Why the study?
Most machine learning ECG studies focus on raw signal data, yet paper-printed or digital ECG images are more accessible in clinical practice.
Does an image-based deep learning algorithm accurately diagnose abnormalities on 12-lead ECG images?
Does an image-based deep learning algorithm accurately diagnose abnormalities on 12-lead ECG images?
Image-based deep learning models can accurately diagnose ECG abnormalities from generated images within the same dataset distribution, but currently struggle to generalize to external datasets.
Image-based DL may aid ECG diagnosis from prints; hypothesis-generating and needs rigorous validation before clinical adoption.
Background Most 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. Methods Deep 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. Results The results demonstrated excellent AUROC, AUPRC, sensitivity and specificity on holdout test data from datasets used in training, but poorer accuracy on unseen datasets. Discussion This 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.
No takes yet. Share an insight, caveat, or question.
Ao et al. (2022) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: