Plan optimization for a lung patient on a parallel linac-MR system

Abstract

Purpose: In a hybrid Linac-MR system, the magnetic field of the MRI produces a Lorentz force on charged particles that affects dose deposition. This study investigates the importance of these dosimetric perturbations in the fluence optimization and final dose calculation stages of the treatment planning process for lung radiotherapy. Methods: We modified the EGSnrc codes to generate patient-specific beamlets for a simplified Stereotactic Body Radiation Therapy (SBRT) treatment plan of a lung patient. Beamlets were generated for three scenarios: no magnetic field, and in the presence of a 0.56T and 1.5T magnetic field parallel to the beam. Each set of beamlets was imported into a research version of the RayStation treatment planning system (TPS), and were used to perform 1) forward dose calculations for unmodulated beams; 2) fluence optimization and subsequent forward calculation for modulated beams. MATLAB code was written to compare the optimized fluence patterns corresponding to each magnetic field strength. Results: The monitor units to achieve equivalent target coverage for unmodulated plans decreased slightly with increasing magnetic field, with max differences of -0.9% for 0.56T, and -2.4% for 1.5T compared to 0T. For the optimized modulated plans, the mean lung dose decreased by 0.8% and 9.5% for the 0.56 T and 1.5T cases, respectively, compared to the no field case. For modulated plans where optimization was done without the magnetic field, PTV D5 values increased by 0.2% and 2.6% for 0.56T and 1.5T respectively, indicating a longer DVH tail. Conclusion: Our study suggests that modeling the parallel magnetic field is dosimetrically relevant for treatment planning of a lung patient at 1.5T, but is clinically negligible at 0.56T. It also indicated that equivalent target coverage (compared to a conventional linac) can be achieved with better sparing of lung if the magnetic field is modeled during planning and fluence optimization.