Measurements of the relative backscatter contribution to the monitor chamber for modern medical linear accelerators; a multi-center study

• Target charge was measured for determination of backscatter to the monitor chamber.
• Jaw settings had greater impact on backscatter for iX than for the four TB linacs.
• Field size dependence of backscatter was clear for flattened beams on iX and TB.
• No field size dependence of the backscatter on the ES linac or for FFF beams on TB.
• Comparison with MC simulations indicated the target charge method to be reliable.

Conversion to absolute dose in Monte Carlo (MC) simulations of MV radiotherapy beams needs correct modeling of backscatter (BS) to the linear accelerator (linac) monitor chamber. For some linacs the BS depends largely on jaw settings. The backscattered fraction (BSF) of radiation can be determined experimentally by measuring ratios of target charge for a given number of monitor units as a function of jaw settings. This was done using the in-house developed ME40 dosimetry system, which is able to determine the target charge for each radiation pulse from the linac. The BSF measurements were performed for different linac models at five Danish radiotherapy clinics. The investigated linacs were four Varian TrueBeams (TB), one Varian iX (iX) and one Elekta Synergy (ES). BSF measurements were performed for square field side lengths ranging from 1 to 40 cm, using the 10 × 10 cm2 field as reference. The impact of the flattening filter on the BSF was investigated through measurements in flattened as well as flattening filter free (FFF) beams. Furthermore, to investigate the contribution from the upper and lower jaws separately, measurements at one of the clinics also included asymmetric fields. For the iX, the obtained BSF measurements were compared with MC simulations performed using the BEAMnrc user code. For flattened beams on the Varian linacs, the measured BSF exhibited a clear linear correlation with square jaw settings (correlation coefficient r > 0.9 with p < 0.001 in all cases). For the ES, however, no correlation between BSF and jaw settings was found (r = 0.04, p = 0.92). The change in BSF with jaw settings was also found to be negligible for FFF beams on the TB linacs, indicating that the flattening filter has a substantial influence on the BSF. Furthermore, the backscatter effect was found to be more pronounced (up to a factor of 7) for the iX compared to the TB. MC simulations on the iX agreed within 0.4% with BSF measurements, indicating that the target charge measurement method used for determination of BSF is accurate. Furthermore, the similar BSF observed for the four TB linacs included in the study also indicates that the method used for target charge measurements is reproducible. The reproducibility lies mainly in the fact that the method basically has no set-up errors and therefore is user independent.