Coronary Artery Disease Classification from Photoplethysmographic Signals

Abstract

Photoplethysmography is a non-invasive and low-cost modality for assessing blood oxygen and volume variations. It is used extensively by physicians for basic monitoring tasks. These signals, however, as prior work has shown, have a plethora of interesting features and can be used for the diagnosis of several cardiovascular diseases. In this article, we aim to detect Coronary Artery Disease (CAD) using Photoplethysmographic signal features. We outline a simple signal processing method to extract these features. Machine learning-based approaches are then used to train classifiers that detect the presence of cardiac distress based on these features. We observe that the proposed method is effective in detecting CAD in a MIMIC-III, a benchmark data set. The technique can be used for low-cost monitoring of early signs of cardiac diseases.