Heart failure & cardiomyopathies
Real-World Artificial Intelligence–Based Electrocardiographic Analysis to Diagnose Hypertrophic Cardiomyopathy
Authors' conclusion
Prospective real-world application of the AI-ECG algorithm to detect HCM was associated with a high degree of accuracy, varying with the chosen probability threshold. It also enabled the identification of 5% of patients with no prior HCM diagnosis.
Study context
- Population
- All patients undergoing 12-lead electrocardiograms at Cleveland Clinic, Cleveland, Ohio, between February 19, 2024, and November 1, 2024patients, 103,492 electrocardiograms analyzed
- Participants
- 45,873 patients
- Comparison
- AI-ECG diagnostic performance for hypertrophic cardiomyopathy (HCM) at various HCM probability thresholds (≥0.95, ≥0.90, and ≥0.85)
Methodological detailsprospectively analyzed, single center
- Prospectively analyzed
- Single center
Main finding
As reported in the abstract
Flagged potential HCM (unique patients)
1,265
2.7%
Of flagged patients, confirmed HCM
511
40.4%
Of flagged patients, new HCM diagnosis
63
5%
Of flagged patients, alternate diagnosis
691
54.6%
Further findings
AI-ECG diagnostic performance by HCM probability threshold
Sensitivity
Specificity and accuracy Threshold ≥0.85 >98%
Positive predictive value Threshold ≥0.95 66%
- AI-ECG diagnostic performance by HCM probability threshold: ≥0.85 provided the highest sensitivity; PPV highest at ≥0.95
AI-ECG performance by sex and age (subgroup)
- Performed similarly in both men and women
- More sensitive in individuals <50 years but more specific in individuals ≥50 years
Original publication
Summarised from the published abstract; every number is checked against it