Author: Hongyu Jiang 👨🔬
Affiliation: Department of Radiation Oncology, University of Kansas Medical Center 🌍
Purpose: This study presents a method for transforming coordinates from RayStation’s treatment planning system into the Monte Carlo (MC) coordinate system used in GEANT4 for dose calculations in spot scanning proton therapy.
Methods: DICOM CT files from RayStation provided 3D patient geometry, while the DICOM RP file contained treatment plan details, including isocenter position, gantry and couch angles, and proton spot data. A MATLAB program was developed to extract and convert this information into a GEANT4 input file. Instead of rotating the proton source, the gantry angle was implemented by rotating the CT volume, simplifying proton source positioning. Each proton field consisted of multiple spot layers, with all spots in a layer sharing the same nominal energy. Spots were defined by their monitor unit and (X’, Y’) coordinates relative to the beam’s central axis. These coordinates were then transformed into GEANT4 coordinates and projected from the isocenter plane to the proton source plane.
Results: The patient volume was modeled as a rectangular box in GEANT4, with dimensions based on the CT parameters. The isocenter was placed at the coordinate origin, and the Cartesian axes were defined as +X (patient’s left), +Y (posterior), and +Z (superior). The couch and gantry angles were applied by rotating the CT volume in the correct order and direction: first around the Y-axis by the couch angle, then around the Z-axis by the negative gantry angle. The proton source plane was positioned at Y = -40 cm. Proton source sampling parameters were calculated using RayStation commissioning data and the Fermi-Eyges equation. Transformation from RayStation coordinates to GEANT4 was straightforward: Y’ mapped to GEANT4 Z, and X’ mapped to GEANT4 X.
Conclusion: We developed and validated a method for transforming RayStation coordinates into GEANT4 coordinates, establishing a foundation for future MC dose calculation studies.