A simple model for clinical dosimetry with an electronic portal imager

Increasing use of highly conformal, dynamic, and intensity modulated treatment fields has inspired the development of new methods of dose verification. Electronic portal imaging devices (EPIDs) allow measurements to be taken throughout treatment for near-instantaneous verification of such dynamic fields, EPIDs are easy to operate and their application to dosimetry requires little additional effort. Clinical implementation of EPIDs for dosimetry has seen little use, however, because the currently available detectors utilize liquid-ion chamber arrays which have a nonlinear dose rate response. The calculation of dose from EPID measurements has been attempted using complex tables of correction factors measured with an EPID or using a convolution model. These methods are difficult to implement and impractical for frequent clinical use. A simple, clinically useful method using standard empirical data and a virtual-source correction factor is developed for prediction of transmission dose behind a phantom along the central axis. Transmission doses predicted with this model are shown to be accurate with an average difference of 1.5% compared with ion chamber measurements and 2% compared with EPID measurements. This method is also used to calculate dose in a phantom from EPID measurements; this is shown to be accurate compared to actual dose delivered with an average difference of 1.6%. The verification of transmission doses or point doses in a patient or phantom is determined to be possible. Correction factors may be used to apply this method to points off the central axis and in the presence of a variety of beam modifiers