2849

Abstract

Over the last decade, kilovoltage cone beam computed tomography (CBCT) has developed into a potentially significant technology for radiotherapy treatment simulation and verification. The CBCT capability on a conventional simulator is of added advantage if it can supplement the image acquisition task of a CT simulator (e.g., when the CT bore is too small for the patient or treatment setup). However, due to less accurate density information derived from CBCT, there has been reluctance by clinicians to use it for 3D treatment planning. In this work, we evaluate the utility of the CBCT function on a commercial simulator (Acuity IX) for 3D treatment planning by comparing plans based on volumetric and density data acquired with CBCT scans with those based on CT simulator (Picker PQ5000) scans. To assess the suitability of the CBCT simulator for such a study, we initially evaluated its imaging performance by comparing it with a CT simulator. Tests of density resolution, temporal stability, contrast resolution, scan uniformity, and noise, showed results that were comparable for the CBCT and the CT simulator, suggesting that the former could be used for radiotherapy treatment planning purposes. Owing to differences in the image reconstruction processes, CT numbers generated by both imaging modalities differ. Hence, we determined separate calibration curves, relating CT number to electron density for a range typically observed in treatment planning, using a Catphan 600 phantom. We performed scans of a Rando anthropomorphic phantom for three sites: head, thorax, and pelvis. For CBCT, a full fan mode was used for head scans (diameter < 15 cm) and a half fan mode for diameters > 15 cm (thorax and pelvis). Virtual target volumes of common disease areas were contoured according to the ICRU 62 criteria on the CBCT scan. These contours were transferred to the CT images (registered with CBCT) in order to have the same target volumes. The scans and contours were used to generate treatment plans using typical beam arrangements and dose prescriptions. A current constraint in using this CBCT device is the limitation in the length of the volume scan being 17 cm (a pending software upgrade will reduce this limitation by allowing contiguous scans to be merged). Figure 1 displays the isodose distribution for a prostate plan generated using (a) CBCT and (b) CT data. The figure also gives an idea of the image quality obtained in these scans. For the three sites investigated, the percentage difference between the mean target dose coverage obtained by the two modalities was < 0.5%. This small difference in dose demonstrates that CBCT data from an Acuity simulator is a viable alternative to CT simulation data for rendering 3D treatment plans.