AUTHOR=Kim Jongkwang , Shon Byungeun , Kim Sangwook , Cho Jungrae , Seo Jung-Ju , Jang Se Yong , Jeong Sungmoon 

TITLE=ECG data analysis to determine ST-segment elevation myocardial infarction and infarction territory type: an integrative approach of artificial intelligence and clinical guidelines

JOURNAL=Frontiers in Physiology

VOLUME=Volume 15 - 2024

YEAR=2024

URL=https://www.frontiersin.org/journals/physiology/articles/10.3389/fphys.2024.1462847

DOI=10.3389/fphys.2024.1462847

ISSN=1664-042X

ABSTRACT=Abstract
Acute coronary syndrome (ACS) is a leading cause of death from cardiovascular diseases worldwide, and ST-segment elevation myocardial infarction (STEMI) represents a severe form of ACS with high prevalence and mortality rates. This study proposes a novel method using deep learning-based artificial intelligence (AI) algorithms to diagnose the STEMI accurately from 12-lead electrocardiogram (ECG) data and to classify the type of infarction territory in detail. Based on an ECG database of 888 MI patients, the study enhanced the generalization capability of the AI model through five-fold cross-validation. The developed ST-segment elevation (STE) detector accurately detected STE, a crucial indicator in the clinical ECG diagnosis of STEMI, in each of the 12 leads. Utilizing this detector in the AI model to differentiate between STEMI and non-ST-segment elevation myocardial infarction (NSTEMI) showed significant results, with an average AUROC of 0.939 and AUPRC of 0.977. Moreover, this detector demonstrated accurate differentiation capabilities for each infarction territory in the ECG, such as anterior myocardial infarction (AMI), inferior MI (IMI), lateral MI (LMI), and suspected left main disease. These results suggest that using AI technology in ECG diagnoses can play an important role in the rapid treatment and prognosis improvement of STEMI patients.