Computed tomography (CT) is an imaging technique widely used for diagnostic evaluation of acute pathology in different organ systems. In recent years, four-dimensional CT angiography (4D-CTA) has become clinically available for visualization of the cerebral vasculature and brain perfusion. This acquisition is of added value for the diagnostic evaluation and follow-up of neurovascular disorders31, including acute cerebral stroke. Acute stroke is the second leading cause of death worldwide according to the world health organization (WHO), with 5.7 million deaths each year. Overall, the visualization, machine learning and deep learning algorithms for 4D-CTA described in this thesis, and their results, provide a bases for further development of automated analysis of 4D-CTA to detect cerebral vascular pathology. Full cerebral vasculature branches can be automatically segmented and labeled, contrast bolus arrival times can be visualized in a simplified and normalized manner for the evaluation of different neurovascular disorders, and vascular occlusions can automatically be detected at image-level. These systems provide the benchmark for future development of algorithms for 4D-CTA and can be of benefit for radiologists, clinicians and patients in the diagnosis of neurovascular disorders.

Download Thesis