Combination of Classifiers to Detect Grade of Glioblastoma using MRS

Abstract

Glioma is classified into Low Grade Glioma and High Grade Glioma. Because of difference in patients’ survival, it is important to classify patients into each grade. In this study, 214 glioma patients were classified into these two grades using Choline and N-acetylaspartate as extracted metabolites from Magnetic Resonance Spectroscopy and combination of Support Vector Machine (SVM), Multi-Layer Perceptron (MLP), and Artificial Neuro-Fuzzy Inference System (ANFIS) as classifiers. Results showed that using ANFIS had the best accuracy between classifiers and combination of MLP and ANFIS had the best accuracy between different combinations of classifiers. Moreover, results showed that, concentration of Choline and N-acetylaspartate are suitable features to discriminating low and high grade glioma.