Dose calculation accuracy using cone-beam CT (CBCT) for pelvic adaptive radiotherapy

This study is to evaluate the dose calculation accuracy using Varian's cone-beam CT (CBCT) for pelvic adaptive radiotherapy. We first calibrated the Hounsfield Unit (HU) to electron density (ED) for CBCT using a mini CT QC phantom embedded into an IMRT QA phantom. We then used a Catphan 500 with an annulus around it to check the calibration. The combined CT QC and IMRT phantom provided correct HU calibration, but not Catphan with an annulus. For the latter, not only was the Teflon an incorrect substitute for bone, but the inserts were also too small to provide correct HUs for air and bone. For the former, three different scan ranges (6 cm, 12 cm and 20.8 cm) were used to investigate the HU dependence on the amount of scatter. To evaluate the dose calculation accuracy, CBCT and plan-CT for a pelvic phantom were acquired and registered. The single field plan, 3D conformal and IMRT plans were created on both CT sets. Without inhomogeneity correction, the two CT generated nearly the same plan. With inhomogeneity correction, the dosimetric difference between the two CT was mainly from the HU calibration difference. The dosimetric difference for 6 MV was found to be the largest for the single lateral field plan (maximum 6.7%), less for the 3D conformal plan (maximum 3.3%) and the least for the IMRT plan (maximum 2.5%). Differences for 18 MV were generally 1-2% less. For a single lateral field, calibration with 20.8 cm achieved the minimum dosimetric difference. For 3D and IMRT plans, calibration with a 12 cm range resulted in better accuracy. Because Catphan is the standard QA phantom for the on-board imager (OBI) device, we specifically recommend not using it for the HU calibration of CBCT.