Dissertation Abstract

Tomographic reconstruction from x-ray projections and its applications to image-guided radiation therapy

Represented by the development of the helical-scan technology, cone-beam scanners, and various sophisticated reconstruction algorithms, X-ray computed tomography (CT) has seen significant advances during the last decade. The use of CT for guiding the process of radiation therapy is currently an active area of research in CT. The major purpose of CT in image-guided radiation therapy (IGRT) is to provide high-quality images at both the treatment planning and dose delivery stages for better achieving the objective of radiation therapy.

The research in this thesis represents a collection of novel reconstruction algorithms and scanning configurations for CT scanners that are currently employed for IGRT. These algorithms and scanning configurations provide strategies for improving image quality, reducing scanning effort, lowering radiation dose, and enhancing the utility of the CT systems.

We have developed and evaluated novel reconstruction algorithms with improved noise and spatial-resolution properties over existing algorithms for full- and half-scan CT. We have presented and evaluated reconstruction algorithms with improved numerical properties for CT scans with asymmetric scanning configurations, which can he used for enlarging the size of the field of view (FOV). We have proposed novel asymmetric configurations and reconstruction algorithms for enhancing the spatial resolution in reconstructed images. We have investigated iterative reconstruction algorithms for few-view megavoltage CT with the purpose of reducing radiation dose and scanning time. We have developed a helical cone-beam scanning capability on a simulator CT system and applied the backprojection-filtration (BPF) algorithm to reconstructing 3D images from data obtained with this system. Finally, we have proposed algorithms in circular cone-beam CT for region of interest (ROI) reconstruction from reduced-scan and/or transversely-truncated data, which could be used for further reducing scanning time and radiation dose and solving the truncation problem caused by the limited FOV size.

The imaging configurations and reconstruction algorithms developed in this thesis also have important implications for diagnostic imaging, small animal imaging; and other medical or industrial applications.

This dissertation citation and abstract are published with permission of ProQuest Information and Learning. Further reproduction is prohibited without permission.