Optimal (68)Ga-PSMA and (18)F-PSMA PET window levelling for gross tumour volume delineation in primary prostate cancer

Item does not contain fulltext
PURPOSE: This study proposes optimal tracer-specific threshold-based window levels for PSMA PET-based intraprostatic gross tumour volume (GTV) contouring to reduce interobserver delineation variability. METHODS: Nine (68)Ga-PSMA-11 and nine (18)F-PSMA-1007 PET scans including GTV delineations of four expert teams (GTV(manual)) and a majority-voted GTV (GTV(majority)) were assessed with respect to a registered histopathological GTV (GTV(histo)) as the gold standard reference. The standard uptake values (SUVs) per voxel were converted to a percentage (SUV%) relative to the SUV(max). The statistically optimised SUV% threshold (SOST) was defined as those that maximises accuracy for threshold-based contouring. A leave-one-out cross-validation receiver operating characteristic (ROC) curve analysis was performed to determine the SOST for each tracer. The SOST analysis was performed twice, first using the GTV(histo) contour as training structure (GTV(SOST-H)) and second using the GTV(majority) contour as training structure (GTV(SOST-MA)) to correct for any limited misregistration. The accuracy of both GTV(SOST-H) and GTV(SOST-MA) was calculated relative to GTV(histo) in the 'leave-one-out' patient of each fold and compared with the accuracy of GTV(manual). RESULTS: ROC curve analysis for (68)Ga-PSMA-11 PET revealed a median threshold of 25 SUV% (range, 22-27 SUV%) and 41 SUV% (40-43 SUV%) for GTV(SOST-H) and GTV(SOST-MA), respectively. For (18)F-PSMA-1007 PET, a median threshold of 42 SUV% (39-45 SUV%) for GTV(SOST-H) and 44 SUV% (42-45 SUV%) for GTV(SOST-MA) was found. A significant pairwise difference was observed when comparing the accuracy of the GTV(SOST-H) contours with the median accuracy of the GTV(manual) contours (median, - 2.5%; IQR, - 26.5-0.2%; p = 0.020), whereas no significant pairwise difference was found for the GTV(SOST-MA) contours (median, - 0.3%; IQR, - 4.4-0.6%; p = 0.199). CONCLUSIONS: Threshold-based contouring using GTV(majority)-trained SOSTs achie