Authors of a study published in Heart Rhythm aimed to develop electrocardiogram (ECG) deep learning (DL) models to identify high risk imaging features in patients with HCM, potentially reducing the need for resource-intensive cardiovascular magnetic resonance (CMR) imaging. Richard T. Carrick, MD, PhD, and colleagues used data from 1,930 patients with HCM to create ECG-DL models predicting systolic dys function, massive hypertrophy, apical aneurysm, and extensive late gadolinium enhancement. These models were externally validated with 233 patients from a HCM center in India. The ECGDL models demonstrated reliable identification of high-risk features during testing (c-statistic, 0.72-0.93) and validation (c-statistic, 0.68-0.91). A screening strategy combining echocardiography and ECG-DL guided CMR reduced CMR recommendations by 61% while maintaining 97% sensitivity for detecting high-risk features. The negative predictive value for the absence of highrisk features without ECG-DL recommendation was 99.5%. The researchers concluded that ECG-DL models can effectively identify high risk HCM features.