Monte Carlo characterisation of proton minibeams.
At a Glance
Section titled “At a Glance”| Metadata | Details |
|---|---|
| Publication Date | 2020-01-01 |
| Journal | University of Canterbury Research Repository (University of Canterbury) |
| Authors | James Thomas Eagle |
| Citations | 1 |
Abstract
Section titled “Abstract”Proton therapy has been shown to offer clinical advantages over standard photon based external beam modalities. Additionally, there is evidence from small animal experiments showing that the recovery of healthy tissue is improved with the use of spatially modulated beams, referred to as microbeams or minibeams which consist of an array of parallel narrow beamlets. Using Monte Carlo modelling, this thesis explores the potentially powerful combination of minibeams and proton therapy by characterising the physical and biological dose when generating proton minibeam by means of a collimator. It is known that patient motion has the potential to spread dose to regions between the minibeams thereby, reducing or removing the spatial modulation. A thorough characterisation of the impact of such motion on lateral spread of dose was performed for a wide range of motion amplitudes, relevant for small animal irradiation (60 µm to 10 mm). Clinical beams appropriate for human therapy were also investigated to determine the impact motion has on proton minibeams. Motion was shown to detrimentally impact all proton minibeam dose distributions. However, simulations for irradiating a brain undergoing pulsations of up to 300 µm an array of a 0.3 mm wide beamlets, (centre to centre (CTC) spacing of 1 mm), showed that the dose between the beamlets, known as the valley region, only increased by a maximum of 2% for both mouse and human treatments. It was further found that increasing the CTC spacings up to 10.77 mm reduced the impact of motion on the valley region, for motions up to as high as 1.1 mm. Increasing the width of the beamlet reduced the effect of motion on dose delivered within the beamlet tracks, however it increased dose in the valley regions, and therefore it is not recommended. Having established motion does not negate the usefulness of proton minibeam therapy, a comprehensive Monte Carlo framework was developed in Tool for particle simulation (TOPAS) to simulate the University of Washington’s preclinical proton radiation platform, which is currently the only existing preclinical proton minibeam beamline. A detailed and novel analysis of the impact each beamline component has on the resulting dose distribution was performed. Unknowns in the physical dimensions of beamline components were identified and estimated by adjusting physical dimensions in the simulated beam transport system. The final beamline model was experimentally verified against measurements with an ion chamber and diamond detector to have a Bragg peak depth consistent with an energy of 44.95 MeV (a depth of 18.1 mm). The decrease in energy from the nominal 50.5 MeV (depth of 22.3 mm) was shown to be due to proton interactions with the graphite degrader, Kapton exit window, and the monitoring ion chamber. It was further shown that these components reduced the Bragg peak depth by 2.0 mm, 0.3 mm, and 2.0 mm, respectively. Other components were also investigated and found to have minimal impact on the beam energy. At the University of Washington proton minibeams are currently produced by means of a physical collimator designed and manufactured during this thesis. A comprehensive analysis and characterisation of collimator material and physical dimensions was carried out with the aim of defining the ideal proton minibeam collimator designed for this beamline, i.e. one that maximises spatial modulation and limits scattered dose. Results indicated that there exists an optimum collimator thickness and a 20 - 25 mm thick collimator was found to be ideal. Increasing the thickness of the collimator beyond this was undesirable, as the collimator would then reduce the number of primary protons whilst maintaining the same number of scattered protons. It was shown of the 50.5 MeV proton beam with a nickel collimator that the first 15 mm of the collimator was responsible for producing the most neutrons, electrons, and photons. The ideal width of the collimator slot width for the UW proton beam was found to be between 0.3 - 0.4 mm as it produced beamlets with a high spatial modulation, commonly characterised as the peak to valley dose ratio (PVDR), whilst only reducing dose at the Bragg peak by approximately 60% of an open (non-collimated) beam. It was also shown that although a high-density collimator should be used to maximise the PVDR, care needs to be taken to reduce neutron production. Of the 78 materials investigated, Group 4 transition metals, particularly nickel, provided the highest PVDR. Nickel was also found to result in the lowest neutron yield, 2.7 times less than that of tungsten, making it an ideal material for minibeam collimators. To fully understand the therapeutic benefit of collimated proton beamlets requires knowledge of the Relative Biological Effectiveness (RBE) and the resulting RBE weighted dose (RWD), of all primary and scattered particles emanating from the proton minibeam collimator. These were investigated for four previously identified materials. While at 50.5 MeV the neutron dose from the different materials varied by a factor of 3.2, overall, it was an insignificant component of the total RWD. The RBE of protons was shown to increase by up to a value of three on the distal edge of the Bragg peak, shifting the Bragg peak depth deeper by 0.1 mm. The RBE in the valleys was found to be up to a value of 1.5 at the surface. This has the consequence of reducing the effective normal superficial tissue sparing from proton minibeams. Summarising the main results from this thesis, it has been shown that proton minibeams for small animal irradiation are feasible even when the target is undergoing internal motion. Having accurately simulated the preclinical proton radiation platform (PPRP) beamline at the University of Washington it was identified that the most suitable design for a physical collimator for 50.5 MeV protons is to construct from either nickel, stainless steel, copper, and brass. The thickness of the collimator should be 25 mm to sufficiently block protons without introducing any additional scatter. This thesis also provides evidence for supporting an increase in the commonly used value for proton RBE as a function of depth in tissue. This result is imperative for planning treatment and therefore should be implemented when considering dose effects for any animal and ultimately human therapies.