Monte Carlo simulation and experimental validation of magnetic field effects on proton dose distributions

Given the physical properties and the sensitivity to morphological variations of proton therapy, it could greatly benefit from integration with magnetic resonance (MR) imaging. Such integration raises several challenges, as both systems mutually interact with each other. The problem of magnetic field induced dose distortions has been predicted by Monte Carlo (MC) simulation in previous studies, but no experimental validation has been performed yet. We present and compare simulated and measured dose distributions in a realistic magnet setup.

2D dose distributions of proton pencils beams (80-180MeV) traversing the field of a 1T NdFeB permanent magnet while depositing energy in a PMMA slab phantom were simulated using the Geant4 toolkit and measured using EBT3 radiochromic films. The Geant4 model was validated against depth-dose measurements performed with a multi-layer ionisation chamber. The magnetic vector field was calculated using finite-element modelling and validated experimentally using a Hall probe. Deflected beam trajectories and depth-dose curves were extracted from the 2D dose distributions and compared. Demagnetization and radioactivation of the magnet material were simulated and monitored during measurement.

The range predicted by the MC model agreed with the Giraffe measurements within 0.5mm, and calculated and measured magnetic field data agreed within 2%. The lateral beam deflection was clearly visible on EBT3 films and ranged from 1mm to 10mm for 80MeV and 180MeV, respectively. Simulated and measured range and deflection agreed within 1mm for all studied energies. Demagnetization and radioactivation effects were negligible.

For the first time, MC simulations of magnetically deflected proton beams inside tissue-equivalent material have been experimentally validated with dose measurements. The results indicate that the magnetic field induced proton beam deflection is both measurable and accurately predictable. This demonstrates the feasibility of accurate dose calculation as well as measurement within the framework of MR-integrated proton therapy.