Aortic aneurysm is a silent, asymptomatic enlargement of the aorta that can lead to life-threatening events. Current clinical guidelines rely on aortic diameter to assess the risk of complications and to time prophylactic interventions. However, the aorta may grow slowly for years or expand rapidly just before the occurrence of adverse events. Size-based criteria are a poor measure of disease progression, so new markers are needed to improve risk stratification in patients with aortic aneurysm.

Advanced magnetic resonance imaging (MRI) permits the non-invasive assessment of cardiovascular anatomy and function. Four-dimensional (4D) Flow MRI measures the three-dimensional blood velocity field over the cardiac cycle, and enables the quantification of blood forces acting on the aortic wall. These forces induce remodelling processes associated with physiological changes and diseases. One such force is the frictional force on the aortic wall, known as wall shear stress (WSS). WSS regulates normal vascular function, but its role in disease progression is unclear. Understanding blood flow forces in both healthy and diseased aortas may provide new insights into the mechanisms of aortic disease, and may help in identifying novel markers to assist clinical decision-making. This thesis aims to enhance the analysis of WSS with 4D Flow MRI in the context of aortic disease.

First, aortic wall motion is incorporated into the WSS analysis. 4D Flow MR image processing methods are mostly manual and time-consuming. Consequently, WSS analysis is mainly restricted to one cardiac phase or neglects aortic wall motion when assessing WSS metrics encompassing the whole cardiac cycle. This thesis tailored semi-automated atlas-based segmentation and surface registration methods to aortic applications, to obtain a moving aortic wall surface.

Second, a robust framework for statistical analysis to compare cohorts based on WSS metrics is proposed. WSS patterns can be visualised and quantified on three-dimensional WSS maps representing the aortic wall surface. These maps are used to infer local differences in WSS metrics between cohorts. This thesis introduces permutation tests in cardiovascular MRI to facilitate such local inter-cohort comparisons.

Third, a deep learning method is proposed to reduce the acquired data in cardiovascular 4D Flow MRI. Increasing the spatial resolution of 4D Flow MR images would be beneficial for estimating WSS metrics. By reducing the acquired data, the lengthy acquisition time of 4D Flow can be shortened or the temporal resolution improved. Alternatively, the time saved can be reinvested to increase the spatial resolution of the images, thereby enhancing WSS assessment.

Finally, the results obtained throughout the work performed in this thesis demonstrate the utility of 4D Flow MRI-based WSS markers in selected cohorts. Markers of altered blood flow were identified in patients with abdominal aortic aneurysm, and in dilated patients with bicuspid valves and aortic regurgitation. Moreover, 4D Flow MRI-based WSS metrics correlated to circulating biomarkers of inflammation and collagen synthesis in patients with mildly dilated aorta and tricuspid valves. These findings showcase the potential of 4D Flow MRI in improving the mechanistic understanding behind aortic diseases and their risk assessment.