<p> </p>
<p>Cerebral aneurysms are presented in 3-5% of the population and account for approximately 10% of all strokes. The clinical decision on treating unruptured aneurysms should not be taken lightly because a majority of the asymptomatic cerebral aneurysm will not rupture, while both endovascular and microsurgical treatments carry the risk of morbidity and mortality. Thus, there is a need for objective risk assessment to reliably predict the high-risk aneurysms to intervene. Recent studies have found that the blood flow hemodynamic metrics such as pressure and wall shear stress (WSS) are related to the growth and rupture of the aneurysms. 4D flow magnetic resonance imaging (MRI) measures time-resolved three-dimensional velocity fields in the aneurysms <em>in vivo</em>, allowing for the evaluation of hemodynamic parameters. This work presents the developments of flow-physics constrained data enhancement and augmentation methods for 4D flow MRI to assist the risk stratification of cerebral aneurysms. First, a phase unwrapping and denoising method is introduced to enhance the dynamic range and accuracy of 4D flow MRI velocity measurement by incorporating the divergence-free constraint of incompressible flow. Moreover, methods are developed to improve the estimation of hemodynamic parameters from 4D flow data including pressure and WSS. The pressure reconstruction method is also applied to the flow data acquired using particle imaging velocimetry (PIV) and particle tracking velocimetry (PTV) and shows superior performance as compared to the existing methods by solving the pressure Poisson equation. We also proposed a framework to estimate the uncertainty of the PIV/PTV based pressure estimation by propagating the velocity uncertainty. In addition, a multi-modality approach is introduced to enhances the resolution and accuracy of 4D flow data with sparse representation, which improves the reliability of the hemodynamic evaluation. Finally, we present a method to measure the left ventricular flow propagation velocity from cardiac imaging to help in assessing the diastolic function. </p>
| Identifer | oai:union.ndltd.org:purdue.edu/oai:figshare.com:article/19650963 |
| Date | 25 April 2022 |
| Creators | Jiacheng Zhang (12455544) |
| Source Sets | Purdue University |
| Detected Language | English |
| Type | Text, Thesis |
| Rights | CC BY 4.0 |
| Relation | https://figshare.com/articles/thesis/Data_Augmentation_and_Enhancement_for_Cardiovascular_4D_Flow_MRI/19650963 |
Page generated in 0.0022 seconds