S2149
Physics - Inter-fraction motion management and daily adaptive radiotherapy
ESTRO 2026
Digital Poster Highlight 3047 Structure Guided Dose Accumulation for Improved Prediction of Adverse Events in Prostate SBRT Owen McLaughlin 1 , Nicola Hill 2 , Orla A Houlihan 3 , Niamh L Clarke 4 , Christina E Agnew 5 , Stephen J McMahon 1 , Suneil Jain 1,2 , Conor K Mcgarry 5,1 1 Johnston Cancer Research Centre, Queen's University Belfast, Belfast, United Kingdom. 2 Department of Clinical Oncology, Northern Ireland Cancer Centre, Belfast Health and Social Care Trust, Belfast, United Kingdom. 3 Department of Clinical Oncology, Beacon Centre, Musgrove Park Hospital, Somerset NHS Foundation Trust, Somerset, United Kingdom. 4 Radiation Oncology, Peter MacCallum Cancer Centre, VIC, Australia. 5 Radiotherapy Physics, Northern Ireland Cancer Centre, Belfast Health and Social Care Trust, Belfast, United Kingdom Purpose/Objective: Dose accumulation is challenging in the pelvis due to large structure deformations between fractions [1]. This study evaluated a structure-guided dose accumulation methodology for prostate cancer patients and compared its ability to predict adverse events (AE) with planned dose. Material/Methods: A cohort of high-risk prostate cancer patients were treated with five fraction prostate and pelvic node SBRT (n=25) or prostate only SBRT (n=15) with a PTV dose of 36.25 Gy [2]. Patients received a SpaceOAR hydrogel spacer. Planning structures were delineated on CT images fused with diagnostic MR scans. Bladder and rectum were retrospectively delineated on each fractions’ pre-treatment CBCT using Eclipse (Varian, v17.0.0).Dose grids were recalculated on CBCTs using a standard CT calibration curve and the Acuros external beam dose calculation algorithm (v16.1.0). Structure guided deformable registration was performed in Velocity (v4.1) to generate deformable registrations for bladder and rectum contours between pre-treatment CBCTs and CT. CBCT dose grids were resampled to planning CT coordinates using registrations, and summed. Planned and accumulated dose-volume histograms (DVHs) were calculated in Eclipse.DVH metrics were tested using univariable logistic regression for significance (p<0.05) in predicting RTOG late grade 1+ gastrointestinal (GI) and grade 2+ genitourinary (GU) AE. Median DVH metrics with bootstrapped confidence intervals (CI) were calculated for patients grouped by toxicity status. Logistic regression models using planned and accumulated dose were compared using area-under the curve (AUC) with 95% CI. Results: Planned bladder V21Gy–V24Gy(%), D30cc and D35cc were significant predictors of GU AE. No significant
for CBCT protocol optimization. Deformable propagation and machine learning-based
segmentation performed differently across organs-of- interest (see Figure 1b). Dose recalculation on cCBCT demonstrated high consistency with rescanned planning CT, with Gamma passing rates (2%, 2mm) at the 90% dose level reaching up to 99% (see Figure 2).Retrospective cCBCT-based recalculation results show significant volumetric, geometric and dosimetric changes in target and organ-of-interest, supporting the decision of plan adaptation. All cCBCT generation workflows succeeded except for one case, where generation failed due to a major anatomical change from complete lung atelectasis to full re-expansion.
Conclusion: The corrected CBCT algorithm demonstrated strong potential for enabling ART in conventional c - arm linac. The workflow achieved high HU accuracy and dosimetric reliability, supporting feasibility for ART in non-integrated settings. Further work will focus on implementing this workflow for triggering prospective ART. Keywords: CBCT-based, dose-guided, online-ART
Made with FlippingBook - Share PDF online