31 EPID response modeling with Monte Carlo for Novalis TrueBeam STx

Introduction Portal Dosimetry is a very convenient tool to perform patient QA for VMAT plan with TrueBeam STx Linac. On the other hand, it is not sensitive to gantry rotation. Monte Carlo modeling of EPID response could offer an alternative to the PDIP (portal dosimetry prediction algorithm) in Eclipse. It also opens wider perspectives: response prediction with phantom or patient. Preliminary work focuses on PSF (point spread function) fitting, and verification against measurements. Methods As we cannot irradiate EPID with a very narrow ( Monte Carlo calculation of the EPID response is performed with PenEasy. A Phase Space is obtained after collimator from PRIMO. This phase space can then be handled by PenEasy (including EPID spread function), with simple object in the beam path. Results The sharpest transition from blocked to open beam is obtained using metal blocks behind EPID. A.Latorre showed that best fit to experimental results is obtained with modified Lorentzian model (with FWHM = 0.74 mm). This model fits well the steepest part of the penumbra. Minor deviation ( 5%) occurs in the furthest region under open beam. This model does not fit under the blocked area (probably due to block transmission/scattering). Ongoing step is to measure EPID response with simple regular shaped PMMA blocks. Conclusions Monte Carlo simulation (combining PRIMO, PenEasy and fitted EPID spread function) offers the perspective of predicting EPID response for TrueBeam STx linac. It can be used as an alternative to PDIP algorithm. Using it with anisotropic phantom could make Portal Dosimetry sensitive to gantry rotation. Using it in vivo may allow to check patient positioning against CT positioning in TPS.