An artificial intelligence-powered electrocardiograph detected atrial fibrillation in ECGs with 90 percent accuracy, even if heart rhythm had returned to normal by the time of the screening, according to a new study.
The study was published Aug. 1 in The Lancet and led by researchers at Mayo Clinic in Rochester, Minn. The scientists trained the AI using approximately 450,000 ECGs from the Mayo Clinic's digital data vault, then tested it on the normal-rhythm ECGs of more than 36,000 patients.
The AI system was able to identify with remarkable accuracy the more than 3,000 patients in the test set who had experienced atrial fibrillation. The findings suggest that routine ECGs, which are noninvasive and now widely available via smartwatches and other consumer products, can reliably detect atrial fibrillation that is either impending or recent and asymptomatic.
"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," said senior author Paul Friedman, MD, chair of Mayo Clinic's cardiovascular medicine department. "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."
More articles on AI:
Microsoft leverages blockchain for AI initiatives
Google's AI anticipates 90% of kidney failures up to 48 hours prior to symptoms, study shows
HonorHealth to deploy AI-based system to improve patient flow