Detection and Severity Assessment of Peripheral Occlusive Artery Disease via Deep Learning Analysis of Arterial Pulse Waveforms: Proof-of-Concept and Potential Challenges.

Sooho Kim,Jin-Oh Hahn,Byeng Dong Youn

doi:10.3389/fbioe.2020.00720

Abstract

Toward the ultimate goal of affordable and non-invasive screening of peripheral occlusive artery disease (PAD), the objective of this work is to investigate the potential of deep learning-based arterial pulse waveform analysis in detecting and assessing the severity of PAD. Using an established transmission line model of arterial hemodynamics, a large number of virtual patients associated with PAD of a wide range of severity and the corresponding arterial pulse waveform data were created. A deep convolutional neural network capable of detecting and assessing the severity of PAD based on the analysis of brachial and ankle arterial pulse waveforms was constructed, evaluated for efficacy, and compared with the state-of-the-art ankle-brachial index (ABI) using the virtual patients. The results suggested that deep learning may diagnose PAD more accurately and robustly than ABI. In sum, this work demonstrates the initial proof-of-concept of deep learning-based arterial pulse waveform analysis for affordable and convenient PAD screening as well as presents challenges that must be addressed for real-world clinical applications.

Highlights

Peripheral artery occlusive disease (PAD) is a highly prevalent vascular disease associated with high morbidity and mortality risks
We evaluated our deep learning (DL)-based pulse waveform analysis (PWA) approach to peripheral occlusive artery disease (PAD) diagnosis and compared its efficacy with the state-of-the-art ankle-brachial index (ABI) technique, in terms of PAD detection and severity assessment efficacy, using the test dataset constructed in section “Creation of Virtual PAD Patients.”
We evaluated our approach and ABI technique using the 20,000 arterial blood pressure (BP) and BF waveform data of these 2000 virtual patients by (i) classifying each arterial BP and BF waveform data sample into healthy or PAD category based on the PAD severity predicted by the deep convolutional neural network (CNN) when the brachial and ankle BP waveforms in the sample were inputted and the ABI value computed from the waveforms, (ii) aggregating the classification results across all the 20,000 data samples associated with all the 2000 virtual patients, and (iii) computing the sensitivity and specificity as well as the accuracy of PAD detection

Summary

Introduction

Peripheral artery occlusive disease (PAD) is a highly prevalent vascular disease associated with high morbidity and mortality risks. It was estimated that >8 million and >200 million people were suffering from PAD in the United States (in 2000) (Allison et al, 2007) and globally (in 2010) (Fowkes et al, 2013), and the number of PAD patients is projected to sharply increase with societal aging It makes a significant adverse impact on morbidity and quality of life, and carries significant mortality implications as a powerful predictor of coronary artery disease and cerebrovascular disease (Golomb et al, 2006). PAD diagnosis necessitates angiography techniques (Guthaner et al, 1983; Romano et al, 2004; Cavallo et al, 2019) These techniques are not ideally suited to affordable and convenient PAD detection and severity assessment. It is often criticized for its limited accuracy and robustness in diagnosing PAD (Nelson et al, 2012)

Objectives

Methods

Results

Discussion

Conclusion