The prediction of transmitted dose distributions using a 3D treatment planning system

Abstract

Patient dose verification is becoming increasingly important with the advent of new complex radiotherapy techniques such as conformal radiotherapy (CRT) and intensity-modulated radiotherapy (IMRT). An electronic portal imaging device (EPID) has potential application for in vivo dosimetry. In the current work, an EPID has been modelled using a treatment planning system (TPS) to predict transmitted dose maps. A thin slab of RW3 material used to initially represent the EPID. A homogeneous RW3 phantom and the thin RW3 slab placed at a clinical distance away from the phantom were scanned using a CT simulator. The resulting CT images were transferred via DICOM to the TPS and the density of the CT data corresponding to the thin RW3 slab was changed to 1 g/cm3. Transmitted dose maps (TDMs) in the modelled EPID were calculated by the TPS using the collapsed-cone (C-C) convolution superposition (C/S) algorithm. A 6 MV beam was used in the simulation to deliver 300 MU to the homogenous phantom using an isocentric and SSD (source-to-surface) technique. The phantom thickness was varied and the calculated TDMs in the modelled EPID were compared with corresponding measurements obtained from a calibrated scanning liquid-filled ionisation chamber (SLIC) EPID. The two TDMs were compared using the gamma evaluation technique of Low et al. The predicted and measured TDMs agree to within 2 % (averaged over all phantom thicknesses) on the central beam axis. More than 90 % of points in the dose maps (excluding field edges) produce a gamma index less than or equal to 1, for dose difference (averaged over all phantom thicknesses), and distance-to-agreement criteria of 4 %, 3.8 mm, respectively. In addition, the noise level on the central axis in the predicted dose maps is less than 0.1 %. We found that phantom thickness changes of approximately 1 mm, which correspond to dose changes on the central beam axis of less than 0.6 %, can be detected in the predicted transmitted dose distributions.