Session E28: Biological Fluid Dynamics: Physiological Cardiac Flows

Intraventricular Vector Flow Mapping with Data Fusion and Uncertainty Quantification

Left ventricular (LV) flow patterns contribute to diastolic suction, energetic efficiency, and cardiovascular homeostasis. Echocardiographic vector flow mapping (VFM) allows clinicians to non-invasively quantify LV flow nearly in real time. However, despite recent advances, VFM is limited by assuming planar flow and exclusively relying on uncertainty-prone boundary conditions to uniquely define (i.e., regularize) the velocity field. Here, we present a VFM method rooted in Bayesian inference that allows us to flexibly incorporate multi-modality acquisitions (e.g., echo-PIV and Doppler acquisitions and/or multi-velocity encoding Doppler acquisitions) to achieve robust regularization. The algorithm enforces priors based on flow physics and incorporates the uncertainty of the input data. Of note, we include the LV wall boundary condition uncertainty obtained from a deep learning segmentation method. We apply the new VFM method to two ground-truth synthetic velocity fields, one of which is divergence-free while the other is not. We simulated Doppler acquisitions and contrast agent bubbles on these data to perform echo-PIV and feed the VFM algorithm. We also apply the new VFM algorithm to clinical echo-PIV and Doppler acquisitions, including multi-VENC sequences.   

Not Participating

Left Ventricular Assist Devices (LVADs) are rotary continuous flow pumps that are implanted to support cardiac function for end-stage heart failure. This study investigates hemodynamic patterns in the left ventricle that may promote thrombus formation. Measurements were conducted in a patient-specific thin-walled silicone LV model in a mock circulation loop with refraction-index-matched blood-mimicking fluid. Flow fields were obtained using stereo particle image velocimetry (PIV), and residence time analysis based on washout was conducted using planar laser-induced fluorescence (PLIF). Flow fields and washout were analyzed for varying levels of native contractility with and without LVAD support. Larger regions of stagnation in the LV were measured when ventricle contraction was severely impaired, independent of LVAD support level. High LV contraction reduced thrombogenic flow patterns. Reducing stagnation in the LV is intimately linked to LV contractility, while higher levels of LVAD support play a secondary role. Increased LV contractility also led to large changes in intraventricular pressure, which in turn produced greater heterogeneity of flow patterns throughout the cardiac cycle. This effect is associated with better washout of the LV and reduced risk of clot formation.   

In-vitro coupled left atrioventricular-aortic hemodynamic simulator for systemic circulation

We developed a physiologically accurate in-vitro hydraulic setup that mimics the hemodynamic of coupled atrioventricular-vascular system. This unique experimental simulator has three major components: 1) the aortic system, 2) left ventricle (LV), and 3) left atrium (LA). The aortic system is a human scale artificial aorta with the aortic arch, main branches of the aorta, and cerebral vessels. The LV simulator is composed of a thin-walled transparent silicone ventricle-like sac, installed in a container connected to a programmable piston-in-cylinder pump that generates contraction patterns for the LV. The LA is also a thin-walled transparent silicone sac in a different container which can operate on both passive and active (contraction) modes. LV contractility, LV compliance, preload, cardiac output, heart rate, total arterial resistance, and total arterial compliance are easily controlled in this setup. Our system generates physiologically accurate pressure and flow waveforms in the LV, LA, cerebral vessels, and various locations along the aorta. Our setup can be used for studying LV-arterial coupling in which both LV and aorta affect each other. This system will be very useful in understanding underlying hemodynamics mechanisms of cardiovascular and cerebrovascular diseases.   

Non-Newtonian Patient-specific Numerical Study of Left Atrial Hemodynamic

Ischemic strokes due to cardiac thromboembolism are a leading cause of mortality in patients with atrial fibrillation (AF). During AF, disturbed atrial beating creates stagnant regions where clots can form, typically in the left atrial appendage (LAA). Blood has non-Newtonian rheology due to red blood cell aggregate (rouleaux) formation. Aggregation requires residence times of 1–10 s at shear rates <100 s^-1. These conditions are met in the LAA and, consonantly, rouleaux are observed clinically as spontaneous echocardiographic contrast. Yet, previous CFD studies consider Newtonian rheology. We explore how non-Newtonian rheology affects LA flow in 6 patient-specific anatomies from 4D-CT imaging (3 in sinus rhythm, 3 in AF). We implement a semi-implicit Carreau-Yasuda model incorporating hematocrit (Hct), shear, and residence time to account for rouleaux formation. We consider a wide range of Hct values (37-55) and compare vs. Newtonian simulations on the same patients. We found that non-Newtonian effects can significantly affect LAA blood viscosity even at low Hct, altering hemodynamics LAA stasis predictions by CFD.   

A Doppler-exclusive non-invasive computational diagnostic, monitoring, and predictive framework for patients with mixed and complex cardiovascular diseases

Hemodynamics quantification is critically useful for accurate and early diagnosis, but we still lack proper diagnostic methods for many cardiovascular diseases. Furthermore, as most interventions intend to recover the healthy condition, the ability to monitor and predict hemodynamics following interventions can have significant impacts on saving lives. Predictive methods are rare, enabling prediction of effects of interventions, allowing timely and personalized interventions and helping critical clinical decision making about life-threatening risks based on quantitative data. We developed an innovative non-invasive Doppler-based diagnostic, monitoring and predictive tool for patient with complex cardiovascular diseases enabling quantifying both local and global hemodynamics (called myCardiac). myCardiac also predicts the breakdown of the effects of each disease constituents on the heart function. Presently, neither of these can be obtained noninvasively in patients and when invasive procedures are undertaken, the collected metrics cannot be by any means as complete as the ones myCardiac provides. myCardiac purposefully uses a limited number of noninvasive input parameters all of which can be measured using Doppler echocardiography and sphygmomanometer. myCardiac was validated against clinical cardiac catheterization, Magnetic resonance imaging and Doppler echocardiography measurements.   

The effect of timing of the LVAD speed modulation on intraventricular flow patterns. An in-vitro PIV study.

In the past decades, LVADs have become a common therapy for advanced heart failure patients. Despite a lower incidence of in-pump thrombosis, stroke remains the main complication. We hypothesize that the intraventricular flow patterns modified by the LVAD lead to stagnation, platelet activation and thrombogenecity. Pulsatility modes, periodically modulating the LVAD speed, aim to break down stagnation areas in the pump and the left ventricle (LV). However, their effect on flow patterns and risk of stroke is still under investigation. This study investigates the effect on flow patterns of the synchronicity of the pulsatility mode with the LV contraction cycle. 3C-2D velocity fields were obtained from PIV measurements performed in an LV silicone model implanted with an LVAD.  Different delays between the onset of the modulation and peak systole were experimented. This speed modulation showed a strong effect on LVAD flow rate, with lower minimums when the modulation happened during LV relaxation. In this condition, an increased stagnation was also observed from the velocity fields. These results show that the effect of the speed modulation is strongly dependent on native LV contractility, highlighting the need for physiological and LVAD parameters joint optimization to reduce thrombosis.   

On the correlation between Eulerian and Lagrangian Hemostasis Indices in the Left Atrium

Eulerian indices derived from wall-shear stress (WSS) and kinetic energy (KE) are often used to study hemostasis in the cardiac chambers. The calculation of these indices is convenient as it does not require integrating over multiple heartbeats. The averaged WSS, oscillatory shear index, relative residence time (RRT), and endothelial cell activation potential are among the most common. Also, KE below a certain threshold is used to indicate blood stasis. Most of these indices were conceptualized for tube-shaped vessels where low, oscillatory WSS and KE reveal disturbed-flow regions of recirculation and stasis. However, flow in the left atrium (LA) is highly 3D, and the correlation between Lagrangian and Eulerian indices has not yet been analyzed, let alone established. Here, we examine the correlation between Eulerian and Lagrangian indices in the LA, focusing on the left atrial appendage (LAA), where clots form preferentially. Our results show a low to moderate correlation between Eulerian indices and their Lagrangian counterparts (e.g., RRT and actual residence time), suggesting that care must be exercised when interpreting Eulerian indices in the LA and the LAA.   

Pulmonary vein flow split affects left atrial appendage stasis in patient-specific CFD simulations

Blood flow inside the left atrium (LA) is associated with severe cardiovascular events like embolic stroke. Patient-specific CFD simulations predict LA hemodynamics and blood stasis in the left atrial appendage (LAA), the preferred thrombosis site. In these simulations, patient-specific LA geometry and transmitral flow rates can be extracted from 4D anatomical images. However, the flow rates through the pulmonary vein (PV) inlets are a source of uncertainty because they need to be modeled. We studied how the flow split between the left and right PVs affects LA blood flow in 14 patients (11 in sinus rhythm, 3 in AF) using an in-house immersed boundary solver. We considered even flow split as customary in the literature as well as the more physiological 45-55% and 40-60% splits. We found that PV flow split affects the LA body's mean flow pattern qualitatively and the LAA residence time (RT) quantitatively. However, the LAA RT dependence on PV flow split was comparable to its beat-to-beat variability, suggesting that the PV flow split might not be a critical source of uncertainty as long as this parameter is kept within the general physiological range.   

GPU accelerated simulations of the whole heart open new research paths

The predicting capabilities of cardiovascular computational models depend on the accurate solution of the hemodynamics, the realistic characterization of the hyperelastic and electric tissue properties along with the correct description of their interaction. The resulting fluid–structure–electrophysiology interaction (FSEI) thus requires an immense computational power, usually available only in large supercomputing centers.      

Our group has made an effort in porting the original CPU code to CUDA and the GPU–accelerated FSEI algorithm can now tackle complex simulations of the whole cardiac dynamics - including the four chambers with the valves, connected with the thoracic aorta, the pulmonary veins/arteries, the inferior/superior   venae cavae - within a few hours and desktop workstations, thus strongly reducing the time–to–solution and the need for sophisticated infrastructures. This approach is particularly appealing if the tool has to support medical decisions which require quick solutions possibly obtained by local computational resources.  

The accuracy of the numerical tool and the unprecedented computational speed-up further open the way for running simulation campaigns to study the response to pathologies of synthetic populations or for virtual testing medical devices through an uncertainty quantification analysis.