Author:

Abstract:

X-ray Computed Tomography (CT) is one of the most important means in current medical imaging. The theory was originally proposed by Godfrey N Hounsfield in 1967. Today, the CT scanners have evolved enormously, in both system design and imaging performance. Cone beam computed tomography (CBCT) is a relatively new CT technique with applications in diagnostic radiology, image-guided surgery and image-guided radiotherapy. The most prominent feature of CBCT is the image acquisition geometry, where volumetric data are acquired within a single rotation of the paired source and detector. As compared to conventional CT, CBCT has a lower dose requirement and the images are produced (reconstructed) with truly isotropic data samples in three-dimensional space and are of a higher spatial resolution. Clinically, CBCT was first introduced for angiography in interventional radiology and later explored as an image-guidance tool in external radiotherapy. Over the last decade, CBCT of the head and neck with dedicated imaging systems has also been increasingly available.Whether and where can CBCT replace the traditional radiography and CT has become a hot topic in recent years. On the one hand, commercial CBCT systems show large difference from each other, such as the wide range of exposure setting. On the other hand, quality test methodologies used in general radiology still have to be adapted before applied to CBCT. As every emerging technique, regulatory policies and standardization measures for CBCT are yet to be established or improved. As every imaging modality based on ionizing radiation, radiation safety is an issue too. Given all these reasons, investigation and optimization methods and tools are of great interest in current researches. The Monte Carlo simulation technique, for instance, seems to be powerful for CBCT studies by offering the possibility to look into the detailed imaging process from a physics point of view. However, real CBCT systems are very complicated in design architecture. Basic Monte Carlo simulations might therefore be limited in both accuracy and efficiency. The large difference of the CBCT systems demands also high flexibility of the simulation platform.The objective of this thesis study was to develop an accurate, efficient and flexible simulation tool with general applicability to CBCT, and to apply the simulation tool for quality assurance, system evaluation and optimization of dedicated CBCT of the head and neck. The work consists of five parts, including one preliminary study, one in-depth study of the simulation technique and three selected applications. In the preliminary study, we learned experiences about the basics of CBCT simulation and about different approaches to represent the imaging object within the simulation. In the next study, we developed and experimentally validated a so-called hybrid simulation technique, which accounts for the complete CBCT imaging chain using combined methods and in an efficient manner. The application of this technique, as illustrated in the three subsequent studies, coversthe system design, the associated dose and the resulting image quality. We managed to demonstrate the usefulness of the simulation technique in exploring the current CBCT systems beyond their provided hardware design, in systematic evaluation of the dose distribution in typical CBCT examinations as well as in identifying the causation of certain image artefacts and in assisting software algorithms to ultimately improve the image quality. We believe the hybrid simulation technique and the associated new methodologies have a great value not only for CBCT researches but also for medical imaging researches in general.