Purpose: Characterization of a multileaf collimator (MLC) quality assurance (QA) technique for the positional accuracy of a double-focused and double-stacked MLC on a 0.35 T MR-linac using an ionization chamber array (ICA) is presented. As the experimental evaluation of the proposed QA technique may suffer from imperfect positioning of the ICA and gantry sag, Monte Carlo (MC) methods were employed.
Methods: A previously benchmarked GEANT4 MC model of a 0.35 T MR-linac was used. Measured and simulated beam profiles for 10x10 cm² and 24x24 cm² field sizes were compared using gamma analysis for further model validation. Simulations were performed for the X1 and X2 banks of the MLC. The leaves of the bank under investigation occluded half of the inline detector array, creating a half-beam block, while the other bank created a 6.64 cm field size. Positional errors ranging from 0.5-2.5 mm along the crossline direction were introduced in 0.5 mm increments to correlate detector response with the MLC positional offset.
Results: The measured and simulated beam profiles agreed with a 100% pass rate using a gamma criterion of 2%/2mm. A linear relationship was found between the introduced MLC offset and detector response with the coefficient of determination for the X1 and X2 banks being 0.9987 and 0.9991, respectively. The change in detector response with MLC offset was 22.7%/mm and 17.4%/mm for the X1 and X2 banks, respectively, where this difference in sensitivity was determined to be statistically significant (p < 0.001).
Conclusion: The proposed QA method was found to be sensitive to MLC positional offset and has potential to be useful for routine machine QA. Furthermore, differences in sensitivity to MLC offset were found for each MLC bank, which could be the result of the 0.35 T magnetic field influencing the secondary electron trajectories.