Implementation of a Compact Spot-Scanning Proton Therapy System in a GPU Monte Carlo Code to Support Clinical Routine

Abstract

The purpose of this work was to implement a fast Monte Carlo dose calculation tool, Fred, in the Maastro proton therapy center in Maastricht (Netherlands) to complement the clinical treatment planning system. Fred achieves high accuracy and computation speed by using physics models optimized for radiotherapy and extensive use of GPU technology for parallelization. We implemented the beam model of the Mevion S250i proton beam and validated it against data measured during commissioning and calculated with the clinical TPS. The beam exits the accelerator with a pristine energy of around 230 MeV and then travels through the dynamically extendable nozzle of the device. The nozzle contains the range modulation system and the multi-leaf collimator system named adaptive aperture. The latter trims the spots laterally over the 20 × 20 cm2 area at the isocenter plane. We use a single model to parameterize the longitudinal (energy and energy spread) and transverse (beam shape) phase space of the non-degraded beam in the default nozzle position. The range modulation plates and the adaptive aperture are simulated explicitly and moved in and out of the simulation geometry dynamically by Fred. Patient dose distributions recalculated with Fred were comparable with the TPS and met the clinical criteria. Calculation time was on the order of 10–15 min for typical patient cases, and future optimization of the simulation statistics is likely to improve this further. Already now, Fred is fast enough to be used as a tool for plan verification based on machine log files and daily (on-the-fly) dose recalculations in our facility.