Dosimetric effects of removing the flattening filter in radiotherapy treatment units

The aim of this work was to investigate the dosimetric effects of removing the flattening filter from conventional C-arm medical linear accelerators. In conventional linear accelerators used for radiotherapy, a flattening filter is positioned in the beam line to provide a uniform lateral dose profile at a specified depth in water. However, for some radiotherapy treatments, a uniform lateral dose profile is not necessary, e.g. stereotactic treatments with small fields or treatments with intensity modulated fields. In this work, a comprehensive set of measurements and Monte Carlo simulations for a modified Elekta Precise linear accelerator, operating with and without a flattening filter, were performed and the differences were evaluated. For an Elekta Precise linac, it was found that by removing the flattening filter the dose could be delivered approximately twice as fast as when the flattening filter is in the beam line, under certain conditions. The scatter produced in the treatment head was reduced by 30 %–45 % when the flattening filter was removed and the variation of scattered radiation with field size was also reduced. Removal of the flattening filter resulted in a softer photon energy spectra which leads to a steeper absorbed dose fall-off with depth and less lateral variation across the field. By increasing the acceleration potential of the linac, the depth–dose profiles become more similar to those of the equivalent conventional photon beam and thus the output will also be increased. The suitability of two beam quality measures, TPR20,10 and %dd(10)x, in predicting water to air mass collision stopping-power ratios sw,air for flattening filter-free photon beams was also investigated. These quality measures are used in reference dosimetry for the determination of absorbed dose in water. It was shown that the relationship between TPR20,10 and sw,air used in a current international code of practice for reference dosimetry, overestimates the stopping-power ratio by approximately 0.3 % for flattening filter-free photon beams, while the relationship between %dd(10)x and sw,air, used in the North American code of practice is more accurate. A new beam quality metric, consisting of both TPR20,10 and TPR10,5 was evaluated. It was found that this new beam quality specifier more accurately predicted stopping power ratios for flattening filter-free photon beams. A beam quality specifier defined by the first two moments (describing the mean and variance) of the spectral distribution was also investigated and found to accurately predict stopping-power ratios for beams without a flattening filter