Image Based Deep Learning in 12-Lead ECG Diagnosis

BackgroundMost studies on machine learning classification of electrocardiogram (ECG) diagnoses focus on processing raw signal data rather than ECG images. In clinical practice, it is often the case where ECGs printed on paper or only digital images are easily accessible. This study aims to evaluate the accuracy of image based deep learning algorithms on 12 lead ECG diagnosis. MethodsDeep learning models were trained on various 12-lead ECG datasets and evaluated for accuracy by testing on holdout test data as well as data from datasets not seen in training. ResultsThe results demonstrated excellent AUROC, AUPRC, sensitivity and specificity on holdout test data from datasets used in training, but poorer accuracy on unseen datasets. DiscussionThis study demonstrates feasibility of image based deep learning algorithms in ECG diagnosis, and identifies directions for future research in order to develop clinically applicable deep-learning models in ECG diagnosis.