A new artificial intelligence (AI) tool has reported its role in detecting hypertrophic cardiomyopathy (HCM) by analyzing standard 12-lead electrocardiograms (ECGs). HCM affects about 1 in 500 people; however, it often goes undiagnosed, especially in the young, leading to sudden cardiac death. Early detection is essential. This study suggests that AI can assist the clinicians in ealry diagnosis. The study was published in Circulation: Heart Failure.

The device, called Viz HCM, uses deep learning algorithms to screen ECGs and provide a simple binary output indicating whether HCM is suspected or not. Researchers tested the tool on a large dataset comprising 293 confirmed HCM cases and 2,912 non-HCM cases from three hospitals. The reserchers confirmed the diagnoses with chart review, cardiac imaging, and diagnostic codes.

The AI system successfully produced results for 99.3% of HCM-positive and 99.8% of HCM-negative cases. It detected HCM with a sensitivity of 68.4%, meaning it correctly identified nearly seven out of ten cases. Its specificity was 99.1%, indicating it rarely produced false positives. The area under the curve (AUC) was 0.975, a strong indicator of diagnostic accuracy.

Given the low general population prevalence of HCM, the positive predictive value was 13.7%, but its negative predictive value was 99.9%, making it a reliable tool for ruling out the disease. AI maintained consistent performance across different technical subgroups and demographics.