Deep learning identifies ECG biomarker linked to sudden cardiac death

A team of researchers applied deep‑learning techniques to electrocardiogram data to search for predictors of sudden cardiac death. Their analysis uncovered a distinct ECG biomarker that correlates with increased risk. The findings are reported in a Nature article (doi reference) published in 2026. The study used a large cohort of patient recordings to train and validate the model. Results suggest the biomarker could enable earlier identification of high‑risk individuals. The authors propose integrating the marker into routine ECG screening workflows. Clinical trials are needed to confirm its predictive power in broader populations. The discovery may influence future guidelines for cardiac risk assessment.