Ewha Medical Researchers Develop Deep Learning Model to Predict Atrial Fibrillation
- 작성처
- Date2026.07.31
- 2260
Ewha Medical Researchers Develop Deep Learning Model to Predict Atrial Fibrillation
A research team from the College of Medicine at Ewha Womans University has developed a deep learning-based artificial intelligence (AI) model that predicts the risk of developing atrial fibrillation with a standard electrocardiogram (ECG). Findings concerning this model, which could support clinical decision-making, were published in npj Digital Medicine (Impact Factor: 18.0), a leading international journal in digital medicine.
Atrial fibrillation is a common type of cardiac arrhythmia that increases the risk of stroke, heart failure, and death, but early diagnosis is often challenging because it may be asymptomatic or occur intermittently. The research team developed an AI model that can predict the risk of developing atrial fibrillation within three days to one month using only standard ECG results from patients with no apparent symptoms. The team then validated the model’s performance with data from more than 110,000 patients, including datasets from Ewha Womans University Medical Center, Beth Israel Deaconess Medical Center in Boston, Massachusetts, and seven university hospitals in Korea.
The validation results showed that the model achieved an AUROC (area under the receiver operating characteristic curve) of 0.87 in internal validation and 0.75 in external validation, demonstrating strong predictive capability. In addition, a clinical simulation involving 70 physicians from Korea and the United States showed that access to AI-generated predictions significantly improved clinicians’ ability to identify patients at risk of atrial fibrillation and increased the negative predictive value of their clinical assessments. It also enabled physicians to more accurately categorize patients without atrial fibrillation, thereby reducing unnecessary monitoring. These improvements were observed regardless of specialty or years of clinical experience and were particularly pronounced among general cardiologists.
The study demonstrates the AI-ECG model’s potential to serve as a digital tool not only for predicting disease but also for supporting clinical decisions, enhancing diagnostic accuracy and improving the efficiency of patient management in real-world clinical practice.