AI detects signs of AFib in asymptomatic patients

Advanced artificial intelligence (AI) models can be trained to identify abnormal heart rhythms in patients before they even start showing symptoms, according to a new analysis published in JAMA Cardiology.[1]

Specialists from Cedars-Sinai Medical Center, the University of California and Stanford University collaborated with two Veterans Affairs health networks in California on the research. The group’s goal was to improve outcomes for patients with atrial fibrillation (AFib), one of the most common heart rhythm disorders in the world, by detecting their disease as early as possible.

The study focused on outpatient electrocardiogram (ECG) data from more than 900,000 U.S. Veterans originally treated from 1987 to 2022. The mean age was 62.4 years old, and 93.6% were men. While 80% of those ECGs were used to train the algorithm, 10% were used for validation and the remaining 10% were used for testing purposes.

Overall, 3.1% of patients who underwent an ECG went on to experience AFib within the next 31 days. The group found that its AI model was able to accurately identify which patients were most likely to develop AFib. The model achieved an area under the ROC curve of 0.86 and an accuracy of 0.78 when tested on the Veterans Affairs data. Researchers then tried it out on an external dataset from Cedars-Sinai Medical Center, and it achieved an area under the ROC curve of 0.93 and an accuracy of 0.87.

“This research allows for better identification of a hidden heart condition and informs the best way to develop algorithms that are equitable and generalizable to all patients,” senior author David Ouyang, MD, a cardiologist with the Smidt Heart Institute at Cedars-Sinai, said in a prepared statement.

“This study of veterans was geographically and ethnically diverse, indicating that the application of this algorithm could benefit the general population in the U.S.,” added Sumeet Chugh, MD, a fellow cardiologist at Cedars-Sinai and director of its division of AI in medicine.  

Ouyang, first author Neal Yuan, MD, and the rest of the study’s authors are planning to continue their work on this algorithm in additional to developing additional AI models for other purposes.

Click here to read the full study in JAMA Cardiology.

Michael Walter
Michael Walter, Managing Editor

Michael has more than 18 years of experience as a professional writer and editor. He has written at length about cardiology, radiology, artificial intelligence and other key healthcare topics.

Around the web

Ron Blankstein, MD, professor of radiology, Harvard Medical School, explains the use of artificial intelligence to detect heart disease in non-cardiac CT exams.

Eleven medical societies have signed on to a consensus statement aimed at standardizing imaging for suspected cardiovascular infections.

Kate Hanneman, MD, explains why many vendors and hospitals want to lower radiology's impact on the environment. "Taking steps to reduce the carbon footprint in healthcare isn’t just an opportunity," she said. "It’s also a responsibility."