A new study has shown that artificial intelligence (AI) can help doctors better detect serious heart attacks by analyzing ECGs, or electrocardiograms.
The AI system improved accuracy, especially for heart attacks that don’t always show up clearly on ECGs.
It also helped reduce the number of false alarms, according to research published in the Journal of the American College of Cardiology: Cardiovascular Interventions and presented at TCT 2025 in San Francisco.
The type of heart attack studied is called STEMI, or ST-segment elevation myocardial infarction. This is a dangerous condition where a major blood vessel in the heart gets blocked, stopping blood flow and putting the heart muscle at risk.
The best treatment is to quickly reopen the artery using a procedure called percutaneous coronary intervention (PCI). Doctors aim to do this within 90 minutes of diagnosis. Any delay increases the risk of death by three times.
Dr. Robert Herman, the lead author of the study and a heart expert at AZORG Hospital in Belgium, explained why the study is important. He said that using AI can help spot true heart attacks earlier while reducing unnecessary emergency responses. This helps medical teams focus on real emergencies and speeds up care for those who truly need it.
This was one of the first large studies in real-world settings to see how well AI could work in emergency rooms. Researchers looked at 1,032 patients who had symptoms suggesting a STEMI and triggered emergency heart attack responses. The data came from three PCI hospitals in different parts of the world, collected between January 2020 and May 2024.
Each patient’s ECG was analyzed by a new AI system called the STEMI AI ECG Model, also known as Queen of Hearts. The model was trained to find blocked arteries, even in patients with unusual ECG results. It also helped tell the difference between true heart attacks and harmless conditions that look similar.
In total, 601 patients (58%) actually had a STEMI, confirmed by heart scans and blood tests. The remaining 431 patients (42%) were false alarms. The AI model identified 553 of the 601 true heart attacks—much better than the standard method, which caught only 427.
Even more impressively, the AI system had a false alarm rate of just 7.9%, compared to 41.8% with standard triage. That’s a five-fold reduction in unnecessary emergency responses.
Dr. Timothy D. Henry, one of the senior authors and a heart specialist in Cincinnati, said this AI tool could be especially helpful in hospitals that don’t perform PCI. It could help doctors quickly figure out who needs to be moved to a PCI center, ensuring faster and more accurate treatment.
Dr. Mohamad Alkhouli, a cardiologist at the Mayo Clinic, praised the study in an editorial. He said the researchers did a great job creating a useful AI tool for one of the trickiest parts of treating heart attacks—deciding when to activate emergency heart care teams. Still, he cautioned that the AI model needs more testing in different groups of patients.
He added that using AI in medicine isn’t just about proving it works. It’s also about being ready to use it properly, safely, and alongside human judgment—especially when every second counts.
In conclusion, this study shows that AI could play a big role in the future of emergency heart care. By helping doctors make quicker and more accurate decisions, AI could save lives and reduce stress on hospitals.
If you care about heart disease, please read studies that herbal supplements could harm your heart rhythm, and how eating eggs can help reduce heart disease risk.
For more health information, please see recent studies that apple juice could benefit your heart health, and results showing yogurt may help lower the death risks in heart disease.
Copyright © 2025 Knowridge Science Report. All rights reserved.
