FET PET in the evaluation of indeterminate brain lesions on MRI: Differentiating glioma from other non-neoplastic causes - A pilot study

Abstract

We aimed to determine the utility of FET PET in the management of indeterminate CNS lesions found on MRI. We performed a retrospective analysis of patients with FET PET at a single tertiary institution from 2011 to 2015. FET PET images were processed using usual methods and measurements taken including SUVmax, TBRmax, and analysis of dynamic series where available (Kipeak, Vdpeak, as well as tumor:background ratio for these variables). Correlation studies were performed using ANOVA between cohorts of high-grade histology, low-grade histology, and benign histology/stable on observation. Thirty-five patients were included, of whom 34 were suitable for analysis with median follow-up of 5months. The positive predictive value of FET PET in this cohort was 83.3%. FET SUVmax differentiated between patients with high-grade (mean SUV 3.38, 95% CI 2.21-4.55), low-grade (1.88, 95% CI 1.33-2.43) and benign/observation (1.42, 95% CI 1.13-1.71) cohorts (p=0.0003). Similarly, tumour to brain ratio was significant (p<0.0001). Kipeak distinguished between high grade and observation cohorts (p=0.036), as did KiTBR (p=0.025). Vd peak was not significantly different in these two cohorts (p=0.057) but Vd TBR was (p=0.041). In conclusion, FET PET demonstrated a high positive predictive value for glioma in patients with indeterminate brain lesions on MRI. The combination of negative FET and negative FDG PET scans may predict an indolent clinical course. Confirmatory trials are needed to establish the potential value of FET PET in guiding surgical management in this cohort.