Session W09: Biological Fluid Dynamics: Physiological Large Vessels (10:00am - 10:45am CST)

Deep learning based ultrasound for measuring blood flow and vascular wall movement simultaneously in artery

Several cardiovascular diseases (CVDs) are closely related with vascular stiffness and interactions of blood flow and vessel wall dynamics. However, conventional techniques cannot accurately measure local vascular stiffness and their interactions. In this study, we propose deep learning based simultaneous measurement of flow-wall dynamics (DL-SFW) for measuring flow velocity and vascular strain with high resolution. Its performance is verified by comparing with conventional velocimetry for tissue-mimicking phantoms. The DL approach is found to improve relative errors in the measurements of velocity, wall shear stress (WSS), and strain. In vivo feasibility is demonstrated by applying DL-SFW to murine carotid artery with different pathologies: aging and diabetes mellitus (DM). Its measurement results are compared with other velocimetry and elastography. With DL-SFW, the abnormal flow-wall dynamic interactions of aging and DM are figured out compared with control group. The corresponding histological analysis shows the effect of the abnormal interactions between blood flow and wall dynamics on the vessel. This technique provides useful information with high-resolution and accurate diagnosis of CVDs.   

Data-driven blood flow modeling with sparse representation

In modeling cardiovascular flows, we are often faced with the difficulty of dealing with incomplete, low-resolution, and noisy data. Experimental blood flow measurement techniques are limited in their spatial and temporal sampling resolutions and acquisition noise is often inevitable. A fundamental problem is how to recover high-quality and high-resolution data from these measurements. In this talk, we discuss how we can leverage the hidden structure in data to tackle these problems. We will discuss compressed sensing, optimal sparse sensor placement, machine learning reduced-order modeling, and matrix completion techniques. For each method, we will demonstrate a simple example related to blood flow modeling. Our results show how sparse modeling can combine sparse data sampling with sparse representation in a hidden basis to recover hemodynamic data from imperfect measurements.   

Hemodynamic data assimilation using model order reduction and Kalman filter

Obtaining high-fidelity blood flow data is a challenging task. High-resolution patient-specific computational fluid dynamics (CFD) simulations are sensitive to uncertainty in boundary conditions and model parameters. On the other hand, imaging methods such as time resolved phase contrast magnetic resonance imaging (a.k.a 4D-Flow MRI) suffer from low spatio-temporal resolution and acquisition noise. To overcome these limitations, we propose a data assimilation method based on dynamic mode decomposition (DMD), reduced-order modeling (ROM), and Kalman filtering. Our method leverages DMD to build the predictor operator. Subsequently, a Kalman filter is used to merge CFD and 4D-Flow MRI observations. Our reduced-order approach improves prior ensemble Kalman filter models, which required several expensive ensemble simulations. To test our method, we generate synthetic 4D-Flow MRI data and consider CFD simulations with uncertainty in model parameters. We apply our method to different blood flow problems ranging from Womersley flow to blood flow in cerebral aneurysms. We demonstrate the accuracy of our model and improvement in quantifying hemodynamics.   

Experimental Measurement of Pulsatile Blood Pressure in 3D Printed Stenosed Arteries.

Quantification of proximal ($P_{a)}$ and distal $P_{d}$ pressure to arterial stenosis is critical to assess the hemodynamic severity of stenosis via fractional flow reserve ($P_{d}$/$P_{a})$ or trans-stenotic pressure gradient ($P_{a}$-$P_{d})$. Invasive measurement via catheterization requires patient exposure to risk and high medical costs. We built up a pulsatile flow loop, mimicking the blood flow in the human circulatory system, to measure vascular hemodynamics. The loop is equipped with a pulsatile heart pump, elements of resistance and compliance, and measurement devices. The stenosed artery is segmented from computed tomography angiography data, then 3-D printed out and mounted in the loop. The pressure ($P_{a}$ and $P_{d})$ is measured by medical-grade transducers through an in-house built filter/amplifier. The data acquisition system collects pressure and flow-rate signals simultaneously and visualizations are live via LabVIEW. The system can accommodate rigid or flexible 3D printed arteries and real human arteries. The pulsatile flow loop provides a unique capability to validate non-invasive computed pressure and to quantify patient-specific pathophysiological properties of diseased arteries.   

Clot Permeability, Agonist Transport, and Platelet Binding Kinetics in Arterial Thrombosis

During the process of arterial thrombosis, the formation of stable wall-adherent platelet aggregates requires the rapid formation of large number of interplatelet bonds to sustain the drag force exerted on the thrombus by the background fluid. The magnitude of this force is strongly influenced by the thrombus permeability. We investigate platelet aggregation in coronary-sized arteries using both computational simulations and in vitro experiments. The computational model tracks the formation and breaking of bonds between platelets and treats the thrombus as an evolving porous, viscoelastic material, which moves differently than the background fluid. Fluid and thrombus interact through a Kozeny-Carman drag term, parameterized based on our experimental permeability-porosity measurements. We found that 1) Using physiological parameter values, a stable thrombus growth is possible only if the clot permeability is within the range of our experimental measurement. 2) Under high shear flow, soluble agonist released by platelets is limited to the thrombus and a boundary layer downstream. This limits the thrombus growth into the vessel lumen. 3) Adding to the model binding and activation of unactivated platelets through vWF-mediated processes allows greater thrombus growth.   

Transient stability analysis of steady and pulsatile aneurysmal flows

This study presents a computational investigation of the transient growth and convective instabilities in steady and pulsatile flow in a model abdominal aortic aneurysm with various expansion ratio and bulge lengths. Previous global and transient stability analyses have established fairly high critical Reynolds number for the system to become linearly unstable under steady conditions, according to physiological standards ($Re_c$$>$3000). However, the present study reveals transient instabilities in the system at much lower Reynolds numbers(100$<$$Re$$<$1000), even for steady flow. The linear and non-linear energy evolutions of these growth are presented and compared on the basis of size, Reynolds number and disturbance period ($\tau$). Moreover, for some Reynolds numbers in the range considered, a weak, secondary transient growth develops after the temporal growth maximum is reached, creating a secondary growth envelope.The secondary mode is analyzed on the basis of the energy evolutions. Finally, a Floquet stability analysis of pulsatile conditions is discussed for aneurysms of varying sizes and Reynolds numbers. The results obtained help further the understanding of hemodynamics in aneurysms.   

Characteristics of Pulsatile Flow Through Deformed Aortic Arch Models: A Tomographic PIV and CFD Study.

Romanesque and Gothic aortic arches are two types of deformed arch geometries after surgical repair of coarctation of the aorta. The abnormal geometry, particularly the Gothic type, is commonly associated with systemic hypertension and other cardiovascular complications in part due to irregular hemodynamics. In this study, the pulsatile flow characteristics of deformed arch models were investigated using phase-locked tomographic particle image velocimetry and computational fluid dynamic simulations. A flow waveform of 5 liters/min with a heart rate of 60-100 beats/min was used as boundary conditions for both in vitro pulsatile flow simulator and CFD simulations. The variations of flow patterns and average and fluctuation of velocity were analyzed at specific critical phases of a heart cycle. Results reveal distinct 3D flow characteristics in Romanesque and Gothic arch models. Flow separation occurs in the gothic arch model, causing a jet flow impinging on the outer wall of the descending aorta. Different arch curvatures also result in different secondary flow patterns around the arch. The separation jet and altered secondary flows are linked with irregular wall shear stress that are potentially associated with systemic hypertension and atherosclerotic lesions.   

Multi-scale patient-specific simulations for evaluation of surgical revascularization techniques in coronary artery bypass surgery

Coronary artery bypass graft (CABG) surgery redirects blood flow around sections of diseased coronary arteries to improve myocardial perfusion using arterial and venous grafts. Cardiothoracic surgeons are often faced with a choice of different revascularization configurations and sizes for saphenous vein grafts (SVGs). However, there is a current lack of understanding surrounding how SVG configuration affects hemodynamics, graft performance and patency. We investigated hemodynamic characteristics in native coronary arteries and vein grafts of varying configurations using computational CABG models. We constructed patient-specific anatomic models and performed virtual surgery by modifying SVG geometry to simulate single, Y, and sequential graft configurations and SVG diameters ranging from 2 mm to 5 mm. Our simulation results demonstrate that coronary artery flows are insensitive to the choice of the SVG revascularization geometry. The wall shear stress of SVG notably increases when the diameter decreases, following an inverse power scaling with diameter, consistent with a Poiseuille flow assumption. For a given diameter, the spatially averaged wall shear stress on the vein graft increases from the single, to the Y, and the sequential graft configuration.   

Quantification of Arterial Flow Using Planar Digital Subtraction Angiography Image Data with Applications to Hepatic Circulation

Digital subtraction angiography (DSA) is an imaging technique used to visualize blood flow in vessels using a contrast medium injected into the bloodstream. Spatiotemporal variations in image pixel intensities correspond to local flow and transport of the contrast medium in a vessel. DSA imaging is routinely used in a variety of procedures involving cerebral circulation, hepatic circulation, and applications in dialysis. Quantitative analysis of image pixel intensity data can provide valuable information on local flow, vascular network, and branching patterns. However, accounting for dynamic contrast agent movement, image noise due to breathing motion, and planar representation of 2D vasculature, can be challenging. In this presentation, we will describe a computational framework for quantification of flow and vasculature information from planar DSA image sequences. The framework is based on a combination of image processing operations and discretized matrix equations derived from pixel intensity values. We will illustrate the framework using example DSA sequences of the human hepatic circulation, and discuss applications of this technique in broader large artery hemodynamics modeling.   

Quantification of the Hemodynamic Environment around Large Arterial Blood Clots

Pathological blood clotting (thrombosis) is a primary cause or complication in stroke and other cardiovascular diseases. Blood flow and transport in the thrombus neighborhood are intimately connected to disease progression and treatment efficacy. Coherent flow structures around a thrombus governs the transport of coagulation proteins and thrombolytic drug. Forces along thrombus boundary can influence clot growth, and drive permeation of biochemical species. The objective of this work is to characterize the complex hemodynamic environment around an arterial thrombus of heterogeneous microcomposition. Specifically in our study, this objective involved understanding the role of key parameters like thrombus shape, microstructure, and wall disease state on local flow phenomena. We use a hybrid particle-continuum based finite element framework to model pulsatile flow in two-dimensional and three-dimensional arterial thrombus configurations. Flow-mediated transport is further characterized using Lagrangian analysis of coherent flow structures and transport patterns. Using parametric simulations, we illustrate the influence of varying clot geometry, clot microstructure, and extent of wall disease states on flow and transport processes in arterial thrombus neighborhood.   

A two-dimensional blood flow model with arbitrary cross sections.

A two-dimensional model for blood flows with arbitrary cross sections will be presented. The model consists of a hyperbolic system of balance laws for conservation of mass and balance of momentum in the axial and angular directions. The main properties of the system will be discussed and a well-balanced central-upwind scheme will be presented. Important features of the model are inherited at the discrete level by the numerical scheme. For instance, the model is equipped with an entropy function and an entropy inequality that can help us choose the physically relevant weak solutions, and a large class of semi-discrete entropy-satisfying numerical schemes will be described. The merits of the scheme will be tested in a variety of scenarios with applications to problems such as stenoses and aneurysms. In particular, numerical results of a simulation using an idealized aorta model will be shown.   

Simulation of arterial hemorrhage: Two-way coupling of a local high fidelity SPH model with a full body lumped parameter physiology model

Full body computational blood flow simulations allow researchers and clinicians to evaluate the impacts of arterial hemorrhage on patient vitals. In this work, we demonstrate a novel method of simulating hemorrhage by coupling a three-dimensional (3D) high fidelity smoothed particle hydrodynamics (SPH) model at the site of damage with a full body lumped parameter model (https://pulse.kitware.com/). While SPH is commonly used for free surface flows, we adapted the method for arterial flow by implementing buffer regions as dynamic inlet and outlet boundary conditions. Our framework uses a two-way coupling system between the SPH and lumped parameter models at the 3D boundary conditions and hemorrhage location. We validated our framework with pipe flow and demonstrated a realistic scenario within Interactive Medical Simulation Toolkit (iMSTK - https://www.imstk.org/), by simulating 3D flow in an image-derived femoral artery with an artificially introduced hemorrhage. Our framework was able to simulate the impacts of hemorrhage on patient vitals within realistic physiological ranges and create high quality SPH visualizations. Our framework that couples local high-fidelity and global low-fidelity models has the potential to enhance the accuracy of physics-based surgical simulations.   

Investigation of endothelial cell adhesion in a biomimetic 3D printed blood vessel scaffold

Human umbilical vein endothelial cell (HUVEC) responses in a biomimetic 3D bioprinted blood vessel model were investigated in vitro under complex (physiological and pathological) flows with the goal of characterizing the conditions driving mechanotransduction. Cylindrical vessel scaffolds (12.7mm ID, 76.2mm L) were fabricated using polylactic acid (PLA) material. The vessel inner walls were cultured with HUVEC (~1.25 X 106 cells mL-1) in a bioreactor system using endothelial cell growth medium (ECM, Cell Applications Inc). A flow loop facility capable of generating steady and physiological, pulsatile flows was decontaminated prior to vessel installation. The ECM was circulated in the flow loop for 2 hours, subjecting the HUVECs to wall shear stresses from the flow forcing. Real-time flow conditions were monitored using catheterized pressure sensors and an ultrasonic flow rate sensor. Post-facto analysis performed on dissected vessels using confocal microscopy indicated the successful adhesion of HUVECs to the vessel walls. These results pave the way for HUVEC proliferation studies in realistic blood vessel constructs and physiological flows by the synergistic coupling of high-fidelity in vitro measurements with post-facto monitoring of cell biochemical response mechanisms.