Quality assurance of contouring and dosimetric consistency in the GETUG-AFU 31 salvage prostate stereotactic reirradiation trial.

Abstract

Salvage stereotactic body radiotherapy (SBRT) represents a promising approach for managing intraprostatic recurrence following radiotherapy. The multicenter phase I/II GETUG-AFU 31 trial evaluated 6 × 6 Gy SBRT, with delineation and dosimetric dummy runs performed to ensure consistent target and organ-at-risk (OAR) delineation and treatment plan quality across centers. We report the outcomes of this pre-accrual quality reassurance process. Eighteen centers received CT, PET, and multiparametric prostate MRI (mpMRI) datasets of recurrent lesions and contoured the gross target volume (GTV), clinical target volume (CTV), prostate, bladder/rectum (whole and wall), penile bulb, and urethra + 3 mm. The Dice similarity coefficient (DSC) and average Hausdorff distance (HD avg) were calculated against an expert reference. Each center subsequently generated a 6 × 6-Gy plan on another anonymized dataset. Dose-volume indices were compared with the protocol limits. Major deviations were subsequently revised. Of the 18 centers, 5 (28 %) required resubmission. Variability was high for the GTV (median DSC: 0.63 [interquartile range {IQR}: 0.45-0.70; HD avg: >1.8 mm), CTV: (0.79 [0.71-0.82]; HDavg: >1.8 mm), and urethra: (0.10 [0.05-0.26]; HD avg: >3.6 mm). Larger structures showed higher agreement: bladder, 0.93; rectum, 0.85; prostate, 0.85. All revised plans satisfied constraints: CTV D_median > 38.1 ± 1.4 Gy; rectal-wall V_27Gy averaged 0.33 ± 0.23 cc (≤1 cc), bladder-wall V_27Gy averaged 0.88 ± 0.32 cc (≤2 cc); urethral D_0.035cc was 37.0 ± 1.0 Gy (≤39 Gy). The reproducibility of the target volume and urethral delineation is essential in the context of salvage radiotherapy, for which formal guidelines are anticipated.