Artificial intelligence may soon be a frontline ally in spotting heart attacks. PMcardio’s Queen of Hearts algorithm just showed major promise in a large, multi-center U.S. validation study presented at the TCT 2025 conference and published in JACC: Cardiovascular Interventions.
The AI tool correctly identified 92 percent of true heart attacks on first ECG readings, a sharp jump from 71 percent with traditional triage. Even more impressive, it slashed false alarms from 42 percent to just 8 percent. That kind of precision could mean fewer unnecessary procedures and faster treatment for real emergencies.
Researchers at Harvard’s Beth Israel Deaconess Medical Center, UC Davis, and UTHealth Houston reviewed data from over a thousand emergency cases where heart attacks were suspected. The results suggest AI could dramatically improve both accuracy and speed, two critical factors in cardiac survival.
Dr. Timothy D. Henry from The Christ Hospital said the findings prove AI ECG analysis can help doctors diagnose faster, especially for patients transferred from small or rural hospitals. “Only 17 percent of those patients currently receive timely treatment,” he said.
The new study also builds on results from DIFOCCULT-3, a massive 6,000-patient clinical trial spanning 18 hospitals in Turkey. Early data suggest the AI-assisted group received care up to five hours sooner than standard methods, and short-term outcomes improved as a result.
PMcardio’s system is not a black box. Its visual explanation feature highlights which ECG regions drive each decision, making it easier for doctors to verify and trust the algorithm’s calls. With up to 40 percent of heart attacks showing atypical ECG patterns, that transparency could make a lifesaving difference.
