Medical Physics Seminar – Monday, March 22, 2010
The Application of kV Cone Beam CT Images for Adaptive Radiotherapy
Yi Rong, Ph.D.
Research Assistant
Department of Medical Physics
UW-School of Medicine and Public Health
Madison, Wisconsin
This work was designed to validate the feasibility of kV Cone Beam CT imaging system for online/offline Adaptive Radiotherapy to improve the quality of radiotherapy on an individual basis. We evaluated the importance of using CBCT image for patient localization in terms of tumor control and normal tissue avoidance. Significant tumor control loss would occur in patients if no image guidance had been performed. Shifting or rotating prior to treatment may not be sufficient for correcting anatomy variations in their shape and size, thus adaptive radiotherapy is essential for accurate dose delivery. Precise calibration of Hounsfield Units (HU) to electron density (HU-density) is essential to dose calculation. On-board kV CBCT imaging is used predominantly for patients’ positioning, but will potentially be used for dose calculation. The impacts of varying three imaging parameters (mAs, source-imager distance (SID) and cone angle) and phantom size on the HU number accuracy and HU-density calibrations for CBCT imaging were studied. We proposed a site-specific calibration method to achieve higher accuracy in CBCT image based dose calculation. Three configurations of the CIRS water equivalent electron density phantom were used to simulate sites including head, lungs, and lower body (abdomen/pelvis). The planning CT scan was used as the baseline for comparisons. CBCT scans of these phantom configurations were performed using Varian Trilogy™ system in a pre-calibrated mode with fixed tube voltage (125kVp), but varied mAs, SID and cone angle. An HU-density curve was generated and evaluated for each set of scan parameters. Three HU-density tables generated using different phantom configurations with the same imaging parameter settings were selected for dose calculation on CBCT images for an accuracy comparison. Changing mAs or SID had small impact on HU numbers. Yet reducing the cone angle significantly decreases the HU discrepancy. The HU-density table was also affected accordingly. By performing dose comparison between CT and CBCT image based plans, results showed that using the site-specific HU-density tables to calibrate CBCT images of different sites improves the dose accuracy to ~2%. Our phantom study showed that CBCT imaging is a feasible option for dose computation in adaptive radiotherapy approach if the site-specific calibration is applied.