Acute Myocardial Infarction Detection Using Deep Learning-Enabled Electrocardiograms
Problem
Acute myocardial infarction (AMI) needs rapid, accurate diagnosis from 12-lead ECGs so the culprit lesion can be treated early.
Why it matters
AMI carries a poor prognosis, making early accurate diagnosis and intervention critical.
What the AI did
A residual network was trained on 21,837 PTB-XL 12-lead ECGs (15,285 training, 6,552 validation) and tested on 205 external ECGs to automatically diagnose AMI.
Results
Performance verified via AUC, precision, sensitivity, specificity and F1; exact values not stated in source (abstract truncated)
Impact
Evidence supports a validated research model on an external test set; clinical adoption is not stated.
Record
| Venue | Frontiers in Cardiovascular Medicine |
|---|---|
| Published | 2021 |
| Work type | article |
| Authors | Xiehui Chen, Wenqin Guo, Lingyue Zhao, Weichao Huang, Lili Wang, Aimei Sun, Lang Li, Fanrui Mo |
| Citations | 36 (OpenAlex, at import time) |
| References | 31 |
| Model or method | residual network |
| Organization | not stated in source |
| Primary source | https://doi.org/10.3389/fcvm.2021.654515 |
| Open access | gold |
| Keywords | acute myocardial infarction, ECG, deep learning, residual network, diagnosis |
Importance rating
significant — thousands to millions affected
Abstract states AMI carries a poor prognosis and that accurate diagnosis and early intervention of the culprit lesion are of extreme importance; a major acute cardiac condition, though no epidemiological figures are given.