Comparison between the TRS-398 code of practice and the TG-51 dosimetry protocol for flattening filter free beams

Dosimetry protocols for external beam radiotherapy currently in use, such as the IAEA TRS-398 and AAPM TG-51, were written for conventional linear accelerators. In these accelerators, a flattening filter is used to produce a beam which is uniform at water depths where the ionization chamber is used to measure the absorbed dose. Recently, clinical linacs have been implemented without the flattening filter, and published theoretical analysis suggested that with these beams a dosimetric error of order 0.6% could be expected for IAEA TRS-398, because the TPR20,10 beam quality index does not accurately predict the stopping power ratio (water to air) for the softer flattening-filter-free (FFF) beam spectra. We measured doses on eleven FFF linacs at 6 MV and 10 MV using both dosimetry protocols and found average differences of 0.2% or less. The expected shift due to stopping powers was not observed. We present Monte Carlo k_Q calculations which show a much smaller difference between FFF and flattened beams than originally predicted. These results are explained by the inclusion of the added backscatter plates and build-up filters used in modern clinical FFF linacs, compared to a Monte Carlo model of an FFF linac in which the flattening filter is removed and no additional build-up or backscatter plate is added.