Deep Learning-Based Heart Sound Analysis for Left Ventricular Diastolic Dysfunction Diagnosis.

Yang Yang,Hui Wang,Yi-Neng Zheng,Xing-Ming Guo

doi:10.3390/diagnostics11122349

Abstract

The aggravation of left ventricular diastolic dysfunction (LVDD) could lead to ventricular remodeling, wall stiffness, reduced compliance, and progression to heart failure with a preserved ejection fraction. A non-invasive method based on convolutional neural networks (CNN) and heart sounds (HS) is presented for the early diagnosis of LVDD in this paper. A deep convolutional generative adversarial networks (DCGAN) model-based data augmentation (DA) method was proposed to expand a HS database of LVDD for model training. Firstly, the preprocessing of HS signals was performed using the improved wavelet denoising method. Secondly, the logistic regression based hidden semi-Markov model was utilized to segment HS signals, which were subsequently converted into spectrograms for DA using the short-time Fourier transform (STFT). Finally, the proposed method was compared with VGG-16, VGG-19, ResNet-18, ResNet-50, DenseNet-121, and AlexNet in terms of performance for LVDD diagnosis. The result shows that the proposed method has a reasonable performance with an accuracy of 0.987, a sensitivity of 0.986, and a specificity of 0.988, which proves the effectiveness of HS analysis for the early diagnosis of LVDD and demonstrates that the DCGAN-based DA method could effectively augment HS data.

Highlights

Left ventricular diastolic dysfunction (LVDD) is a clinical syndrome characterized by inadequate active relaxation and decreased cardiac output, resulting in elevated enddiastolic pressure and possibly alterations in cardiac function [1]
When the manifestations of heart failure (HF) such as dyspnea, edema, and fatigue gradually appear in the LVDD population, it will move towards an irreversible stage and cause progression to HFpEF [6,7], and around half of all HF hospital admissions are accounted for by patients with HFpEF [8]
Since the deep convolutional generative adversarial networks (DCGAN) model can generate a large number of new samples that are closer to real samples while maintaining the validity of the semantic features [20,47], it is used to generate more samples to expand datasets in this paper

Summary

Introduction

Left ventricular diastolic dysfunction (LVDD) is a clinical syndrome characterized by inadequate active relaxation and decreased cardiac output, resulting in elevated enddiastolic pressure and possibly alterations in cardiac function [1]. The overall incidence of LVDD in the general population is approximately 30% [2,3,4,5] and has a positive correlation with all-cause mortality. When the manifestations of heart failure (HF) such as dyspnea, edema, and fatigue gradually appear in the LVDD population, it will move towards an irreversible stage and cause progression to HFpEF [6,7], and around half of all HF hospital admissions are accounted for by patients with HFpEF [8]. Early diagnosis of LVDD is of great significance for preventing the deterioration of cardiac function as well as timely treatment

Methods

Discussion

Conclusion