Monte Carlo simulation of MLC-shaped TrueBeam electron fields benchmarked against measurement

Modulated electron radiotherapy (MERT) and combined, modulated photon/electron radiotherapy (MPERT) have received increased research attention, having shown capacity for reduced low dose exposure to healthy tissue and comparable, if not improved, target coverage for a number of treatment sites. Accurate dose calculation tools are necessary for clinical treatment planning, and Monte Carlo (MC) is the gold standard for electron field simulation. With many clinics replacing older accelerators, MC source models of the new machines are needed for continued development, however, Varian has kept internal schematics of the TrueBeam confidential and electron phase-space sources have not been made available. TrueBeam electron fields are not substantially different from those generated by the Clinac 21EX, so we have modified the internal schematics of the Clinac 21EX to simulate TrueBeam electrons. BEAMnrc/DOSXYZnrc were used to simulate 5x5 and 20x20 cm$^2$ electron fields with MLC-shaped apertures. Secondary collimating jaws were set 0.5 cm beyond the MLC periphery and to 40x40 cm$^2$. Complete accelerator models agreed with diode measurement within 2%/2 mm at 12 and 20 MeV, and within 3%/3 mm at 6 MeV. Comparisons of measured depth and profile data showed dose-dependencies on jaw position; simulated energy fluences scored just above the phantom showed that, for small apertures, dose dependencies on jaw position are dominated by changes in the in-field energy fluence (not scattered by jaws or MLC) at 6 MeV, while at 20 MeV dose dependencies are dominated by the scattered component. Our models reproduce these jaw position dependencies, which is an asset as there is no consensus on the optimal position for jaws in modulated electron field delivery. Good agreement between simulation and measurement and flexibility in jaw position make our models appropriate for use in MERT and MPERT planning.