IMG-88. Prognostic relevance of preoperative FET PET in patients with newly diagnosed glioblastoma

Abstract BACKGROUND The present study investigated the prognostic relevance of O-(2-[18F]fluoroethyl)-L-tyrosine (FET) PET parameters in newly diagnosed IDH-wildtype glioblastoma. METHODS Fifty patients with newly diagnosed and histomolecularly characterized glioblastoma according to the WHO 2021 classification who had undergone FET PET imaging prior to diagnostic biopsy or surgery and subsequent postoperative radiotherapy with concomitant and adjuvant temozolomide (n=36) or temozolomide plus lomustine (n=14) were retrospectively analyzed. The nnUNet-based JuST_BrainPET tool was used for segmentation of the FET PET tumor volume based on a tumor-to-brain ratio (TBR) of ≥1.6. All segmentations were visually checked. Quantitative PET parameters, i.e., maximum and mean TBR values, and metabolic tumor volumes (MTV), were correlated with overall survival (OS) using Cox regression models. Additional clinical parameters included age (range, 26-82 years), MGMT promoter methylation status (methylated in 58% of patients), RANO resection class (range, 2-4), treatment regimen, and postoperative Karnofsky Performance Status (range, 60-100%) and NANO score (range, 0-7 points), RESULTS In univariate Cox regression, preoperative MTV (hazard ratio [HR], 1.10; 95% CI, 1.04-1.17; p=0.002) and unmethylated MGMT promoter (HR, 2.82; 95% CI, 1.22-6.70; p=0.016) were the only parameters significantly associated with shorter OS. In multivariate analysis, MTV remained prognostic (p=0.003), as well as MGMT promoter methylation status (p=0.031). Model comparison using Akaike’s Information Criterion favored MTV over MGMT promoter methylation as the better overall prognostic fit. CONCLUSION These data support the integration of the FET PET tumor volume as prognostic biomarker in glioblastoma risk stratification. Further studies with larger datasets are needed to substantiate our findings.