Influence of spatial redistribution of heterogeneities in proton beam characteristics

The purpose of this study was to investigate the fundamental properties of spot-scanning proton beams and compare them to Monte Carlo (MC) simulations, both with and without CT calibration, using spatially diverse combinations of materials.

A heterogeneous phantom was created by spatially distributing titanium, wax, and thermocol to generate six scenarios of heterogeneous combinations. Proton pencil beams ranging in energy from 100 to 226.2 MeV were directed perpendicular to each heterogeneous combination, and the exit proton was measured using a Lynx scintillation detector and a Zebra Multi-Layer-Ionization-Chamber for depth dose and spot profile measurements, respectively. The identical measurement configuration was duplicated in the RayStation-TPS. The measured and simulated RayStation-MC beam characteristics were compared.

The results showed that at 100 MeV, the mean standard deviation of spot size was 5.66 ± 0.27 mm, while at 226.2 MeV, it rapidly decreased to 3.37 ± 0.07 mm. The physical phantom showed a larger perturbation difference between measurement and MC simulation than the virtual phantom. MC overestimates ranges up to 1.5 % in virtual phantoms, but underestimates ranges up to 5 % in physical phantoms. Range perturbations over 1 mm occurred in 35.7 % of virtual phantom measurements and in 85.7 % of physical phantom measurements.

Despite using a CT artefact reduction approach and an accurate Monte-Carlo dose calculation algorithm, perturbations in proton characteristics were still observed. It is essential to be aware of the limits of the TPS in managing such heterogeneous combinations. It is recommended to perform more validation checks on heterogeneous combinations than on individual materials.