Distinct ECG Phenotypes Identified in Hypertrophic Cardiomyopathy Using Machine Learning Associate With Arrhythmic...

Area of Science:

• Cardiology
• Biomedical Engineering
• Computational Biology

Background:

• Ventricular arrhythmias are a primary cause of sudden cardiac death (SCD) in hypertrophic cardiomyopathy (HCM).
• Current risk stratification methods for HCM lack robust electrophysiological biomarkers.
• ECG analysis offers potential for identifying distinct HCM phenotypes and assessing SCD risk.

Purpose of the Study:

• To identify distinct HCM phenotypes using computational ECG analysis.
• To correlate these phenotypes with clinical risk factors and cardiac magnetic resonance (CMR) imaging findings.
• To evaluate the potential of ECG-based phenotyping for SCD risk stratification in HCM.

Main Methods:

• High-fidelity 12-lead Holter ECGs from 85 HCM patients and 38 controls were analyzed.
• Mathematical modeling and computational clustering were employed to identify ECG phenotypes.
• Clinical data and CMR imaging were used to assess hypertrophy extent and distribution within identified subgroups.

Main Results:

• Three HCM phenotypes were initially identified based on QRS morphology, with no significant differences in arrhythmic risk or hypertrophy distribution.
• Incorporating T wave analysis revealed four phenotypes, including a group (1A) with normal QRS and primary T wave inversion.
• Phenotype 1A exhibited significantly higher HCM Risk-SCD scores and a predominance of coexisting septal and apical hypertrophy compared to other groups.

Conclusions:

• Computational ECG phenotyping can classify HCM patients into distinct subgroups based on ECG features.
• Primary T wave inversion, independent of QRS abnormalities, is associated with increased SCD risk and specific hypertrophy patterns in HCM.
• ECG-based phenotyping holds promise as a novel, independent tool for SCD risk stratification in HCM patients.