Development and validation of a TOPAS Monte Carlo model for a 0.5 T bi-planar Linac-MR

MR-guided radiotherapy enables real-time imaging and adaptive treatment but may introduce magnetic field effects that alter dose deposition. Accurate dose calculation in such settings requires detailed Monte Carlo (MC) modeling.

To develop and validate a detailed MC model of the 0.5 T bi-planar Linac-MR with an integrated, custom-designed multileaf collimator (MLC) module in TOPAS.

A MC Model of the 0.5 T bi-planar Linac-MR with a 6 MV FFF beam, commercialized as the Aurora-RT (MagnetTx Oncology Solutions, Canada), is developed in TOPAS. A custom 3D magnetic field vector map, tracking with gantry angle, is incorporated into this TOPAS model. An electron source is used for X-ray generation, and all components of the linac head from the target downward are modeled in detail. The MLCs are modeled from stereolithography (STL) design files and controlled via an empirically driven mechanism developed in this work. Water tank measured percent depth dose (PDD) curves and profiles (3 $\times$ 3  ${\mathrm{cm}}^2$ to 25 $\times$ 25  ${\mathrm{cm}}^2$ ) are compared to MC simulated data to optimize the electron source energy and radial distribution. Additionally, output factors are simulated and compared to measurements. MLC transmission at 10 cm depth in solid water is simulated and measured using GAFChromic EBT3 film. MLC positioning accuracy is evaluated by comparing off-axis MLC-defined field profiles measured at 10 cm depth in water. The MC model's dose calculation accuracy is further evaluated by comparing measured and simulated surface doses using EBT3 film, and PDDs in slab phantoms using parallel plate chambers. Surface dose is measured by placing films at the surface and 5 cm depth in a solid water phantom. PDDs are measured and simulated in the following slab phantom configurations: polystyrene and polystyrene-bone-lung-polystyrene.

A 5.5 MeV electron source energy and a 1.3 mm radial distribution (FWHM) provides the best match between measurement and MC. Simulated PDDs pass 1% $|$ 1 mm gamma criteria at 100% compared to measurements for all fields investigated. Simulated profiles at various depths for all fields (3 $\times$ 3  ${\mathrm{cm}}^2$ to 25 $\times$ 25  ${\mathrm{cm}}^2$ ) score $>$ 96.7% in 2% $|$ 2 mm compared to measurement. Evaluated output factors are in good agreement with measurement (within 1%) for all fields except 3 $\times$ 3  ${\mathrm{cm}}^2$ (within 1.5%). The 2% $|$ 2 mm gamma pass-rates for MLC defined off-axis fields are $>$ 98%. The maximum mean Distance-To-Agreement (DTA) in the penumbra region (1% criteria) and mean dose difference in central region for inline and crossline profiles, are 1 mm and 1%, respectively. MC simulated MLC transmission at central axis (0.28% - 0.29%) is in good agreement with measurement (0.28% - 0.44%). Film surface dose relative to D_max is 74.4% in measurement and 73.6% in simulation. Lastly heterogeneous phantom PDDs passed 1% $|$ 1 mm gamma criteria at $>98$ % compared to measurement.

The developed TOPAS MC model of the 0.5 T Linac-MR demonstrates high accuracy for dose verification in magnetic fields. The MLC module, including its coordinate positioning mechanism, is fully validated for open aperture applications. This MC model provides a reliable framework for dose simulations in a magnetic field.