Determination of magnetic field correction factors for dosimetry in MR-integrated proton therapy

Objectives: For the integration of magnetic resonance imaging (MRI) into proton therapy (PT), a 0.22 T MRI was installed at the pencil beam scanning beam line at OncoRay. As a next step, dosimetry in the magnetic field has to be established. This work aims to study the influence of the static field (B0) of the MRI on ionisation chamber (IC) responses for proton beams through measurements and Monte Carlo (MC) simulations. Materials & methods: A Semiflex 0.3 and a PinPoint 3D IC were positioned in a water phantom placed in the MR imager isocenter. The absolute dose at five proton energies (70, 110, 150, 190, 226.7 MeV) was measured within the entrance plateau of the depth-dose curve using a 10×10 cm² homogeneous irradiation field. The correction factor kB→,M,Q was obtained by dividing the measured dose with/without B0. For the MC simulations, beam-commissioning data (depth-dose profiles in water, beam spot sizes in air) were used to create a MC beam model in TOPAS (1.5×105 particles). A 3D map of the scanner’s magnetic field (MF) was calculated with COMSOL and used in the simulations to mimic the experimental setup. Results: The MF correction factor kB→,M,Q showed systematic energy-dependent differences between dose readings with and without B0. For the Semiflex 0.3, kB→,M,Q was 0.9926, 0.9942, 0.9941, 0.9959 and 1.0036 for 70, 110, 150, 190 and 226.7 MeV, respectively. For the same energies, kB→,M,Q for the PinPoint 3D was 0.9920, 0.9931, 0.9938, 0.9952 and 0.9969. For all energies, the standard deviations of kB→,M,Q were smaller than 0.002 for both ICs. MC simulations of the Semiflex 0.3 response to 110 MeV showed no statistically significant B0 effect with kB→,M,Q of 0.997 (95% CI: 0.9893, 1.0049). Conclusion: Measurements showed a small but significant influence of the MRI scanner’s B0 field on the IC response, which was beam energy-dependent. Further investigations should clarify the necessity of dosimetric correction factors for the MR-integrated PT.