Room: Track 3
EPID-based in vivo dosimetry can verify the accuracy of the delivered treatment plan. However, the EPID image is contaminated with patient-generated scattered photons. If this component can be accurately estimated, it’s effect can be removed and therefore the resulting in vivo patient dose estimate will be more accurate. Our group is investigating a ‘tri-hybrid’ algorithm combining analytical, Monte Carlo, and pencil beam scatter kernel methods to provide fast but accurate estimates of patient-generated photon scatter.
Analytical methods are used to solve exactly for singly-scattered photon fluence. For the multiply scattered photon component, a modified Monte Carlo (MC) simulation method was applied, using only a few thousand histories. From each second and higher-order interaction site in the MC simulation, energy fluence entering all pixels of the imager scoring plane was calculated using analytical methods. For the bremsstrahlung and positron annihilation component, a convolution/superposition approach was employed using pencil beam scatter kernels superposed on the incident fluence. The kernel library is pre-generated for a variety of thicknesses and air gaps.
The total patient-scattered photon fluence entering the EPID was compared with full Monte Carlo simulation (EGSnrc) for validation. Three geometric phantoms (water, half-water/half-lung, CT thorax) were tested with a 6 and 18 MV, 10 x10 cm² photon beam, SSD 100 cm, source-to-imager distance 150 cm, and 40x40 cm² imager.
For these tests, the proposed tri-hybrid method was demonstrated it to be in good agreement with full Monte Carlo simulation, with the average and standard deviation across the entire EPID within 0.5% and 0.8% respectively. The method takes approximately 80 seconds to execute on a single (non-parallel) CPU, while the full MC simulations took about 32 hours.
For examples studied here, the proposed tri-hybrid method demonstrates accuracy approaching full Monte Carlo simulation but with a considerably reduced calculation time.