Implementation and validation of a fast Monte Carlo code as a secondary dose calculation engine for proton therapy with narrow beams

This study presents the implementation and validation of the fast, simplified open-source MC code MCsquare as a secondary dose calculation engine for intensity-modulated proton therapy with narrow beams (the Gaussian-shaped beam spot with standard deviations as small as 1-2 mm) produced by a synchrotron-based system with minibeam modification. A proton therapy system was modeled with MCsquare, using commissioning data, and with TOPAS, an explicit multi-particle MC code. The computed dose distributions were compared with the patient-specific quality assurance (QA) measurements (203 measurements for 94 treatment fields), using gamma analysis with criteria of 3% and 3 mm. The dose distributions in the patient geometry defined by computed tomography (CT) images were simulated with MCsquare and TOPAS and compared. For the main beam, the gamma passing rates of the patient-specific QA averaged 99.4% and 97.9% for MCsquare and 99.2% and 98.5% for TOPAS, with and without range shifter use, respectively. For minibeams, the rate was 100% for both MC codes. The dose distributions calculated with TOPAS and MCsquare on the patient's CT were identical, within the statistical error of the simulation. The simulation time with MCsquare varied between 1 and 25 min per plan on a 16-core workstation with a 2% statistical error. The fast, simplified MCsquare and the slower TOPAS using explicit multi-particle transport produced statistically identical dose distributions. The results support using MCsquare as a secondary dose engine for narrow beams.