Key result
Multimodal deep-learning model predicts hs-TnT elevation with ~0.88 external AUROC using ECG, age, and sex.
Why the study?
Prompt diagnosis of cardiac conditions using ECGs and troponin testing can improve outcomes, motivating the development of a multi-modal deep-learning model using ECG data, age, and sex to predict high-sensitivity troponin elevation during chest-pain triage.
Does a multi-modal deep-learning model incorporating 12-lead ECG, age, and sex accurately predict high-sensitivity troponin elevation within 36 hours in adults presenting to the emergency room with chest pain or dyspnoea?
Observational (n=70,484)
Yes
Does a multi-modal deep-learning model incorporating 12-lead ECG, age, and sex accurately predict high-sensitivity troponin elevation within 36 hours in adults presenting to the emergency room with chest pain or dyspnoea?
A novel, externally validated multi-modal deep-learning model combining triage 12-lead ECGs with patient age and sex accurately predicts high-sensitivity troponin elevation (external AUROC 0.8765), offering a rapid tool to assist emergency room chest-pain and dyspnoea triage.
Answers are grounded in this paper and cite their sources.
May accelerate ER triage for suspected AMI; extends multimodal deep learning to hs-TnT prediction from ECG.
Aims: Electrocardiograms (ECGs) and troponin (Tn) testing are essential tools for the diagnosis and management of cardiac conditions. Prompt diagnosis using these tools can significantly improve patient outcomes. Methods and results: The objective of this study was to design and create a deep-learning model capable of predicting high-sensitivity troponin (hs-Tn) elevation in patients undergoing chest-pain triage. We developed a novel, multi-modal, externally validated deep-learning model that incorporates ECG data, age, and sex to predict high-sensitivity troponin-T elevation. The dataset used for this study was multi-centre and externally validated, drawing from data collected in two emergency rooms. The study population included all patients presenting to the ER with either chest pain or dyspnoea during the study period, where an ECG was recorded and a Tn test was performed. The model was trained on a dataset comprising 35 821 ECGs, with a positive fraction of 35.7%. It achieved an internal area under the receiver operating characteristic (AUROC) of 0.8958 ± 0.0040 (95% CI) and an AUROC of 0.8765 ± 0.0110 in external validation. The model's Score-CAM saliency maps demonstrated high activation from the ST-segment, indicating that the model draws information from relevant ECG segments. Conclusion: This study presents new opportunities for enhancing triage processes, enabling more rapid and accurate alerts to physicians regarding acute myocardial infarctions. The primary benefit of predicting Tn elevation lies in the objectivity of the label compared with compounded clinical outcomes and diagnoses.
No takes yet. Share an insight, caveat, or question.
Hilgendorf et al. (2025) conducted an observational in Suspected acute myocardial infarction / Myocardial injury (chest pain or dyspnoea) (n=70,484). Multi-modal deep-learning model (12-lead ECG, age, and sex) vs. Uni-modal deep-learning model (ECG only) and clinical high-sensitivity troponin testing was evaluated on Prediction of high-sensitivity troponin-T (hs-TnT) elevation (>14 ng/L) within 36 hours of ER admission (external validation AUROC) (AUROC 0.8765 ± 0.0110 (internal AUROC 0.8958 ± 0.0040), 95% CI 0.8655-0.8875). A multi-modal deep-learning model incorporating ECG, age, and sex predicted high-sensitivity troponin-T elevation with an internal AUROC of 0.8958 and external AUROC of 0.8765.
Synapse has enriched 10 related papers. Consider them for comparative context: