Usefulness of Machine Learning-Based Detection and Classification of Cardiac Arrhythmias With 12-Lead Electrocardiograms

被引:46
作者
Chang, Kuan-Cheng [1 ,2 ]
Hsieh, Po-Hsin [3 ]
Wu, Mei-Yao [4 ,5 ]
Wang, Yu-Chen [1 ,6 ,7 ]
Chen, Jan-Yow [1 ,2 ]
Tsai, Fuu-Jen [8 ]
Shih, Edward S. C. [9 ]
Hwang, Ming-Jing [9 ]
Huang, Tzung-Chi [3 ,10 ,11 ]
机构
[1] China Med Univ Hosp, Div Cardiovasc Med, 2 Yude Rd, Taichung 40447, Taiwan
[2] China Med Univ, Grad Inst Biomed Sci, Taichung, Taiwan
[3] China Med Univ, Dept Biomed Imaging & Radiol Sci, Taichung, Taiwan
[4] China Med Univ, Sch Postbaccalaureate Chinese Med, Taichung, Taiwan
[5] China Med Univ Hosp, Dept Chinese Med, Taichung, Taiwan
[6] Asia Univ Hosp, Div Cardiovasc Med, Taichung, Taiwan
[7] Asia Univ, Dept Biotechnol, Taichung, Taiwan
[8] China Med Univ Hosp, Dept Med Res, Taichung, Taiwan
[9] Acad Sinica, Inst Biomed Sci, Taipei, Taiwan
[10] China Med Univ Hosp, Artificial Intelligence Ctr, Taichung, Taiwan
[11] Asia Univ, Dept Bioinformat & Med Engn, Taichung, Taiwan
关键词
BIDIRECTIONAL LSTM; DEEP; CARDIOLOGY; COMMITTEE; IMAGES;
D O I
10.1016/j.cjca.2020.02.096
中图分类号
R5 [内科学];
学科分类号
1002 ; 100201 ;
摘要
Background: Deep-learning algorithms to annotate electrocardiograms (ECGs) and classify different types of cardiac arrhythmias with the use of a single-lead ECG input data set have been developed. It remains to be determined whether these algorithms can be generalized to 12-lead ECG-based rhythm classification. Methods: We used a long short-term memory (LSTM) model to detect 12 heart rhythm classes with the use of 65,932 digital 12-lead ECG signals from 38,899 patients, using annotations obtained by consensus of 3 board-certified electrophysiologists as the criterion standard. Results: The accuracy of the LSTM model for the classification of each of the 12 heart rhythms was >= 0.982 (range 0.982-1.0), with an area under the receiver operating characteristic curve of >= 0.987 (range 0.987-1.0). The precision and recall ranged from 0.692 to 1 and from 0.625 to 1, respectively, with an F-1 score of >= 0.777 (range 0.777-1.0). The accuracy of the model (0.90) was superior to the mean accuracies of internists (0.55), emergency physicians (0.73), and cardiologists (0.83). Conclusions: We demonstrated the feasibility and effectiveness of the deep-learning LSTM model for interpreting 12 common heart rhythms according to 12-lead ECG signals. The findings may have clinical relevance for the early diagnosis of cardiac rhythm disorders.
引用
收藏
页码:94 / 104
页数:11
相关论文
共 24 条
[1]   Detection of ST segment deviation episodes in ECG using KLT with an ensemble neural classifier [J].
Afsar, Fayyaz A. ;
Arif, M. ;
Yang, J. .
PHYSIOLOGICAL MEASUREMENT, 2008, 29 (07) :747-760
[2]   Deep Learning for Brain MRI Segmentation: State of the Art and Future Directions [J].
Akkus, Zeynettin ;
Galimzianova, Alfiia ;
Hoogi, Assaf ;
Rubin, Daniel L. ;
Erickson, Bradley J. .
JOURNAL OF DIGITAL IMAGING, 2017, 30 (04) :449-459
[3]   Deep learning approach for active classification of electrocardiogram signals [J].
Al Rahhal, M. M. ;
Bazi, Yakoub ;
AlHichri, Haikel ;
Alajlan, Naif ;
Melgani, Farid ;
Yager, R. R. .
INFORMATION SCIENCES, 2016, 345 :340-354
[4]   Clinical applications of machine learning in cardiovascular disease and its relevance to cardiac imaging [J].
Al'Aref, Subhi J. ;
Anchouche, Khalil ;
Singh, Gurpreet ;
Slomka, Piotr J. ;
Kolli, Kranthi K. ;
Kumar, Amit ;
Pandey, Mohit ;
Maliakal, Gabriel ;
van Rosendael, Alexander R. ;
Beecy, Ashley N. ;
Berman, Daniel S. ;
Leipsic, Jonathan ;
Nieman, Koen ;
Andreini, Daniele ;
Pontone, Gianluca ;
Schoepf, U. Joseph ;
Shaw, Leslee J. ;
Chang, Hyuk-Jae ;
Narula, Jagat ;
Bax, Jeroen J. ;
Guan, Yuanfang ;
Min, James K. .
EUROPEAN HEART JOURNAL, 2019, 40 (24) :1975-+
[5]   Comparison of Deep Learning Approaches for Multi-Label Chest X-Ray Classification [J].
Baltruschat, Ivo M. ;
Nickisch, Hannes ;
Grass, Michael ;
Knopp, Tobias ;
Saalbach, Axel .
SCIENTIFIC REPORTS, 2019, 9 (1)
[6]   Detection of tuberculosis patterns in digital photographs of chest X-ray images using Deep Learning: feasibility study [J].
Becker, A. S. ;
Bluthgen, C. ;
Van, V. D. Phi ;
Sekaggya-Wiltshire, C. ;
Castelnuovo, B. ;
Kambugu, A. ;
Fehr, J. ;
Frauenfelder, T. .
INTERNATIONAL JOURNAL OF TUBERCULOSIS AND LUNG DISEASE, 2018, 22 (03) :328-+
[7]   Deep Learning in Mammography Diagnostic Accuracy of a Multipurpose Image Analysis Software in the Detection of Breast Cancer [J].
Becker, Anton S. ;
Marcon, Magda ;
Ghafoor, Soleen ;
Wurnig, Moritz C. ;
Frauenfelder, Thomas ;
Boss, Andreas .
INVESTIGATIVE RADIOLOGY, 2017, 52 (07) :434-440
[8]   Framewise phoneme classification with bidirectional LSTM and other neural network architectures [J].
Graves, A ;
Schmidhuber, J .
NEURAL NETWORKS, 2005, 18 (5-6) :602-610
[9]   A Novel Connectionist System for Unconstrained Handwriting Recognition [J].
Graves, Alex ;
Liwicki, Marcus ;
Fernandez, Santiago ;
Bertolami, Roman ;
Bunke, Horst ;
Schmidhuber, Juergen .
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 2009, 31 (05) :855-868
[10]   Common errors in computer electrocardiogram interpretation [J].
Guglin, ME ;
Thatai, D .
INTERNATIONAL JOURNAL OF CARDIOLOGY, 2006, 106 (02) :232-237