JS03.4.A THE DIAGNOSTIC ACCURACY OFO-(2-[18F]-FLUOROETHYL)-L-TYROSINE POSITRON EMISSION TOMOGRAPHY ([18F]FET PET) IMAGING IN GLIOBLASTOMA IN DIFFERENTIATING BETWEEN EARLY TUMOR RECURRENCE AND PSEUDOPROGRESSION: A SINGLE CENTER STUDY AND A SYSTEMATIC

Abstract BACKGROUND The accuracy of O-(2-[18F]-fluoroethyl)-L-tyrosine ([18F]FET) PET is controversial when differentiating glioblastoma from pseudoprogression within 6 months after the STUPP regimen. We evaluated the diagnostic accuracy of a 20-minute [18F]FET PET within 6-month post-chemoradiation in detecting disease recurrence, and the results were compared with [18F]FET PET performed in the period of 7-12-month post-chemoradiation. A systemic review was conducted for comparison. MATERIAL AND METHODS One-hundred-sixty patients with MRI-suspected first glioblastoma recurrence who underwent [18F]FET PET during two time periods, early 1-6-month (n=101) and late 7-12-month (n=59) following chemoradiotherapy, were retrospectively studied. In the subgroup of patients (n=61), [18F]FET PET scan at radiotherapy planning was available for assessment of metabolic changes during treatment course using the PET RANO 1.0 criteria. PET scans were analyzed semi-quantitatively with maximum and mean tumor-to-normal brain-ratios (TBRmax, TBRmean) and biological tumor volume (BTV). Accuracies for disease recurrence were calculated using histopathology within 3 months or clinicoradiological follow-up without treatment change up to 6 months after [18F]FET PET. Cox proportional hazards models with overall survival (OS) was used to evaluate the prognostic value of [18F]FET PET. A literature search was performed using PubMed/Medline, EMBASE and Cochrane until December 2024 to identify cases resembling our cohort. RESULTS PET imaging performed within 6-month post-treatment showed only a trend towards significant differences in [18F]FET uptake in distinguishing disease recurrence from pseudoprogression with diagnostic accuracies of 68% for TBRmax and 60% for TBRmean except for a small subgroup from the early period (1-3 months) with significant but low accuracies of 74% and 69%. The BTV showed a significant high sensitivity of 94% but a specificity of only 30%. Using the PET RANO 1.0 criteria, patients were classified as complete responders (n=1), partial responders (n=31) and non-responders (n=29) and changes in [18F]FET uptake failed to predict the subsequent diagnosis (p=0.159). Multivariate analysis showed that BTV was inversely associated with OS (p=0.021). [18F]FET PET performed after 6-month post-chemoradiation showed significant high accuracies of 90% for TBRmax and 92% for TBRmean. The literature search yielded five studies including 71 patients. The pooled diagnostic accuracies of [18F]FET PET performed within 6-month post-treatment yielded 86% for TBRmax and 80% for TBRmean. Upon adding our results, the pooled diagnostic accuracies dropped to 78% and 59%. CONCLUSION The accuracy of [18F]FET PET in differentiating glioblastoma progression from pseudoprogression within 6-month post-chemoradiation may not be reliable and should be reserved for later time points.