A Novel Classification Model Based on Hyperspectral Imaging for Predicting Response to Tacrolimus in Patients With Primary Membranous Nephropathy

Abstract

At present, the research to predict the efficacy of tacrolimus (TAC) mainly focuses on serological indexes and urine analysis. Because these indicators are affected by many factors, they cannot accurately predict the therapeutic effect of primary membranous nephropathy (PMN) patients. In this study, a novel classification model (RCN) based on hyperspectral imaging combined with one-dimensional convolutional neural networks (1D CNN) and relevance vector machine (RVM) was proposed for predicting patients' response to TAC. Based on the treatment outcomes of corticosteroids combined with TAC, the patients were divided into a remission group and a nonremission group. Through the analysis of hyperspectral data of pathological slices of patients in both the remission group and the nonremission group, the research results show that the model can effectively extract key features from the spectral data and achieve high classification performance, and it can predict the therapeutic effect of TAC in PMN patients.