Introduction. Atherosclerotic plaque rupture may occur without warning leading to severe clinical events such as heart attack and stroke. The mechanisms causing plaque rupture are not well understood. It is hypothesized that mechanical forces may play an important role in the plaque rupture process and that image-based computational mechanical analysis may provide useful information for more accurate plaque vulnerability assessment. The objectives of this dissertation are: a) develop in vivo magnetic resonance imaging (MRI)-based 3D computational models with fluid-structure Interactions (FSI) for human atherosclerotic carotid plaques; b) perform mechanical analysis using 3D FSI models to identify critical stress/strain conditions which may be used for possible plaque rupture predictions. Data, Model, and Methods. Histological, ex vivo/ in vivo MRI data of human carotid plaques were provided by the University of Washington Medical School and Washington University Medical School. Blood flow was assumed to be laminar, Newtonian, viscous and incompressible. The Navier-Stokes equations with arbitrary Lagrangian-Eulerian (ALE) formulation were used as the governing equations for the flow model. The vessel and plaque components were assumed to be hyperelastic, isotropic, nearly-incompressible and homogeneous. The nonlinear Mooney-Rivlin model was used to describe the nonlinear properties of the materials with parameter values chosen to match available experimental data. The fully-coupled FSI models were solved by a commercial finite element software ADINA to obtain full 3D flow and stress/strain distributions for analysis. Validation of the computational models and Adina software were provided by comparing computational solutions with analytic solutions and experimental data. Several novel methods were introduced to address some fundamental issues for construction of in vivo MRI-based 3D FSI models: a) an automated MRI segmentation technique using a Bayes theorem with normal probability distribution was implemented to obtain plaque geometry with enclosed components; b) a pre-shrink process was introduced to shrink the in vivo MRI geometry to obtain the no-load shape of the plaque; c) a Volume Component-Fitting Method was introduced to generate a 3D computational mesh for the plaque model with deformable complex geometry, FSI and inclusions; d) a method using MRI data obtained under in vitro pressurized conditions was introduced to determine vessel material properties. Results. The effects of material properties on flow and wall stress/strain behaviors were evaluated. The results indicate that a 100% stiffness increase may decrease maximal values of maximum principal stress (Stress-P1) and maximum principal strain (Strain-P1) by about 20% and 40%, respectively; flow Maximum-Shear-Stress (FMSS) and flow velocity did not show noticeable changes. By comparing ex vivo and in vivo data of 10 plaque samples, the average axial (25%) and inner circumferential (7.9%) shrinkages of the plaques between loaded and unloaded state were obtained. Effects of the shrink-stretch process on plaque stress/strain distributions were demonstrated based on six adjusted 3D FSI models with different shrinkages. Stress-P1 and Strain-P1 increased 349.8% and 249% respectively with 33% axial stretch. The effects of a lipid-rich necrotic core and fibrous cap thickness on structure/flow behaviors were investigated. The mean values of wall Stress-P1 and Strain-P1 from lipid nodes from a ruptured plaque were significantly higher than those from a non-ruptured plaque (112.3 kPa, 0.235 & 80.1 kPa, 0.185), which was 40.2% and 26.8% higher, respectively (p<0.001). High stress/strain concentrations were found at the thin fibrous cap regions. These results indicate that high stress concentrations and thin fibrous cap thickness might be critical indicators for plaque vulnerability. Conclusion. In vivo image-based 3D FSI models and mechanical image analysis may have the potential to provide quantitative risk indicators for plaque vulnerability assessment.

Recent advances in deep learning have greatly improved the ability to generate analysis models from medical images. In particular, great attention is focused on quickly generating models of the left ventricle from cardiac magnetic resonance imaging (cMRI) to improve the diagnosis and prognosis of millions of patients. However, even state-of-the-art frameworks present challenges, such as discontinuities of the cardiac tissue and excessive jaggedness along the myocardial walls. These geometrical features are often anatomically incorrect and may lead to unrealistic results once the geometrical models are employed in computational analyses. In this research, we propose an end-to-end pipeline for a subject-specific model of the heart's left ventricle from Cine cMRI. Our novel pipeline incorporates the uncertainty originating from the segmentation methods in the estimation of cardiac indices, such as ejection fraction, myocardial volume changes, and global radial and longitudinal strain, during the cardiac cycle. First, we propose an anatomically-guided deep learning model to overcome the common segmentation challenges while preserving the advantages of state-of-the-art frameworks, such as computational efficiency, robustness, and abstraction capabilities. Our anatomically-guided neural networks include a B-spline head, which acts as a regularization layer during training. In addition, the introduction of the B-spline head contributes to achieving a robust uncertainty quantification of the left ventricle inner and outer walls. We validate our approach using human short-axis (SA) cMRI slices and later apply transfer learning to verify its generalization capabilities in swine long-axis (LA) cMRI slices. Finally, we use the SA and LA contours to build a Gaussian Process (GP) model to create inner and outer walls 3D surfaces, which are then used to compute global indices of cardiac functions. Our results show that the proposed pipeline generates anatomically consistent geometries while also providing a robust tool for quantifying uncertainty in the geometry and the derived cardiac indices.