Automated, Portable, and Low-Cost System for Home Screening of Peripheral Arterial Disease

Peripheral artery disease (PAD) is the manifestation of atherosclerosis where peripheral arteries are narrowed by the deposition of lipid and cholesterol and formation of fat fiber plaques on their walls. Occlusion and narrowing of the arteries may lead to several cardiovascular diseases and therefore in-time diagnosis and treatment of PAD is essential. There are several invasive and noninvasive methods to diagnosis of PAD, however, still an automated, easy-to-use, and affordable device for in-home monitoring of cardiovascular health is lacking. In this study, an oscillometric system is used to record arterial wall oscillations at different external pressures in lower and upper limbs. Wavelet transform is used to extract features from the amplitude of the recorded arterial wall oscillations and an adaptive neuro-fuzzy inference system (ANFIS) is used to identify PAD based on the extracted features. Because of the ANFIS high computational cost, linear discriminant analysis (LDA) is applied to the features to reduce their dimension before feeding them to ANFIS. The performance of proposed method is compared versus the conventional ankle-brachial test on a dataset of oscillometric recording obtained from 14 patients with PAD and 14 healthy individuals. Based on a five-fold cross validation, our method achieved an accuracy of 82% in detecting PAD. The results show promise toward the development of in-home cardiovascular monitors for cardiovascular health assessment.