P09.11.B 18F-FET-PET MULTIPARAMETRIC ANALYSIS FOR DIFFERENTIATING DISEASE RELAPSE AND RADIATION-RELATED CHANGES IN HIGH-GRADE GLIOMA PATIENTS

Abstract BACKGROUND In operated high-grade glioma (HGG) patients, 18F-FET-PET represents a useful tool to differentiate radiation-related changes (RRC) from disease recurrence after chemoradiotherapy. However, the most reliable 18F-FET-PET-derived parameter to aid in diagnosis is still debated. The aim of our study was to retrospectively calculate the most accurate parameter in our dataset and compare its value with the European Guidelines reference. MATERIAL AND METHODS Between 2019 and 2024, 45 consecutive HGG patients underwent 46 dynamic 18F-FET-PET scans after surgical resection and adjuvant chemoradiotherapy. A tumour-to-background ratio (TBR) cut-off value of 1.6 was used to automatically segment tumour volume (MTV) in visually positive studies. A standard VOI (VOIst) was centred on the region of maximum tumour uptake. Dividing the mean uptake of VOIst and MTV by background allowed calculation of TBRmax and TBRmean, respectively. Time-activity curves (TACs) of VOIst were plotted and classified by shape: increasing (Type 1), plateau (Type 2), or decreasing (Type 3); with Type 1 considered negative, and Type 2 or Type 3 as positive. Histological data and/or MRI imaging three months after 18F-FET-PET were used as gold standards to distinguish between relapse and RRC. RESULTS Of the 46 scans, 34 (74%) were classified as relapse and 12 (26%) as RRC based on follow-up (18/46) and histology (28/46). ROC curve analysis of SUVmax, MTV, TBRmax, and TBRmean yielded AUCs of 0.838, 0.563, 0.809, and 0.813, respectively. Optimal cut-off values, sensitivities, and specificities were as follows: SUVmax (2.34; 77%; 83%), MTV (3 cm³; 63%; 67%), TBRmax (1.49; 82%; 83%), and TBRmean (1.73; 88%; 67%). TAC-pattern dichotomous analysis showed a sensitivity of 71% and specificity of 83%. Using the guideline-recommended TBRmax cut-off of 1.9, sensitivity and specificity would have been 68% and 83%, respectively. CONCLUSION In high-grade glioma patients, the 18F-FET-PET-derived parameter TBRmax proved to be an accurate imaging tool for the differential diagnosis between radiation-related changes and disease relapse. However, the optimal cut-off value should still be determined based on larger population studies.