The International Electrotechnical Commission (IEC) has previously defined standard rotation operators for positive gantry, collimator and couch rotations for the radiotherapy DICOM coordinate system that is commonly used by treatment planning system
s. Coordinate transformations to the coordinate systems of commonly used Monte Carlo (MC) codes (BEAMnrc/DOSXYZnrc and VMC++) have been derived and published in the literature. However, these coordinate transformations disregard patient orientation during the computed tomography (CT) scan, and assume the most commonly used head first, supine orientation. While less common, other patient orientations are used in clinics - Monte Carlo verification of such treatments can be problematic due to the lack of appropriate coordinate transformations. In this work, a solution has been obtained by correcting the CT-derived phantom orientation and deriving generalized coordinate transformations for field angles in the DOSXYZnrc and VMC++ codes. The rotation operator that includes any possible patient treatment orientation was determined using the DICOM Image Orientation tag (0020,0037). The derived transformations of the patient image and beam direction angles were verified by comparison of MC dose distributions with the Eclipse treatment planning system, calculated for each of the eight possible patient orientations.
Flattening filter free (FFF) beams due to their non-uniformity, are sub-optimal for larger field sizes. The purpose of this study was to investigate the incident electron beam distributions that would produce flat FFF beams without the use of flatten
ing filter. Monte Carlo (MC) simulations with BEAMnrc and DOSXYZnrc codes have been performed to evaluate the feasibility of this approach. The dose distributions in water for open 6MV beams were simulated using Varian 21EX linac head model, which will be called flattening filter (FF) model. Flattening filter has then been removed from FF model, and MC simulations were performed using (1) 6 MeV electrons incident on the target, (2) 6 MeV electron beam with electron angular distributions optimized to provide as flat dose profiles as possible. Configuration (1) represents FFF beam while configuration (2) allowed producing a flat FFF (F4) beam. Optimizations have also been performed to produce flattest profiles for a set of dose rates (DRs) in the range from 1.25 to 2.4 of the DR of FF beam. Profiles and percentage depth doses PDDs from 6MV F4 beams have been calculated and compared to those from FF beam. Calculated profiles demonstrated improved flatness of the FFF beams. In fact, up to field sizes within the circle of 35 cm diameter the flatness of F4 beam at dmax was better or comparable to that of FF beam. At 20 cm off-axis the dose increased from 52% for FFF to 92% for F4 beam. Also, profiles of F4 beams did not change considerably with depth, and for large fields out-of-field dose was reduced by about a factor of two compared to FF beam. PDDs from F4 beams were similar to those of FFF beam. The DR for the largest modeled (44 cm diameter) F4 beam was higher than the DR from FF beam by a factor of 1.25. It was shown that the DR can be increased while maintaining beam flatness, but at the cost of reduced field size.
Monte Carlo (MC) methods provide the most accurate to-date dose calculations in heterogeneous media and complex geometries, and this spawns increasing interest in incorporating MC calculations into treatment planning quality assurance process. This i
nvolves MC dose calculations for clinically produced treatment plans. To perform these calculations, a number of treatment plan parameters specifying radiation beam and patient geometries need to be transferred to MC codes, such as BEAMnrc and DOSXYZnrc. Extracting these parameters from DICOM files is not a trivial task, one that has previously been performed mostly using Matlab-based software. This paper describes the DICOM tags that contain information required for MC modeling of conformal and IMRT plans, and reports the development of an in-house DICOM interface, through a library (named Vega) of platform-independent, object-oriented C++ codes. The Vega library is small and succinct, offering just the fundamental functions for reading/modifying/writing DICOM files in a C++ program. The library, however, is flexible enough to extract all MC required data from DICOM files, and write MC produced dose distributions into DICOM files that can then be processed in a treatment planning system environment. The library can be made available upon request to the authors.
