S1884
Physics - Dose prediction/calculation, optimisation and applications for photon and electron planning
ESTRO 2026
References: 1 Royal Marsden NHS Foundation Trust. A Feasibility Study of Dose De-Escalation in Prostate Radiotherapy Using the Magnetic Resonance Linear Accelerator (MRL). clinicaltrials.gov; 2024. https://clinicaltrials.gov/study/NCT057094962 Cooper S, et al. DESTINATION 1 preliminary dosimetry analysis. Clinical Oncology. 2025; 38. doi:10.1016/j.clon.2024.10.0153 Eiben B, et al. 4505 Delivered dose reconstruction for organs with differential motion. Radiotherapy and Oncology. 2025; 206: p.S3352-S3354. doi:10.1016/S0167- 8140(25)03424-34 Persson E, et al. Real-time motion- including dose estimation of simulated multi-leaf collimator-tracked magnetic resonance-guided radiotherapy. Medical Physics. 2024; 51(3): p.2221- 2229. doi:10.1002/mp.16798 Keywords: Delivered dose reconstruction, MR-guided RT Mini-Oral 2378 Automated Grid Optimisation for Spatially Fractionated (Lattice) Radiotherapy Luca Mulcahy 1,2 , Marie-Claude Biston 1,2 , David Sarrut 2 , Marie-Pierre Sunyach 1 , Thomas Baudier 2 , Vanina Isnardi 1 1 Radiotherapy, Centre Léon Bérard, Lyon, France. 2 CREATIS, Université de Lyon, Lyon, France Purpose/Objective: Lattice Radiotherapy (LRT) shows promise for treating large, radioresistant tumors, but current treatment planning workflow remains operator-dependent and lacks standardization. (1). This work aimed to develop and evaluate an automated treatment-planning- system (TPS) independent algorithm for generating clinically acceptable LRT sphere geometries while improving reproducibility and efficiency. Material/Methods: Between 2021-2025, 21 sarcoma patients were treated at our hospital using lattice technique. Treatment planning involved the use of MIM Maestro (MIM software inc.) to manually position the spheres, following the geometry described by Duriseti et al.(2). A method was developed to generate a configurable three-dimensional grid of high-dose spheres and optimise its placement within the GTV. Given a grid centered on the GTV centroid, the algorithm optimises its placement-through rotation and translation-to maximise the number of spheres fully contained within the target volume. Voxel-based inclusion testing selects the configuration with the highest score, and resulting spheres are exported as an RTSTRUCT DICOM file for clinical planning. Additional functionality includes optional enlarged spheres for
reconstructed the dose to a static patient anatomy (static). Results: Figures 1 and 2 show dose reconstruction results for the D90%(GTV+4mm) and the D2%(rectum), respectively. Rectal-dose differences between the ISO- shift and CDRP-method were observed for individual fractions indicating an effect of differential motion. Out of 66 fractions 55/57(83%/86%) fulfilled the primary endpoint according to the ISO-shift/CDRP- method. In comparison, 70% of fractions fulfilled this endpoint when evaluated on the pMRI3.
Conclusion: Our point-cloud-based dose-reconstruction
framework enables precise delivered doses to be computed from each structures’ frame of reference and thus accounts for differential inter-fraction motion whilst utilising beam-on imaging and logged delivery data. This method identifies a larger number of fractions fulfilling the primary endpoint than dose- recalculations on the post-treatment anatomy, suggesting the latter could provide a worst-case estimate. However, precise delivered doses will be critical for assessing the clinical impact of new advanced treatment regimes.
Made with FlippingBook - Share PDF online