Aims: A majority of acute coronary syndromes (ACS) present without typical ST elevation. One-third of non–ST-elevation myocardial infarction (NSTEMI) patients have an acutely occluded culprit coronary artery [occlusion myocardial infarction (OMI)], leading to poor outcomes due to delayed identification and invasive management. In this study, we sought to develop a versatile artificial intelligence (AI) model detecting acute OMI on single-standard 12-lead electrocardiograms (ECGs) and compare its performance with existing state-of-the-art diagnostic criteria. Methods and results: An AI model was developed using 18 616 ECGs from 10 543 patients with suspected ACS from an international database with clinically validated outcomes. The model was evaluated in an international cohort and compared with STEMI criteria and ECG experts in detecting OMI. The primary outcome of OMI was an acutely occluded or flow-limiting culprit artery requiring emergent revascularization. In the overall test set of 3254 ECGs from 2222 patients (age 62 ± 14 years, 67% males, 21.6% OMI), the AI model achieved an area under the curve of 0.938 [95% confidence interval (CI): 0.924–0.951] in identifying the primary OMI outcome, with superior performance [accuracy 90.9% (95% CI: 89.7–92.0), sensitivity 80.6% (95% CI: 76.8–84.0), and specificity 93.7 (95% CI: 92.6–94.8)] compared with STEMI criteria [accuracy 83.6% (95% CI: 82.1–85.1), sensitivity 32.5% (95% CI: 28.4–36.6), and specificity 97.7% (95% CI: 97.0–98.3)] and with similar performance compared with ECG experts [accuracy 90.8% (95% CI: 89.5–91.9), sensitivity 73.0% (95% CI: 68.7–77.0), and specificity 95.7% (95% CI: 94.7–96.6)]. Conclusion: The present novel ECG AI model demonstrates superior accuracy to detect acute OMI when compared with STEMI criteria. This suggests its potential to improve ACS triage, ensuring appropriate and timely referral for immediate revascularization.

Timely reperfusion is critical in reducing mortality in ST-segment elevation myocardial infarction (STEMI). Although ECG-guided cardiac catheterization laboratory (CCL) activation based on first medical contact recognition improves system-level response, diagnostic uncertainty-particularly in atypical presentations-contributes to false-positive activations (FPAs) and reperfusion delays. To evaluate the diagnostic performance and operational impact of artificial intelligence (AI)-based ECG analysis in real-world STEMI triage across a multicenter U.S. registry. We retrospectively analyzed 1,032 patients with suspected STEMI who triggered emergent CCL activation at three geographically diverse PCI centers (January 2020-May 2024). Index ECGs underwent standard triage and blinded retrospective AI ECG analysis (Queen of Hearts, PMcardio) trained to detect acute coronary occlusion and benign mimics. The reference standard was an angiographically confirmed culprit lesion with positive enzymes. Diagnostic accuracy, subgroup analyses, and FPA reclassification were compared. Of 1,032 emergent CCL activations, 601 (58.2%) had confirmed STEMI. The AI ECG model outperformed standard triage, demonstrating higher index ECG sensitivity 553/601 (92.0%, 95% CI: 89.7-94.1) vs. 427/601 (71.0%, 95% CI: 67.4-74.6), reducing FPA rates 34/431 (7.9%, 95% CI: 6.4-9.6) vs. 180/431 (41.8%, 95% CI: 38.9-44.7), and improving specificity 431/531 (81.0%, 95% CI: 77.2-84.5) vs. 154/531 (29.0%, 95% CI: 24.8-33.4) (all p < 0.001). AI ECG model AUC was 0.94 (95% CI: 0.92-0.95), maintaining consistent performance across clinically challenging subgroups (e.g., atrial fibrillation, bundle branch block, STEMI equivalents). The AI ECG model reclassified 277/306 (91%) of biomarker-negative FPAs correctly. AI-based ECG analysis significantly improved STEMI detection reduced false activations, and enhanced recognition of non-conventional presentations. This supports integration of AI-based ECG analysis into acute chest pain pathways.

The electrocardiogram (ECG) is one of the most accessible and comprehensive diagnostic tools used to assess cardiac patients at the first point of contact. Despite advances in computerized interpretation of the electrocardiogram (CIE), its accuracy remains inferior to physicians. This study evaluated the diagnostic performance of an artificial intelligence (AI)-powered ECG system and compared its performance to current state-of-the-art CIE. An AI-powered system consisting of 6 deep neural networks (DNN) was trained on standard 12‑lead ECGs to detect 20 essential diagnostic patterns (grouped into 6 categories: rhythm, acute coronary syndrome (ACS), conduction abnormalities, ectopy, chamber enlargement and axis). An independent test set of ECGs with diagnostic consensus of two expert cardiologists was used as a reference standard. AI system performance was compared to current state-of-the-art CIE. The key metrics used to compare performances were sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and F1 score. A total of 932,711 standard 12‑lead ECGs from 173,949 patients were used for AI system development. The independent test set pooled 11,932 annotated ECG labels. In all 6 diagnostic categories, the DNNs achieved high F1 scores: Rhythm 0.957, ACS 0.925, Conduction abnormalities 0.893, Ectopy 0.966, Chamber enlargement 0.972, and Axis 0.897. The diagnostic performance of DNNs surpassed state-of-the-art CIE for the 13 out of 20 essential diagnostic patterns and was non-inferior for the remaining individual diagnoses. Our results demonstrate the AI-powered ECG model's ability to accurately identify electrocardiographic abnormalities from the 12‑lead ECG, highlighting its potential as a clinical tool for healthcare professionals.

Background: Some authors advocate that ECGs with conventional computer algorithm (CCA) interpretations of “normal” need not be immediately reviewed. However, such ECGs may actually manifest findings of acute coronary occlusion myocardial infarction (OMI). We sought to determine if such cases can be detected by artificial intelligence (AI). Methods: We studied a retrospective series (2014–2024) of cases with ≥1 pre-angiography ECGs with a proven OMI outcome with a CCA ECG interpretation of “normal”. The OMI outcome was defined as (1) the diagnosis of acute type I MI, (2) an angiographic culprit with intervention, and (3) one of the following, (a) TIMI-0-2 flow, or (b) TIMI-3 or unknown flow, with high peak troponin or new wall abnormality. Each ECG as retrospectively interpreted by the PMcardio OMI AI ECG model. The primary analysis was the performance of AI in diagnosing "OMI" among these CCA “normal” ECGs. Results: Forty-two patients with OMI met the inclusion criteria. The first ECG was interpreted as “normal” by the CCA in 88% of cases; AI interpreted 81% as OMI and 86% as abnormal. Of the 78 total ECGs interpreted by the CCA, 73% were diagnosed as “normal”. Of this 73%, AI identified 81% as abnormal and 72% as OMI. Conclusion: The Conventional Computer Algorithm may interpret an ECG manifesting OMI as “normal”. AI not only recognized these as abnormal, but in 81% of patients, correctly recognized OMI on the first ECG and recognized 72% of all the CCA “normal” ECGs as OMI. It was rare for AI to diagnose a normal ECG for any OMI patient.