Associations between echocardiographic traits and AI-ECG predictions of heart failure
The study investigated if predictions made by an AI-ECG model for heart failure risk match traditional measurements from echocardiograms. The findings showed that global longitudinal strain (GLS) is the strongest predictor, followed by other systolic measures. Importantly, these correlations are particularly strong in patients with preserved left ventricular ejection fraction (LVEF), suggesting the AI captures diastolic dysfunction features.