Powerful Medical wins FDA De Novo for Queen of Hearts heart-attack ECG AI

Powerful Medical, Inc. has received U.S. Food and Drug Administration (FDA) De Novo authorization for its STEMI AI ECG Model, commercially associated with its Queen of Hearts technology, creating a new Class II regulatory pathway for artificial-intelligence software intended to help identify acute myocardial infarction patterns requiring urgent cardiology review. FDA records show application DEN250044 was granted on September 3 after nearly a year of review, with the agency assigning the new cardiovascular product code SHS. Unlike a standard 510(k), the De Novo pathway is used when a novel low-to-moderate-risk device lacks a suitable predicate, meaning Powerful Medical has effectively established a regulatory category that future competing products may be able to follow. The commercial opportunity is unusually broad because the software operates on ordinary 12-lead electrocardiograms, potentially inserting an AI layer into emergency workflows without requiring hospitals to purchase an entirely new diagnostic hardware platform.

The intended use extends beyond simply displaying an algorithmic probability. Powerful Medical says the model is designed to detect STEMI and STEMI-equivalent patterns, prioritize patients and notify the appropriate clinician or cardiology team when urgent assessment may be needed. Several U.S. health systems had reportedly been waiting for FDA authorization before taking the platform live, giving the company potential launch customers rather than beginning commercialization without an installed clinical pipeline.

Why does the Queen of Hearts model challenge the traditional STEMI approach?

Conventional emergency pathways rely heavily on ST-segment elevation criteria to identify patients likely to have an acutely blocked coronary artery and move them rapidly toward catheterization. The weakness is that some true coronary occlusions do not satisfy traditional STEMI thresholds, while other ECG patterns can mimic STEMI and trigger unnecessary cath-lab activation. Powerful Medical has trained Queen of Hearts around a broader concept of occlusion myocardial infarction, attempting to identify ECG morphology associated with an acutely obstructed culprit vessel rather than depending only on classic ST-elevation rules.

A multicenter U.S. registry involving 2,523 patients with suspected STEMI found that the AI model identified 93.8% of patients who had acute myocardial infarction with an angiographic culprit lesion. Sensitivity was 96.3% when blood flow through the culprit artery was severely reduced to TIMI grade zero or one, while the model correctly classified 79.7% of patients without acute myocardial infarction as negative. The reported area under the receiver-operating-characteristic curve was 0.952, although the study was retrospective and its authors explicitly called for prospective validation.

Those numbers explain the potential workflow value without implying that AI should replace cardiologists. Missing an occluded artery can delay revascularization, while unnecessary cath-lab activation consumes specialist resources and can expose patients to invasive procedures. A system capable of improving both sensitivity and specificity could therefore have clinical and hospital-economic value, but the real test will come when the FDA-authorized version operates prospectively across diverse U.S. health systems.

What does the evidence show in harder non-ST-elevation heart attacks?

One of the more clinically interesting studies evaluated Queen of Hearts among 224 high-risk patients classified as non-ST-elevation acute coronary syndrome who underwent emergent invasive coronary angiography. Occlusion myocardial infarction was ultimately present in 129 patients, or 58% of the cohort, even though conventional STEMI patients had been excluded. The AI identified 58% of those occlusions from the initial ECG with specificity and positive predictive value of 78%, while an AI-augmented triage approach reduced false positives from 42% under standard care to 22%.

The same study also exposes an important limitation because initial-ECG sensitivity was only 58% in this difficult NSTE-ACS population. Serial ECGs reduced false negatives, reinforcing that the algorithm should not be used as a standalone rule-out test when clinical suspicion remains high. The authors concluded that the model could improve identification of patients requiring urgent angiography but specifically emphasized continued clinical judgment.

That nuance matters commercially because successful clinical AI rarely works by replacing a physician decision completely. A tool that reliably prioritizes difficult cases, reduces false alarms and brings expert-like ECG interpretation to hospitals without continuous specialist coverage may create substantial value even if physicians remain responsible for every final treatment decision. Powerful Medical’s task will be demonstrating that this supporting role improves outcomes and workflow consistently outside research settings.

Why could De Novo authorization give Powerful Medical a first-mover advantage?

Most AI medical devices reach the U.S. market through 510(k) substantial-equivalence submissions, while relatively few receive De Novo classification each year. Powerful Medical had no existing FDA predicate with the same acute-MI indication, so the agency’s authorization establishes the first formal pathway for this type of AI-ECG triage software. That allows Queen of Hearts to enter the market with a regulatory distinction that competitors previously could not cite.

The advantage will not be permanent because De Novo authorization also creates a predicate against which future products can potentially pursue 510(k) clearance. Powerful Medical therefore needs to turn regulatory leadership into installed customers, integration agreements and clinical evidence before rival algorithms can use the category it helped establish. The company says its wider PMcardio platform is already used by more than 120,000 clinicians internationally, although U.S. availability of individual modules depends on their specific regulatory status.

FDA records also indicate that the STEMI AI ECG Model received authorization with a Predetermined Change Control Plan. Such plans can allow certain prespecified modifications to AI-enabled medical software to occur within an FDA-reviewed framework rather than requiring an entirely new marketing submission for every planned change, provided the manufacturer stays within the authorized boundaries. That capability can become commercially important in machine-learning products because software improvement cycles can occur far more frequently than traditional hardware generations.

How might U.S. reimbursement affect adoption of the Queen of Hearts platform?

Powerful Medical has already explored the Medicare reimbursement pathway rather than assuming hospitals will fund clinical AI entirely from operating budgets. In its fiscal 2027 New Technology Add-on Payment application, the company described an expected technology cost of approximately $175 per analyzed patient and the federal review calculated a potential maximum add-on payment of $113.75 under the application assumptions. Those figures came from the reimbursement application while FDA authorization was still pending and should not be mistaken for confirmation that a final NTAP payment has been granted at those exact economics.

The submission nevertheless provides unusual visibility into how Powerful Medical is thinking about commercialization. At $175 per analyzed patient, the platform could create a usage-based revenue model that scales with emergency-department volume rather than relying only on fixed annual software subscriptions. Hospitals would then evaluate whether reimbursement plus potential savings from fewer unnecessary cath-lab activations and earlier treatment can justify the remaining software cost.

That economic case becomes particularly relevant outside major academic centers. Smaller hospitals and emergency departments may not have a cardiologist continuously available to interpret difficult ECGs, yet they still need to decide rapidly whether a patient should be transferred or sent to a catheterization laboratory. AI decision support can have greater operational value in those environments if integration remains simple and false alarms are kept low.

Can Powerful Medical turn clinical publications into a defensible commercial moat?

Powerful Medical lists more than 30 clinical studies and numerous conference abstracts across its research program, including recent work on coronary occlusion, inflammatory ECG mimics and emergency-department triage. That breadth can help hospital committees evaluate the software because adoption decisions increasingly require evidence extending beyond algorithm-development datasets. However, much of the literature predates U.S. authorization or evaluates investigational model versions, so real-world evidence from the commercial FDA-authorized configuration will become increasingly important.

The company also remains privately held, meaning revenue, cash position and valuation are not publicly available. That makes it difficult to assess how much commercial runway Powerful Medical has for a national U.S. rollout or how aggressively it can invest in integration, field support and payer strategy. The technology appears to have crossed an important regulatory and clinical threshold, but healthcare AI adoption is rarely determined by accuracy alone.

Queen of Hearts now has something most aspiring cardiac AI algorithms lack: a novel FDA authorization tied directly to urgent acute-coronary-syndrome triage. The next stage will determine whether that regulatory lead converts into routine use across American emergency departments. If it does, Powerful Medical may have created not simply another ECG algorithm but a new software layer inside one of medicine’s most time-sensitive care pathways.