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

Abstract

RESULTS

ROC curve analysis for ⁶⁸Ga-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 ¹⁸F-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).

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.

CONCLUSIONS

Threshold-based contouring using GTV_majority-trained SOSTs achieves an accuracy comparable with manual contours in delineating GTV_histo. The median SOSTs of 41 SUV% for ⁶⁸Ga-PSMA-11 PET and 44 SUV% for ¹⁸F-PSMA-1007 PET form a base for tracer-specific window levelling.

TRIAL REGISTRATION

Clinicaltrials.gov ; NCT03327675; 31-10-2017.

METHODS

Nine ⁶⁸Ga-PSMA-11 and nine ¹⁸F-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.