Purpose: An accurate and fast model of an electronic portal imaging device (EPID) in Monte Carlo simulation is necessary for transit image calculation, wherein structure resolution is one of the limitations. To resolve this issue, we aim to introduce an effective atomic-number scaling method for EPID modeling in EGSnrc/DOSXYZnrc.
Methods: A solid water phantom at the isocenter and an aS1200 EPID at 150 cm SDD were modeled in a single DOSXYZnrc CT phantom input file (*.egsphant). The resolution of the phantom was set to 2.5 mm for reasonable memory usage and computation speed. The density of EPID voxels was scaled by effective atomic number, Zeff(component)/Zeff(water) at 1.75MeV (average energy on the EPID surface) for each component in EPID. Transit images for phantom thicknesses (10, 20, and 30 cm) and field sizes (5x5, 10x10, and 15x15 cm²) were simulated and compared with measurements on Varian Edge. The simulated and measured images were normalized at 10x10cm² under a 20cm phantom.
Results: Using the atomic-number scaling method, excellent agreement was achieved between simulated and measured transit images. The agreement was consistent for varying field sizes, but changed with phantom thickness with an average difference of -2.43%, 0%, and 1.80% for 10cm, 20 cm (normalization), and 30cm, respectively. Image profiles agreed well with the measurements after off-axis correction on the calculation, as the simulated images correspond to raw EPID images. Simulated penumbras also showed excellent agreement with the measurements.
Conclusion: Effective atomic-number scaled EPID models showed good agreement with measurements under homogeneous phantoms. Agreement slightly varies with phantom thickness, possibly due to the limitation of scaling the water base material by effective atomic numbers instead of directly modeling the actual higher Z materials, affected by incident photon energy.
TH- External Beam- Photons: portal dosimetry, in-vivo dosimetry and dose reconstruction