Graph Convolutional Network Approach for Predicting the Severity of Carpal Tunnel Syndrome

Abstract

This study proposes a framework based on graph convolutional network to predict the severity of carpal tunnel syndrome. The data of 100 patients diagnosed and treated for carpal tunnel syndrome (CTS) were included in this study resulting in 164 operated hands. The collected data include patient generalities, data from the clinical examination of the stage of CTS, electrophysiological study (EPS) and the data from BCTQ questionnaire. The data was prepared for modelling the graph.  A weighted graph that stores not only the features of each case but also the relationship between the cases is created.  We compared the model accuracy of graph convolutional network to different machine learning algorithms. The results showed that this model achieves higher result of 90 % accuracy.