Novel AI ECG model outperforms standard triage for acute coronary occlusion

Using artificial intelligence (AI) to analyze electrocardiograms (ECG) improved detection of severe heart attacks, including those that presented with unconventional symptoms, or atypical ECG patterns, and reduced false positives, according to a study published in JACC: Cardiovascular Interventions and simultaneously presented at TCT 2025 in San Francisco.

ST-segment elevation myocardial infarction (STEMI) is a severe type of heart attack where a major coronary artery is blocked, preventing blood flow to the heart muscle. Quickly restoring blood flow, or reperfusion, using percutaneous coronary intervention is the standard of care; however, delays in achieving the guideline-recommended time to reperfusion still persist, especially at hospitals and centers not specializing in PCI and in rural areas. Time to reperfusion longer than 90 minutes is associated with threefold higher rates of mortality.

"AI-driven ECG interpretation can bring the best of both worlds – identify true heart attacks early while reducing unnecessary activations," said Robert Herman, MD, PhD, lead author of the study and a cardiovascular researcher at AZORG Hospital in Aalst, Belgium.

Improving the accuracy of triage at the first medical contact can streamline emergency care, reduce fatigue and strain on clinical teams, and ensure that patients who truly need urgent intervention receive it without delay."

Robert Herman, Cardiovascular Researcher, AZORG Hospital

In one of the first large, real-world evaluations of an AI–based ECG model for STEMI triage in the emergency setting, researchers retrospectively looked at 1,032 patients with suspected STEMI who triggered emergency reperfusion protocols. Data was from three geographically diverse primary PCI centers between January 2020 and May 2024. Each patient's initial ECG underwent analysis by the STEMI AI ECG Model (Queen of Hearts) trained to detect acute coronary occlusion, including STEMI equivalents and differentiate from benign mimics.

Angiography and biomarkers confirmed that 601 (58%) were STEMIs and 431 (42%) were false positives. The AI ECG model did better than standard triage, detecting 553 of 601 confirmed STEMIs vs. 427 detected by standard triage on the initial ECG. AI ECG had a false positive rate of 7.9% vs. 41.8% for standard triage, representing a fivefold reduction.

"These results indicate that AI-enhanced STEMI diagnosis at the first medical contact has the potential to shorten time to treatment and reduce false activations," said Timothy D. Henry, MD, FACC, senior author of the study, The Carl and Edyth Lindner Family Distinguished Chair in Clinical Research and Medical Director of The Carl and Edyth Lindner Center for Research and Education at The Christ Hospital in Cincinnati. "This technology may be especially valuable in optimizing the transfer of STEMI patients from non-PCI centers to ensure timely and appropriate care."

In an accompanying editorial comment, Mohamad Alkhouli, MD, MBA, cardiologist at the Mayo Clinic, said the researchers should be "commended for developing an operational AI model aimed at addressing one of the most complex and error-prone aspects of interventional cardiology practice-STEMI activation."

However, he emphasized that the AI model employed in the study should be interpreted with caution, as it was originally developed to detect occluded arteries rather than STEMI and necessitates further prospective validation across diverse patient populations.

"The true challenge is not proof of accuracy alone, but readiness-to integrate, regulate, and interpret AI as a complement to human judgment, particularly in high-stakes, time-sensitive clinical settings," Alkhouli said.