The Silent Signal: Unmasking Myocardial Ischemia in a Resting Heartbeat with Machine Learning

Ischemic heart disease (IHD) remains the leading cause of morbidity and mortality worldwide, imposing a staggering burden on healthcare systems and societies. To assess the diagnostic capabilities of the single lead electrocardiography in the diagnosis of IHD using the machine learning model. A prospective, non-randomized, minimally invasive, single-center, case-control study enrolled male and female participants aged ≥40 years. All participants underwent resting single-lead electrocardiography (SLECG) and pulse wave recording using a portable Cardio-Qvark® device, and stress computed tomography myocardial perfusion imaging with vasodilation test. Based on coronary computed tomography perfusion (CTP) results, 80 participants were stratified into two groups: Group 1 with stress-induced myocardial perfusion defect (n = 31) and Group 2 without stress-induced myocardial perfusion defect (n = 49). Statistical processing carried out using the R programming language v4.2, Python V.3, Statistica 12 programme. (StatSoft, Inc. (2014). STATISTICA (data analysis software system), version 12. www.statsoft.com.), and SPSS IBM version 28. P considered statistically significant at <0.05. The best model performance in the diagnosis of IHD was Random Forest model. The model showed a diagnostic accuracy based on the parameters of the SLECG, AUC 0.988 [95 % confidence interval (CI); 0.967-1.000], Sensitivity 0.871 [95 % CI; 0.739-0.971], Specificity 0.959 [ 95 % CI; 0.894-1.000]. This study demonstrates that machine learning analysis of resting single-lead ECG signals, acquired via the portable Cardio-Qvark® device, achieves near-perfect accuracy (AUC 0.988) in diagnosing ischemic heart disease validated against myocardial perfusion imaging.