Recently, there is growing interest in operating medical linear accelerators without a flattening filter. Due to the superposition of multiple intensity patterns, uniform beam profiles are no longer required for IMRT. For SBRT non-uniform PTV doses are applied per definition and for the small field sizes required, the peaked shape of the lateral beam pro les is not very pronounced. The main rational for the use of flattening filter free (FFF) beams is the substantially increased dose rate. Other advantages are reduced scatter and leaf transmission. Consequently, leakage radiation, which mainly contributes to peripheral dose (PD) at larger distances of the field edge, is reduced.
Two Elekta Precise linear accelerators were modified to provide 6 and 10 MV photon beams with and without a flattening filter. In the FFF mode a 6-mm thick copper filter was rotated into the beam line to reduce electron contamination and to stabilize the beam. It acts as a beam stopper in case the target breaks. The aim of the present work was a comprehensive dosimetric characterization of unflattened beams including the determination of beam quality variations across the field as well as the measurement of PD for advanced radiotherapy techniques. The measured data were complemented by Monte Carlo (MC) simulations performed in collaboration with the Skane University Hospital, Lund. More specifically, dosimetric data including percentage depth dose, lateral profiles, scatter factors, surface dose and leaf transmission was acquired. Subsequently, the data was implemented into the respective planning system and served as experimental basis for the tuning and verification of a MC model of the accelerator head. Beam quality variations were assessed by the determination of half-value layers (HVL) as a function of the off-axis ray angle. For the 6 MV beams the data was verified at Saint Luke's Hospital, Dublin, with the same linear accelerator but different dosimetric equipment.
Treatment head leakage radiation was measured according to IEC recommendations. Peripheral doses were determined for SBRT and IMRT plans that were delivered to the relevant anatomic region of an anthropomorphic phantom which was extended by a solid water slab phantom. Dosimetric measurements were performed with three different detector types, which were positioned within the slab phantom along the longitudinal isocentric axis.
Depths of dose maxima, dmax, for flattened and unflattened beams did not deviate by more than 2 mm and penumbral widths agreed within 1 mm. In FFF mode the collimator exchange effect was on average found to be 0.3% for rectangular fields. Between maximum and minimum field size head scatter factors of unflattened beams showed 40 and 56% less variation for 6 and 10 MV beams than conventional beams. Phantom scatter factors for FFF beams differed up to 4% from published reference data. For field sizes smaller than 15 x 15 cm2, surface doses were increased for unflattened beams, with maximum differences of 7% at 6 MV and 25% at 10 MV for 5 x 5 cm2 fields. For a 30 x 30 cm2 field, relative surface dose decreased by about 10% for FFF beams. Leaf transmission on the central axis was 0.3 and 0.4% lower for unflattened 6 and 10 MV beams, respectively. The difference between on-axis HVL and at an off-axis ray angle of 10deg was 11% for flattened 6 and 10 MV photon beams. The results agreed within 2% with a commonly adopted generic off-axis energy correction. For unflattened beams the variation was less than 5% in the whole range of off-axis ray angles up to 10deg. The difference in relative HVL data was less than 1% for unflattened beams at 6 and 10 MV. The results of the MC calculations showed an increase in dose output per initial electron at the central axis of a factor of 1.79 and 2.66 for the 6 and 10 MV beams, respectively. The amount of scattered photons from the accelerator head was reduced by 31.7% for the 6 MV beam and 47.6% for the 10 MV beam. With unflattened beams treatment head leakage was reduced by 52% for 6 and 65% for 10 MV. Thus, peripheral doses were in general smaller for treatment plans calculated with unflattened beams.
At 20 cm distance from the field edge PD was on average reduced by 23 and 31% for 6 and 10 MV SBRT plans. For the 10 MV IMRT plans the average reduction was 16% for a prostate and 18% for a head & neck case.
Due to the removal of the conically shaped flattening filter dose rates are substantially increased. The energy spectra of unflattened beams are softer and vary less across the field, thus, a simplification of dose calculation and increased dosimetric accuracy is expected. Removing the flattening filter leads to reduced PD for advanced treatment techniques.
The application of higher beam energies contributed to a further reduction of PD. As IMRT increases out-of-field doses compared to conformal radiotherapy, it is definitely beneficial to seek for a reduction of the patients exposure to dose outside the treatment field. In future projects, the dose calculation accuracy and the achievable plan quality of unflattened beams need to be evaluated for current treatment planning systems.