Session U05: Hemodynamics in Large Vessels

Direct Numerical Simulations of the pulsatile flow in real patients' coronary arteries with bifurcations

This study aims to predict the plaque growth along the arterial walls of coronary arteries. Real patient arterial geometry with volumetric plague composition is obtained using virtual histology-intravascular ultrasound (VH-IVUS) data. Direct Numerical Simulations (DNS) are performed with our in-house CFD code that solves the Navier-stokes and continuity equations using a finite difference scheme coupled with a Runge-Kutta algorithm for the time integration. The flow physics at the bifurcation is studied and the assumption on Murray's hypothesis are tested for a wide range of cases, varying the dimensions of the two daughter branches and their angle. One of his main assumptions is that the pressure is considered to be equal at the outlets downstream of a bifurcation. In a realistic scenario, the pressure at the two ends of an artery bifurcation can be different based on the downstream organs. A correction to the model is proposed based on our numerical results. Results on Leukocyte adhesion are compared with patient data to assess the validity of our proposed model.   

Development and validation of numerical blood flow model providing medical decision support for the management of patients with coronary artery disease

Coronary artery disease and atherosclerotic plaques in the major arteries are one of the leading causes of death worldwide. According to the American Heart Association the cost of cardiovascular disease in ten years time will exceed $1 Trillion. We developed an integrated diagnostic assistance package, which consists of a reduced-order numerical model used to simulate blood flow through a cardiovascular network and corresponding IT interface. Proposed methodology uses combination of Computed Tomography Coronary Angiography measurements, supplemented by computed Fractional Flow Reserve (FFR) – a non-invasive personalized test, which is not only reducing the rate of death, but is also cost-effective compared to stress ECG testing of chest pain patients. The numerical simulation model employs a set of 1-D conservation equations for the pulsating flow with additional modules accounting for the effects of variable length stenosis and effective obstruction, vessels curvature and bifurcations and variations of blood viscosity. Validation of numerical model is carried out against controlled experiments in 3D-printed cardiovascular model and in-vivo measurements. Utilization of in-vivo and in-vitro experimental data allows for a better control and fine-tuning of the model parameters.   

Simulation of a Pro-Atherogenic High-Risk Carotid Artery Bifurcation Geometry

Pathological blood flow-induced shear stress is associated with atherosclerosis, a cardiovascular disease that is a leading cause of deaths in the US. Physiological flow CFD simulations are used to investigate the following wall-shear-stress metrics: time-averaged-wall-shear-stress (TAWSS), oscillatory shear index (OSI) and relative residence time (RRT) in carotid artery bifurcation models. Two carotid artery geometries were adopted from patient-averaged anatomies with different internal, external, and common carotid artery features (ICA, ECA and CCA, respectively). The geometry associated with high risk of disease has a larger ICA angle, asymmetric branching angle and lower ICA/CCA diameter ratio than the low-risk case. Unsteady flow simulations employ a physiological inflow waveform and resistive outflow conditions. Low TAWSS regions, generally associated with plaques, were found at the outer ICA sinus wall for both geometries. The high-risk geometry exhibits flow separation and low WSS with high multi-directionality in the ICA. Furthermore, high OSI and RRT were observed at the ICA and ECA outer walls, downstream of the bifurcation for the high-risk geometry. Results demonstrate the utility of OSI and RRT, to augment classical TAWSS, as indicators of atherosclerosis risk.

    

Parametric Investigation on Stroke Risks from Carotid Artery Disease

Embolic Stroke of Undetermined Source continues to account for a significant proportion of all ischemic strokes with a limited ability to identify stroke etiology. Disambiguating embolic stroke etiology is a critical step in improving treatment efficacy and reducing incidence of recurring stroke events. A prominent source of emboli are the carotid arteries, where buildup of atherosclerotic plaques can generate thromboemboli that move into the cerebral arteries. Currently, it is common notion to presume carotid embolization causes stroke events mainly in middle cerebral artery on the same hemisphere (ipsilateral). There have been reported cases where anterior or posterior strokes, and contralateral strokes (on opposite brain hemispheres) occur, which challenges this notion. The role of proximal collateral flow, the influence of stenosis severity and laterality, and embolic particle size effects – these remain challenging to understand in a patient-specific setting. This inspired our in silico investigation, where we develop a patient-specific hemodynamics and embolus transport model based on a heart-brain arterial network, and conduct a set of parametric investigations on flow and carotid embolus distribution in the Circle of Willis.

    

Understanding Particle Transport In Human Vascular Network Using In Vitro Benchtop Flow Modeling

The transport and distribution of particles in the human vascular network play several major roles in physiological phenomena in health and disease. Tracking particles in vasculature in vivo remains a methodological challenge, while in silico modeling involves key assumptions and limitations regarding particle size and flow environment. Here we discuss the development of a physiologically realistic in vitro benchtop flow-loop to track the transport and distribution of particles across human vasculature. Specifically, we illustrate a study on tracking representative embolic particles through a 3D printed phantom of the human common carotid bifurcation and summarize the trends in resulting particle distribution. We will discuss observations on anatomically representative particles driven at different flow speeds and with different temporal wave forms through: (a) idealized phantom models; and (b) anatomically accurate arterial phantom models; and demonstrate how these factors affect the distribution of embolic particles through the common carotid artery. The in vitro design and results are compared and contrasted against an in silico model for embolus transport across the same phantom models, to assess the functionality and accuracy of our approach.   

Response of a collapsible tube carrying flow with and without stenosis

Arterial stenosis is the constriction in arteries caused by the growth of plaque and can lead to many cardiovascular health issues. A deformable tube with internal constriction and internal flow is a very complicated fluid-structure interaction problem. We experimentally investigate the dynamics of a collapsible latex tube carrying flow and subjected to a constant external pressure of 700 Pa, with a thermocol ball (25 mm diameter) pasted to the internal surface of the flexible wall. Tube response is measured using high-speed photography with two locations of the constriction, namely L/2 and L/4 from the upstream end (L is the length of the collapsible tube). Without constriction, the tube exhibits collapsed states, self-excited milking (periodic/aperiodic) oscillations and distended states, as the Reynolds number increases. With constriction, aperiodic high-amplitude milking oscillations are completely absent. However, high-frequency low-amplitude fluttering is observed at higher Reynolds number when the constriction is at L/2, due to increased tension in the tube wall. Flow visualization shows the presence of strong recirculation upstream of the constriction, indicating significant viscous dissipation and pressure drop across the stenosis, suppressing the milking oscillations.

    

In Vitro Flow Experiments for Aortic Dissection Disease

Aortic dissection is an uncommon but often deadly degeneration and weakening of the aorta. The layers that constitute the aortic vessel wall dissect and delaminate allowing blood to flow in the inter-layer space. This forms a parallel blood conduit known as the false lumen. Acute cases typically present to the ER, and a diagnosis must be made in a timely and accurate manner. Speed and accuracy don't normally go hand in hand. Diagnosis and disease management can benefit greatly from a thorough understanding of the hemodynamics in relation to disease morphology. To this end, multiple in vitro flow models representing patent and non-patent disease geometries are investigated. Particle image velocimetry (PIV) is used to measure the flow behavior in the true lumen and in the false lumen. As a surrogate for diagnostic X-Ray tomography, fluorescent dye is injected into the incoming blood flow and measured using a laser-induced fluorescence (LIF) setup at the dissection site. The results from the two measurement tools provide complimentary information that can greatly inform on disease diagnosis, and aid in improving patient outcomes.   

An In Silico Case Study on Patient-Specific Hemodynamics During Transarterial Radioembolization of Liver Cancer

In 2020, liver cancer was the fourth most prevalent form of cancer worldwide. Transarterial radioembolization (TARE) is a transcatheter procedure in which radioactive particles are delivered to a liver tumor. Pretreatment mapping studies utilize a surrogate particle to observe drug distribution as a function of catheter placement. However, discrepancies in distribution between the pretreatment tumor dose and the actual treatment tumor dose have been attributed to a mismatch in particle morphology and catheter placement between procedures. In this study, we develop an in silico model of a patient hepatic arterial network to understand the fundamental mechanisms behind these discrepancies. This model integrates flow information derived from patient medical imaging to quantify angiogenic effect on blood flow. Hemodynamics is solved using SimVascular, and particle trajectories are calculated using a Lagrangian model previously developed in house. Particle distribution to tumor-feeding vessel indicates size-dependent effects, potentially leading to mismatch between pretreatment and treatment dosing scenarios. With this framework in place, optimal catheter placement and differing injection rates can also be implemented to observe how clinical conditions can alter treatment accuracy. 

    

Transition to turbulence in oscillatory flows in stenosed pipes

We studied oscillatory flow in stenosed cylindrical pipesin axisymmetric and to quantify and characterize the onset of turbulence in zero mean oscillatory flows. Direct numerical simulation (DNS) in various degrees of stenosis, flow frequencies and Reynolds numbers were conducted amounting to about 180 simulations. The canonical studies provide basic insight into the factors that lead to onset and sustainment of turbulence. DNS were conducted using the lattice Boltzmann method (LBM) solver Musubi. Stenosis with area reduction of 75%, 60%, 50% and 25% were studied in both axisymmetric and eccentric configurations. Meshes of up to 2.8 billion cells were created and simulations were conducted  on 300’000 CPU cores of the SuperMUC-NG petascale system in Munich, GERMANY. We found: 1. The flow transitions only in higher degrees of stenosis namely 50%, 60% and 75%, where Re~1800 is the approximate threshold for transition. 2. The flow reversal stabilizes the flow field for all pulsation frequencies and degrees of stenoses. 3. A higher pulsation frequency leads to earlier breakdown of flow – a phenomenon that is seen mostly for lower stenoses degrees. 4. The eccentricity of the stenosis causes higher fluctuations.

    

Multiparametric hemodynamic analyses for carotid artery flow using ultrafast ultrasound flow imaging without a contrast agent

Non-invasively obtaining hemodynamic information on the carotid artery is a significant challenge for medical imaging techniques due to their spatial and temporal limitation. The plane-wave imaging enabled high-precision characterization of complex blood flows obtained at a very high frame rate (>1000 fps). However, quantitative hemodynamic analyses in the carotid artery have not been shown yet. This study provides ultrafast and high-resolved ultrasound flow imaging for a human carotid artery without a contrast agent and quantitative hemodynamic analyses. The plane-wave imaging sequences of the carotid artery were obtained from healthy volunteers by a Vantage system and a customized linear ultrasonic probe at a 1 kHz frame rate. A spatio-temporal singular-value-decomposition filtered blood flow images were processed using the particle image velocimetry. The given flow velocity field estimated the hemodynamic indexes. During the peak systolic phase, reversed flow and vortices were observed in the carotid bulb due to its dilated vascular structure. The presence of a low WSS region mirrored that of the recirculating flow region. These results suggest that the proposed technique shows the good delineation of fast and complex carotid artery flow without a contrast agent.

    

Statistical analysis of symptomatic and asymptomatic patients with plaques clogged carotid artery using CFD

Carotid atherosclerosis (stenosis) is responsible for 10-20% of all strokes. Symptoms of stroke include weakness of extremities or speech problems. Up to 77% of strokes occur without warning symptoms; a critical challenge when selecting asymptomatic patients for carotid surgery. Carotid stenosis is measured using ultrasound but cannot determine which asymptomatic patients will develop stroke. The goal of this research is to use computational fluid dynamics (CFD) to identify patients at risk for future stroke.  

We present a numerical investigation of flow patterns of symptomatic and asymptomatic patients of carotid disease. CT imaging delineates plaque morphology and location, which has been used to construct CFD models for more than 60 patients (half symptomatic and half asymptomatic). An open-source CFD package, OpenFOAM, is used to simulate blood flow and to determine possible symptom-relevant parameters such as wall shear stress (WSS), pressure, turbulent kinetic energy, and velocity magnitude within a 2 cm vicinity of the carotid apex. Preliminary results suggest that WSS in the ICA is higher in the symptomatic group, suggestive of WSS as an indicator of stroke risk. Further analysis will be employed on CFD characteristics between symptomatic and asymptomatic patient groups.

    

Computational Modeling of Transmural Flow and Shear Stress on Smooth Muscle Cells in Human Carotid Artery

Determining the methods to stop the interstitial (transmural) flow-initiated plaque formation depends critically on our understanding of interstitial flow and the shear stress on vascular smooth muscle cells (VSMCs). In this work, a three-layered biphasic model representing the human common carotid artery (CCA) is employed. Numerical methods are used to estimate the fluid flux through the media under lumen blood pressure in FEBio. The material consists of Darcy’s flow with a strain-dependent permeability for the interstitial fluid and the HGO model for the fiber-reinforced solid material with an orthotropic hyperelastic behavior. As the geometry, symmetric thick-walled concentric rings are chosen and are constituted in open ring form to have circumferential residual stress. Shears stress around VSMCs, and the intensity of the fluid passing through the VSMCs are calculated. The effect of aging is investigated by using two different geometries with different physical properties.