Researchers from the Mayo Clinic have developed a new artificial intelligence (AI)-based approach to screening for hypertrophic cardiomyopathy.
The research team, publishing in the Journal of the American College of Cardiology, developed a convolutional neural network (CNN) that was validated using 12-lead electrocardiograms (ECGs) from 2,448 patients with verified diagnoses of hypertrophic cardiomyopathy (as well as more than 51,000 age- and sex-matched cohorts without hypertrophic cardiomyopathy). They looked particularly for the ability of the CNN to detect hypertrophic cardiomyopathy, and then proceeded to test it on another cohort of patients (612 patients with hypertrophic cardiomyopathy and 12,788 controls).
According to the study results, the area under curve (AUC) of the CNN in the validation dataset was 0.95 (95% CI, 0.94 to 0.97) at the optimal probability threshold of 11% for having hypertrophic cardiomyopathy. When the 11% probability threshhold was applied to the testing dataset, the AUC for the CNN was 0.96, indicating a strong ability to detect patients with hypertrophic cardiomyopathy.
“The good performance in patients with a normal ECG is fascinating,” Peter Noseworthy. MD, a study author and cardiologist at the Mayo Clinic, said in a press release. “It’s interesting to see that even a normal ECG can look abnormal to a convolutional neural network. This supports the concept that these networks find patterns that are hiding in plain sight.”
When testing the AI subgroups, the reserachers looked at a group diagnosed with left ventricular hypertrophy. The CNN AUC for detection of this condition was 0.95. The AUC for the subgroup with normal ECGs was also 0.95. For aortic stenosis, the AUC was 0.94.
“This is a promising proof of concept, but I would caution that, despite its powerful performance, any screening test for a relatively uncommon condition is destined to have high false positive rates and low positive predictive value in a general population,” Konstantinos Siontis, MD, co-author on the study and a cardiologist at Mayo Clinic, added. “We still need to better understand which particular populations will benefit from this test as a screening tool.”
#ECG read by Artificial Intelligence: is this the future of hypertrophic #cardiomyopathy (HCM) detection? Dr. @noseworthypeter and @MayoClinic colleagues develop an #AI approach for HCM screening based on 12-lead electrocardiography in the latest #JACC. https://t.co/hZeIv94HNY pic.twitter.com/zy39feDMxG
— JACC Journals (@JACCJournals) February 18, 2020
http://twitter.com/J_Design328/statuses/1229880691900309505
99% Negative predictive value in excluding #HCM: excellent screening capability for #AI -based ECG reading https://t.co/0blZkuB74D
— Santo Dellegrottaglie (@SantoDellegrot1) February 18, 2020
ECG-based detection of HCM by an AI algorithm can be achieved with high diagnostic performance, particularly in younger patients.
🔗https://t.co/p6RV9bG0wF pic.twitter.com/mpl1AcE7UW— Vicente Pallares (@vic_pallares) February 19, 2020
© 2025 Mashup Media, LLC, a Formedics Property. All Rights Reserved.