AI predicts the risk of atrial fibrillation and is expected to be used in other heart diseases
Dec 06, 2024
|
Professor Information Young and Yoo Hee-tae of cardiology at Severance Cardiovascular Hospital, instructor Cho Seung-hoon, Professor Yoo Seung-chan of the Medical Life System Information Class at Yonsei University School of Medicine, and graduate student Um Soo-jung of the Graduate School of Life System Information announced on the 6th that they have developed an ECG aging analysis artificial intelligence deep learning model to predict the risk and early incidence of atrial fibrillation.
An electrocardiogram that records the heartbeat as an electrical signal is used to diagnose heart disease. Recently, artificial intelligence technology has been developed, and attempts are being made to predict heart conditions based on electrocardiogram analysis.
The research team developed an artificial intelligence deep learning model that can analyze the degree of ECG aging by learning about 1.5 million ECG databases owned by Severance Hospital. This was compared and analyzed with about 700,000 electrocardiogram data from six countries, and the procedure to verify the effectiveness of the AI model was completed.
In particular, this study is significant in that it verified whether AI learned with Korean electrocardiogram databases produces the same results even if the races are different through prominent overseas institutions such as the Mayo Clinic in the United States and the British Bio Bank.
Using the verified AI model, the research team analyzed the degree of ECG aging and the risk of atrial fibrillation in about 280,000 people who underwent ECG tests in four multinational cohorts.
As a result of the analysis, group A (50,108) with aging ECG had a 1.86 times higher risk of developing atrial fibrillation than group B (235,504). In addition, the risk of developing atrial fibrillation before reaching the age of 66 or older was 2.07 times higher in group A than in group B.
It was also confirmed that the incidence of atrial fibrillation increased by 3% and the risk of early onset by 4% as the age of measurement of the electrocardiogram increased by 1 year older than the actual age of the person.
The research team analyzed that the correlation between the degree of electrocardiogram aging and the occurrence of atrial fibrillation was confirmed for the results of this study. Accordingly, it is evaluated that early prediction and prevention of not only atrial fibrillation but also other heart diseases caused by aging will be possible based on the degree of ECG aging.
Professor Information Young Cho not only made it possible to predict the occurrence of non-invasive atrial fibrillation through AI developed to understand the degree of electrocardiogram aging, but also showed the best predictability among existing measurements"As electrocardiography is an important biomarker for diagnosing heart disease, we expect this study to be used not only for atrial fibrillation but also for predicting other heart diseases."
The results of this study were recently published in the European Heart Journal (IF 39.3), a prestigious journal in cardiology.
|
This article was translated by Naver AI translator.
