Abstract
Aims
To assess the evidence for deep learning (DL) analysis of resting electrocardiograms (ECG) to predict structural cardiac pathologies such as left ventricular systolic dysfunction, myocardial hypertrophy, and ischaemic heart disease.
Methods and Results
A systematic review was conducted to identify published original articles on end-to-end DL analysis of resting ECG signals for the detection of structural cardiac pathologies. Studies were excluded if the ECG was acquired by ambulatory, stress, intracardiac, or implantable devices, and if the pathology of interest was arrhythmic in nature. After duplicate reviewers screened search results, 12 articles met the inclusion criteria and were included. Three articles used DL ECG to detect left ventricular systolic dysfunction, achieving an area under the curve (AUC) of 0.89-0.93 and accuracy of 98%. One study used DL ECG to detect left ventricular hypertrophy, achieving an AUC of 0.87 and accuracy of 87%. Six articles used DL ECG to detect acute myocardial infarction, achieving an AUC of 0.88-1.00 and accuracy of 83-99.9%. Two articles used DL ECG to detect stable ischaemic heart disease, achieving an accuracy of 95-99.9%. DL algorithms, particularly those that used convolutional neural networks, outperformed rules-based algorithms and other machine learning algorithms.
Conclusions
DL is a promising technique to analyze resting ECG signals for the detection of structural cardiac pathologies, which has clinical applicability for more effective screening of asymptomatic populations and expedited diagnostic work-up of symptomatic patients at risk for cardiovascular disease.
The Effectiveness of a Deep Learning Model to Detect Left Ventricular Systolic Dysfunction from Electrocardiograms
