Key result
An ECG cloud computing analysis algorithm achieved an AUC of 0.9317 for classifying normal and abnormal ECG records, and 98.51% sensitivity and 98.26% specificity for atrial fibrillation.
Why the study?
Does a cloud-based ECG analysis algorithm accurately classify normal/abnormal ECGs and atrial fibrillation?
Does a cloud-based ECG analysis algorithm accurately classify normal/abnormal ECGs and atrial fibrillation?
Effect estimate: AUC 0.9317
A cloud-based ECG analysis algorithm demonstrated high sensitivity and specificity for detecting atrial fibrillation and abnormal ECGs, comparable to state-of-the-art methods.
The severe challenges of the fast aging population and the prevalence of cardiovascular diseases highlight the needs for effective solutions supporting more accurate and affordable medical diagnosis and treatment. Recent advances in cloud computing have inspired numerous designs of cloud-based health care services. In this paper, we developed a cloud-computing platform monitored by physicians, which can receive 12-lead ECG records and send back diagnostic reports to users. Aiming to lessen the physicians' workload, we implemented an analysis algorithm that can identify abnormal heart rate, irregular heartbeat, abnormal amplitude, atrial fibrillation and abnormal ECG in it. A large number of testing samples were used to evaluate performance. Our algorithm achieved a TPR95 (specificity under the condition of negative predictive value being equal to 95 percent) of 68.5 percent and 0.9317 AUC (area under the ROC curve) for classification of normal and abnormal ECG records and a sensitivity of 98.51 percent and specificity of 98.26 percent for atrial fibrillation classification, comparable to the state-of-the-art results for each subject. The proposed ECG cloud computing service has been applied in Hunan Jinshengda Aerial Hospital Network and it now can receive and analyze ECG records in real time.
No takes yet. Share an insight, caveat, or question.
Jin et al. (2018) studied Cardiovascular diseases. ECG cloud computing service and analysis algorithm was evaluated on Classification of normal and abnormal ECG records (AUC 0.9317). An ECG cloud computing analysis algorithm achieved an AUC of 0.9317 for classifying normal and abnormal ECG records, and 98.51% sensitivity and 98.26% specificity for atrial fibrillation.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: