Entry skin dose reduction in an inline MRI-linac using an electron contamination deflector coupled with a helium...

A unique feature of inline MRI-linac prototype systems is the changes to entry surface or skin doses. For the high field (1 T) Australian MRI-linac prototype, a strong surface dose increase is observed. This is a by-product of secondary electron contamination being focused down the MRI bore due to the magnetic field being parallel with the x-ray beam direction. It has been anticipated that future clinical treatments will utilize a 2 cm thick water equivalent bolus layer in air above the patient's skin to absorb the electron contamination that causes the large dose hot spot. However, this will reduce the maximum dose rate of the beam and negate the skin dose-sparing effect typical of MV x-ray beams.

To explore through Monte Carlo (MC) modeling and experimental (Exp) measurements, the skin dose changes when employing a small bespoke permanent magnet electron contamination deflector (ECD) in combination with a helium gas volume (HV) located in the MRI bore. The ECD will purge electron contamination from the beam before it reaches the MRI bore, while the HV minimizes the reintroduction of the contamination above the patient.

3D magnetic field models of the Australian MRI-linac coupled with and without the ECD were created in COMSOL Multiphysics, and the magnetic field maps were imported into Geant4 MC simulations. The Geant4 simulations included the 6 MV linac, multileaf collimator (MLC), ECD, HV and a 30 $\times 30\times$ 30 ${\mathrm{cm}}^3$ water phantom positioned at the MRI isocenter. Field sizes of 2.4 $\times$ 2.4  ${\mathrm{cm}}^2$ , 7.2 $\times$ 7.2  ${\mathrm{cm}}^2$ , and 12.4 $\times$ 12.4  ${\mathrm{cm}}^2$ were simulated and surface skin doses were calculated at 70 $\mu m$ depth. Simulations were performed with and without the MRI field, ECD, and HV to characterize the changes introduced by each component. Exp measurements were also performed using real prototypes of the ECD and HV in the identical environment in order to validate our simulation results. Dose measurements were conducted using a MOSkin™ detector and Gafchromic^® EBT3 films.

The skin dose hot spot maximum is measured to be 122.5% (2.4 $\times$ 2.4  ${\mathrm{cm}}^2$ ), 194.8% (7.2 $\times$ 7.2 ${\mathrm{cm}}^2$ ), and 260.5% (12.4 $\times$ 12.4 ${\mathrm{cm}}^2$ ) of the dose at $d_{max}$ (1.5 cm depth) with no magnetic field (MC MRI off). Introducing the ECD reduced these values by around 30%. As expected, the reintroduction of electron contamination below the ECD was observed due to the beam requiring transport through another 120 m of air before reaching the phantom surface. Finally, including the HV further reduced the skin dose hot spots by approximately 60% of their original values. The latter results also demonstrated a considerable reduction in the cross-sectional area of the skin dose hot spot.

The inclusion of an ECD coupled with a HV in an inline MRI-linac prototype system has significantly reduced the intensity of the skin dose increases. It appears to be superior solution over a generic 2 cm water bolus for all field sizes and also will retain some skin dose-sparing effect. The methods used in this work are expected to be feasible with other inline MRI-linac designs.