Session L19: Minisymposium II: Intra-Ventricular Fluid Mechanics in Heart Failure

A Risk-based Framework for the use of Computational Modeling of Medical Devices to Support Regulatory Decisions

Computational models and simulations (CM&S) are used in multiple stages within the Total Product Life Cycle (discovery and ideation, regulatory decision, product launch, and post market monitoring) of medical devices.  The Food and Drug Administration (FDA) has been proactive in establishing a risk-based framework for the use of CM&S to support regulatory decisions utilizing research from the field of Verification, Validation, and Uncertainty Quantification (VVUQ). This framework includes an American Society of Mechanical Engineering Standard (ASME V&V 40) that includes code and calculation verification as well as validation of CM&S for medical devices. This talk will present: 1) an overview of this risk-based framework; 2) examples of the use of CM&S in regulatory; 3) a summary of an in-depth example of the execution of a VVUQ plan for a computational model of heart-device interactions in the form a benchtop mock-circulatory loop with a left-ventricular assist device (SDSU-CS-LVAD); and 4) a comparison of the intraventricular flow dynamics of the SDSU-CS-LVAD and clinical results. Overall, this presentation should provide an overview of a concrete framework that is enabling innovative advances of CM&S in the TPLC of medical devices.   

Intra-ventricular blood flow dynamics for diagnosis and prognosis - How far are we from clinical translation?

Blood flow dynamics within the heart are known to play a major role in the quantitative assessment of cardiac function. Investigations of the intra-ventricular flow structures have identified various biomarkers with the potential to outperform conventional clinical characterizations such as the ejection fraction. A plethora of studies on the in-silico, in-vitro and in-vivo scale have emphasized the versatility of intra-ventricular flow analyses. Particularly, computational approaches have the potential to facilitate personalized cardiovascular therapies. Still, clinical translation is a major obstacle which needs to be tackled and there is a large gap between academic research and application.

 

Within this talk I am going to present two distinct examples of our work using computational modelling for the description of intra-ventricular flows. The first example investigates the role of inflow cannula orientation on the thrombosis risk during left ventricular assist device support. The second example focusses on the patient-specific analysis of blood flow in the heart of neonates and children with congenital heart diseases using real-time three-dimensional echocardiography.

Although the research questions are very different, similar problems including model development, implementation, and result interpretation were identified. At last, some solution approaches to these problems will be presented. The talk concludes with open questions towards a more community-based and integrated modelling of intra-ventricular flows.   

Direct-numerical simulations of cardiac flows – what have we learnt?

In this talk, we will provide an overview of our group's endeavors in utilizing direct numerical simulations (DNSs) to analyze the functionality and dysfunctions observed in left ventricular flows. The Reynolds numbers of the flow in the left-ventricle are in the range of about 4000, which make them perfectly suitable for DNS based analysis. Our simulations have been carried out using our in-house sharp-interface immersed boundary solver, ViCar3D. The presentation will encompass the following investigations: (a) Examination of mixing and washout phenomena in both healthy and diseased ventricles, shedding light on their impact. (b) Exploration of the role played by intraventricular flows in the dissipation of viscous energy, uncovering their significance. (c) Evaluation of the potential influence of trabeculae on left ventricular hemodynamics, offering insights into their involvement. (d) Introduction of a novel metric developed to assess the risk of LV thrombosis in individuals with cardiac diseases, facilitating better risk management. (e) Study of the dynamics exhibited by both natural and bioprosthetic aortic valves, emphasizing their influence on aortic hemodynamics. Through this talk, we aim to highlight significant findings from our work and their implications for the advancement of cardiovascular health.   

Intraventricular Fluid Mechanics in Advanced Heart Failure patients implanted with Left Ventricular Assist Devices

LVAD improvements have reduced pump-related complications and open a way for long-term treatment of heart failure that can compete with heart transplantation. Stroke of unknown origins has remained unacceptably high, bringing the focus on intraventricular hemodynamics, now that pump hemodynamics are milder and better understood. Our labs have studied the fluid mechanics in ventricles with advanced heart failure, implanted with an LVAD at the apex. The role of recirculation near the apical pocket created by the LVAD inflow cannula, low velocities in the LV outflow track due to closed aortic valve, pump flow rate dependency on preload dependencies of different LVADs in the market, and the speed modulation schemes recently introduced by manufacturers to mimic pulsatility in these continuous centrifugal pumps, have been studied by a combination of computational simulations and in-vitro experiments with time-resolved stereo PIV. We describe the general features of flow in a ventricle that doesn’t significantly change shape or volume (consistent with advanced heart failure implanted with a continuous flow pump), and characterize the recirculation regions, quantifying stasis to perform comparisons that could help understand the impact of hemodynamics on thrombosis. Different Eulerian metrics are defined for the whole ventricle, moving away from the paradigm of Wall Shear Stress, and a novel Lagrangian strategy to quantify platelet trajectories and their exposure to stasis inside the ventricle is presented.   

Computational modeling of intra-ventricular hemodynamics and left ventricle thrombosis

Left ventricular thrombosis (LVT) poses a significant risk factor for systemic thromboembolism after myocardial infarction. Post-infarction remodeling can permanently affect the ventricular wall's contractility leading to ventricular dysfunction. Abnormal intra-ventricular hemodynamics resulting from ventricular dysfunction combined with increased hypercoagulability can induce a procoagulant state in the ventricle, increasing the risk for LVT. Despite advancements in cardiovascular medicine, diagnosing, preventing, and treating LVT remains challenging. Computational fluid dynamic simulations offer a promising avenue to investigate intra-ventricular hemodynamics and LVT. We employ fluid-structure interaction methods informed by cine MRI data to compute intra-ventricular flows. Image registration allows us to capture ventricular wall displacement enabling the dynamic simulation of the ventricular flows. Based on the computed flow field, we perform reaction-advection-diffusion simulations to quantify the activation and transport of platelets and agonists. We simulate various post-infarction remodeling scenarios to examine the impact of ventricle infarction size and location on intra-ventricular hemodynamics and transport and their role in LVT development. The insights obtained from these simulations may help improve the early detection and management of LVT.   

Reduced Order Models of Intraventricular Flow for Interpretable Phenotypic Classification of Heart Failure Patients

Multi-dimensional, high-resolution cardiac flow imaging is increasingly available in the clinical setting, but the adoption of flow-related metrics into clinical decision support systems is lagging behind this technological growth. A bottleneck in this adoption is deriving explainable metrics that identify disease states. Heart failure (HF) is no exception. 

Our hypotheses to address this challenge are: 1) Left ventricular (LV) flow patterns contain enough data to phenotype patients with different HF subtypes; 2) Applying machine learning (ML) to data-driven reduced-order models (ROMs) of LV flow can discover simple, interpretable metrics for LV flow phenotyping. We analyzed >200 color-Doppler echocardiograms of healthy volunteers (VOL) and patients with dilated (DCM) or hypertrophic (HCM) cardiomyopathy. We applied proper orthogonal decomposition (POD) to build ROMs of LV flow. We tested classifiers using purposely simple ML analyses of the ROMs and requiring different supervision levels. We ran vector flow mapping to visualize 2D flow patterns (e.g., diastolic vortex) and interpret the ML results. To compare data across patients and blind classifiers to chamber geometry, all Doppler sequences were mapped into a common cubic domain using anatomical landmarks and events of the cardiac cycle.

The POD-based ROMs stably represents each cohort's flow patterns through 10-fold cross-validation. The 1st POD mode captures >80% of the flow kinetic energy (KE) and represents the LV filling/emptying jets. The 2nd mode represents the diastolic vortex; its KE contribution ranges from 8% (HCM) to 39% (DCM). Semi-unsupervised classification using patient-specific ROMs reveals that the KE ratio of these two modes, the vortex-to-jet (V2J) ratio, is a simple, interpretable metric that accurately clusters DCM, HCM, and VOL patients. Receiver operating characteristic curves using V2J as threshold have areas under the curve equal to 0.81, 0.91, & 0.95 for the classification problems HCM vs. VOL, DCM vs. VOL, & DCM vs. HCM, respectively.

The favorable classification accuracy, obtained with overtly uninvolved ML methods and using readily available input data, opens possibilities to amplify the role of cardiac flow imaging in the early detection and management of HF and the development of personalized treatment strategies.