SUMMARY

Computed tomography (CT) has become an important imaging modality in clinical, and has been extended into a wide range of usage with the development of novel CT scanner such as dual energy CT (DECT) and cone-beam CT (CBCT). However, the imaging dose during CT scanning increasingly raises public concern nowadays. Iterative reconstruction algorithms have shown promise on improving quality of CT image with reduced radiation. In the first part of study, a novel iterative reconstruction method is developed to enable a new data acquisition scheme for potential reduction in imaging dose and engineering cost of DECT. As low-dose imaging protocols using iterative reconstruction algorithms becomes more popular, a practical algorithm of noise STD map and NPS calculation on an iteratively reconstructed CT image could be a useful complementary tool in clinical practice. In the second part of this study, a practical method for pixel-wise calculation of noise statistics on an iteratively reconstructed CT image is developed, which enables accurate calculation of noise STD for each pixel and NPS. In addition to dose issue, the widespread use of CBCT in radiation therapy is also hindered by significant image artifacts, most of which are shading artifacts due to scatter contamination. A number of methods have been recently proposed to correct such artifacts for an improved image quality. However, their direct benefit to advance radiation therapy is unclear. In the third part of this study, the performance of deformable image registration between planning CT (pCT) and CBCT affected by the shading correction on current clinical CBCT images is investigated.