Key result
The Conv-BiLSTM-Attn multimodal deep learning model detected acute myocardial infarction from 12-lead ECGs with an AUROC of 0.848 (95% CI 0.84-0.86), outperforming benchmark models.
Why the study?
Most prior deep learning studies for automated ECG analysis in acute myocardial infarction relied on small, curated datasets with limited external validation, limiting clinical applicability.
Does a multimodal deep learning model (Conv-BiLSTM-Attn) improve the detection of acute myocardial infarction from 12-lead ECGs compared to benchmark models?
Population
96 813 patients (145 656 ECGs) across three Swedish hospitals
Comparison
Conv-BiLSTM-Attn deep learning model vs benchmark models
Design
Multi-centre model development and cross-hospital external validation study
Loading...
Supports accurate AMI detection via DL on routine ECGs; leaves open need for prospective validation before clinical adoption.
Observational (n=96,813)
Yes
Does a multimodal deep learning model (Conv-BiLSTM-Attn) improve the detection of acute myocardial infarction from 12-lead ECGs compared to benchmark models?
Effect estimate: AUROC 0.848 (95% CI 0.84-0.86)
A multimodal deep learning model integrating 12-lead ECG signals and demographic data accurately detects acute myocardial infarction and demonstrates robust generalizability across different hospitals.
A 2025 study conducted an observational in Acute myocardial infarction (n=96,813). Conv-BiLSTM-Attn multimodal deep learning model vs. Benchmark models was evaluated on Acute myocardial infarction detection (AUROC) (AUROC 0.848, 95% CI 0.84-0.86). The Conv-BiLSTM-Attn multimodal deep learning model detected acute myocardial infarction from 12-lead ECGs with an AUROC of 0.848 (95% CI 0.84-0.86), outperforming benchmark models.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: