7. Dosimetric effects of setup error, range and dose calculation uncertainties: comparison between Pencil Beam and Monte Carlo algorithm

Abstract

Purpose (1) To compare the dosimetric effect of setup error and range uncertainties with the effect of using two dose calculation algorithms (Pencil Beam (PB) vs. Monte Carlo (MC)) on protontherapy treatment plans, and (2) To assess which uncertainties affect the most the difference between the nominal and actual dose distribution. Methods 12 treatment plans designed for clinical treatment at the Proton Therapy Centre in Trento were selected, covering four anatomical sites: brain, head and neck, chordoma of the spine, and craniospinal axis. The plans, which were initially optimized with the PB algorithm, were recalculated with MC with a statistical uncertainty of 1%. Then, we evaluated the effect of geometrical (setup) and range uncertainties on the planned dose distribution with an in–house software. This software simulates 8 setup errors and 2 range errors (16 calculations per plan in total), in order to estimate near worst-case scenarios for target volume and organs at risks (OAR). We evaluated the differences between MC vs. PB distribution and the robustness of PBS plans using the same dosimetric indices used to optimize the plan (e.g. D1 and D99 for the target volume D1 for serially responding complications, etc.). Results In Table 1 , we show CTV and OAR dosimetric indices for the plans optimized with PB, then recalculated with MC, and also the worst-case scenario after robustness analysis. CTV D99 in MC plans are smaller than in PBS plans. Furthermore, except for plans of intracranial treatments, MC dose distributions are in general worse than the worst-case scenario (Fig. 1). Consequently, sparing of OARs with MC is obtained. For instance, the lacrimal gland mean dose is lower by 9% and 50% in MC vs PB for the static case and the worst case scenario, respectively. Conclusions In 2 out of the 4 disease sites we studied, the effect of using two dose calculation algorithms is greater than the effect of setup error and range uncertainties especially in CTV coverage. Considering MC as a gold standard, the minimum dose to the CTV isn’t guaranteed when plans are designed with PB.