Abstract 13581: Prognostic Value of Left Ventricular Global Longitudinal Strain for Early Prediction of Cancer Therapy-Related Cardiac Dysfunction

Abstract

Introduction: This prospective study aimed to determine if MRI-imaging and Artificial Intelligence (AI)-based computation of global longitudinal strain (GLS) in the left ventricle (LV) is an early predictor of cancer therapy-related cardiac dysfunction (CTRCD) in breast cancer patients. Hypothesis: To show with risk analysis that GLS is an independent prognosticator of CTRCD when compared to LV ejection fraction (LVEF) with the patient cohort administered antineoplastic chemotherapy treatment. Methods: Motion-encoded data with Displacement Encoding with Stimulated Echoes (DENSE) MRI were acquired on 32 breast cancer patients at baseline and 3- and 6-months follow-ups after chemotherapy. Automated image segmentation for chamber quantification and phase-unwrapping in the LV occurred with two DeepLabV3+ fully convolutional networks (FCNs), and 3D strains computed with the meshfree Radial Point Interpolation Method. CTRCD (including adverse cardiac events and cardiotoxicity) risk analyses with the measured clinical, chamber quantification, and strain parameters were conducted with univariable and bivariable Cox proportional hazard (PH) regressions. Results: GLS worsened from baseline to the 3- and 6-months follow-ups (-19.1 ± 2.1% vs. -16.0 ± 3.1% vs. -16.1 ± 2.9%, P<0.001) (Fig. 1). With univariable Cox regression, the 3-months follow-up GLS was independently associated as an agonist (hazard ratio [HR]-per-SD: 2.1; 95% CI: 1.4-3.1, P<0.001) and LVEF as protector (HR-per-SD: 0.8; 95% CI: 0.7-0.9, P<0.001) for CTRCD occurrence. In bivariable regression, GLS (HR-per-SD: 2.0; 95% CI: 1.2-3.4, P=0.01) was an independent predictor of CTRCD regardless of LVEF (HR-per-SD: 1.0; 95% CI: 0.9-1.2, P=0.9) as covariate. Conclusions: The end-point risk analyses show that LV GLS is an early CTRCD predictor independent of LVEF in breast cancer patients treated with antineoplastic chemotherapy. GLS should be a standard metric for the clinical diagnosis of CTRCD.