ESTRO 2024 - Abstract Book

S4630

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

ESTR0 2024

18. Vanhavere, F., Loos, M., Plompen, A. J. M., Wattecamps, E., & Thierens, H. (1998). A combined use of the BD-PND and BDT bubble detectors in neutron dosimetry. Radiation Measurements, 29(5), 573-577. 19. Takada, M., Kitamura, H., Koi, T., Nakamura, T., & Fujitaka, K. (2004). Measured proton sensitivities of bubble detectors. Radiation protection dosimetry, 111(2), 181-189.

1348

Proffered Paper

Clinical application of a probabilistic robustness evaluation tool for proton therapy treatments

Francesco Fracchiolla 1 , Lamberto Widesott 1 , Roberto Righetto 1 , Carlo Algranati 1 , Dante Amelio 2 , Annalisa Trianni 1 , Stefano Lorentini 1 1 UO Fisica Sanitaria Trento Hospital APSS, Proton Therapy Center, Trento, Italy. 2 UO Protonterapia Trento Hospital APSS, Proton Therapy Center, Trento, Italy

Purpose/Objective:

The study has two main purposes:

1. to implement a probabilistic Robustness Evaluation (pRE) tool for pencil beam proton therapy treatments that explicitly simulates all the treatment-related uncertainties and the mitigation effect of the fractionation 2. to find the best combination of systematic uncertainty parameters to be defined in the worst-case Robustness Evaluation (wRE) approach (implemented in commercial TPS) which best surrogates the pRE results.

Material/Methods:

Ten years of yearly, monthly and daily QA results were analysed to determine mechanical, geometrical and dosimetric uncertainties of our proton pencil beam scanning delivery system. In particular, gantry and couch isocentricity and rotational uncertainties together with absolute spot positioning uncertainties were derived. Moreover, for 157 patients, post-treatment imaging was acquired and compared with setup images to determine the intra-fraction patient positioning error when thermoplastic facial masks were used as immobilization devices. All of these uncertainties were randomly sampled for each fraction of the treatment from their probability distribution function. The patient positioning uncertainty was simulated in all x, y, and z directions in order to sample the entire 3D space for each scenario dose: 8 scenarios along the cardinal axis. Range uncertainty was simulated as the only pure systematic error with a value of ±3.5%: 2 scenarios. This workflow was repeated 10 times in order to increase the statistical significance of the simulations in the dosimetric space: 160 dose scenarios were simulated for each patient. Each scenario was the results of N (N = number of fractions) dose calculations. The pRE calculation time for each patient was registered.

21 patients (10 skull base chordoma and 11 head and neck) were planned with robust multiple field optimization (MFO) approach. For each plan four different Robustness Evaluations with the worst-case approach (wRE)