koreabiomed.com

AI predicts atrial fibrillation risk through ECG aging analysis: study

AI-powered analysis of electrocardiogram (ECG) aging can predict the risk of developing atrial fibrillation (AF), according to researchers from Severance Cardiovascular Hospital and Yonsei University College of Medicine.

From left, Professors Joung Bo-young and Yu Hee-tae of Severance Cardiovascular Hospital's cardiology division, and Professors You Seng-chan and Cho Seung-hoon of Yonsei University's Department of Biomedical Systems Informatics, developed an AI deep learning model that predicts the risk and early onset of atrial fibrillation through ECG aging analysis. (Courtesy of Severance Hospital)

From left, Professors Joung Bo-young and Yu Hee-tae of Severance Cardiovascular Hospital's cardiology division, and Professors You Seng-chan and Cho Seung-hoon of Yonsei University's Department of Biomedical Systems Informatics, developed an AI deep learning model that predicts the risk and early onset of atrial fibrillation through ECG aging analysis. (Courtesy of Severance Hospital)

The team, including Professors Joung Bo-young and Yu Hee-tae from Severance Cardiovascular Hospital’s cardiology division, alongside Professors Cho Seung-hoon, You Seng-chan, and Master's graduate Eom Su-jeong from Yonsei University's Department of Biomedical Systems Informatics, developed an AI deep learning model that predicts the risk and early onset of AF through ECG aging analysis.

The research, published in November in the European Heart Journal (Impact Factor: 39.3), is notable for its validation across different ethnic groups. The AI, trained on a Korean ECG database, was tested for consistency using data from prominent institutions such as Mayo Clinic and UK Biobank.

Correlation between AI-ECG age and actual age (Source: European Heart Journal)

Correlation between AI-ECG age and actual age (Source: European Heart Journal)

The team trained the AI model using approximately 1.5 million ECGs from Severance Hospital’s database and validated it with around 700,000 ECGs from six countries.

The AI model was then applied to ECG data from 280,000 individuals across four multinational cohorts. The analysis found that people with aged ECGs (Group A, 50,108 people) had a 1.86 times higher risk of developing AF compared to those with normal ECGs (Group B, 230,504 people). The risk of early onset AF before the age of 66 was also 2.07 times higher in Group A.

Additionally, the study found that for each year the ECG aging exceeded actual age, the risk of AF increased by 3 percent, and the risk of early onset rose by 4 percent.

The research team concluded that their findings confirm the link between ECG aging and AF development, suggesting that ECG aging could serve as a tool for early prediction and prevention of other heart diseases associated with aging.

“The AI developed to assess ECG aging allows for non-invasive prediction of atrial fibrillation and demonstrates the highest predictive accuracy among existing methods,” said Professor Joung. “Since ECG is a key biomarker in heart disease diagnosis, we expect this research to be applied to predict not only AF but also other heart diseases.

tweet

Related articles

Cancer immunotherapy that reflects Asian patients' characteristics will arrive soon

Undue dependence on government came back as pressure: medical professors

New clue found to predict death because of ANCA-associated vasculitis

Researchers find biomarkers impacting hepatocyte survival and repairs

Korean headache society pushes for CGRP as 1st-line migraine treatment

Kim Ji-hye jkim404@docdocdoc.co.kr

See Other Articles

Copyright © KBR Unauthorized reproduction, redistribution prohibited

Read full news in source page