S2181
Physics - Inter-fraction motion management and daily adaptive radiotherapy
ESTRO 2026
Digital Poster 4321 Anatomical predictors of bladder constraint violations in post-prostatectomy SBRT: secondary analysis from POPART trial Valeria Faccenda 1,2 , Denis Panizza 1,2 , Martina Scotti 3 , Federica Ferrario 4 , Riccardo Ray Colciago 2 , Elena De Ponti 1,2 , Stefano Arcangeli 2,4 1 Medical Physics, Fondazione IRCCS San Gerardo dei Tintori, Monza, Italy. 2 School of Medicine and Surgery, University of Milan Bicocca, Milan, Italy. 3 Physics Department, University of Milan, Milan, Italy. 4 Radiation Oncology, Fondazione IRCCS San Gerardo dei Tintori, Monza, Italy Purpose/Objective: Post-prostatectomy SBRT planning requires strict adherence to bladder constraints to minimize the risk of urinary morbidity. Anatomical variability and difficulty maintaining a full bladder, especially in patients with pre-existing urinary dysfunction, may compromise plan robustness and impact delivered dose. Anticipating planning challenges is therefore crucial to minimize side-effects and avoid treatment delays due to repeated simulations. As a secondary analysis of the POPART trial (NCT04831970), we investigated whether anatomical features at planning could predict bladder dose constraint violations both on planned and delivered doses. Material/Methods: Sixty consecutive patients treated with salvage SBRT to the prostate bed in 5 fractions were retrospectively analyzed. Planning features included bladder volume, PTV volume, and their spatial relationships. A decision tree model was developed to identify predictors of bladder constraint violations, and its performance was assessed by accuracy, sensitivity, specificity, and kappa statistics. For a subset of 10 patients, delivered dose was estimated via synthetic CT (sCT) dose recalculations, created from CBCT using Mozi ARTAssistant module (Manteia Technologies). Variations in anatomical parameters and concordance between planned and delivered dose were evaluated. Results: The strongest anatomical predictor of constraint violations was the bladder–PTV overlap: values <13 cc consistently indicated low risk, whereas ≥ 13 cc predicted a high probability of constraint violations. In the intermediate range (7–13 cc), additional parameters refined risk stratification: larger PTV (>79 cc) combined with smaller bladder volumes (<66 cc) markedly increased the probability of planning challenges. The final model achieved 90% accuracy (95% CI: 0.70–0.96), with kappa = 0.76, sensitivity = 0.86, and specificity = 0.93. In the sCT subset, average variations across five fractions compared to planning CT were − 2% for PTV, +31% for bladder volume, and
dose difference of 0.6% relative to original reference plans (see Table 1). All five prospective patients were safely treated on a 1.5 T MRI-linac using the phantom- based workflow. The median (inter-quartile range) time for the initial fraction increased by 14.0 (10.4- 14.3) minutes compared with the conventional pathway, without deterioration in target coverage or organ sparing.
Conclusion: A simulation-free, phantom-based reference planning workflow for MRI-guided oART in intermediate-risk prostate cancer is feasible and clinically implementable. The approach removes the need for dedicated patient-specific simulation imaging while maintaining dosimetric quality, at the expense of a modest increase in treatment time at the first fraction. References: [1]: Lorenzen EL, et al. An open-source nnU-net algorithm for automatic segmentation of MRI scans in the male pelvis for adaptive radiotherapy. Front Oncol. 2023;13:1285725. Keywords: sim-free, online adaptive radiotherapy, prostate
Made with FlippingBook - Share PDF online