Prediction of Coronary Artery Calcium Using Deep Learning of Echocardiograms.

Researchers

Journal

Modalities

Models

Abstract

Coronary artery calcification (CAC), often assessed by computed tomography (CT), is a powerful marker of coronary artery disease (CAD) that can guide preventive therapies. CTs, however, are not always accessible or serially obtainable. It remains unclear if other widespread tests such as transthoracic echocardiograms (TTEs) can be used to predict CAC.Using a dataset of 2881 TTE videos paired with coronary calcium CTs, we trained a video-based artificial intelligence (AI) convolutional neural network to predict CAC scores from parasternal long axis (PLAX) views. We evaluated the model’s ability to classify patients from a held-out sample as well as an external site sample into zero CAC and high CAC (CAC ≥400 Agatston units) groups by receiver operating characteristic (ROC) and precision-recall curves. We also investigated whether such classifications prognosticated significant differences in 1-year mortality rates by log-rank test of Kaplan-Meier curves.TTE AI models had high discriminatory abilities in predicting zero CAC (ROC AUC 0.81 (95% CI 0.74-0.88), F1 0.95) and high CAC (AUC 0.74 (0.68-0.8), F1 0.74). This performance was confirmed in an external test dataset of 92 TTEs ((AUC 0.75 (0.65-0.85), F1 0.77), (AUC 0.85 (0.76-0.93), F1 0.59), respectively). Risk stratification by TTE-predicted CAC performed similarly to CT CAC scores in prognosticating significant differences in 1-year survival in high CAC patients (CT CAC≥400 vs. CT CAC<400 p=0.03, TTE-predicted CAC≥400 vs. TTE-predicted CAC<400 p=0.02).A video-based deep learning model successfully used TTE videos to predict zero CAC and high CAC with high accuracy. TTE-predicted CAC prognosticated differences in 1-year survival similar to CT CAC. Deep learning of TTEs holds promise for future adjunctive CAD risk stratification to guide preventive therapies.Copyright © 2022. Published by Elsevier Inc.