SU-E-T-218: In Vivo 3D Dose Verification for IMRT Using Electronic Portal Imaging Device (EPID)

Abstract

Purpose: To provide in vivo 3D dose verification for IMRT treatment sessions using an FFT‐based dose calculation method for heterogeneous media. Methods: Dosimetriccalibration of the EPID is performed using an ionization chamber with mini‐phantom. Optical scattering within the EPID is corrected to minimize systemic error. The dose‐response of EPID, i.e. EPID reading as a function of dose, is linear with dose. A fast ray‐tracing algorithm is developed to calculate the full 3D radiological length distribution within the heterogeneous phantom/patient and up to the EPID active imaging plane. Attenuation of the primary beam as a function of radiological length is approximated by a linear quadratic exponential function and its off‐axis variation is determined experimentally by transmittance measurements in order to account for the polyenergetic MV beams. An iterative algorithm is used to find a self‐consistent solution separating the total dose measured by the EPID into primary and scatter contributions through a phantom/patient. Calculation of the scatter contribution from phantom to EPID is achieved by first deconvolving the effect of EPID's internal lateral scattering, followed by the convolution of a 2D scatter kernel accounting for the patient scatter to EPID. Results: The dose‐response for the amorphous‐silicon (aSi) EPID has been obtained. The fast ray‐tracing algorithm has been developed and tested on both phantom and Rando phantom patient geometry. The linear quadratic parameters for our clinical 6MV beam have been estimated both on the central axis and off‐axis, and used for initial validation of this exit dose verification method. First result has shown that the estimated patient dose is within 5% of that calculated by the treatment planning system. Conclusions: With a fast kernel‐based 3D dose calculation algorithm, aSi EPID can potentially be used for dose verification as a routine patient QA procedure.