ESTRO 2024 - Abstract Book

S4673

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

ESTR0 2024

Material/Methods:

Non-clinical optimization and evaluation functions considering variable proton RBE and related metrics, including LET, proton track-ends (TE), and the dirty dose (DD) concept were implemented in research versions of the RayStation treatment planning system (TPS). The latter functions serve as surrogates for enhanced RBE, TE representing the number of protons stopping within a volume [1] and DD representing the dose deposited by protons with an LET exceeding a specified threshold. [2] To gain insights into the clinical adoption, cutting-edge research endeavours, and prospects related to variable proton RBE, a comprehensive survey was distributed to 24 individuals from 19 institutions (10 in Europe and 9 in the United States). These institutions have actively used the aforementioned research functionalities for several years. The online questionnaire encompassed 25 questions, addressing various aspects of proton RBE considerations in proton treatment planning. Nineteen individuals from 16 distinct institutions participated in the survey. The preferred clinical reporting of LET was dose-averaged LET (LET d ) (94% of the individuals preferred this), normalized to unit density in water (60%), for all primary and secondary protons (60%). However, some individuals rather preferred to include the LET contribution from primary protons only (13%) or all ions (27%) and to score it in the local medium (40%). This supports previously reported inconsistencies in LET reporting. [3,4] Nevertheless, the preference of most of the responders aligns with both a recent publication on LET harmonization in Europe [4] and the clinical LET reporting in RayStation. While all institutions clinically prescribe RBE=1.1, each institute also employs strategies to mitigate potentially enhanced RBE effects in organs at risk (OARs), focusing on intracranial serially structured OARs such as the brainstem, chiasm, and optical structures. A subset of around 10% of the institutions also implements measures to exploit RBE in the tumour. Common planning approaches for mitigating potentially enhanced RBE in OARs include beam avoidance (72%), additional beams (72%), alternative beam arrangements (61%), and reducing target doses in regions adjacent to OARs (44%). For plan evaluation, 67% of the responders assessed beam angles, spot maps and Bragg peak positions to mitigate enhanced RBE values in OARs. Notably, 42% of the responders also evaluate LET d and variable RBE distributions clinically using scripting capabilities. In the realm of research, all participating institutions have explored functions extending beyond RBE=1.1, focusing on serially structured OARs. When rating the usefulness of the functions (scale: 1-5, with 5 being the most useful), linear RBE-LET, LET d , DD and TE functions received high and roughly equivalent ratings (averages: 4-5) for both optimization and evaluation (Figure 1). In contrast, LQ-based RBE functions scored substantially lower for optimization (average: 2.2), which stems from the observations that it primarily reduced the dose rather than the LET d which some responders were seeking. Nevertheless, the evaluation score of LQ-based RBE models was like the other functions (average: 4.1) and most institutions would prefer McNamara et al. [5] (69%) and Wedenberg et al. [6] (38%) along with linear LET-RBE models (50%) in clinical TPSs (Figure 2) since all responders think that at least a subset of proton therapy patients could benefit from optimization and evaluation with functions beyond RBE=1.1. Results: