SCATTER CORRECTION AND 4D RECONSTRUCTION OF CONE BEAM COMPUTED TOMOGRAPHY (CBCT) WITH A MOVING BLOCKER

Abstract

Come-beam computed tomography (CBCT) is being widely used as an image guidance tool in radiation therapy nowadays. The quality of CBCT image is important to localize and delineate the tumor and to define relevant volumes when the patient is in the treatment position. However, there are several factors that can deteriorate the image quality of CBCT. The large amount of scatter signal due to the use of the flat panel detector (FPD) is a major factor and greatly degrades the quality of reconstruction images. Another major factor is the respiratory motion of the patient, which can cause severe motion blurring artifacts. Although four dimensional (4D) CBCT has been developed to reduce the motion artifacts, it increases the radiation dose in order to get satisfactory reconstruction images as the conventional phase-by-phase reconstruction for 4D CBCT requires more projection data than 3D CBCT. In this dissertation, we propose some methods to address these problems.

After the brief introduction of CBCT, the factors that deteriorate the CBCT image quality and the existing studies are reviewed. Next, the moving blocker scatter correction method for CBCT is introduced and its optimization using different deconvolution methods is investigated through a simulation study and a slanted-edge experimental study. A robust moving blocker scatter correction method is further developed, which uses multiple adjacent projection views to estimate a more accurate scatter distribution. Finally, 4D CBCT with a moving blocker is developed for simultaneous dose reduction and scatter correction. The effectiveness of the proposed scatter correction and 4D reconstruction methods is demonstrated either by simulations of digital phantoms or real phantom and clinical CBCT data.