3230 - Daily CBCT-Based Dose Accumulation Unveils Spatial Dependency of Dominant Intraprostatic Lesion Coverage and Urinary Toxicity in Hypofractionated Prostate Radiotherapy

Purpose/Objective(s):

This study aimed to quantify the spatial discrepancy and toxicity outcomes between planned and delivered doses to dominant intraprostatic lesions (DILs) and organs-at-risk (OARs) in hypofractionated prostate radiotherapy, utilizing a novel deformable dose accumulation framework integrating multi-modal imaging (MRI/PSMA-PET) and daily CBCT.

Materials/Methods:

In a cohort of 23 prostate cancer patients (575 CBCT series) undergoing moderate hypofractionated radiotherapy (70Gy/25f) with 5mm PTV margins (0mm posteriorly). Fused multiparametric MRI and PSMA-PET/CT datasets were utilized for precise delineation of DILs, CTV, rectum, and urethra on planning CT (pCT). A dual-phase image registration protocol was implemented: initial rigid registration with daily couch shift correction enabled fractional dose transfer. Then structure propagation via intensity-based deformable registration. After the same physician modified the structures, the contour-based hybrid DIR was performed. Cumulative dose was reconstructed through 25-fraction DVF stacking. Daily CBCT image guidance started with bone-matching, then prostate matching, and adjustments based on the rectum's anterior wall was adopted. Radiobiological modeling incorporated EUD (a=-4 for CTV) and Lyman-Kutcher NTCP models (n=0.09, m=0.13, TD50 = 78.5Gy for Grade>2 GI toxicity; n=0.12, m=0.24 , TD50 = 81.7Gy for Grade>2 GU toxicity). Dice similarity coefficient (DSC) and mean distance agreement (MDA) were used to evaluate the DIR accuracy. Statistical comparisons (planned vs. delivered doses) employed using paired t-test if they were normally distributed, otherwise using Wilcoxon rank test, p < 0.05 (statistical software).

Results:

Median DSC were all more than 0.9 (CTV=0.93, Rectum=0.95, Bladder=0.94), the MDA were all less than 3mm. The median D98 of the CTV was 66.59Gy compared to the planned 69.27Gy, with a V95 of 98.08% versus 99.41%. Despite a statistically significant decrease, dosimetric requirements were still met. The median dose to 100% of DILs was 71.41Gy, but two patients showed >5% underdose due to the DILs proximity to the rectal anterior wall, highlighting that DILs coverage depends on its position. Rectum protection achieved (NTCP=0.24% vs planned 0.30%, p=0.32). The bladder's high-dose area (V70~V73.5) decreased, while V40~V60 increased significantly. GU toxicity (NTCP=15.62%) was 64x higher than rectum toxicity and requires more attention.

Conclusion:

Daily rectal anterior wall-based matching with the reduced PTV margin, effectively protects the rectum while ensuring sufficient CTV coverage. Nevertheless, DIL coverage critically depends on lesion location relative to rectal interface. Care should be taken to avoid GU toxicity, and, when necessary, adaptive protocols should be adopted in hypofractionated prostate radiotherapy.