Single-lead ECG-based machine learning model for noninvasive detection of left ventricular hypertrophy in hypertension

OBJECTIVE: Left ventricular hypertrophy (LVH) is classified as concentric or eccentric based on left ventricle relative wall thickness. Using machine learning techniques on basic clinical parameters and features from a single-lead electrocardiogram (ECG), we detected LVH in a hypertensive population without cardiovascular disease. METHODS: We enrolled 812 subjects with hypertension with no indications of cardiovascular disease. Based on left ventricular mass index and relative wall thickness, the subjects were classified into two groups; i) normal geometry and concentric remodeling were classified as no-LVH, whereas ii) concentric and eccentric hypertrophy were classified as having LVH. We trained a Random Forest to distinguish between the two categories. For comparison, we also trained a logistic regression and a convolutional neural network model. We performed feature importance and interaction analysis using SHAP to interpret the model's predictions for feature importance and feature interactions. RESULTS: Our model was able to distinguish subjects with no-LVH from the ones with LVH, with an ROC/AUC of 0.82 (95% CI: 0.71-0.91) and an average precision of 0.61. At threshold 0.3, specificity is 81% and sensitivity is 58%. Age, corrected QT interval, T wave duration, and being female were the most important features that contributed to the model's predictions. CONCLUSION: Using an analysis of single-lead ECG combined with clinical data, we demonstrated strong performance in detecting LVH. The identification of key ECG features, such as corrected QT interval and T wave morphology, underscores the clinical value of single-lead ECG analysis. These findings are particularly important in the era of wearable devices, where accessible, noninvasive screening for cardiac conditions such as LVH can be integrated into everyday health monitoring.