High resolution entry and exit Monte Carlo dose calculations from a linear accelerator 6 MV beam under the influence of transverse magnetic fields.

A current concern with 6 MV transverse field MRI-linac hybrid systems is the predicted increases in skin dose (both the entry and exit sides) caused by the effects of the magnetic field on secondary electrons. In this work high resolution GEANT4 Monte Carlo simulations have been performed at the beam central axis in the entry and exit regions of a water phantom to predict surface (0 microm depth) and skin (70 microm depth) doses when placed in such a hybrid system. A 30 x 30 x 20 cm3 water phantom with 10 microm thick voxels has been simulated by being irradiated perpendicularly with a 6 MV photon beam (Varian 2100C) of sizes of 5 x 5, 10 x 10, 15 x 15, and 20 x 20 cm2. Uniform transverse magnetic fields of 0.2, 0.75, 1.5, and 3 T with varying thickness above the phantom have been investigated. Simulations with and without lepton contamination have been performed. In the entry region the high resolution scoring has yielded unexpected surface and skin doses. There is a small amount of nonpurged air-generated lepton contamination that originates immediately above the phantom surface and delivers its dose over very short longitudinal distances in the entry region. At 0.2 T the surface and skin doses are not accurately predicted using lepton-contamination-free simulations and extrapolated lower resolution scoring. Lepton-free simulations are up to 7% of Dmax lower than simulations with leptons. However, compared to 0 T, entry skin dose is reduced at 0.2 and 0.75 T but increases to 28%-31% of Dmax at 3 T. For skin doses at the central axis in the exit region, high resolution scoring shows relative increases of 38%-106%, depending on the magnetic field strength and field size. These values are also up to 20% higher than lower resolution results. The shape of the exit dose profiles varies unpredictably and so extrapolation of low resolution data is insufficient. In order to achieve accurate Monte Carlo skin dosimetry in a transverse field MRI-linac system, the authors recommend using high resolution scoring. In systems of 0.2 T the inclusion of air-generated lepton contamination is also recommended.