DL-LVEF: Deep-Learning Measurement of the Left Ventricular Ejection Fraction from Echocardiographic Images

Left ventricular ejection fraction (LVEF) is a commonly used index of cardiac functionality. Thus, accuracy in its measurement is fundamental. LVEF measure is usually manually performed by clinicians from echocardiographic images. Use of automatic algorithms could make LVEF measurement more objective. Thus, the aim of the present work is to present DL-LVEF, a new automatic algorithm for LVEF measurement based on deep learning identification and segmentation of the left ventricular endocardium performed by combining the YOLOv7 algorithm and a U-Net. To this aim, the CAMUS database was used, which includes 1800 echocardiographic images acquired from 450 patients with annotated LVEF values and manual segmentation of the left ventricular endocardium. The database was divided into training dataset (70%) and testing dataset (30%). In both datasets, measured and annotated LVEF values (%) were found to be highly correlated (p=0.96 and p=0.89, respectively) and not statistically different (52.6% vs. 52.6% and 54.6% vs. 53.9%, respectively); mean absolute error was 4% and 5%, respectively. Thus, DL-LVEF provided objective and accurate LVEF measurement. Future DL-LVEF evolutions will also provide segmentation of other cardiac anatomical structures and, thus, will allow measurement of other clinically relevant cardiac indexes.