Performance of Artificial Intelligence–Powered ECG Analysis in Suspected ST-Segment Elevation Myocardial Infarction

Q: What are the key findings of this study?

In suspected STEMI, the AI-ECG model accurately identified occlusion myocardial infarction with an AUCROC of 0.952, including a 93.8% sensitivity for patients with an angiographic culprit.

Q: What is the clinical evidence from this study?

Study design: Cohort. Population: Suspected ST-segment elevation myocardial infarction (STEMI) (n=2523). Intervention: Artificial intelligence-powered ECG analysis (Queen of Hearts AI-ECG model) vs. Coronary angiographic findings and troponin elevation (ground truth). Primary outcome: Detection of occlusion myocardial infarction (OMI) by AI-ECG model (AUCROC 0.952, 95% CI 0.924-0.966).

Q: What does this research mean for the field?

An artificial intelligence-powered ECG model accurately identifies occlusion myocardial infarction in patients with suspected STEMI, achieving an AUC of 0.952 and 93.8% sensitivity for angiographic culprit lesions. Novelty: ClaimNovelty.INCREMENTAL. Consensus alignment: ConsensusAlignment.NEUTRAL.

Scott W. Sharkey; Robert Herman; Dawn Witt; Frank Aguirre; Mehmet Yildiz; David M. Larson; Avinash Murthy; Heather Rohm; Stephen W. Smith; Will Belzer; Jenny Chambers; Ellen Cravero; Seth Bergstedt; Greg Kerola; David Farmer; Andrew Willett; H. Pendell Meyers; Julia Harris; Christopher M VanHove; Timothy D. Henry

doi:10.1016/j.jacadv.2026.102671

Key result

AI-ECG model accurately detects occlusion myocardial infarction in suspected STEMI with an AUC of ~0.95.

AUCROC 0.952
95% CI 0.924-0.966
n=2,523

Figures

Graphical abstract published by JACC Advances — Graphical abstract · Published by JACC Advances · CC BY

Why the study?

Does an AI-powered ECG analysis model accurately detect occlusion myocardial infarction in patients with suspected STEMI?

Key Points

To evaluate the diagnostic accuracy of a novel artificial intelligence electrocardiogram model for detecting acute coronary occlusion in patients with cardiac catheterization laboratory activation for suspected STEMI.
Retrospectively analyzed 2,523 consecutive patients across 3 U.S. sites in the Midwest STEMI Consortium (2018-2022) with cardiac catheterization laboratory activations.
Categorized patients into 3 clinical cohorts: acute myocardial infarction (AMI) with culprit (n=2,076), AMI without culprit (n=314), and no-AMI (n=133).
Assessed an artificial intelligence model trained to classify catheterization-activating ECGs as occlusion myocardial infarction, OMI(+) versus OMI(-).
The model achieved an overall AUCROC of 0.952 (95% CI: 0.924-0.966), with a sensitivity of 93.8% (95% CI: 92.7%-94.8%) for AMI with culprit and a specificity of 79.7% (95% CI: 71.9%-86.2%) for no-AMI.
Sensitivity across pre-PCI initial TIMI flow grades was 96.3% for TIMI 0/1, 93.1% for TIMI 2, and 86.9% for TIMI 3 (P < 0.001).
In the AMI without culprit cohort, the model flagged OMI(+) in 78% of takotsubo syndrome cases, 61% of MINOCA cases, and 67% of myopericarditis cases.

Study Design

Type

Cohort (n=2,523)

Multicenter

Yes

Structured PICO

Does an AI-powered ECG analysis model accurately detect occlusion myocardial infarction in patients with suspected STEMI?

P

Population

2,523 patients with cardiac catheterization laboratory activation for suspected STEMI from a multicenter U.S. registry, predominantly male (68.3%) with a median age of 63, assessed retrospectively.

E

Exposure

Retrospective analysis using a CE-certified PMcardio AI-ECG model (Queen of Hearts, eOMI v1) trained to identify acute coronary occlusion on standard 12-lead ECGs.

C

Comparator

Ground truth diagnosis based on coronary angiographic findings (presence of culprit) and troponin elevation.

O

Outcome

Diagnostic accuracy (AUCROC) of the AI-ECG model to identify occlusion myocardial infarction.surrogate

Main Result

Effect estimate: AUCROC 0.952 (95% CI 0.924-0.966)

AI-ECG analysis highly accurately identifies acute coronary occlusion in patients with suspected STEMI, potentially improving timely reperfusion and reducing unnecessary catheterization laboratory activations.

Limitations

Retrospective study limited to the Midwestern region of the United States
Majority of patients were Caucasian
Only the CCL-activating ECG was analyzed
Excluded patients with AMI who did not meet STEMI ECG criteria (no CCL activation)
CCL-activating ECG and coronary angiogram were not simultaneous
Retrospective design limited to the Midwestern US
Lack of diversity (88% Caucasian)
Only the single CCL-activating ECG was analyzed per patient
Excluded patients with AMI not meeting traditional STEMI criteria
Dynamic nature of acute coronary thrombosis implies ECG and angiogram temporal disconnect

Discussion

No takes yet. Share an insight, caveat, or question.

Implication

Supports prospective validation of AI-ECG for OMI triage in suspected STEMI; leaves open effects on activations and outcomes.

Abstract

BACKGROUND: Artificial intelligence (AI)-based electrocardiogram (ECG) analysis has emerged as a promising adjunct to human ECG interpretation in suspected ST-segment elevation myocardial infarction (STEMI). OBJECTIVES: To expand knowledge in this evolving field, the authors retrospectively analyzed the performance of a novel AI-ECG model in patients with cardiac catheterization laboratory activation for suspected STEMI. METHODS: Consecutive patients were gathered from a multicenter U.S. STEMI registry (2018-2022) and categorized into 3 clinical cohorts based on the presence or absence of angiographic culprit and troponin elevation: acute myocardial infarction (AMI) with culprit, AMI without culprit, and no-AMI. Cardiac catheterization laboratory-activating ECGs were analyzed using an AI-ECG model trained to identify acute coronary occlusion and classified as occlusion myocardial infarction, OMI(+) or not, OMI(-). RESULTS: The study included 2,523 patients, 68.3% male, with a median age of 63 years. AMI with culprit was present in 2076 (82.3%), AMI without culprit in 314 (12.4%), and no-AMI in 133 (5.3%). Among AMI with culprit patients, the model correctly identified 93.8% as OMI(+). Sensitivity for TIMI flow 0/1, 2, and 3 was 96.3%, 93.1%, and 86.9% respectively; P < 0.001. The model correctly identified 79.7% of no-AMI patients as OMI(-). The AUCROC was 0.952 (95% CI: 0.924-0.966). The AMI without culprit cohort included takotsubo syndrome OMI(+) = 78%, MI with nonobstructive coronary arteries OMI(+) = 61%, and myopericarditis OMI(+) = 67%. CONCLUSIONS: In suspected STEMI, this AI-ECG model correctly identified nearly all patients with acute coronary obstruction and most of those without AMI. If prospectively validated, this approach could improve management of patients with suspected AMI.

Cite This Study

Sharkey et al. (2026) conducted a cohort in Suspected ST-segment elevation myocardial infarction (STEMI) (n=2,523). Artificial intelligence-powered ECG analysis (Queen of Hearts AI-ECG model) vs. Coronary angiographic findings and troponin elevation (ground truth) was evaluated on Detection of occlusion myocardial infarction (OMI) by AI-ECG model (AUCROC 0.952, 95% CI 0.924-0.966). In suspected STEMI, the AI-ECG model accurately identified occlusion myocardial infarction with an AUCROC of 0.952, including a 93.8% sensitivity for patients with an angiographic culprit.

synapsesocial.com/papers/69c2294caeb5a845df0d383b https://doi.org/10.1016/j.jacadv.2026.102671

View Full Paper

Ask AI

Bookmark

Share

Performance of Artificial Intelligence–Powered ECG Analysis in Suspected ST-Segment Elevation Myocardial Infarction

Figures

Key Points

Study Design

Structured PICO

Main Result

Limitations

Discussion

Implication

Abstract

Cite This Study

Performance of Artificial Intelligence–Powered ECG Analysis in Suspected ST-Segment Elevation Myocardial Infarction

Figures

Key Points

Study Design

Structured PICO

Main Result

Limitations

Discussion

Implication

Abstract

Cite This Study

Keep reading

Keep reading