TH‐AB‐BRB‐09: Effect of Hydrogel Spacer On Achievable Rectal DVH for IMRT in Prostate Cancer

Purpose: Implanted tissue spacers show great promise in reducing radiation dose to organs at risk (OAR) and/or permitting dose escalation to target volumes in IMRT plans. Dosimetric benefits have been demonstrated in a randomized study of such a device despite variations in plan optimization. The purpose of this analysis was to quantify the dosimetric effect of a tissue spacer using a validated prostate DVH prediction model trained with previous high-quality clinical prostate plans. Methods: Imaging and treatment planning for men undergoing IMRT for localized stage T1-T2 prostate cancer was performed before and after hydrogel spacer implantation between the prostate and anterior rectum. Pre- and post-implant plans were independently optimized to satisfy dose-volume criteria for rectum, bladder, and penile bulb. Image segmentation data were used to predict DVHs achievable using IMRT for rectum and bladder with institutionally-validated model parameters [Medical Physics 39, 7446 (2012); doi: 10.1118/1.4761864]. Predicted and actual clinical DVHs (cDVH) for rectum were used to compare pre- and post-implant plans. Results: 101 pairs of plans from 5 clinical sites treating at least 14 spacer subjects each were analyzed. An average ± s.d. absolute reduction of 9.4 ± 6.5 % in V60 and 7.0 ± 4.6 % in V70 was seen for pDVH. Corresponding reductions in cDVH were 10.3 ± 6.2 % and 8.0 ± 4.3, respectively. Standard deviations of corresponding pDVH and cDVH values were similar. Correlation coefficients between predicted and actual Vx values were computed for pre-implant, post-implant, and pre-post Vx differences. Correlations between pDVH and cDVH were significantly positive, but not uniform among clinical sites, suggesting differences in plan optimization. Conclusion: Predicted DVHs showed an absolute difference of nearly 10% in rectal volume receiving ≥ 60 Gy (50.4% relative reduction in V60) with use of peri-rectal spacer, consistent with analysis of clincal pre- and post-implant plans. This work was supported by a research grant from Augmenix, Inc.