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

Area of Science:

• Cardiology
• Artificial Intelligence
• Medical Diagnostics

Background:

• Artificial intelligence (AI)-based electrocardiogram (ECG) analysis is a developing tool for interpreting suspected ST-segment elevation myocardial infarction (STEMI).
• Human interpretation of ECGs can be time-consuming and prone to error, necessitating advanced diagnostic aids.

Purpose of the Study:

• To evaluate the performance of a novel AI-ECG model in identifying acute coronary occlusion.
• To assess the model's accuracy in patients undergoing cardiac catheterization for suspected STEMI.

Main Methods:

• A retrospective analysis of 2,523 patients from a multicenter U.S. STEMI registry (2018-2022).
• Patients were categorized into acute myocardial infarction (AMI) with culprit, AMI without culprit, and no-AMI cohorts.
• An AI-ECG model classified ECGs for acute coronary occlusion (OMI(+) or OMI(-)).

Main Results:

• The AI-ECG model achieved 93.8% accuracy in identifying OMI(+) among patients with AMI and culprit.
• Sensitivity for detecting coronary occlusion (TIMI flow 0/1) was 96.3%.
• The model correctly identified 79.7% of no-AMI patients as OMI(-), with an AUCROC of 0.952.

Conclusions:

• The AI-ECG model demonstrated high accuracy in identifying acute coronary obstruction in suspected STEMI cases.
• The model also effectively identified patients without acute myocardial infarction.
• Prospective validation could lead to improved patient management for suspected AMI.