Design, fabrication, commissioning, and dosimetric verification of a GRID collimator for proton SFRT on a compact...

Proton GRID radiotherapy (RT) is an extension of photon spatially fractionated radiation therapy (SFRT) techniques for bulky invasive tumors. It has been hypothesized that GRID RT improves the therapeutic ratio by minimizing normal tissue toxicities associated with treating bulky volumes while inducing abscopal radiobiologic effects. However, compact, synchrocyclotron-based proton therapy machines with large spot sizes pose unique technical challenges in implementing proton GRID RT.

The purpose of this work is to (a) design and model a collimating brass aperture within the RayStation treatment planning system (TPS), (b) validate the designed aperture by creating a commissioning plan and measuring the absolute and relative proton dose distributions delivered, and (c) perform a robustness analysis to determine the allowable mechanical tolerances and uncertainties during treatment delivery.

A custom (27 × 21.5 × 5 cm³) brass collimator (.decimal) was designed and fabricated with divergently-matched circular holes of 15-mm-diameter arranged in a hexagonal pattern. In-house RayStation scripts were developed to import the computer-aided design (CAD) model of the collimator into the TPS, and accurately orient and position the collimator as "support structures" for each beam angle requiring the collimator on a given plan. Divergently-matched cylindrical optimization structures were then created with 5, 10 and 15 mm diameters. Commissioning plans were created to deliver uniform proton physical doses (50 cGy and 8 Gy) through each aperture to 5-15 cm depth within a water phantom. One dimensional (1D) and 2D proton dose measurements were performed with various available radiation detectors, including: (a) PPC05 parallel-plate ion chamber, (b) MatriXX ion chamber array, (c) Lynx scintillation detector, and (d) Gafchromic EBT3 radiochromic films. Gamma analyses at 3%/3 mm criteria were performed for the 2D dose measurements acquired with the MatriXX and Lynx detectors. Finally, mounting errors were simulated within the TPS by artificially displacing the brass aperture along the crossline, inline and snout extension directions to determine the minimum allowable mechanical deviations between the TPS and the actual mounted aperture position.

Experimental measurements showed the best dosimetric agreements with the TPS calculations for optimization cylinder diameters of 10 mm with gamma passing rates of at least 97.9%. 1D absolute proton dose measurements with an ionization chamber showed agreement within 2.11% of TPS calculations once correcting for partial-volume averaging. Simulated mounting or setup errors within the TPS indicated a lateral positional requirement of ± 1.5 mm and a longitudinal snout positional requirement of ± 3 cm to achieve gamma passing rates of at least 90% (institutional standards).

We have commissioned a brass collimator consisting of milled divergent apertures for clinical SFRT treatments via a proton GRID technique. This process included an assessment of dosimetric sensitivity of aperture positioning error, and also dosimetric evaluation of the aperture model as a brass support structure within the TPS. Future works entail the creation of clinical SFRT plans using different planning techniques and their respective dose comparisons.