3-D Imaging of Coronary Vessels Using C-arm CT

Cardiovascular disease has become the number one cause of death worldwide. Its prevention, diagnosis and therapy are therefore highly important topics in today’s clinical practice and research. For the diagnosis and therapy of coronary artery disease, interventional C-arm-based fluoroscopy is an imaging method of choice. It delivers 2-D X-ray images from almost arbitrary directions, but a 2-D projection image is naturally limited in its depiction of complex spatial relations. While the C- arm systems are capable of rotating around the patient and thus allow a CT-like 3-D image reconstruction, their long rotation time of about five seconds leads to strong motion artefacts in 3-D coronary artery imaging. Several methods to estimate the coronary motion and compensate for it during 3-D image reconstruction can be found in the literature. All have their specific properties, advantages and disadvantages, which are discussed in the first part of this thesis.

Then, a novel method is introduced that is based on a 2-D–2-D image registra- tion algorithm, henceforth called rmc (Registration-based Motion Compensation). It is embedded in an iterative algorithm for motion estimation and compensation. rmc does not require any complex segmentation or user interaction and is thus fully automatic, which is a very desirable feature for interventional applications. Motion estimation and compensation becomes more difficult when projection data from the whole heart cycle is used from the beginning. rmc overcomes this by successively increasing the utilised amount of projections in a bootstrapping process.

Throughout the remainder of this thesis, rmc is first evaluated in a simulation study using a simple numerical phantom, then on the publicly available Cavarev platform (employing an anthropomorphic phantom), and finally in a study using 58 human clinical datasets. Through the simulation study, approximations for the inherent error of the investigated algorithms were established. In addition, evidence that the missing depth information of a 2-D motion model is not a limiting factor for coronary artery imaging was found. The Cavarev experiments investigated the effect of different filter kernel choices during the execution of rmc. For the quantitative evaluation on human clinical data, a new software called CoroEval was introduced to the scientific community.

Overall, it could be shown from both the quantitative results as well as the human observer ratings that rmc can be successfully applied to a large set of clinical data without user interaction or parameter changes, and with a high robustness against initial 3-D image quality, while delivering results that are at least up to the current state of the art, and better in many cases.