AI-Based ECG Analysis for Early Detection of Occlusion Myocardial Infarction Using Coronary Angiography Results.
LR, L. & HW, C. (2026). AI-Based ECG Analysis for Early Detection of Occlusion Myocardial Infarction Using Coronary Angiography Results.. Studies in health technology and informatics. https://doi.org/10.3233/SHTI260818
LR L, HW C. AI-Based ECG Analysis for Early Detection of Occlusion Myocardial Infarction Using Coronary Angiography Results.. Studies in health technology and informatics. 2026; doi: 10.3233/SHTI260818
LR L, HW C. AI-Based ECG Analysis for Early Detection of Occlusion Myocardial Infarction Using Coronary Angiography Results.[J]. Studies in health technology and informatics. 2026. DOI: 10.3233/SHTI260818.
@article{lr2026,
author = {Liu LR and Chiu HW},
title = {AI-Based ECG Analysis for Early Detection of Occlusion Myocardial Infarction Using Coronary Angiography Results.},
journal = {Studies in health technology and informatics},
year = {2026},
doi = {10.3233/SHTI260818},
note = {PMID: 42393980},
}
TY - JOUR AU - Liu LR AU - Chiu HW TI - AI-Based ECG Analysis for Early Detection of Occlusion Myocardial Infarction Using Coronary Angiography Results. T2 - Studies in health technology and informatics PY - 2026 DO - 10.3233/SHTI260818 AN - PMID:42393980 ER -
Myocardial Infarction remains a high-mortality cardiovascular disease. While ST-Elevation Myocardial Infarction (STEMI) is the traditional intervention standard, many Occlusion Myocardial Infarction (OMI) cases are misclassified as Non-ST Elevation Myocardial Infarction (NSTEMI), leading to delayed treatment and poor outcomes. This study collected clinical and 12-lead ECG data from AMI patients over five years, categorized into OMI and Non Occlusion Myocardial Infarction (NOMI) groups based on angiography and cardiac enzymes. We developed a ResNet-1D deep learning model to identify OMI from Electrocardiograms (ECG) signal patterns. The model achieved an OMI recall of 0.75 and specificity to 0.53. Our findings suggest that while ST-elevation remains a primary OMI indicator, machine learning can effectively assist clinicians in detecting hidden OMI cases within the NSTEMI population, providing critical diagnostic support alongside clinical analysis.