Room: Room 205
Purpose: Megavoltage computed tomography (MVCT) imaging has been used for patient setup in TomoTherapy. However, there is a concern that the MVCT cannot include all body structures due to the limited field-of-view (FOV) of MVCT. The purpose of this study was to expand the FOV of MVCT based on an iterative reconstruction (IR) algorithm using prior information of treatment planning kilovoltage CT (kVCT).
Methods: A cheese phantom with various density plugs was placed at the outside of the conventional FOV and scanned. The MVCT images were reconstructed with expanded FOV by four reconstruction algorithms, i.e., filtered back projection (FBP), IR without prior information, IR with prior information of original kVCT image, IR with prior information of converted kVCT image. The pixel values of the converted kVCT were calculated from kV-MV conversion formula determined by scanning a cheese phantom with kVCT and MVCT. For evaluation, root mean square error (RMSE) was calculated by subtracting a registered MVCT image, which was scanned at the center of FOV, from each reconstructed image.
Results: RMSE of the image reconstructed by FBP, IR without prior information, IR with information of original kVCT image, IR with information of converted kVCT image, was 214.0, 109.6, 60.9, and 53.2 HU, respectively. The outer region of the conventional FOV could be clearly visualized based on the IR algorithm with prior information of converted kVCT image.
Conclusion: Our results showed that MVCT images with expanded FOV can be potentially reconstructed based on IR algorithm using prior information of treatment planning kVCT.