Predicting troponin biomarker elevation from electrocardiograms using a deep neural network
Authors:
- Lukas HilgendorfLHLukas Hilgendorf
- Petur PeturssonPPPetur Petursson
- Vibha GuptaVGVibha Gupta
- Truls RamunddalTRTruls Ramunddal
- Erik AnderssonEAErik Andersson
- Peter LundgrenPLPeter Lundgren
- Christian DworeckCDChristian Dworeck
- Charlotta LjungmanCLCharlotta Ljungman
- Jan BorenJBJan Boren
- Aidin RawshaniARAidin Rawshani
- Elmir OmerovicEOElmir Omerovic
- Gustav SmithGSGustav Smith
Open Access
- 30 October 2024
- journal article
- research article
- Published by BMJ in Open Heart
- Vol. 11 (2) , e002937
Abstract
Background: Elevated troponin levels are a sensitive biomarker for cardiac injury. The quick and reliable prediction of troponin elevation for patients with chest pain from readily available ECGs may pose a valuable time-saving diagnostic tool during decision-making concerning this patient population. Methods and results: The data used included 15 856 ECGs from patients presenting to the emergency rooms with chest pain or dyspnoea at two centres in Sweden from 2015 to June 2023. All patients had high-sensitivity troponin test results within 6 hours after 12-lead ECG. Both troponin I (TnI) and TnT were used, with biomarker-specific cut-offs and sex-specific cut-offs for TnI. On this dataset, a residual convolutional neural network (ResNet) was trained 10 times, each on a unique split of the data. The final model achieved an average area under the curve for the receiver operating characteristic curve of 0.7717 (95% CI±0.0052), calibration curve analysis revealed a mean slope of 1.243 (95% CI±0.075) and intercept of −0.073 (95% CI±0.034), indicating a good correlation between prediction and ground truth. Post-classification, tuned for F1 score, accuracy was 71.43% (95% CI±1.28), with an F1 score of 0.5642 (95% CI±0.0052) and a negative predictive value of 0.8660 (95% CI±0.0048), respectively. The ResNet displayed comparable or surpassing metrics to prior presented models. Conclusion: The model exhibited clinically meaningful performance, notably its high negative predictive accuracy. Therefore, clinical use of comparable neural networks in first-line, quick-response triage of patients with chest pain or dyspnoea appears as a valuable option in future medical practice.Keywords
Funding Information
- Wallenberg Centre for Molecular and Translational Medicine
- ALF (ALFGBG-998297)
- Swedish Society for Medicine (SLS-1000926)
This publication has 16 references indexed in Scilit:
- Meta-Analysis of the Performance of AI-Driven ECG Interpretation in the Diagnosis of Valvular Heart DiseasesThe American Journal of Cardiology, 2023
- Deep learning augmented ECG analysis to identify biomarker-defined myocardial injuryScientific Reports, 2023
- Type 2 myocardial infarction: a diagnostic and therapeutic challenge in contemporary cardiologyInternal and Emergency Medicine, 2022
- Self-supervised representation learning from 12-lead ECG dataComputers in Biology and Medicine, 2022
- Artificial intelligence-enhanced electrocardiography in cardiovascular disease managementNature Reviews Cardiology, 2021
- Artificial intelligence algorithm for detecting myocardial infarction using six-lead electrocardiographyScientific Reports, 2020
- Classification of ECG Signals Based on LSTM and CNN2020
- Fourth universal definition of myocardial infarction (2018)European Heart Journal, 2018
- Evolution of the management of acute myocardial infarction: a 20th century sagaThe Lancet, 1998
- Review of neural network applications in medical imaging and signal processingMedical & Biological Engineering & Computing, 1992