MCNPX Dual Resolution Simulation of DOse Distribution for Proton Therapy

• Main Authors: Shao Chih Wei, 魏劭至
• Other Authors: T. C. Chao
• Format: Others
• Published: 2013
• Online Access: http://ndltd.ncl.edu.tw/handle/04426824065704237049

碩士 === 長庚大學 === 醫學影像暨放射科學系 === 101 === The purpose of this study is to access proton dose distribution in dual resolution phantoms using MCNPX 2.6.0. The dual resolution phantom uses higher resolution in ROI (Bragg peak, area near large dose gradient or heterogeneous interfaces) and lower resolution in the rest. In this study, MCNPX 2.6.0 was installed in Ubuntu 10.04 with MPI for parallel computing. To record the energy deposition, we used *F8 and FMesh1. *F8 records the energy deposition directly but needs to be defined by each cell. FMesh1 is convenient for large voxel number phantoms but indirectly convert dose deposition from fluence. To simulate dose distribution in dual resolution phantoms, 60 and 120 MeV narrow proton beam were incident into coarse, dual and fine resolution phantoms with pure water, water-bone-water and water-air-water interfaces. The Fisher’s least significant difference test was used to evaluate the statistical significant in this study. The results showed that FMesh1 may be not so accurate in recording the dose from nuclear interaction. Other secondary particle like alpha, deuteron, tritium had no significant difference between *F8 and FMesh1. In dual resolution simulations, the doses in coarse resolution phantoms are underestimated due to partial volume effect. The dose distributions in dual or high resolution phantoms agreed well each other and dual resolution phantoms saved at least 10 times simulation time comparing to fine resolution phantom. The differences between dual resolution phantoms and fine resolution phantoms are small but statistically significant concluded by fisher’s LSD test. In the ideal pencil proton beam simulation, we found the efac parameter is too large to simulate dose distribution accurately; the simulation uncertainties would be underestimated by improper efac setting.