Accurate and rapid identification of heart attacks remains critical for improving survival in emergency settings. Published in JACC: Cardiovascular Interventions, this multicenter U.S. registry study evaluated the diagnostic performance of artificial intelligence–based electrocardiogram analysis for STEMI triage.

The analysis included 1,032 patients who triggered emergent cardiac catheterization laboratory activation between January 2020 and May 2024. The artificial intelligence model was trained to distinguish true occlusions from benign electrocardiogram mimics.

Confirmed STEMI was present in 601 patients (58.2%). Artificial intelligence analysis achieved higher sensitivity than standard triage (92.0% vs. 71.0%) and markedly reduced false-positive activations (7.9% vs. 41.8%). Specificity improved from 29.0% to 81.0%, with an AUC of 0.94. Performance remained consistent across complex cases, including atrial fibrillation and bundle branch block.

By accurately identifying true infarctions and reducing unnecessary activations, enhanced ECG interpretation strengthens real-time decision-making in emergency settings. It supports faster reperfusion and more efficient use of cardiac catheterization resources.