2D-echocardiography vs cardiac MRI strain using deep learning: a prospective cohort study in patients with HER2-positive breast cancer undergoing trastuzumab

Abstract

Abstract Funding Acknowledgements Type of funding sources: Private hospital(s). Main funding source(s): Promotiefonds Albert Schweitzer Hospital Background Subclinical cardiotoxicity due to trastuzumab could be recognized by repeated measurements of global longitudinal/radial strain (GLS/GRS) which could be performed with cardiac magnetic resonance (CMR) or two-dimensional speckle tracking echocardiography (2DE-ST). Although CMR is the gold standard for cardiac function evaluation, it is not used for cardiac monitoring. 2DE-ST might be a reasonable alternative. Purpose To study the predictive value of early 2DE-ST for left ventricular ejection fraction (LVEF) changes during trastuzumab for HER2-positive breast cancer. Methods HER2-positive breast cancer patients receiving trastuzumab, with or without anthracycline, underwent 2DE-ST at baseline and after 3 and 6 months (m) trastuzumab. Cardiac magnetic resonance (CMR) imaging (with ST) was performed at baseline and 6m. We studied the correlation between 2DE-ST- and CMR-derived GLS and GRS. We then associated baseline and 3m 2DE-ST with later CMR-LVEF, and with cardiotoxicity, defined as CMR-LVEF <45% and/or absolute decline >10% during trastuzumab. Results 47 patients were included. GLS measurements based on 2DE-ST and CMR showed weak correlation (Pearson’s r = 0.33; P = 0.041); GRS measurements were uncorrelated (r = 0.09; P = 0.979). 2DE-LVEF at baseline and 3m, and 2DE-STE-GLS at 3m were predictive of CMR-LVEF at 6m (Table 1). In contrast, the change in 2DE-ST-GLS at 3m was predictive of the change in CMR-LVEF at 6m, whereas the change in 2DE-LVEF was not. Importantly, the 11 patients (28%) who developed cardiotoxicity had larger 2DE-ST-GLS change at 3m than those who did not (median 5.2% versus 1.7%; odds ratio 1.81, 95% confidence interval 1.11–2.93; P = 0.016). Conclusion Although correlations between 2DE-ST and CMR are weak, ST-measurements appeared useful to improve the performance of 2DE in predicting LVEF changes after 6m of trastuzumab. Table 1 2DE CMR-LVEF after 6m TZT Mean difference (95% CI) Change in CMR-LVEF after 6m TZT Mean difference (95% CI) Cardiotoxicity OR (95% CI) Before TZT LVEF, % 0.85 (0.42, 1.27)* 0.32 (-0.16, 0.80) 0.88 (0.75, 1.02) ST-GLS, % -0.42 (-1.31, 0.46) -0.28 (-1.14, 0.58) 1.13 (0.87, 1.46) 3 Months TZT LVEF, % 0.59 (0.30, 0.88)* 0.29 (-0.04, 0.61) 0.85 (0.74, 0.98)* ST-GLS, % -1.14 (-2.07, -0.19)* -0.62 (-1.54, 0.30) 1.36 (0.94, 1.84) Change at 3 months TZT LVEF, % 0.30 (-0.11, -0.71) 0.21 (-0.19, 0.61) 0.90 (0.80, 1.01) ST-GLS, % -1.17 (-2.14, -0.20)* -1.10 (-2.02, -0.18)* 1.81 (1.11, 2.93)* * P-value <0.05