Abstract
This work reports a method based on correlation functions to convert EPID transit signals into in vivo dose values at the isocenter point, D iso, of dynamic IMRT beams supplied by Varian linac. Dose reconstruction for intensity-modulated beams required significant corrections of EPID response, due to the X-ray component transmitted through multileaf collimator. The algorithm was formulated using a set of simulated IMRT beams. The beams were parameterized by means of a fluence inhomogeneity index, FI, introduced to describe the degree of beam modulation with respect to open beams. This way, all dosimetric parameters involved in D iso reconstruction algorithm, such as the correlation functions, the correction factor for EPID to phantom distance and the modulated tissue maximum ratios, were determined as a function of the FI index. Clinical IMRT beams were used to irradiate a homogeneous phantom, and for each beam, the agreement between the reconstructed dose, D iso, and the dose computed by TPS, D iso,TPS, was well within 5 %. Moreover, the average ratios, R, between the D iso, and D iso,TPS, resulted equal to 1.002 ± 0.030. Thirty-five IMRT fields of 5 different patients undergoing radiotherapy for head-neck tumors were tested and the results were displayed on a computer screen after 2 min from the end of the treatment. However, 350 in vivo tests supplied an average ratio R equal to 1.004 ± 0.040. The in vivo dosimetry procedure here presented is among the objectives of a National Project financially supported by the Istituto Nazionale di Fisica Nucleare for the development of in vivo dosimetry procedures (Piermattei et al. in Nucl Instrum Methods Phys Res B 274:42-50, 2012) connected to the Record-Verify system of the radiotherapy center.