Prediction of mortality from 12-lead electrocardiogram voltage data using a deep neural network
| Autor: | Joseph B. Leader, Martin C. Stumpe, Brandon K. Fornwalt, Sushravya Raghunath, Ashraf T. Hafez, Dustin N. Hartzel, Christopher M. Haggerty, H. Lester Kirchner, Aalpen A. Patel, Kipp W. Johnson, Christopher W. Good, Dominik Beer, Linyuan Jing, Alvaro E. Ulloa Cerna, Arun Nemani, David P. vanMaanen, Brian P. Delisle, Amro Alsaid, Tanner Carbonati, Joshua V. Stough, John M. Pfeifer, Katelyn Young |
|---|---|
| Rok vydání: | 2019 |
| Předmět: |
0301 basic medicine
Adult Male medicine.medical_specialty Risk Assessment General Biochemistry Genetics and Molecular Biology 03 medical and health sciences Electrocardiography 0302 clinical medicine Cardiologists Deep Learning Internal medicine Cause of Death medicine Humans cardiovascular diseases Mortality Cause of death Aged Proportional Hazards Models Retrospective Studies Aged 80 and over medicine.diagnostic_test Artificial neural network business.industry Proportional hazards model Deep learning Hazard ratio Retrospective cohort study General Medicine Middle Aged Prognosis 030104 developmental biology ROC Curve 030220 oncology & carcinogenesis Area Under Curve Cardiology Female Artificial intelligence Neural Networks Computer business Risk assessment Algorithms |
| Zdroj: | Nature medicine. 26(6) |
| ISSN: | 1546-170X |
| Popis: | The electrocardiogram (ECG) is a widely used medical test, consisting of voltage versus time traces collected from surface recordings over the heart1. Here we hypothesized that a deep neural network (DNN) can predict an important future clinical event, 1-year all-cause mortality, from ECG voltage–time traces. By using ECGs collected over a 34-year period in a large regional health system, we trained a DNN with 1,169,662 12-lead resting ECGs obtained from 253,397 patients, in which 99,371 events occurred. The model achieved an area under the curve (AUC) of 0.88 on a held-out test set of 168,914 patients, in which 14,207 events occurred. Even within the large subset of patients (n = 45,285) with ECGs interpreted as ‘normal’ by a physician, the performance of the model in predicting 1-year mortality remained high (AUC = 0.85). A blinded survey of cardiologists demonstrated that many of the discriminating features of these normal ECGs were not apparent to expert reviewers. Finally, a Cox proportional-hazard model revealed a hazard ratio of 9.5 (P |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full text from SpringerLink