Exit EPID Image Prediction Below Heterogeneous Phantoms Using Monte Carlo Codes
J Yoon1*, J Kim2, J Jung1, I Yeo3, (1) East Carolina University, Greenville, NC, (2) University of Pittsburgh Cancer Institute, Pittsburgh, PA, (3) Loma Linda Univ Medical Center, Loma Linda, CASU-E-T-160 Sunday 3:00PM - 6:00PM Room: Exhibit Hall
Purpose: To predict accurate exit EPID images below lung or bone phantoms by appropriate Monte Carlo modeling. It has been evaluated that XVMC and EGSnrc Monte Carlo codes are suitable for dose calculation in tissue equivalent material at both patient and EPID levels. However, the EPID image prediction is challenging due to its complex structure and material composition of high atomic number.
Methods: A 6MV beam and phantom were modeled using BEAMnrc and XVMC codes. The accuracy of exit dose was validated through comparison with 2-D ion chamber dose under a 2 cm build-up layer. After the validation, Varian aS1000 EPID has been modeled to calculate the EPID response. The material composition of EPID has been used for modeling in EGSnrc user code, DOSRZnrc, whereas homogeneous layer modeling by density scaling of the composition has been applied in XVMC. Dose images at 150 cm were calculated under combinations of plastic water, lung and bone flat phantoms for various field sizes. The results were compared to EPID measurement.
Results: Ion chamber array measurements at EPID level agreed with the XVMC and BEAMnrc/DOSRZnrc calculations within ~3%. EPID images calculated by DOSRZnrc were in good agreement (~1%) with the measurement in all cases. The results from XVMC were within ~2% difference in plateau region, but penumbra slopes were sluggish and dose response at off-axis was slightly lower.
Conclusion: EPID dose images were calculated through EPID composition modeling by XVMC and BEAMnrc/DOSRZnrc. Both Monte Carlo calculations agreed with the measurements in all settings, except penumbra slopes and off-axis responses in XVMC calculations. The EPID model by XVMC can be used for faster and less accurate calculations, whereas the other for more accurate, time spending calculations.