ESTRO 2025 - Abstract Book

S3480

Physics - Optimisation, algorithms and applications for ion beam treatment planning

ESTRO 2025

Conclusion: This study demonstrates that the proposed energy layer-wise proton beam deflection compensation strategy implemented in the MC-based TPS incorporating the MR magnetic field map can accurately recover SFUD distributions in the presence of the MR magnetic field.

Keywords: MRiPT, proton therapy, pencil beam scanning

References: [1] Hoffmann A. et al . MR-guided proton therapy: a review and a preview. Radiat. Oncol. 2020;15(1):129.

1948

Digital Poster Multi-model radiobiological optimization of treatment plans in carbon ions radiotherapy (CIRT) Giovanna Martino 1 , Antonio Carlino 1 , Eugen Hug 2 , Markus Stock 1,3 , Piero Fossati 2,3 1 Medical Physics, MedAustron Ion Therapy Center, Wiener Neustadt, Austria. 2 Radiation Oncology, MedAustron Ion Therapy Center, Wiener Neustadt, Austria. 3 Department of Basic and Translational Oncology and Haematology, Karl Landsteiner University of Health Sciences, Krems, Austria Purpose/Objective: In Carbon Ions Radiotherapy (CIRT) the RBE-weighted dose D(RBE) is predicted by models such as the modified Microdosimetric Kinetic Model (mMKM), applied in Japan [1], and the Local Effect Model (LEM-I), adopted in Europe and China [2&3]. At MedAustron carbon-ions treatment plans are optimized using LEM-I. For patients treated with schedules derived from Japanese experience (16 fractions in 4 weeks for H&N and sarcoma) the dose prescription and constraints for OARs are obtained from those used in Japan corrected to account for the differences between LEM-I and mMKM [4]. In our TPS (RayStation 11B) LEM-I-optimized plans can be recomputed using mMKM. We evaluate each plan using both RBE-models, aiming at a homogeneous target coverage and proper organ sparing, in both scenarios. Material/Methods: A LEM-I-optimized plan shows a steep target DVH, centered around the prescribed D(RBE). For the same plan recomputed using mMKM the dose homogeneity of the target is lost because of the higher RBE assigned by the mMKM to the high LETd component of the dose. In order to increase the homogeneity of mMKM recomputed RBE-weighted-dose we reoptimize iteratively. Hot and cold mMKM areas are contoured and used as help structures in the second round of optimization. We increase LEM I prescription to 103-105% of initial nominal value in mMKM cold-spots and we reduce it to 95% of the nominal value in mMKM hotspots. Results: For lateralized H&N targets we remove the mMKM hotspot (corresponding to high LETd) typically affecting mucosa or other OARs located in the deep portion of the target volumes without the need of contralateral beams and preserving the target coverage with 95% of the prescription [Fig.1]. For large non-H&N targets we achieve a homogeneous target coverage in the mMKM recomputation, while avoiding the underdose in the middle of the target and reducing the mMKM hotspot close to the OARs. Forcing more particles to stop in the initial mMKM cold-spots and less particles to stop in the hotspots we also achieve an improved and more homogeneous LETd distribution within the target [Fig.2].