Artificial intelligence prediction of age from echocardiography as a marker for cardiovascular disease

While chronological age is a universal risk predictor across most populations and diseases, distinguishing between biologically older from younger individuals may identify individuals with accelerated or delayed cardiovascular aging. This study presents a deep learning model to predict age from echocardiogram videos, leveraging 2,610,266 videos from 166,508 studies from 90,738 patients. Leveraging multi-view echocardiography, our model achieved a mean absolute error (MAE) of 6.76 (6.65-6.87) years and a coefficient of determination (R²) of 0.732 (0.72-0.74) on Cedars-Sinai Medical Center test set. Consistent performance was observed across four external validation cohorts. Predictions were associated with increased risk of coronary artery disease, heart failure, and stroke, and captured discontinuities before and after a heart transplant. Guided back propagation highlighted the model's focus on the mitral valve, mitral apparatus, and basal inferior wall, underscoring the potential of computer vision-based assessment of echocardiography in enhancing cardiovascular risk assessment and aging.