Development of the GUI system “PAWP ViQE” for Supporting Cardiology Examination by Explainable Regression CNN

Abstract

The heart failure is related to death. To early detect the heart failure, it is required to measure Pulmonary Artery Wedge Pressure (PAWP) by the right heart catheterization. However, there are various risks of complications at carrying out this examination. Therefore, we developed a Regression Convolutional Neural Network (R-CNN) to estimate PAWP from a chest radiograph. Moreover, we integrated it into the GUI system called ”PAWP ViQE” to make it available to physicians. In this paper, we reported the development process.