ESTRO 2025 - Abstract Book

S3483

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

ESTRO 2025

References: [1] Inaniwa T et al. Treatment planning for a scanned carbon beam with a modified microdosimetric kinetic model. Phys Med Biol. 2010;55:6721-6737. [2] Scholz M., Kraft G. Track structure and the calculation of biological effects of heavy charged particles. Adv Sp Res. 1996;18:5-14. [3] Scholz M. et al. Computation of cell survival in heavy ion beams for therapy. Radiat Environ Biophys. 1997;36:59 66. [4] Fossati P et al. Dose prescription in carbon ion radiotherapy: a planning study to compare NIRS and LEM approaches with a clinically-oriented strategy. Phys Med Biol. 2012;57:7543-7554.

1985

Poster Discussion Increasing robustness against FLASH model uncertainties by using pencil-beam delivery pattern optimization in proton therapy Manon C. van Zon 1,2 , Sebastiaan Breedveld 1 , Mischa S. Hoogeman 1,2 , Steven J.M. Habraken 1,2,3 1 Department of Radiotherapy, Erasmus MC Cancer Institute, Rotterdam, Netherlands. 2 Department of Medical Physics and Informatics, Holland Proton Therapy Center, Delft, Netherlands. 3 Department of Radiation Oncology, Leiden University Medical Center, Leiden, Netherlands Purpose/Objective: Ultra-high dose rates have been demonstrated to trigger the FLASH effect. In proton therapy with pencil-beam scanning, the spatiotemporal structure of dose delivery - as described by the pencil-beam scanning dose rate (PBS DR) - highly depends on the pencil-beam delivery pattern [1,2]. So far, substantial uncertainties with respect to PBS DR thresholds for FLASH remain. Typically, PBS-DRs>40 Gy/s are considered to trigger the FLASH effect, however, it is surmised that the FLASH effect is already triggered at lower PBS-DRs [3]. We investigate the robustness [4] of delivery pattern optimization for different PBS-DR levels against FLASH-weighted dose (FWD). Material/Methods: Treatment plans for ten lung cases (PTVs 33-229 cc) were generated using a 54 Gy/3 fractionation scheme with three 244 MeV co-angular proton transmission beams. In every fraction, one beam was delivered. Plans were scaled to D 95%,PTV =100%D pres . We modeled the FLASH effect via the FLASH enhancement ratio (FER) in which (i) dose>8 Gy is required to trigger the FLASH effect (FER>1) and (ii) the FER is a function of the PBS-DR, for which we used both a threshold and logistic function (Figure 1a). PBS-DR distributions were optimized using delivery pattern optimization. Here, the fitness function minimizes the mean FWD, the voxel-wise ratio between dose and FER, in a 20 mm ring surrounding the GTV. Hence, increased FER results in reduced FWD. Evaluation was performed on the FWD in all healthy-tissue voxels receiving >8 Gy using the FER threshold model with thresholds at 30, 40 and 50 Gy/s. Results: Figure 1b shows, for both functions in Figure 1a, the population mean FWD volume histogram for evaluations at the threshold model with thresholds at 30, 40 and 50 Gy/s. It shows that the logistic model leads to improved FWD when assumed that the FLASH effect is already triggered at PBS-DR>30 Gy/s, while not resulting in worse FWD for higher PBS-DR thresholds. In Figure 1c, the difference in optimized FWD between the logistic and threshold model is evaluated on V 80%Dpres (grey line). It shows improved median V 80%Dpres when using the logistic model, especially when evaluated at PBS-DR>30 Gy/s. Figure 2 shows for both functions in Figure 1a the optimized delivery patterns and corresponding PBS-DR distributions. It shows an increased volume of PBS-DRs>40 Gy/s for the logistic model.