Encrypted login | home

Program Information

The Development of a Tetrahedral Phantom and Its Monte Carlo Simulation for Proton Therapy

no image available
T Chao

T.C. Chao1,2*, H.K. Chang1 , C.C. Lee1,2 , C.J. Tung1 , (1) Chang Gung University, Taoyuan, Taiwan, (2) Chang Gung Memorial Hospital, Linkou, Taiwan


SU-I-GPD-T-135 (Sunday, July 30, 2017) 3:00 PM - 6:00 PM Room: Exhibit Hall

Purpose: To develop a tetrahedral phantom and to apply it into GEANT4 Monte Carlo for proton dose simulation. Although voxel phantoms are widely used in Monte Carlo simulation of dose calculation, they have several limitations including (1) non-infinitesimal resolution and (2) hard to be deformed. Polygonal mesh phantoms can be used to model smooth surface and can be deformed. These phantoms, however, takes longer time for Monte Carlo simulation and cannot account for tissue heterogeneity. Tetrahedral phantoms are deformable, more simulation efficient than Polygonal mesh phantoms, and can deal with spatial and temporal heterogeneous. That is, tetrahedral phantoms can be easily converted into a series of 4D phantom by applying deformation vector fields to the four vertices of each tetrahedron.

Methods: A tetrahedral phantom was developed using DICOM CT images. 3D doctor, Rhinocero 5, tetget, and in-house codes, which can reduce volume of large tetrahedrons and account for tissue heterogeneity, were implemented to convert medical images into polygonal surfaces, and finally a tetrahedral phantom. This phantom was then imported into GEANT4 to simulate proton dose distribution of a 5 cm* 5 cm, 190 MeV proton beam.

Results: 3D doses were obtained and analyzed to DVHs, isodose curves, depth doses and dose profiles. Currently, the simulation efficiency is not optimized yet. It took few days to get a simulation with dose uncertainty less than 1% in an 8 core computer.

Conclusion: A tetrahedral phantom has been successfully constructed using clinical CT images. This phantom can be implemented into GEANT4 for ideal beam simulation. Realistic beam parameters are necessary for a more realistic dose simulation. Message Passing Interface for parallel computing maybe used to speed up this simulation.

Contact Email: