A multitask deep learning model utilizing electrocardiograms for major cardiovascular adverse events prediction
Abstract Deep learning analysis of electrocardiography (ECG) may predict cardiovascular outcomes. We present a novel multi-task deep learning model, the ECG-MACE, which predicts the one-year first-ever major adverse cardiovascular events (MACE) using 2,821,889 standard 12-lead ECGs, including traini...
Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2025-01-01
|
| Series: | npj Digital Medicine |
| Online Access: | https://doi.org/10.1038/s41746-024-01410-3 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1841559040365166592 |
|---|---|
| author | Ching-Heng Lin Zhi-Yong Liu Pao-Hsien Chu Jung-Sheng Chen Hsin-Hsu Wu Ming-Shien Wen Chang-Fu Kuo Ting-Yu Chang |
| author_facet | Ching-Heng Lin Zhi-Yong Liu Pao-Hsien Chu Jung-Sheng Chen Hsin-Hsu Wu Ming-Shien Wen Chang-Fu Kuo Ting-Yu Chang |
| author_sort | Ching-Heng Lin |
| collection | DOAJ |
| description | Abstract Deep learning analysis of electrocardiography (ECG) may predict cardiovascular outcomes. We present a novel multi-task deep learning model, the ECG-MACE, which predicts the one-year first-ever major adverse cardiovascular events (MACE) using 2,821,889 standard 12-lead ECGs, including training (n = 984,895), validation (n = 422,061), and test (n = 1,414,933) sets, from Chang Gung Memorial Hospital database in Taiwan. Data from another independent medical center (n = 113,224) was retrieved for external validation. The model’s performance achieves AUROCs of 0.90 for heart failure (HF), 0.85 for myocardial infarction (MI), 0.76 for ischemic stroke (IS), and 0.89 for mortality. Furthermore, it outperforms the Framingham risk score at 5-year MACEs and 10-year mortality prediction. Over 10-year follow-ups, the model-predicted-positive group exhibits significantly higher MACE incidences than the model-predicted-negative group (relative incidence ratio: HF: 15.28; MI: 7.87; IS: 4.74; mortality: 13.18). Using solely ECGs, ECG-MACE effectively predicts one-year events and exhibits long-term anticipation. It provides potential applications in preventive medicine. |
| format | Article |
| id | doaj-art-0d4722468b3c4d61bbd4301f5221255e |
| institution | Kabale University |
| issn | 2398-6352 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | npj Digital Medicine |
| spelling | doaj-art-0d4722468b3c4d61bbd4301f5221255e2025-01-05T12:47:20ZengNature Portfolionpj Digital Medicine2398-63522025-01-018111010.1038/s41746-024-01410-3A multitask deep learning model utilizing electrocardiograms for major cardiovascular adverse events predictionChing-Heng Lin0Zhi-Yong Liu1Pao-Hsien Chu2Jung-Sheng Chen3Hsin-Hsu Wu4Ming-Shien Wen5Chang-Fu Kuo6Ting-Yu Chang7Center for Artificial Intelligence in Medicine, Chang Gung Memorial HospitalCenter for Artificial Intelligence in Medicine, Chang Gung Memorial HospitalDepartment of Cardiology, Chang Gung Memorial HospitalCenter for Artificial Intelligence in Medicine, Chang Gung Memorial HospitalKidney Research Center, Department of Nephrology, Chang Gung Memorial HospitalDepartment of Cardiology, Chang Gung Memorial HospitalCenter for Artificial Intelligence in Medicine, Chang Gung Memorial HospitalCollege of Medicine, Chang Gung UniversityAbstract Deep learning analysis of electrocardiography (ECG) may predict cardiovascular outcomes. We present a novel multi-task deep learning model, the ECG-MACE, which predicts the one-year first-ever major adverse cardiovascular events (MACE) using 2,821,889 standard 12-lead ECGs, including training (n = 984,895), validation (n = 422,061), and test (n = 1,414,933) sets, from Chang Gung Memorial Hospital database in Taiwan. Data from another independent medical center (n = 113,224) was retrieved for external validation. The model’s performance achieves AUROCs of 0.90 for heart failure (HF), 0.85 for myocardial infarction (MI), 0.76 for ischemic stroke (IS), and 0.89 for mortality. Furthermore, it outperforms the Framingham risk score at 5-year MACEs and 10-year mortality prediction. Over 10-year follow-ups, the model-predicted-positive group exhibits significantly higher MACE incidences than the model-predicted-negative group (relative incidence ratio: HF: 15.28; MI: 7.87; IS: 4.74; mortality: 13.18). Using solely ECGs, ECG-MACE effectively predicts one-year events and exhibits long-term anticipation. It provides potential applications in preventive medicine.https://doi.org/10.1038/s41746-024-01410-3 |
| spellingShingle | Ching-Heng Lin Zhi-Yong Liu Pao-Hsien Chu Jung-Sheng Chen Hsin-Hsu Wu Ming-Shien Wen Chang-Fu Kuo Ting-Yu Chang A multitask deep learning model utilizing electrocardiograms for major cardiovascular adverse events prediction npj Digital Medicine |
| title | A multitask deep learning model utilizing electrocardiograms for major cardiovascular adverse events prediction |
| title_full | A multitask deep learning model utilizing electrocardiograms for major cardiovascular adverse events prediction |
| title_fullStr | A multitask deep learning model utilizing electrocardiograms for major cardiovascular adverse events prediction |
| title_full_unstemmed | A multitask deep learning model utilizing electrocardiograms for major cardiovascular adverse events prediction |
| title_short | A multitask deep learning model utilizing electrocardiograms for major cardiovascular adverse events prediction |
| title_sort | multitask deep learning model utilizing electrocardiograms for major cardiovascular adverse events prediction |
| url | https://doi.org/10.1038/s41746-024-01410-3 |
| work_keys_str_mv | AT chinghenglin amultitaskdeeplearningmodelutilizingelectrocardiogramsformajorcardiovascularadverseeventsprediction AT zhiyongliu amultitaskdeeplearningmodelutilizingelectrocardiogramsformajorcardiovascularadverseeventsprediction AT paohsienchu amultitaskdeeplearningmodelutilizingelectrocardiogramsformajorcardiovascularadverseeventsprediction AT jungshengchen amultitaskdeeplearningmodelutilizingelectrocardiogramsformajorcardiovascularadverseeventsprediction AT hsinhsuwu amultitaskdeeplearningmodelutilizingelectrocardiogramsformajorcardiovascularadverseeventsprediction AT mingshienwen amultitaskdeeplearningmodelutilizingelectrocardiogramsformajorcardiovascularadverseeventsprediction AT changfukuo amultitaskdeeplearningmodelutilizingelectrocardiogramsformajorcardiovascularadverseeventsprediction AT tingyuchang amultitaskdeeplearningmodelutilizingelectrocardiogramsformajorcardiovascularadverseeventsprediction AT chinghenglin multitaskdeeplearningmodelutilizingelectrocardiogramsformajorcardiovascularadverseeventsprediction AT zhiyongliu multitaskdeeplearningmodelutilizingelectrocardiogramsformajorcardiovascularadverseeventsprediction AT paohsienchu multitaskdeeplearningmodelutilizingelectrocardiogramsformajorcardiovascularadverseeventsprediction AT jungshengchen multitaskdeeplearningmodelutilizingelectrocardiogramsformajorcardiovascularadverseeventsprediction AT hsinhsuwu multitaskdeeplearningmodelutilizingelectrocardiogramsformajorcardiovascularadverseeventsprediction AT mingshienwen multitaskdeeplearningmodelutilizingelectrocardiogramsformajorcardiovascularadverseeventsprediction AT changfukuo multitaskdeeplearningmodelutilizingelectrocardiogramsformajorcardiovascularadverseeventsprediction AT tingyuchang multitaskdeeplearningmodelutilizingelectrocardiogramsformajorcardiovascularadverseeventsprediction |