S1999
Physics - Dose prediction/calculation, optimisation and applications for photon and electron planning
ESTRO 2026
Conclusion: Using generated gEUD bowel objectives in a RapidPlan prostate and nodes model enables comparable target coverage and improved OAR sparing relative to manually optimised clinical plans. This approach enhances planner control and adaptability, addressing a common limitation of models based solely on relative volume constraints. The method demonstrates potential for wider adoption in knowledge-based planning for complex pelvic M. Hussein et al., “Clinical validation and benchmarking of knowledge-based IMRT and VMAT treatment planning in pelvic anatomy,” Radiotherapy and Oncology, vol. 120, no. 3, 2016, doi: 10.1016/j.radonc.2016.06.022.[2] A. Fogliata et al., “RapidPlan knowledge based planning: Iterative learning process and model ability to steer planning strategies,” Radiation Oncology, vol. 14, no. 1, 2019, doi: 10.1186/s13014-019-1403-0. Keywords: knowledge-based planning, RapidPlan, prostate Poster Discussion 4468 Swin-UNETR for Out-of-Field Dose Estimation from In-Field Dose in External Beam Radiotherapy MOHAMMED EL AICHI 1,2 , NATHAN BENZAZON 1,2 , MEISSANE M'HAMDI 1,2 , IBRAHIMA DIALLO 1,2 , RODRIGUE ALLODJI 3 , FLORENT DE VATHAIRE 4 , ERIC DEUTSCH 1,2 , MARIA VAKALOPOULOU 5 , CHARLOTTE ROBERT 1,2 1 Département de Radiothérapie, Institut Gustave Roussy, Villejuif, France. 2 Radiothérapie Moléculaire et Innovations Thérapeutiques, U1030, Inserm, Gustave Roussy, Université Paris-Saclay, Villejuif, France. 3 Centre de Recherche en Épidémiologie et Santé des Populations, U1018, Inserm, Gustave Roussy, radiotherapy. References: [1] Université Paris-Saclay, Villejuif, France. 4 Centre de Recherche en Épidémiologie et Santé des Populations, U1018, Inserm, Gustave Roussy, Université Paris- Saclay,, Villejuif, France. 5 Mathématiques et Informatique pour la Complexité et les Systèmes, Université Paris-Saclay, CentraleSupélec, Gif-sur- Yvette, France Purpose/Objective: Low (<2 Gy) and very low (<150 mGy) radiation doses delivered outside the treatment field in radiotherapy may affect the immune system, potentially contributing to radiation-induced lymphopenia, which is strongly associated with worse clinical outcomes. This is particularly relevant in radio-immunotherapy, where preserving immune function is crucial for treatment efficacy. Out-of-field dose modeling for
60 Gy to the high-dose prostate target and 47Gy to the low-dose prostate target and pelvic nodes in 20 fractions. Target dose-volume objectives were fixed. Dose-volume objectives for all OARs were model- generated, including line objectives for the bladder and rectum, and gEUD objectives for the bowel. Planners were permitted to adjust objectives during plan optimisation. Model-generated plans were produced for a test cohort of 20 patients and compared against their corresponding clinical plans. Plan quality was assessed by comparing PTV coverage (V95%), homogeneity index (HI), conformity, and standard dose-volume metrics to bowel, bladder, rectum, and femoral heads. Statistical analysis was performed using paired Wilcoxon signed-rank tests. Results: PTV coverage, homogeneity, and conformity were non- inferior between model-generated and clinical plans (p > 0.05), except for high-dose PTV coverage which was on average 0.11% less for model plans although still clinically acceptable. Use of gEUD-based bowel objectives achieved improved OAR sparing, with statistically significant average mean dose reductions for bowel, bladder and rectum compared with clinical plans (p < 0.05). Bowel gEUD objectives provided greater flexibility for planner adjustment in cases where clinical goals were initially unmet, reducing the need for iterative re-optimisation. 14/20 plans were completed in a single optimisation.
Made with FlippingBook - Share PDF online