Diagnosis of atrial fibrillation based on AI-detected anomalies of ECG segments

Early detection of atrial fibrillation (AF) is crucial for its effective management and prevention. Various methods for detecting AF using deep learning (DL) based on supervised learning with a large labeled dataset have a remarkable performance. However, supervised learning has several problems, as...

Full description

Bibliographic Details
Main Authors: Sanghoon Choi, Kyungmin Choi, Hong Kyun Yun, Su Hyeon Kim, Hyeon-Hwa Choi, Yi-Seul Park, Segyeong Joo
Format: Article
Language:English
Published: Elsevier 2024-01-01
Series:Heliyon
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S240584402310805X
_version_ 1797336983107796992
author Sanghoon Choi
Kyungmin Choi
Hong Kyun Yun
Su Hyeon Kim
Hyeon-Hwa Choi
Yi-Seul Park
Segyeong Joo
author_facet Sanghoon Choi
Kyungmin Choi
Hong Kyun Yun
Su Hyeon Kim
Hyeon-Hwa Choi
Yi-Seul Park
Segyeong Joo
author_sort Sanghoon Choi
collection DOAJ
description Early detection of atrial fibrillation (AF) is crucial for its effective management and prevention. Various methods for detecting AF using deep learning (DL) based on supervised learning with a large labeled dataset have a remarkable performance. However, supervised learning has several problems, as it is time-consuming for labeling and has a data dependency problem. Moreover, most of the DL methods do not provide any clinical evidence to physicians regarding the analysis of electrocardiography (ECG) for classification or detection of AF. To address these limitations, in this study, we proposed a novel AF diagnosis system using unsupervised learning for anomaly detection with three segments, PreQ, QRS, and PostS, based on the normal ECG. Two independent datasets, PTB-XL and China, were used in three experiments. We used a long short-term memory (LSTM)-based autoencoder to train the segments of the normal ECG. Based on the threshold of anomaly scores using mean squared error (MSE), it distinguished between normal and AF segments. In Experiment A, the best score was that of PreQ, which detected AF with an AUROC score of 0.96. In Experiment B and C for cross validation of each dataset, the best scores were also of PreQ, with AUROC scores of 0.9 and 0.95, respectively. To verify the significance of the anomaly score in distinguishing between AF and normal segments, we utilized an XG-Boosted model after generating anomaly scores in the three segments. The XG-Boosted model achieved an AUROC score of 0.98 and an F1 score of 0.94. AF detection using DL has been controversial among many physicians. However, our study differentiates itself from previous studies in that we can demonstrate evidence that distinguishes AF from normal segments based on the anomaly score.
first_indexed 2024-03-08T09:02:47Z
format Article
id doaj.art-b680d40df17e4f4390aa57ddb9fa3ddd
institution Directory Open Access Journal
issn 2405-8440
language English
last_indexed 2024-03-08T09:02:47Z
publishDate 2024-01-01
publisher Elsevier
record_format Article
series Heliyon
spelling doaj.art-b680d40df17e4f4390aa57ddb9fa3ddd2024-02-01T06:32:34ZengElsevierHeliyon2405-84402024-01-01101e23597Diagnosis of atrial fibrillation based on AI-detected anomalies of ECG segmentsSanghoon Choi0Kyungmin Choi1Hong Kyun Yun2Su Hyeon Kim3Hyeon-Hwa Choi4Yi-Seul Park5Segyeong Joo6Department of Biomedical Engineering, University of Ulsan College of Medicine, Seoul, Republic of Korea; Asan Medical Institute of Convergence Science and Technology, Asan Medical Center, Seoul, Republic of KoreaDepartment of Biomedical Engineering, University of Ulsan College of Medicine, Seoul, Republic of Korea; Asan Medical Institute of Convergence Science and Technology, Asan Medical Center, Seoul, Republic of KoreaDepartment of Biomedical Engineering, University of Ulsan College of Medicine, Seoul, Republic of Korea; Asan Medical Institute of Convergence Science and Technology, Asan Medical Center, Seoul, Republic of KoreaDepartment of Biomedical Engineering, University of Ulsan College of Medicine, Seoul, Republic of Korea; Asan Medical Institute of Convergence Science and Technology, Asan Medical Center, Seoul, Republic of KoreaDepartment of Biomedical Engineering, University of Ulsan College of Medicine, Seoul, Republic of Korea; Asan Medical Institute of Convergence Science and Technology, Asan Medical Center, Seoul, Republic of KoreaDepartment of Biomedical Engineering, University of Ulsan College of Medicine, Seoul, Republic of Korea; Asan Medical Institute of Convergence Science and Technology, Asan Medical Center, Seoul, Republic of KoreaDepartment of Biomedical Engineering, University of Ulsan College of Medicine, Seoul, Republic of Korea; Asan Medical Institute of Convergence Science and Technology, Asan Medical Center, Seoul, Republic of Korea; Corresponding author. Asan Medical Institute of Convergence Science and Technology, Asan Medical Center, University of Ulsan College of Medicine, 388-1 Pungnap-dong, Songpa-gu, Seoul, Republic of Korea.Early detection of atrial fibrillation (AF) is crucial for its effective management and prevention. Various methods for detecting AF using deep learning (DL) based on supervised learning with a large labeled dataset have a remarkable performance. However, supervised learning has several problems, as it is time-consuming for labeling and has a data dependency problem. Moreover, most of the DL methods do not provide any clinical evidence to physicians regarding the analysis of electrocardiography (ECG) for classification or detection of AF. To address these limitations, in this study, we proposed a novel AF diagnosis system using unsupervised learning for anomaly detection with three segments, PreQ, QRS, and PostS, based on the normal ECG. Two independent datasets, PTB-XL and China, were used in three experiments. We used a long short-term memory (LSTM)-based autoencoder to train the segments of the normal ECG. Based on the threshold of anomaly scores using mean squared error (MSE), it distinguished between normal and AF segments. In Experiment A, the best score was that of PreQ, which detected AF with an AUROC score of 0.96. In Experiment B and C for cross validation of each dataset, the best scores were also of PreQ, with AUROC scores of 0.9 and 0.95, respectively. To verify the significance of the anomaly score in distinguishing between AF and normal segments, we utilized an XG-Boosted model after generating anomaly scores in the three segments. The XG-Boosted model achieved an AUROC score of 0.98 and an F1 score of 0.94. AF detection using DL has been controversial among many physicians. However, our study differentiates itself from previous studies in that we can demonstrate evidence that distinguishes AF from normal segments based on the anomaly score.http://www.sciencedirect.com/science/article/pii/S240584402310805XAnomaly detectionElectrocardiogramAtrial fibrillationClinical evidenceAutoencoder
spellingShingle Sanghoon Choi
Kyungmin Choi
Hong Kyun Yun
Su Hyeon Kim
Hyeon-Hwa Choi
Yi-Seul Park
Segyeong Joo
Diagnosis of atrial fibrillation based on AI-detected anomalies of ECG segments
Heliyon
Anomaly detection
Electrocardiogram
Atrial fibrillation
Clinical evidence
Autoencoder
title Diagnosis of atrial fibrillation based on AI-detected anomalies of ECG segments
title_full Diagnosis of atrial fibrillation based on AI-detected anomalies of ECG segments
title_fullStr Diagnosis of atrial fibrillation based on AI-detected anomalies of ECG segments
title_full_unstemmed Diagnosis of atrial fibrillation based on AI-detected anomalies of ECG segments
title_short Diagnosis of atrial fibrillation based on AI-detected anomalies of ECG segments
title_sort diagnosis of atrial fibrillation based on ai detected anomalies of ecg segments
topic Anomaly detection
Electrocardiogram
Atrial fibrillation
Clinical evidence
Autoencoder
url http://www.sciencedirect.com/science/article/pii/S240584402310805X
work_keys_str_mv AT sanghoonchoi diagnosisofatrialfibrillationbasedonaidetectedanomaliesofecgsegments
AT kyungminchoi diagnosisofatrialfibrillationbasedonaidetectedanomaliesofecgsegments
AT hongkyunyun diagnosisofatrialfibrillationbasedonaidetectedanomaliesofecgsegments
AT suhyeonkim diagnosisofatrialfibrillationbasedonaidetectedanomaliesofecgsegments
AT hyeonhwachoi diagnosisofatrialfibrillationbasedonaidetectedanomaliesofecgsegments
AT yiseulpark diagnosisofatrialfibrillationbasedonaidetectedanomaliesofecgsegments
AT segyeongjoo diagnosisofatrialfibrillationbasedonaidetectedanomaliesofecgsegments