Session H15: Bio: Cardiovascular Flow II

The effect of inlet boundary conditions in image-based CFD modeling of aortic flow

CFD of cardiovascular flow is a growing and useful field, but simulations are subject to a number of sources of uncertainty which must be quantified. Our work focuses on the uncertainty introduced by the selection of inlet boundary conditions in an image-based, patient-specific model of the aorta. Specifically, we examined the differences between plug flow, fully developed parabolic flow, linear shear flows, skewed parabolic flow profiles, and Womersley flow. Only the shape of the inlet velocity profile was varied--all other parameters were held constant between simulations, including the physiologically realistic inlet flow rate waveform and outlet flow resistance. We found that flow solutions with different inlet conditions did not exhibit significant differences beyond $1.75$~inlet diameters from the aortic root. Time averaged wall shear stress~(TAWSS) was also calculated. The linear shear velocity boundary condition solution exhibited the highest spatially averaged TAWSS, about $2.5$\% higher than the fully developed parabolic velocity boundary condition, which had the lowest spatially averaged TAWSS.   

Hemodynamics of the Aortic Jet and Implications for Detection of Aortic Stenosis Murmurs

Cardiac auscultation with a stethoscope has served as the primary method for qualitative screening of cardiovascular conditions for over a hundred years. However, a lack of quantitative understanding of the flow mechanism(s) responsible for the generation of the murmurs, as well as the effect of intervening tissue on the propagation of these murmurs has been a significant limiting factor in the advancement of automated cardiac auscultation. In this study, a multiphysics computational modeling approach is used to investigate these issues. Direct numerical simulation (DNS) is used to explore the fluid dynamics of the jets formed at the aortic valve and the pressure fluctuations generated by the interaction of this jet with the aortic wall. Subsequently, structural wave propagation in the tissue is resolved by a high-order, linear viscoelastic wave solver in order to explore the propagation of the murmurs through a tissue-like material. The implications of these results for cardiac auscultation are discussed.   

Hemodynamics of physiological blood flow in the aorta with nonlinear anisotropic heart valve

The hemodynamic blood flow in cardiovascular system is one of the most important factor, which causing several vascular diseases. We developed a new Curvilinear Immersed Boundary -- Finite Element -- Fluid Structure Interaction (CURVIB-FE-FSI) method to analyze hemodynamic of pulsatile blood flow in a real aorta with nonlinear anisotropic aortic valve at physiological conditions. Hyperelastic material model, which is more realistic for describing heart valve have been incorporated in the CURVIB-FE-FSI code to simulate interaction of aortic heart valve with pulsatile blood flow. Comparative studies of hemodynamics for linear and nonlinear models of heart valve show drastic differences in blood flow patterns and hence differences of stresses causing impact at leaflets and aortic wall.   

Flow Dynamics of Contrast Dispersion in the Aorta

The time profile of the contrast concentration or arterial input function (AIF) has many fundamental clinical implications and is of~importance for many imaging modalities and diagnosis such as MR perfusion, CT perfusion and CT angiography (CTA). Contrast dispersion in CTA has been utilized to develop a novel method- Transluminal Attenuation Flow Encoding (TAFE)- to estimate coronary blood flow (CBF). However, in~clinical practice, AIF is only available in the descending aorta and is used as a surrogate of the AIF at the coronary ostium. In this work we use patient specific computational models of the complete aorta to investigate the fluid dynamics of contrast dispersion in the aorta. The simulation employs a realistic kinematic model of the aortic valve and the dispersion patterns are correlated with the complex dynamics of the pulsatile flow in the curved aorta. The simulations allow us to determine the implications of using the descending aorta AIF as a surrogate for the AIF at the coronary ostium. PE is supported by the NIH Individual Partnership Program. -/abstract- Category: 4.7.1: Biological fluid dynamics: Physiological - Cardiovasc   

Single camera volumetric velocimetry in aortic sinus with a percutaneous valve

Cardiac flows have long been understood to be highly three dimensional, yet traditional in vitro techniques used to capture these complexities are costly and cumbersome. Thus, two dimensional techniques are primarily used for heart valve flow diagnostics. The recent introduction of plenoptic camera technology allows for traditional cameras to capture both spatial and angular information from a light field through the addition of a microlens array in front of the image sensor. When combined with traditional particle image velocimetry (PIV) techniques, volumetric velocity data may be acquired with a single camera using off-the-shelf optics. Particle volume pairs are reconstructed from raw plenoptic images using a filtered refocusing scheme, followed by three-dimensional cross-correlation. This technique was applied to the sinus region (known for having highly three-dimensional flow structures) of an in vitro aortic model with a percutaneous valve. Phase-locked plenoptic PIV data was acquired at two cardiac outputs (2 and 5 L/min) and 7 phases of the cardiac cycle. The volumetric PIV data was compared to standard 2D-2C PIV. Flow features such as recirculation and stagnation were observed in the sinus region in both cases.   

Simulations of blood flow in patient-specific aortic dissections with a deformable wall model

Aortic dissection is a life-threatening condition in which blood penetrates into the vessel wall, creating a second flow channel, often requiring emergency surgical repair. Up to 50{\%} of patients who survive the acute event face late complications like aortic dilatation and eventual rupture. Prediction of late complications, however, remains challenging. We therefore aim to perform accurate and reliable patient-specific simulations of blood flow in aortic dissections, validated by 4D-Flow MRI. Among other factors, this is a computational challenge due to the compliance of the vessel walls and the large degree of membrane deformation between the two flow channels. We construct an anatomic patient-specific model from CT data including both flow channels and the membrane between them. We then run fluid structure interaction simulations using an arbitrary Lagrangian-Eulerian (ALE) formulation within a multiscale variational framework, employing stabilized finite element methods. We compare hemodynamics between a rigid and a deformable wall model and examine membrane dynamics and pressure differences between the two flow channels. The study focuses on the computational and modeling challenges emphasizing the importance of employing a deformable wall model for aortic dissections.   

Uncertainty Quantification in Multi-Scale Coronary Simulations Using Multi-resolution Expansion

Computational simulations of coronary flow can provide non-invasive information on hemodynamics that can aid in surgical planning and research on disease propagation. In this study, patient-specific geometries of the aorta and coronary arteries are constructed from CT imaging data and finite element flow simulations are carried out using the open source software SimVascular. Lumped parameter networks (LPN), consisting of circuit representations of vascular hemodynamics and coronary physiology, are used as coupled boundary conditions for the solver. The outputs of these simulations depend on a set of clinically-derived input parameters that define the geometry and boundary conditions, however their values are subjected to uncertainty. We quantify the effects of uncertainty from two sources: uncertainty in the material properties of the vessel wall and uncertainty in the lumped parameter models whose values are estimated by assimilating patient-specific clinical and literature data. We use a generalized multi-resolution chaos approach to propagate the uncertainty. The advantages of this approach lies in its ability to support inputs sampled from arbitrary distributions and its built-in adaptivity that efficiently approximates stochastic responses characterized by steep gradients.   

Computational fluid dynamics comparisons of wall shear stress in patient-specific coronary artery bifurcation using coronary angiography and optical coherence tomography.

Blood flow dynamics directly influence biology of the arterial wall, and are closely linked with the development of coronary artery disease. Computational fluid dynamics (CFD) solvers may be employed to analyze the hemodynamic environment in patient-specific reconstructions of coronary arteries. Although coronary X-ray angiography (CA) is the most common medical imaging modality for 3D arterial reconstruction, models reconstructed from CA assume a circular or elliptical cross-sectional area. This limitation can be overcome with a reconstruction technique fusing CA with intravascular optical coherence tomography (OCT). OCT scans the interior of an artery using near-infrared light, achieving a 10-micron resolution and providing unprecedented detail of vessel geometry. We compared 3D coronary artery bifurcation models generated using CA alone versus OCT-angiography fusion. The model reconstructed from CA alone is unable to identify the detailed geometrical variations of diseased arteries, and also under-estimates the cross-sectional vessel area compared to OCT-angiography fusion. CFD was performed in both models under pulsatile flow in order to identify and compare regions of low wall shear stress, a hemodynamic parameter directly linked with progression of atherosclerosis.   

Effects of vascular structures on the pressure drop in stenotic coronary arteries

A stenosis, which is a narrowing of a blood vessel, of the coronary arteries restricts the flow to the heart and it may lead to sudden cardiac death. Therefore, the accurate determination of the severity of a stenosis is a critical issue. Due to the convenience of visual assessments, geometric parameters such as the diameter stenosis and area stenosis have been used, but the decision based on them sometimes under- or overestimates the functional severity of a stenosis, i.e., pressure drop. In this study, patient-specific models that have similar area stenosis but different pressure drops are considered, and their geometries are reconstructed from the coronary computed tomography angiography (CCTA). Both steady and pulsatile inflows are considered for the simulations. Comparison between two models that have a bifurcation right after a stenosis shows that the parent to daughter vessel angle results in different secondary flow patterns and wall shear stress distributions which affect the pressure downstream. Thus, the structural features of the lower and upper parts of a stenosis significantly affect the pressure drop.   

Morphology of drying blood pools.

Often blood pools are found on crime scenes providing information concerning the events and sequence of events that took place on the scene. However, there is a lack of knowledge concerning the drying dynamics of blood pools. This study focuses on the drying process of blood pools to determine what relevant information can be obtained for the forensic application. We recorded the drying process of blood pools with a camera and measured the weight. We found that the drying process can be separated into five different: coagulation, gelation, rim desiccation, centre desiccation, and final desiccation. Moreover, we found that the weight of the blood pool diminishes similarly and in a reproducible way for blood pools created in various conditions. In addition, we verify that the size of the blood pools is directly related to its volume and the wettability of the surface. Our study clearly shows that blood pools dry in a reproducible fashion. This preliminary work highlights the difficult task that represents blood pool analysis in forensic investigations, and how internal and external parameters influence its dynamics. We conclude that understanding the drying process dynamics would be advancement in timeline reconstitution of events.