Alternative method to pre-diagnosed coronary artery disease using photoplethysmography: a lesson from COVID-19 pandemic

Abstract

Ischemic heart disease (IHD) is one of the underlying factors that contribute to mortality in COVID-19 infected patients. IHD or coronary artery disease (CAD) is commonly diagnosed using invasive coronary angiography (ICA) or computed tomography angiography (CTA). However, these imaging modalities are costly, operationally complex and hardly accessible, especially during the pandemic. Thus, researchers have great interest in using non-invasive techniques of electrocardiography (ECG) and photoplethysmography (PPG) as alternatives to pre-diagnose the disease. This study focused on the detection of the severity of stenosis in the coronary artery using PPG among newly diagnosed IHD patients. A total of 88 patients of Hospital Canselor Tuanku Muhriz were involved. They were grouped as having severe stenosis if their stenosis percentage are at 70% or more, based on ICA or CTA evidence. A total of 73 time-domain features were analyzed in this study. Five machine learning methods were investigated to categorize the patients using up to 15 selected features. Results showed that the Discriminant Analysis method performed the best with accuracy, sensitivity and specificity of 88.46%, 100% and 70%, respectively. In conclusion, the severity of stenosis in coronary arteries has a high potential of being detected using simple non-invasive tools of PPG.