Damien Coisne

Computational Approach to Estimating the Effects of Blood Properties on Changes in Intra-stent Flow

In this study various blood rheological assumptions are numerically investigated for the hemodynamic properties of intra-stent flow. Non-newtonian blood properties have never been implemented in blood coronary stented flow investigation, although its effects appear essential for a correct estimation and distribution of wall shear stress (WSS) exerted by the fluid on the internal vessel surface. Our numerical model is based on a full 3D stent mesh. Rigid wall and stationary inflow conditions are applied. Newtonian behavior, non-newtonian model based on Carreau-Yasuda relation and a characteristic newtonian value defined with flow representative parameters are introduced in this research. Non-newtonian flow generates an alteration of near wall viscosity norms compared to newtonian. Maximal WSS values are located in the center part of stent pattern structure and minimal values are focused on the proximal stent wire surface. A flow rate increase emphasizes fluid perturbations, and generates a WSS rise except for interstrut area. Nevertheless, a local quantitative analysis discloses an underestimation of WSS for modelisation using a newtonian blood flow, with clinical consequence of overestimate restenosis risk area. Characteristic viscosity introduction appears to present a useful option compared to rheological modelisation based on experimental data, with computer time gain and relevant results for quantitative and qualitative WSS determination.

Tissue Doppler echocardiographic quantification. Comparison to coronary angiography results in Acute Coronary Syndrome patients

BackgroundMultiples indices have been described using tissue Doppler imaging (DTI) capabilities. The aim of this study was to assess the capability of one or several regional DTI parameters in separating control from ischemic myocardium.MethodsTwenty-eight patients with acute myocardial infarction were imaged within 24-hour following an emergent coronary angioplasty. Seventeen controls without any coronary artery or myocardial disease were also explored. Global and regional left ventricular functions were assessed. High frame rate color DTI cineloop recordings were made in apical 4 and 2-chamber for subsequent analysis. Peak velocity during isovolumic contraction time (IVC), ejection time, isovolumic relaxation (IVR) and filling time were measured at the mitral annulus and the basal, mid and apical segments of each of the walls studied as well as peak systolic displacement and peak of strain.ResultsDTI-analysis enabled us to discriminate between the 3 populations (controls, inferior and anterior AMI). Even in non-ischemic segments, velocities and displacements were reduced in the 2 AMI populations. Peak systolic displacement was the best parameter to discriminate controls from AMI groups (wall by wall, p was systematically < 0.01). The combination IVC + and IVR< 1 discriminated ischemic from non-ischemic segments with 82% sensitivity and 85% specificity.ConclusionDTI-analysis appears to be valuable in ischemic heart disease assessment. Its clinical impact remains to be established. However this simple index might really help in intensive care unit routine practice.

2-D left ventricular flow estimation by combining speckle tracking with Navier-Stokes-based regularization: an in silico, in vitro and in vivo study.

Despite the availability of multiple ultrasound approaches to left ventricular (LV) flow characterization in two dimensions, this technique remains in its childhood and further developments seem warranted. This article describes a new methodology for tracking the 2-D LV flow field based on ultrasound data. Hereto, a standard speckle tracking algorithm was modified by using a dynamic kernel embedding Navier-Stokes-based regularization in an iterative manner. The performance of the proposed approach was first quantified in synthetic ultrasound data based on a computational fluid dynamics model of LV flow. Next, an experimental flow phantom setup mimicking the normal human heart was used for experimental validation by employing simultaneous optical particle image velocimetry as a standard reference technique. Finally, the applicability of the approach was tested in a clinical setting. On the basis of the simulated data, pointwise evaluation of the estimated velocity vectors correlated well (mean r = 0.84) with the computational fluid dynamics measurement. During the filling period of the left ventricle, the properties of the main vortex obtained from the proposed method were also measured, and their correlations with the reference measurement were also calculated (radius, r = 0.96; circulation, r = 0.85; weighted center, r = 0.81). In vitro results at 60 bpm during one cardiac cycle confirmed that the algorithm properly measures typical characteristics of the vortex (radius, r = 0.60; circulation, r = 0.81; weighted center, r = 0.92). Preliminary qualitative results on clinical data revealed physiologic flow fields.

Blood flow in stented coronary artery : numerical fluid dynamics analysis

Recent generalization of stent implantation in interventional cardiology require full understanding of blood flow cartography. Interdepency between fluid stresses and in vivo cells covering lumen artery are regularly accused to be one of the instigator of neointimal proliferation (thickening of the inner layer of blood vessels) and mid-term restenosis. This study purpose to numericaly investigate the three dimensional flow in vicinity of an endoprothesis. We used a finite element method to simulate a steady flow of non-Newtonian fluid in a coronary artery using a rigid wall approximation. Results on the velocities, wall shear stress and wall shear stress gradients are presented. Theses simulations allow identification of stagnation site and low wall shear stress area that may be prone to clot formation and neointimal hyperplasia. Intra stent flow knowledge can potentially contribute to optimization of prothesis design and decreasing second intervention rate.

Apical left ventricular myocardial dysfunction is an early feature of cardiac involvement in myotonic dystrophy type 1.

Left ventricular (LV) dysfunction is a major prognostic determinant in myotonic dystrophy type 1 (DM1). Therefore, markers of early‐stage LV impairment may be useful. The aim of this study was to evaluate 2D echocardiographic LV strain in a cohort of DM1 patients with preserved left ventricular ejection fraction (LVEF) and to compare the results with matched controls.

Left ventricular non-compaction and idiopathic dilated cardiomyopathy: the significant diagnostic value of longitudinal strain

Left ventricular non-compaction (LV NC) is characterized by abnormal trabeculations that are mainly at the LV apex. Distinction between LV NC and non-specific dilated cardiomyopathies (DCMs) remains often challenging. We sought to find additive tools comparing the longitudinal strain characteristics of LVNC versus idiopathic DCM in a cohort of patients. 48 cases of LVNC (derivation cohort) were compared with 45 cases of DCM. Global and regional multi-layer (sub-endocardial, mid-wall, and sub-epicardial) LV longitudinal strain analysis was performed. Results were compared to define the best tool for distinguishing LVNC from DCM. A validation cohort (41 LVNC patients) was then used to assess the performance of the proposed diagnostic tools. In the derivation cohort, longitudinal deformation (strain) was greater in LVNC than in DCM patients. Longitudinal shortening was greater in the non-compacted segments than in the compacted ones. A mid-wall strain base-apex gradient had 88.4 % sensitivity and 66.7 % specificity in distinguishing LVNC from DCM (AUC = 0.83; cut-off of −23 or |0.23|%). In a multivariable model, the base-apex mid-wall gradient in an apical 4-chamber view was the only independent echocardiographic criteria (OR = 0.76, CI 95 % [0.66; 0.90], p = 0.0010) allowing the distinction between LVNC and DCM. In the validation cohort, the base-apex mid-wall gradient of strain had 88.4 % sensitivity, 85.7 % negative predictive values for the diagnosis of LVNC. Longitudinal strain, especially the base-apex longitudinal gradient of strain, appears as an additive valuable tool for distinguishing LVNC from DCM.

2D intracardiac flow estimation by combining speckle tracking with navier-stokes based regularization: a study with dynamic kernels

Echocardiographic transducers record two-dimensional (2D) datasets in a sector reference after which a scan-conversion is applied to obtain the images in Cartesian coordinates. To assess left ventricular(LV) flow dynamics by a low dose contrast injection, we recently developed a 2D tracking methodology by combining speckle tracking (ST) with Navier-Stokes based regularization and it has been tested in synthetic ultrasound datasets prior to the scan conversion. However, in clinical settings the estimation becomes challenging due to the inhomogeneous image patterns which are inherently introduced by scan-conversion and are more likely to be locally strengthened by non homogeneous bubble seeding and high velocity gradient. To better deal with that, the aim of this study was hereby to modify the previous method by using a dynamic tracking kernel size. Its performance was first quantified in synthetic scan-converted ultrasound data based on a computational fluid dynamics model of LV flow. The applicability of the approach was tested in an experimental phantom setup with pulsed flow that mimics the normal human heart and simultaneously allows for optical particle image velocimetry as a standard reference technique. Both qualitative and quantitative comparison of the estimated flow fields and reference measurements showed that the modified methodology can correctly characterize the flow field properties and is promising to offer new insights into the flow dynamics inside the left ventricle.

Experimental study of blood laminar flow through a stented artery

The objective of this research is to study the blood flow close to the wall of a stented artery. Indeed, previous works have showed that the restenosis phenomenon is induced by the endothelial cells stimulation due to the wall shear stress values. The coronary angioplasty is responsible of wall shear stress modification, mainly between the stent struts, at the inlet and the outlet of the endoprothesis. That is why, to study the flow disturbances through a stented section, we built an in vitro model reproducing the struts shapes of a marketed endoprothesis. The experimental artery is composed of a see-through square section vein, which reproduce the struts design with a magnitude of 100. A programmable pump provide a steady or a pulsatile flow. By using the velocimetry per imagery of particle (PIV) optical method we have explored the flow between and over the stent branches, in order to assess and to quantify the wall shear stress and to locate the interesting zones.

Determination of the optimal region for interaliasing distance measurement for flow regurgitant rate calculation: a fluid simulation study

BACKGROUND Color Doppler imaging of the convergent region is promising for quantifying valvular regurgitation. Nevertheless, proximal isovelocity surface area method has limitations. We sought to determine the optimal localization to measure the most precise flow rate using a new approach: the interaliasing distance. METHODS A finite volume-based program was used to perform simulations in unsteady flow conditions. Different instantaneous flow rates, leaflet angles, and orifice sizes were tested reproducing physiologic conditions. Relative difference between actual and interaliasing distance flow rate was calculated for each configuration. RESULTS The relationship between the relative error and the aliasing velocity location was described by a third-order polynomial equation. The magnitude of relative error is a function of the flow rate, orifice size, and leaflet angle. CONCLUSION The optimal distance from the orifice to measure the interaliasing distance was when the closer aliasing was between 4 and 8 mm from the orifice.

Localisation et détection de stents sur des images rayons-X bas contraste

Resume Actuellement, de plus en plus d’endoprotheses vasculaires, ou stents, sont implantes pour traiter les stenoses. La qualite du deploiement du stent dans les vaisseaux est supposee etre un important facteur de restenose. Le probleme quotidien rencontre en routine clinique est donc la verification de l’expansion du stent. Pour verifier de maniere non invasive et rapide cette derniere, l’utilisation des images rayons X est preferable, mais les stents possedent un tres faible contraste sur de telles images. Nous proposons de les localiser et de les detecter automatiquement en appliquant sur des images radios, des methodes adaptees a la detection de pliure. Dans cet article, nous comparons dans un premier temps, deux methodes classiques de segmentation, l’une d’elle est fondee sur la valeur des niveaux de gris, et l’autre fondee sur un filtrage adapte. La comparaison est faite sur le taux de fausses detections, les methodes ayant ete testees sur une centaine d’images. Dans un second temps, nous proposons un detecteur multilocal flou afin de detecter de facon plus precise les mailles du stent.