The Food and Drug Administration has authorized an artificial intelligence model that reads a standard 12-lead EKG and flags signs of the most urgent heart attacks, including patterns that conventional EKG criteria can miss.
The decision targets a known gap in emergency heart care. Doctors have long identified the most dangerous heart attacks, in which a coronary artery is completely blocked, by looking for a pattern on the EKG called ST-segment elevation. Patients with that pattern are sent quickly to a cardiac catheterization lab, where the blockage can be opened. But some people with a blocked artery do not show the textbook pattern, and their treatment can be delayed.
For a family in an emergency room with a relative who has chest pain, that early sorting decision shapes what happens next. The newly authorized tool is designed to help with it.
The Regulatory Step and Why It Is Unusual
The FDA granted de novo authorization to the model, which New York-based Powerful Medical markets as Queen of Hearts. The FDA's de novo database lists the device as the STEMI AI ECG Model, with a decision date of Sept. 3. The de novo pathway is used to classify new types of low- to moderate-risk devices that have no existing equivalent on the market.
That route is uncommon for AI. Most AI devices reach the market through the FDA's 510(k) clearance pathway, and fewer than 10 AI devices a year receive de novo classification, which typically requires more clinical evidence, STAT reported.
Robert Herman, the company's chief medical officer and co-founder, told TCTMD that it is "the first FDA De Novo authorization specifically for AI-ECG detection of ACS that requires urgent cardiologist review and intervention." ACS stands for acute coronary syndrome, an umbrella term that includes heart attacks. Herman said the intended use also covers triage, prioritization, and notifying the appropriate clinician or cardiology team.
The model was developed with input from Stephen W. Smith, an emergency physician at Hennepin County Medical Center in Minneapolis. It received FDA breakthrough device designation in March 2025 and obtained CE mark approval in Europe in 2022, according to TCTMD.
The Evidence Behind the Authorization
The main U.S. data come from a retrospective registry study across three emergency heart attack networks: Beth Israel Deaconess Medical Center in Boston, UC Davis in Sacramento, and UTHealth in Houston. Researchers reviewed more than 1,000 patients treated emergently for a suspected heart attack.
According to the company's announcement of the study results, the AI correctly identified 92% of true heart attacks on the first EKG, compared with 71% using standard triage. False alarms fell from about 42% to 8%. The findings were presented at the 2025 Transcatheter Cardiovascular Therapeutics conference and published in JACC: Cardiovascular Interventions.
Both results matter. Higher detection means fewer missed blockages. Fewer false alarms means fewer patients sent for an invasive catheterization they did not need, a procedure that carries its own risks.
Timothy D. Henry, director of clinical research at The Christ Hospital in Cincinnati, who presented the results, said the tool could matter most for "patients transferred from smaller or rural hospitals that don't perform emergency heart procedures, where only 17% currently receive timely treatment." Henry reports research support from Powerful Medical, according to TCTMD.
Powerful Medical says the model is supported by more than 20 peer-reviewed studies involving more than 40,000 patients. The company also points to DIFOCCULT-3, a randomized trial it says is enrolling about 6,000 patients at 18 hospitals in Turkey. The company has reported early data showing faster treatment in the AI-assisted group, but long-term results have not been published.
The Limits Worth Stating Plainly
Queen of Hearts is a triage aid, not a diagnosis, and it is not a consumer product. It runs within hospital and emergency care workflows on a clinical 12-lead EKG, and clinicians make the treatment decisions.
The U.S. validation data come from a retrospective registry review. Investigators compared what the AI would have flagged with what actually happened, rather than randomly assigning patients to AI-assisted or standard care. Several key figures above also come from the company's own announcements. Randomized evidence is still accumulating, and the largest trial cited by the company is taking place outside the United States.
Adoption is another open question. An extra alert helps only if staff act on it. Herman told TCTMD that "we already have multiple US health systems under contract, integrated, and ready to go live," but no public list of participating hospitals exists, and the authorization does not address which insurers will pay for the tool.
MedicalDaily has reported on other efforts to bring AI into heart care, including federally funded AI agents for heart failure care. The same caution applies here. Authorization means a tool met a regulatory standard. It does not guarantee better outcomes at every hospital that uses it.
Chest Pain Guidance That Has Not Changed
Nobody should wait for an AI-equipped hospital, ask for this tool by name, or delay care to find one. The steps that protect heart muscle are the same as before.
Chest pressure, tightness, or pain; pain spreading to the arm, jaw, neck, or back; shortness of breath; cold sweat; nausea; or sudden lightheadedness are reasons to call 911 rather than drive. Emergency medical crews can record an EKG on the way and alert the hospital before arrival, which can save valuable time. Women, older adults, and people with diabetes are more likely to have less typical symptoms, such as fatigue, indigestion-like discomfort, or shortness of breath without chest pain.
People with known coronary artery disease can ask their cardiologist whether the nearest hospital performs emergency catheterization and how transfers work if it does not. That question is worth asking regardless of the software a hospital uses.
The next steps are deployment, payment decisions, and published randomized results. MedicalDaily will report where the tool goes live and what the randomized data show.
Key Questions Answered
What did the FDA authorize? A de novo authorization for Queen of Hearts, an AI model from Powerful Medical that analyzes a 12-lead EKG to flag suspected acute coronary syndrome needing urgent cardiology review.
Why is de novo significant? It is used for new types of devices with no existing equivalent and typically requires more clinical evidence. STAT reported that fewer than 10 AI devices a year receive it.
How accurate was it? In a retrospective U.S. registry study of more than 1,000 patients, the company reported that the AI caught 92% of true heart attacks on the first EKG versus 71% with standard triage, and false alarms fell from about 42% to 8%.
Can patients use this themselves? No. It runs within hospital and emergency care workflows on a clinical EKG, and clinicians make the treatment decisions.
Which patients may benefit most? Researchers point to patients at hospitals that do not perform emergency heart procedures and who need a transfer.
Is it available at my hospital? Unknown. The company says several U.S. health systems are ready to launch it, but no public hospital list exists.
Does this change what I should do for chest pain? No. Call 911 rather than driving, so emergency crews can begin care and alert the hospital early.
