Program Information
Improving On-BeamCT Reconstruction From Transit Imaging of Treatment Beams Via MV-To-KV HUcalibration for Dose Verification
H Lee1*, K Cheong2 , J Jung3 , S Cho1 , I Yeo5 , (1) Korea Advanced Institute of Science and Technology, Yuseong-gu, Daejeon, (2) Hallym Univ Sacred Heart Hospital, Anyang, ,(3) East Carolina Univ, Greenville, NC,(4) Loma Linda Univ Medical Center, Loma Linda, CA
Presentations
SU-I-GPD-J-30 (Sunday, July 30, 2017) 3:00 PM - 6:00 PM Room: Exhibit Hall
Purpose: To reconstruct patient images with KV imaging quality at the moment of radiation delivery using the transit MV images and planning CT images.
Methods: We proposed a method to obtain the patient images using transit images of therapeutic irradiations and planning CT images in AAPM2016 annual meeting. In this study, a conversion relationship between the MV CT and KV CT has been established using the HU calibration phantom (RMI). The phantom was CT-scanned and MV-imaged and reconstructed; the two images were linearly correlated. The steps of the proposed reconstruction are summarized as follows. A lung phantom inserted with a ball marker (3-cm diameter) was CT-scanned. Its transit images were generated by Monte Carlo simulation (A). The phantom with the ball shifted by 1 cm was CT-scanned and irradiated by the treatment beams while acquiring transit imaging (B): the transit image simulates radiation therapy environment, while the planning CT provides a ground truth. The difference between A and B was fed into an image reconstruction engine resulting in a difference image due to the shift. The difference image was converted back into the HU value map, and added to the original planning CT image.
Results: The average difference of the ball area between the difference-applied image and the ground truth CT image and was 71.3HU, while CBCT difference was 460HU as an error .
Conclusion: Through the MV-to-KV calibration, we have successfully demonstrated that the MV CT image can be handled in the planning CT image domain, promising to improve the image quality. We will further investigate various calibration methods.
Contact Email: