Monitor backscatter factors for the Varian 21EX and TrueBeam linear accelerators: measurements and Monte Carlo modelling
Linac backscattered radiation (BSR) into the monitor chamber affects the chamber's signal and has to be accounted for in radiotherapy dose calculations. In Monte Carlo (MC) calculations, the BSR can be modelled explicitly and accounted for in absolute dose. However, explicit modelling of the BSR becomes impossible if treatment head geometry is not available. In this study, monitor backscatter factors (MBSFs), defined as the ratio of the charge collected in the monitor chamber for a reference field to that of a given field, have been evaluated experimentally and incorporated into MC modelling of linacs with either known or unknown treatment head geometry. A telescopic technique similar to that by Kubo (1989 Med. Phys. 16 295-98) was used. However, instead of lead slits, a 1.8 mm diameter collimator and a small (2 mm diameter) detector positioned at extended source to detector distance were used. This setup provided a field of view to the source of less than 3.1 mm and allowed for MBSF measurements of open fields from 1 × 1 to 40 × 40 cm2. For the fields with both X and Y dimensions exceeding 15 cm, a diode detector was used. A pinpoint ionization chamber was used for smaller fields. MBSFs were also explicitly modelled in MC calculations using BEAMnrc and DOSXYZnrc codes for 6 and 18 MV beams of a Varian 21EX linac. A method for deriving the D_ch^forward values that are used in MC absolute dose calculations was demonstrated. These values were derived from measured MBSFs for two 21EX and four TrueBeam energies. MBSFs were measured for 6 and 18 MV beams from Varian 21EX, and for 6 MV, 10 MV-FFF, 10 MV, and 15 MV beams from Varian TrueBeam linacs. For the open field sizes modelled in this study for the 21EX, the measured MBSFs agreed with MC calculated values within combined statistical (0.4%) and experimental (0.2%) uncertainties. Variation of MBSFs across field sizes was about a factor of two smaller for the TrueBeam compared to 21EX Varian linacs. Measured MBSFs and the derived D_ch^forward factors allow for the incorporation of the BSR effect into accurate radiotherapy dose calculations without explicit backscatter modelling.