Found in 1-2% of the population, bicuspid aortic valve (BAV) is the most common congenital cardiovascular disease and is associated with increased risk of aortic valve dysfunction and aortic disease. Current treatment is focused on surgical intervention either after the development of symptomatic aortic stenosis/regurgitation or aortic diameter thresholds. However, guidelines frequently change and surveys have shown that many surgeons follow personal protocols that are often at odds with published guidelines, so there is a need for improved biomarkers to better risk-stratify BAV patients. Magnetic resonance imaging (MRI) using 4D flow, a technique which gives volumetric and temporally resolved velocity information, has the potential to provide hemodynamic biomarkers that may be predictive of disease progression in BAV patients. For instance, wall shear stress on the walls of the aorta has been shown to correlate with extracellular matrix degradation in the ascending aortic wall of BAV patients undergoing aortic resection and replacement. This thesis focused on the evaluation of BAV patients with 4D flow MRI and on streamlining the workflow to facilitate future clinical translation. First, normative values for velocity, helicity, and wall shear stress in the aorta were determined to generate normal aortic atlases to compare patients to and to elucidate normal differences due to age and sex. Second, BAV patients who underwent 4D flow with at least 5 years of follow up were analyzed to show that abnormally elevated wall shear stress in the aorta was correlated with increased diameter at follow up. Third, to avoid the cumbersome and time-consuming manual analysis associated with 4D flow, a pipeline using convolutional neural networks to automatically process these scans was generated. Fourth, a preliminary balanced steady state free precession sequence was developed to improve vessel lumen segmentation and potentially enable measurement of aortic motion and expansion in thesis patients. The normative values and automated workflow developed during this doctoral work could contribute to evaluation and future clinical studies of BAV patients.

Creator

• Scott, Michael Baran

DOI

• https://doi.org/10.21985/n2-cgq2-nh91

Subject

• Biomedical engineering
• Medical imaging

Language

• en

Alternate Identifier

• etdadmin_upload_817047
• http://dissertations.umi.com/northwestern:15540

Keyword

• 4D flow
• MRI
• Aorta

Date created

• 2021-01-01

Resource type

• Dissertation

Rights statement

• In Copyright