Multiparametric prediction of glioma grade and IDH status using 18F-FET PET/CT dynamic and texture analysis

Abstract Background Gliomas are the most frequent malignant brain tumors and are heterogeneous in histology, genetics, and outcome. Mutations in isocitrate dehydrogenase (IDH) have been described by several studies as an independent predictor of better survival. We aimed at evaluating the respective value of O-(2-[18F]fluoroethyl)-L-tyrosine (18F-FET) PET/CT dynamic and texture analysis in patients with untreated glioma to assess grade and IDH mutation status. Methods In total, 73 patients (male: 48, median age: 47 [36–59]) who underwent an 18F-FET PET/CT for initial glioma evaluation were retrospectively included. Twenty patients had a WHO grade II and forty-one a grade III-IV glioma. IDH status was available in 61 patients. Time-activity-curve (TAC) type and 20 parameters (conventional, texture, shape and histogram derived) obtained from static analysis using the LIFEx© software were recorded. Respective performance was assessed using receiver operating characteristic (ROC) curve analysis and stepwise multivariate regression analysis. Results TAC type and several texture parameters derived from static parameters showed satisfactory to good performance to predict glioma grade and IDH status. Both TAC type (stepwise OR: 28.8 [95%CI: 4.73-175.14], p < 0.001) and NGLDM coarseness (stepwise OR: 9.7 x 1019 [95%CI: 1.47–6.5 x 1039], p = 0.048) were independent predictors of glioma grade. However, only TAC type was an independent predictor of IDH status (OR: 3.74 [95%CI: 1.17-12.0], p = 0.026). Conclusions 18F-FET PET/CT dynamic and texture analysis may help predict both glioma grade and IDH status, especially in low-grade tumors.