Coronary disease & ACSDiagnostic or prognostic study
AI-Enabled Electrocardiographic (ECG) Analysis Improves Diagnostic Accuracy and Reduces False STEMI Activations
- Diagnostic accuracy study
- retrospective
- multicentre
- blinded artificial intelligence (AI) ECG analysis
Authors' conclusion
AI-based ECG analysis significantly improved STEMI detection, reduced FPAs, and enhanced the recognition of nonconventional presentations.
Study context
- Population
- Patients with suspected STEMI triggering emergent cardiac catheterization laboratory (CCL) activation at 3 US PCI centres, January 2020 to May 2024601 confirmed STEMI
- Participants
- 1,032 patients
- Comparison
- AI ECG analysis vs standard triage
- Outcome
- Diagnostic accuracy vs reference standard (angiographically confirmed culprit lesion with positive enzymes)
Performance
As reported in the abstract
Diagnostic performance for STEMI
| Test or model | Sensitivity | FPA rate | Specificity | AUC |
|---|---|---|---|---|
| AI ECG | 553 of 601 [92.0%; 95% CI: 89.7%-94.1%] | 34 of 431 [7.9%; 95% CI: 6.4%-9.6%] | 431 of 531 [81.0%; 95% CI: 77.2%-84.5%] | 0.94 (95% CI: 0.92-0.95) |
| Standard triage | 427 of 601 [71.0%; 95% CI: 67.4%-74.6%] | 180 of 431 [41.8%; 95% CI: 38.9%-44.7%] | 154 of 531 [29.0%; 95% CI: 24.8%-33.4%] |
- AI ECG vs Standard triage: P < 0.001 — sensitivity, FPA rate and specificity (P < 0.001 for all)
Further findings
- Reclassification of biomarker-negative false positive activations (FPAs) by AI ECG
- 91% · 277 events
- Reclassification of biomarker-negative FPAs by AI ECG: 277 of 306 correctly reclassified
AI ECG performance in challenging subgroups (AF, bundle branch block, STEMI equivalents)
- Consistent performance across clinically challenging subgroups
Original publication
AI-Enabled ECG Analysis Improves Diagnostic Accuracy and Reduces False STEMI Activations. JACC Cardiovasc Interv. Published 28 October 2025. doi:10.1016/j.jcin.2025.10.018
Summarised from the published abstract; every number is checked against it