eduzhai > Applied Sciences > Engineering >

Remote atrial fibrillation burden estimation using deep recurrent neural network

  • king
  • (0) Download
  • 20210506
  • Save

... pages left unread,continue reading

Document pages: 10 pages

Abstract: The atrial fibrillation burden (AFB) is defined as the percentage of timespend in atrial fibrillation (AF) over a long enough monitoring period. Recentresearch has demonstrated the added prognosis value that becomes available byusing the AFB as compared with the binary diagnosis. We evaluate, for the firsttime, the ability to estimate the AFB over long-term continuous recordings,using a deep recurrent neutral network (DRNN) approach. Methods: The modelswere developed and evaluated on a large database of p=2,891 patients, totalingt=68,800 hours of continuous electrocardiography (ECG) recordings acquired atthe University of Virginia heart station. Specifically, 24h beat-to-beat timeseries were obtained from a single portable ECG channel. The network, denotedArNet, was benchmarked against a gradient boosting (XGB) model, trained on 21features including the coefficient of sample entropy (CosEn) and AFEvidence.Data were divided into training and test sets, while patients were stratifiedby the presence and severity of AF. The generalizations of ArNet and XGB werealso evaluated on the independent test PhysioNet LTAF database. Results: theabsolute AF burden estimation error |E AF|, median and interquartile, on thetest set, was 1.2 (0.1-6.7) for ArNet and 3.1 (0.0-11.7) for XGB for AFindividuals. Generalization results on LTAF were consistent with E AF of 2.6(1.1-14.7) for ArNet and 3.6 (1.0-16.7) for XGB. Conclusion: This researchdemonstrates the feasibility of AFB estimation from 24h beat-to-beat intervaltime series utilizing recent advances in DRNN. Significance: The noveldata-driven approach enables robust remote diagnosis and phenotyping of AF.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×