AI could enable accurate screening for atrial fibrillation

While common, atrial fibrillation often is fleeting. Therefore, it is challenging to diagnose. Atrial fibrillation may not occur during a standard 10-second, 12-lead EKG, and people are often unaware of its presence. Prolonged monitoring methods, such as a loop recorder, require a procedure and are expensive.

Accuracy and timeliness are important in making an atrial fibrillation diagnosis. Left undetected, atrial fibrillation can cause stroke, heart failure and other cardiovascular disease. Knowing that a patient has atrial fibrillation helps direct treatment with blood thinners, notes Paul Friedman, M.D., chair of the Department of Cardiovascular Medicine at Mayo Clinic. Dr. Friedman, who is a cardiac electrophysiologist, is the study's senior author.

"When people come in with a stroke, we really want to know if they had AF in the days before the stroke, because it guides the treatment," says Dr. Friedman. "Blood thinners are very effective for preventing another stroke in people with AF. But for those without AF, using blood thinners increases the risk of bleeding without substantial benefit. That's important knowledge. We want to know if a patient has AF."

Using approximately 450,000 EKGs of the over 7 million EKGs in the Mayo Clinic digital data vault, researchers trained AI to identify subtle differences in a normal EKG that would indicate changes in heart structure caused by atrial fibrillation. These changes are not detectable without the use of AI.

Researchers then tested the AI on normal-rhythm EKGs from a group of 36,280 patients, of whom 3,051 were known to have atrial fibrillation. The AI-enabled EKG correctly identified the subtle patterns of atrial fibrillation with 90% accuracy.

Dr. Friedman says that he is surprised by the findings of this research. If proven out, he said, AI-guided EKGs could direct the right treatment for disease caused by atrial fibrillation, even without symptoms. Moreover, this technology can be processed using a smartphone or watch, making it readily available on a large scale.

"An EKG will always show the heart's electrical activity at the time of the test, but this is like looking at the ocean now and being able to tell that there were big waves yesterday," says Dr. Friedman. "AI can provide powerful information about the invisible electrical signals that our bodies give off with each heartbeat -- signals that have been hidden in plain sight."

"Rather than finding the needle in the haystack by prolonged monitoring, authors basically suggest that AI will be able to judge by looking at the haystack if it has a needle hidden in it," notes Jeroen Hendriks, Ph.D., of the University of Adelaide in Australia, who wrote the study's commentary with Larissa Fabritz, M.D., of the University of Birmingham in the U.K.

Zachi I. Attia and Peter Noseworthy, M.D., are equal first authors of the study. Other authors are Francisco Lopez-Jimenez, M.D.; Samuel Asirvatham, M.D.; Abhishek Deshmukh, M.B.B.S.; Bernard Gersh, M.B., Ch.B., D.Phil.; Rickey Carter, Ph.D.; Xiaoxi Yao, Ph.D.; Alejandro Rabinstein, M.D.; Bradley Erickson, M.D., Ph.D.; and Suraj Kapa, M.D. -- all of Mayo Clinic.