.6%) at base, 8 (28.6%) at mid, and 15 (53.6%) at apex. Eighteen patients (64.2%) demonstrated at least 1 abnormal ECV value. There were trends/significant correlations between cardiac LV doses (V10Gy/V20Gy/V30Gy/V40Gy) and ECV fractions at mid LV (p = 0.095, 0.045, 0.043, 0.035, respectively) and at base (p = 0.14, 0.12, 0.043, 0.083, respectively). There were also trends/significant correlations between cardiac LV max dose/mean dose and ECV fractions: at mid ventricle (p = 0.016 and 0.07) and at base (p = 0.068 and 0.072). There were no dosimetric correlations to LV apex ECV fraction. Smoking history as measured in pack years was also correlated with higher ECV fraction at the LV base/mid/apex (p = 0.13/0.03/0.04), consistent with prior literature. There was no significant difference between ECV fractions and other clinical factors, such as chemotherapy use or hypertension diagnosis. Approximately ∼65% of thoracic radiotherapy patients have abnormal ECV fraction values, but this is confounded by preexisting clinical factors such as smoking. There were correlations between higher LV dose and greater ECV fraction size at the LV base and mid LV. This may represent imaging measurement of radiation-induced myocardial fibrosis. Further prospective study is needed to validate this finding.