S2005
Physics - Dose prediction/calculation, optimisation and applications for photon and electron planning
ESTRO 2026
exploration of trade-offs between target coverage, healthy tissue sparing and robustness. Balancing robustness against varying uncertainty magnitudes reduces sensitivity to the accuracy of the uncertainty model and supports informed decision-making, while unraveling the non-trivial correlation among VROs.The approach was implemented in matRad[2] and tested on a 4DCT lung cancer case for a 7 fields IMRT plan. Uncertainties included target motion across 10 CT- phases and two sets of setup and range errors, one at standard error magnitude and another with 20% increase in standard deviation for range and 3 times larger setup error. The objective space comprised two VROs, each defined on 100 error scenarios, and a dose reduction objective for the lung. Results: Figure 1 illustrates 3D Pareto front projections, highlighting the trade-off between target variance and lung sparing, and correlation between the two VROs, further detailed by DVHs and Standard-Deviation- Volume-Histograms (SDVHs) analyses reported for two selected solutions. The correlation between VRO is indeed non-trivial. Variance reduction can be fully explored while aware of its impact on coverage and OAR sparing. Figure 2 presents the probabilistic evaluation for the two solutions, including the expected dose distribution and coverage passing rates for dose higher than 95% of prescription, for both uncertainty models. The coverage confidence for each voxel is higher for the more robust solution and depends on the uncertainty model analysed.
Conclusion: This study demonstrates how combining MCO with scenario-free RO enhances decision-making and enables incorporation of multiple forms of uncertainty as tradable criteria. The proposed framework increases treatment personalization enabling an informed decision-making process and overcomes key [1] Cristoforetti R, Hardt JJ, Wahl N. Scenario-free robust optimization algorithm for IMRT and IMPT treatment planning. Med Phys. 2025 Jul;52(7):e17905. doi: 10.1002/mp.17905. Epub 2025 May 25. PMID: 40414693; PMCID: PMC12258008.[2] Abbani, N. et al (2025). matRad (v3.2.0). Zenodo. https://doi.org/10.5281/zenodo.17486111 Keywords: Robust Optimisation, Multi-criteria Optimisation Digital Poster 4609 Can target partitioning reduce valley-to-peak dose for SFRT lattice VMAT? Agnes Angerud, Johan Sundström Research, RaySearch Laboratories, Stockholm, Sweden methodological and practical limitations of conventional RO in treatment planning. References: Purpose/Objective: SFRT lattice treatments has been introduced for symptom relief and tumor debulking for large unresectable tumors [1][2]. Using the face-centered- cubic grid described by Duriseti [2] with cube side 6 cm, a 850cc GTV/LTV (lattice target volume) could contain as much as 16 target spheres. Recently Sundström et al. investigated strategies for target division for SRS treatments of patients with many brain metastases [3]. This study investigates if the target partitioning strategy can give lower valleys compared to VMAT with the same number of beams.
Made with FlippingBook - Share PDF online