ESTRO 2024 - Abstract Book

S4681

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

ESTR0 2024

1 Holland Proton Therapy Centre, Radiotherapy, Delft, Netherlands. 2 Leiden University Medical Centre, Radiotherapy, Leiden, Netherlands. 3 Erasmus Medical Centre, Radiotherapy, Rotterdam, Netherlands

Purpose/Objective:

In IMPT treatment of esophageal cancer patients, motion and anatomical changes have a major impact on the delivered dose. To mitigate this impact, generous robustness settings during treatment plan optimization and evaluation are typically employed, resulting in increased exposure of OAR. In addition, repeat CTs are acquired to assess the target coverage and the need to adapt the treatment plan during the treatment course. Interestingly, in our current clinical practice, we observed that with the current robustness settings, adaptations of the treatment plan were not required. This led us to hypothesize that the robustness settings used might be too conservative. Additionally, it remains uncertain whether the use of online adaptive IMPT allows additional reduction in setup robustness settings. Therefore, the aim of this study is to investigate if robustness settings of the treatment plan can be safely reduced, thereby reducing OAR doses. Furthermore, we explore to what extent online plan adaptation can offer benefits for this patient group.

Material/Methods:

Twenty patients with a mid- or distally located esophagus carcinoma, treated with IMPT between October 2021 and March 2022, were retrospectively analysed. This analysis included a planning 4D-CT and 3-5 weekly repeat 4D-CTs (rCTs) for each patient. Four plans were created on the average CT using a setup robustness setting of 8 mm (clinical), 7 mm, 6 mm, and 5 mm and a range robustness setting of 3%. An ITV was used in optimization to incorporate motion of the target. To reduce the interplay effect the spot size was increased by using a range shifter and repainting was applied in case of a breathing amplitude larger than 5 mm. A robust evaluation was performed on 28 scenarios on the end inhale and end exhale phase of the weekly rCTs. To test online plan adaptation, plans with an individual anisotropic setup robustness setting and a range uncertainty of 3% were created and evaluated on the average CT of the rCTs. The setup robustness settings were based on an assessment of all residual errors in the treatment chain, including delineation uncertainties caused by deformable image registration (Table 1). These errors resulted in a setup robustness setting of 4 mm, according to the margin recipe of van Herk et al (1). In addition, an individual anisotropic margin was used for intrafraction motion, based on the movement of the tumour measured on the 4D-CT.

Coverage was considered adequate if 95% of the prescribed dose covered 98% of the CTV-volume over the entire treatment in the voxel-wise minimum dose distribution. A robustness setting was considered adequate, if the