Feasibility of coronary blood flow simulations using mid-fidelity numeric and geometric models.

The fractional flow reserve index (FFR) is currently used as a gold standard to quantify coronary stenosis's functional relevance. Due to its highly invasive nature, the development of noninvasive surrogates based on simulations has drawn much attention in recent years, emphasizing efficient strategies that enable translational research. The focus of this work is twofold. First, to assess the feasibility of using a mid-fidelity numerical strategy (transversally enriched pipe element method, TEPEM), placed between low- and high-fidelity models, for the estimation of flow-related quantities, such as FFR and wall shear stress (WSS). Low-fidelity models, as zero- or one-dimensional models, are computationally inexpensive but in detriment of poorer spatially detailed predictions. On the other hand, high-fidelity models, such as classical three-dimensional numerical approximations, can provide detailed predictions but their transition to clinical application is prohibitive due to high computational costs. As a second goal, we quantify the impact of the length of lateral branches in the blood flow through the interrogated vessel of interest to further reduce the computational burden. Both studies are addressed considering a cohort of 17 coronary geometries. A total of 20 locations were selected to estimate the FFR index for a wide range of Coronary Flow Reserve (CFR) scenarios. Numerical results suggest that the mid-fidelity TEPEM model is a reliable approach for the efficient estimation of the FFR index and WSS, with an error in the order of [Formula: see text] and [Formula: see text], respectively, when compared to the high-fidelity prediction. Moreover, such mid-fidelity models require much less computational resources, in compliance with infrastructure frequently available in the clinic, by achieving a speedup between 30 and 60 times compared to a conventional finite element approach. Also, we show that shortening peripheral branches does not introduce considerable perturbations either in the flow patterns, in the wall shear stress, or the pressure drop. Comparing the different geometric models, the error in the estimation of FFR index and WSS is reduced to less than [Formula: see text] and [Formula: see text], respectively.

Parallel generation of extensive vascular networks with application to an archetypal human kidney model.

Given the relevance of the inextricable coupling between microcirculation and physiology, and the relation to organ function and disease progression, the construction of synthetic vascular networks for mathematical modelling and computer simulation is becoming an increasingly broad field of research. Building vascular networks that mimic in vivo morphometry is feasible through algorithms such as constrained constructive optimization (CCO) and variations. Nevertheless, these methods are limited by the maximum number of vessels to be generated due to the whole network update required at each vessel addition. In this work, we propose a CCO-based approach endowed with a domain decomposition strategy to concurrently create vascular networks. The performance of this approach is evaluated by analysing the agreement with the sequentially generated networks and studying the scalability when building vascular networks up to 200 000 vascular segments. Finally, we apply our method to vascularize a highly complex geometry corresponding to the cortex of a prototypical human kidney. The technique presented in this work enables the automatic generation of extensive vascular networks, removing the limitation from previous works. Thus, we can extend vascular networks (e.g. obtained from medical images) to pre-arteriolar level, yielding patient-specific whole-organ vascular models with an unprecedented level of detail.

Author Correction: Comparison of 1D and 3D Models for the Estimation of Fractional Flow Reserve.

A correction to this article has been published and is linked from the HTML and PDF versions of this paper. The error has been fixed in the paper.

Comparison of 1D and 3D Models for the Estimation of Fractional Flow Reserve.

In this work we propose to validate the predictive capabilities of one-dimensional (1D) blood flow models with full three-dimensional (3D) models in the context of patient-specific coronary hemodynamics in hyperemic conditions. Such conditions mimic the state of coronary circulation during the acquisition of the Fractional Flow Reserve (FFR) index. Demonstrating that 1D models accurately reproduce FFR estimates obtained with 3D models has implications in the approach to computationally estimate FFR. To this end, a sample of 20 patients was employed from which 29 3D geometries of arterial trees were constructed, 9 obtained from coronary computed tomography angiography (CCTA) and 20 from intra-vascular ultrasound (IVUS). For each 3D arterial model, a 1D counterpart was generated. The same outflow and inlet pressure boundary conditions were applied to both (3D and 1D) models. In the 1D setting, pressure losses at stenoses and bifurcations were accounted for through specific lumped models. Comparisons between 1D models (FFR1D) and 3D models (FFR3D) were performed in terms of predicted FFR value. Compared to FFR3D, FFR1D resulted with a difference of 0.00 ± 0.03 and overall predictive capability AUC, Acc, Spe, Sen, PPV and NPV of 0.97, 0.98, 0.90, 0.99, 0.82, and 0.99, with an FFR threshold of 0.8. We conclude that inexpensive FFR1D simulations can be reliably used as a surrogate of demanding FFR3D computations.

A computational framework to characterize and compare the geometry of coronary networks.

This work presents a computational framework to perform a systematic and comprehensive assessment of the morphometry of coronary arteries from in vivo medical images. The methodology embraces image segmentation, arterial vessel representation, characterization and comparison, data storage, and finally analysis. Validation is performed using a sample of 48 patients. Data mining of morphometric information of several coronary arteries is presented. Results agree to medical reports in terms of basic geometric and anatomical variables. Concerning geometric descriptors, inter-artery and intra-artery correlations are studied. Data reported here can be useful for the construction and setup of blood flow models of the coronary circulation. Finally, as an application example, similarity criterion to assess vasculature likelihood based on geometric features is presented and used to test geometric similarity among sibling patients. Results indicate that likelihood, measured through geometric descriptors, is stronger between siblings compared with non-relative patients. Copyright © 2016 John Wiley & Sons, Ltd.

A head-to-head comparison between CT- and IVUS-derived coronary blood flow models.

The goal of this work is to compare coronary hemodynamics as predicted by computational blood flow models derived from two imaging modalities: coronary computed tomography angiography (CCTA) and intravascular ultrasound integrated with angiography (IVUS). Criteria to define boundary conditions are proposed to overcome the dissimilar anatomical definition delivered by both modalities. The strategy to define boundary conditions is novel in the present context, and naturally accounts for the flow redistribution induced by the resistance of coronary vessels. Hyperemic conditions are assumed to assess model predictions under stressed hemodynamic environments similar to those encountered in Fractional Flow Reserve (FFR) calculations. As results, it was found that CCTA models predict larger pressure drops, higher average blood velocity and smaller FFR. Concerning the flow rate at distal locations in the major vessels of interest, it was found that CCTA predicted smaller flow than IVUS, which is a consequence of a larger sensitivity of CCTA models to coronary steal phenomena. Comparisons to in-vivo measurements of FFR are shown.

Computational modeling of blood flow steal phenomena caused by subclavian stenoses.

The study of steal mechanisms caused by vessel obstructions is of the utmost importance to gain understanding about their pathophysiology, as well as to improve diagnosis and management procedures. The goal of this work is to perform a computational study to gain insight into the hemodynamic forces that drive blood flow steal mechanisms caused by subclavian artery stenosis. Such condition triggers a flow disorder known as subclavian steal. When this occurs in patients with internal thoracic artery anastomosed to the coronary vessels, the phenomenon includes a coronary-subclavian steal. True steal can exist in cases of increased arm blood flow, potentially resulting in neurological complications and, in the case of coronary-subclavian steal, graft function failure. In this context, the anatomically detailed arterial network (ADAN) model is employed to simulate subclavian steal and coronary-subclavian steal phenomena. Model results are verified by comparison with published data. It is concluded that this kind of model allows us to effectively address complex hemomdynamic phenomena occurring in clinical practice. More specifically, in the studied conditions it is observed that a regional brain steal occurs, primarily affecting the posterior circulation, not fully compensated by the anterior circulation. In the case of patients with coronary revascularization, it is concluded that there is a large variability in graft hemodynamic environments, which physically explain both the success of the procedure in cases of severe occlusive disease, and the reason for graft dysfunction in mildly stenosed left anterior descending coronary artery, due to alternating graft flow waveform signatures.

A computational approach to generate concurrent arterial networks in vascular territories.

In this work, a computational procedure is proposed to vascularize anatomical regions supplied by many inflow sites. The proposed methodology creates a partition of the territory to be vascularized into nonoverlapping subdomains that are independently supplied by the so-called perforator arteries (inflow sites). Then, in each subdomain, the constrained constructive optimization method is used to generate a network of vessels. The identification of subdomains in a certain vascular territory perfused by many perforator arteries turns out to be a fundamental problem towards understanding the morphological conformation of peripheral beds in the cardiovascular system. The methodology is assessed through two academic examples showing the main structural features of the so-defined vascular territory partition and the corresponding arterial networks. In addition, the vascularization of a three-dimensional sheet-like tissue is presented with potential application in flap planning and design.

A dimensionally-heterogeneous closed-loop model for the cardiovascular system and its applications.

In the present work a computational model of the entire cardiovascular system is developed using heterogeneous mathematical representations. This model integrates different levels of detail for the blood circulation. The arterial tree is described by a one dimensional model in order to simulate the wave propagation phenomena that take place at the larger arterial vessels. The inflow and outflow locations of this 1D model are coupled with lumped parameter descriptions of the remainder part of the circulatory system, closing the loop. The four cardiac valves are considered using a valve model which allows for stenoses and regurgitation phenomena. In addition, full 3D geometrical models of arterial districts are embedded in this closed-loop circuit to model the local blood flow in specific vessels. This kind of detailed closed-loop network for the cardiovascular system allows hemodynamics analyses of patient-specific arterial district, delivering naturally the appropriate boundary conditions for different cardiovascular scenarios. An example of application involving the effect of aortic insufficiency on the local hemodynamics of a cerebral aneurism is provided as a motivation to reproduce, through numerical simulation, the hemodynamic environment in patients suffering from infective endocarditis and mycotic aneurisms. The need for incorporating homeostatic control mechanisms is also discussed in view of the large sensitivity observed in the results, noting that this kind of integrative modeling allows such incorporation.

HeMoLab--Hemodynamics Modelling Laboratory: an application for modelling the human cardiovascular system.

In this work we present HeMoLab (Hemodynamics Modeling Laboratory), a computational environment for modeling the Human Cardiovascular System. Its integrates novel computational tools, running from medical image processing to numerical simulation and visualization. As a simulation tool, it allows to accommodate complex physiological and/or pathophysiological (virtual) scenarios aimed to retrieve detailed information from the numerical computations. Such application makes possible to speed up research in the study and analysis of the cardiovascular system and, to provide a virtual laboratory for medical training and education, and specialized Human Resources development. In order to demonstrate the modeling and simulation capabilities of HeMoLab some cases of use are presented.

Identification of vascular territory resistances in one-dimensional hemodynamics simulations.

The present work deals with the parameter identification problem in outflow models used in one-dimensional simulations of arterial blood flow. Specifically, the resistive elements that define the models used to account for the blood supply to the vascular territories in arterial networks are computed by solving a system of non-linear equations using a Broyden method. This strategy is employed to compute the terminal parameters in the vascular territories of an anatomically detailed model of the arm comprising 67 arterial segments and 16 vascular territories. A comparison with a simple analytical approach, in terms of vascular territory resistances, average blood flows and time-dependent hemodynamic quantities, is performed. Also, a sensitivity analysis is presented to assess the performance of this new approach in normal and abnormal cardiovascular scenarios. This identification procedure allows to correctly set up hemodynamics simulations in highly detailed arterial networks making possible to gain insight in the aspects related to the blood circulation in arterial vessels.

Tuning a lattice-Boltzmann model for applications in computational hemodynamics.

The interest in lattice-Boltzmann models in the computational hemodynamics realm has increased in recent years. In this context, the correct choice of numerical parameters for the appropriate simulation of blood flows in major arteries is a crucial aspect. For this reason, we present three parameter-tuning strategies that allow us to reproduce correctly the pulsatile time-dependent flow of an incompressible fluid under physiological regimes. These strategies are studied for a model based on a single-relaxation-time approach in combination with second order boundary conditions for both velocity and pressure, and proper equilibrium distributions that take care of the incompressible behavior exhibited by the fluid. The implementation is validated with the three-dimensional Womersley flow benchmark. As well, the simulation of blood flows in a curved artery, in an anastomosed vessel, in a patient specific vertebral artery and in an aneurysmal region are presented in order to show how the method and the setting of the numerical parameters are applied to different realistic hemodynamics problems.

On the integration of the baroreflex control mechanism in a heterogeneous model of the cardiovascular system.

The aim of the present work is to describe the integration of a mathematical model for the baroreceptor reflex mechanism to provide regulatory action into a dimensionally heterogeneous (3D-1D-0D) closed-loop model of the cardiovascular system. Such heterogeneous model comprises a 1D description of the arterial tree, a 0D network for the venous, cardiac and pulmonary circulations and 3D patient-specific geometries for vascular districts of interest. Thus, the detailed topological description of the arterial network allows us to perform vasomotor control actions in a differentiated way, while gaining insight about the effects of the baroreflex regulation over hemodynamic quantities of interest throughout the entire network. Two examples of application are presented. Firstly, we simulate the hemorrhage in the abdominal aorta artery and analyze the action of the baroreflex over the system. Secondly, the self-regulated closed-loop model is applied to study the influence of the control action in the hemodynamic environment that determines the blood flow pattern in a cerebral aneurism in the presence of a regurgitating aortic valve.

On the potentialities of 3D-1D coupled models in hemodynamics simulations.

This work comprises a step towards the quantitative and qualitative analysis of coupled local and global hemodynamics phenomena in the arterial system. The aim of this work is to present some numerical examples to put in evidence the importance of the use of 3D-1D coupled models in hemodynamics problems when carrying out simulations of rather complex situations. Accordingly, some cases for which classical modeling cannot be applied are identified and solved. The results obtained here allow us to assess some interrelations between local pointwise quantities (defined at the level of the 3D model) and global mean quantities (defined at the level of the 1D model).

Capnocytophaga.

Capnocytophaga has been found to cause disease in the normal and the immunocompromised patient. Most often this organism is an opportunist that is difficult to isolate but easy to identify and a unique reminder that uncommon infection may respond to common antibiotics.