(Thursday, 7/18/2019) 10:00 AM - 12:00 PM
Room: 304ABC
Purpose: To model realistic CT scanner complexities using the Linear Boltzmann Transport Equation (LBTE) solver to enable rapid, patient-specific CT dose estimation.
Methods: Monte Carlo codes for x-ray transport represent the gold standard method for reliable dose estimations, but they are very time-consuming tools. The method proposed to overcome this limitation is a fast, deterministic Linear Boltzmann Transport Equation (LBTE) solver to estimate CT dose in the diagnostic energy range. The LBTE solver has the flexibility to model the complexities of CT scanners, but the accuracy of the modeling depends on the level of discretization of the parameter space (energy, angle, voxel, etc.). In this study, CT scanner complexities of bowtie filter, helical over-range collimation, and tube current modulation were modeled in the LBTE solver. The scanner effects were modeled by defining distributed, discretized x-ray beams whose spectra and fluence varied across fan angle, gantry angle, and slice position. The results were validated against the Monte Carlo code Geant4, which modeled x-ray beams that varied continuously in space and energy. A voxelated PMMA cylindrical phantom of 32-cm-diameter was modeled as well as an anthropomorphic phantom. The Geant4 simulations used voxels of 1x1x1 mm³ dimension, while the LBTE solver modeled 5x5x3 mm³ voxels.
Results: Results show a good agreement between the two methods, with RMSE < 1% for all the studied configurations. The main advantage of the LBTE solver is the fast execution time: the LBTE solver, running on a GeForce GTX 1080 graphic processing unit, estimated the dose maps in less than 12 seconds for each case.
Conclusion: The LBTE solver accurately estimates CT dose maps, while accounting for scanner complexities and overcoming the time limitations of Monte Carlo codes.
Keywords
Taxonomy
Contact Email
