Baseline left atrial reservoir strain provides independent and incremental prognostic value in predicting chemotherapy related cardiotoxicity

Abstract

Abstract Background Currently, few baseline imaging biomarkers can predict chemotherapy induced cardiotoxicity (CTx). The early identification of CTx is critically important as it determines the prospect of LV function recovery (1). Purpose We aimed to determine if baseline left atrial (LA) reservoir strain (LAS) is predictive of CTx. Methods We performed retrospective analysis of the SUCCOUR study, an international multicentre randomized controlled trial. CTx was defined as an asymptomatic drop of >10% in left ventricular ejection fraction (LVEF) compared to baseline to <55%. LAS analysis was performed using semi-automated speckling tracking of the LA in the four and two chamber views using EchoPAC. Analysis was performed at one year follow up. Cox proportional hazard analysis, c-statistics and Akaike information criterion (AIC) statistical analysis was performed. Results After excluding 78 with inadequate images, 229 were included in the analysis. Most participants were female (n=215, 94%), with a mean age of 54±12 years. Median follow up was 1.02 years (IQR 0.98–1.07). A smoking history was the commonest cardiac risk factor (n=68, 30%). Breast cancer was the main cancer type (n=205, 90%). The mean baseline 3D LVEF, LV global longitudinal strain (GLS), LAS, and LA volume index (LAVI) were 61±4%, 20.6±2.5%, 28.6±7.9%, and 28±9ml/m2, respectively. Participants that developed CTx had a lower LAS at baseline compared to those that did not (25.5±7.4% vs 29.0±7.9% respectively; p=0.027). At 1-year follow-up, 29 patients (13%) developed CTx. On univariable Cox proportional hazard analysis, higher baseline LAS was associated with a lower risk of CTx (HR 0.95, 95% CI 0.90–0.99; p=0.026) but other parameters including baseline GLS was not (p=0.17). In nested Cox models (Figure 1), adding LAS significantly improved the model's predictive accuracy for CTx compared to other clinical and imaging parameters. When LAS was added to the same models it showed improvements in the c-statistics and AIC score (Figure 1). Conclusion Baseline LAS is an independent and incremental predictor of CTx. LAS could be used for baseline risk stratification in oncology patients who receive potentially cardiotoxic chemotherapy. Funding Acknowledgement Type of funding sources: Private grant(s) and/or Sponsorship. Main funding source(s): GE Healthcare