What is the clinical evidence from this study?

Study design: Other. Population: Cardiovascular disease (ECG abnormalities) (n=21801). Intervention: Image-based deep learning model (VGG16). Primary outcome: Diagnostic accuracy (AUROC) for cardiac pathology on holdout test set.

January 9, 2023Open Access

Image based deep learning in 12-lead ECG diagnosis

Key Result

An image-based deep learning model achieved an AUROC of 0.982, sensitivity of 0.950, and specificity of 0.954 for detecting cardiac pathologies on holdout 12-lead ECG images.

Structured PICO

Do image-based deep learning algorithms accurately diagnose conditions from 12-lead ECG images?

Population

21,801 12-lead ECGs from public databases used to train and evaluate a deep learning model for automated diagnosis of cardiac abnormalities.

Intervention

Image-based deep learning models using VGG architecture

Comparator

Best signal and image-based models

Outcome

Diagnostic accuracy (AUROC, AUPRC, sensitivity, specificity)surrogate

Image-based deep learning algorithms can accurately diagnose conditions from 12-lead ECG images, offering a feasible solution for clinical settings where only paper or digital image ECGs are available.

Limitations

Poorer generalization and diagnostic accuracy when applied to unseen external datasets.
Differences in labeling criteria for diagnoses between datasets or variances in ECG quality may have affected external validity.
Did not incorporate additional clinical information known to clinicians at the time of ECG interpretation, such as age, gender, weight, and height.

Abstract

Background: The electrocardiogram is an integral tool in the diagnosis of cardiovascular disease. Most studies on machine learning classification of electrocardiogram (ECG) diagnoses focus on processing raw signal data rather than ECG images. This presents a challenge for models in many areas of clinical practice where ECGs are printed on paper or only digital images are accessible, especially in remote and regional settings. This study aims to evaluate the accuracy of image based deep learning algorithms on 12-lead ECG diagnosis. Methods: Deep learning models using VGG architecture 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. Grad-CAM was utilized to depict heatmaps of diagnosis. Results: The results demonstrated excellent AUROC, AUPRC, sensitivity and specificity on holdout test data from datasets used in training comparable to the best signal and image-based models. Detection of hidden characteristics such as gender were achieved at a high rate while Grad-CAM successfully highlight pertinent features on ECGs traditionally used by human interpreters. 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 image based deep-learning models in ECG diagnosis.

AIに質問

Bookmark

View Full Paper

Cite This Study

Ao et al. (Mon,) conducted a other in Cardiovascular disease (ECG abnormalities) (n=21,801). Image-based deep learning model (VGG16) was evaluated on Diagnostic accuracy (AUROC) for cardiac pathology on holdout test set. An image-based deep learning model achieved an AUROC of 0.982, sensitivity of 0.950, and specificity of 0.954 for detecting cardiac pathologies on holdout 12-lead ECG images.

synapsesocial.com/papers/6a46fd8b8dd8ee6299050984 https://doi.org/https://doi.org/10.3389/frai.2022.1087370

AIに質問

Bookmark

View Full Paper