Prediction of renal transplant rejection and acute tubular necrosis in renal transplant based on SVM

Abstract

Prevention and proper treatment of renal transplant rejection and acute tubular necrosis in kidney are the key to improving the long-term kidney transplant survival rate. Hence, it is important to predict the acute renal graft rejection in early stage. In recent years, there emerged some biomarkers measured through non-invasive techniques that may indicate the acute rejection. In this paper, we apply SVM method to analyze biomarkers, medullary R2* (MR2*) and cortical R2* (CR2*) in transplanted kidney, acquired through BOLD MRI for classification of patients with normally functioning kidney transplants and acute rejection in kidney, including acute allograft rejection and acute tubular necrosis. Furthermore, we use the classification model to predict the acute kidney rejection. The results show that the application of SVM in the analysis of CR2* and MR2* has its potential in prediction of acute rejection in kidney.