Early detection of subclinical left ventricular dysfunction after breast cancer radiation therapy using speckle-tracking echocardiography: association between cardiac exposure and longitudinal strain reduction (BACCARAT study)

Abstract

Abstract Background Breast cancer (BC) radiotherapy (RT) can induce cardiotoxicity, with adverse events often observed many years after BC RT. Subclinical left ventricular (LV) dysfunction can be detected early after BC RT with global longitudinal strain (GLS) measurement based on 2D speckle-tracking echocardiography. This 6-month follow-up analysis from the BACCARAT prospective study aimed to investigate the association between cardiac radiation doses and subclinical LV dysfunction based on GLS reduction. Methods The patient study group consisted of 79 BC patients (64 left-sided BC, 15 right-sided BC) treated with RT without chemotherapy. Echocardiographic parameters, including GLS, were measured before RT and 6 months post-RT. The association between subclinical LV dysfunction, defined as GLS reduction > 10%, and radiation doses to whole heart and the LV were performed based on logistic regressions. Non-radiation factors associated with subclinical LV dysfunction including age, BMI, hypertension, hypercholesterolemia and endocrine therapy were considered for multivariate analyses. Results A mean decrease of 6% in GLS was observed (− 15.1% ± 3.2% at 6 months vs. − 16.1% ± 2.7% before RT, p = 0.01). For left-sided patients, mean heart and LV doses were 3.1 ± 1.3 Gy and 6.7 ± 3.4 Gy respectively. For right-sided patients, mean heart dose was 0.7 ± 0.5 Gy and median LV dose was 0.1 Gy. Associations between GLS reduction > 10% (37 patients) and mean doses to the heart and the LV as well as the V20 were observed in univariate analysis (Odds Ratio = 1.37[1.01–1.86], p = 0.04 for Dmean Heart; OR = 1.14 [1.01–1.28], p = 0.03 for Dmean LV; OR = 1.08 [1.01–1.14], p = 0.02 for LV V20). In multivariate analysis, these associations did not remain significant after adjustment for non-radiation factors. Further exploratory analysis allowed identifying a subgroup of patients (LV V20 > 15%) for whom a significant association with subclinical LV dysfunction was found (adjusted OR = 3.97 [1.01–15.70], p = 0.048). Conclusions This analysis indicated that subclinical LV dysfunction defined as a GLS decrease > 10% is associated with cardiac doses, but adjustment for non-radiation factors such as endocrine therapy lead to no longer statistically significant relationships. However, LV dosimetry may be promising to identify high-risk subpopulations. Larger and longer follow-up studies are required to further investigate these associations. Trial registration ClinicalTrials.gov: NCT02605512, Registered 6 November 2015 - Retrospectively registered