Introduction

In healthy, left anterior descending (LAD) coronary arteries, surface roughness has recently been characterised [1]. Surface roughness can be used as a standard for the development of cardiovascular bio-inspired materials used in the design of novel vascular implants for clinical treatment of vascular diseases. The aim of this study was to inflict mechanical damage to LAD coronary arteries with extreme physiological loading, imitating mechanical rupture of the artery, and assess its effect on surface roughness. Specimens underwent chemical processing, associated with imaging techniques of biological tissue and known to increase the circumferential surface roughness of LAD coronary arteries.

Methods

Six porcine LAD coronary arteries were dissected, with proximal and distal specimens used in the study. Specimens were subjected to uniaxial overloading to replicate mechanical damage. Surface imaging was performed using an Alicona G4 Infinite Focus microscope, pre- and post- chemical processing, involving fixation and dehydration of the biological tissue. Surface roughness was calculated from the 3-dimensional reconstructed optical images, in the circumferential (Ra_C) and longitudinal (Ra_L) directions, along the artery.

Results

A significant increase was noted in the longitudinal and circumferential surface roughness of damaged artery following chemical processing of tissue (p < 0.05).

Table 1 – Mean average result ± standard deviation of specimens (N = 6; n = 12) for surface roughness of healthy and damaged LAD coronary arteries, in processed and unprocessed states. † indicates result is significantly greater (p < 0.05) than both their hydrated damaged and healthy values.

Discussion

The results of this study found a significant increase in Ra_L of damaged LAD coronary artery when comparing processed to unprocessed tissue, which differs from the results of healthy tissue studied in previous work where an increase was only seen in Ra_C. In coronary adventitia, longitudinal stiffness is a direct result of initial fibre alignment, with collagen fibres uniformly stretching in the loading direction [2]. This would support the increase in surface roughness in the longitudinal direction, where collagen may have stretched and deformed under the uniaxial loading conditions. Although mechanical loading damage to coronary arteries may not significantly alter the endothelial surface, it could affect the internal constituents of the coronary artery causing a resultant change in surface roughness. Therefore, this study demonstrates the potential of using surface roughness to assess damage and disease in coronary arteries.

Acknowledgments

Engineering and Physical Sciences Research Council scholarship [M114612B].

References

Burton, Hanna E., et al. "Effects of freezing, fixation and dehydration on surface roughness properties of porcine left anterior descending coronary arteries." Micron 101 (2017): 78-86.

Chen, Huan, et al. "Biaxial deformation of collagen and elastin fibers in coronary adventitia." Journal of Applied Physiology 115.11 (2013): 1683-1693.

Introduction

Novel thoracic stent-graft systems have recently been developed for endovascular treatment of complex thoracic aortic aneurysms and dissections involving the aortic arch. These are designed to fit into the tight curvature of the arch in order to minimise the risk of endoleak and device migration. The implantation of a branched stent graft alters the lumen morphology, which can have a strong influence on flow in the arch and its branches. The objective of this study is to compare blood flow patterns and wall shear stress in the aorta and supra-aortic branches before and after the implantation of a Relay thoracic stent-graft (Bolton Medical, Sunrise, FL, US) through computational fluid dynamic modelling of a patient-specific thoracic aortic aneurysm.

Methods

Pre- and post-operative CT images were used to reconstruct patient-specific geometries. Transitional and pulsatile blood flow was simulated by numerically solving the fundamental mass and momentum conservation equations. Physiological boundary conditions were employed in order to produce clinically relevant results. At the model inlet, in vivo measured velocity waveform of a thoracic aortic aneurysm¹ was scaled to achieve a representative cardiac output of 5 L/min while downstream vasculature was described using a 3-element Windkessel model. The walls were assumed to be rigid where no slip condition was applied.

Results

Aortic flow patterns were significantly altered after the implantation of the stent-graft which caused increased blood velocities in the ascending aorta and the arch, resulting in elevated wall shear stress in the stented region.  The presence of inner tunnels in the main stent-graft body also caused flow derangement and asymmetric wall shear stress in the ascending aorta, where shear range index was up to 6 times higher than in the pre-intervention model. Blood perfusion to the supra-aortic branches was found to be sufficient. Moreover, displacement force on the branched stent-graft was evaluated and was found to be well below the threshold for device migration.

Conclusions

Increased time-averaged wall shear stress in the aortic arch after stent-graft implantation may help reduce the likelihood for thrombus formation, but increased spatial variation of wall shear stress and hemodynamic derangement may have a detrimental effect in the long term.  These findings warrant further investigation of the long-term durability and efficacy of this new device.

References

1. Tan, F. P. P., et al., (2009). Fluid-structure interaction analysis of wall stress and flow patterns in a thoracic aortic aneurysm. International Journal of Applied Mechanics, 1(01), 179–199.

Introduction

In most numerical analyses, vessels are treated as circular tubes, yet the assumption of circularity does not always hold true when examining CTA scans of the aorta. Cross-sections that resemble ellipses can be frequently observed even when thrombus and calcification are excluded. In the present work we statistically analysed a large dataset of patients with abdominal aortic aneurysm (AAA) and used the findings regarding their aneurysmal neck shape into finite element (FE) models to investigate the mechanical implications neck’s cross section has on stenting.

Methods

258 AAA patients were included in the study and for each one the perimeter of the cross section of its aneurysmal neck was examined in 2 locations, one proximal and one distal. For each perimeter, two metrics were used, the circularity factor defined as 4πΑ/P² and the ellipticity defined as R₁/R₂ (A is the area included in the perimeter of the cross section, P the perimeter and R₁, R₂ the minor and major axis of an equivalent ellipse respectively). After identifying the minimum and the mean values of these variables a series of numerical tests was performed. Specifically, a ring stent model was created replicating the structural unit of the Anaconda^TM stent graft (Vascutek, Terumo) and Abaqus^® was used to simulate the manufacturing process, compaction, delivery and final deployment of the ring in vessels of different cross-sections. The quantities of interest were the maximum strain developed on the ring, the chronic outward force (COF) acting radially towards the vessel and the total area of contact between the two.

Results

The analysis of 516 images showed that no correlation exists between the age of the patients or the aneurysmal neck’s length and the ellipticity metric while circularity is only weakly positively correlated with neck diameter. Additionally, both circularity and ellipticity in the male group were statistically significantly higher than in the female group at 5% significance. The minimum ellipticity value was 0.77 and it was further used as a guideline to perform a series of FE simulated stent deployments. Simulations showed that COF and maximum strain increased with the increase of the cross-sectional ovality while the contact area produced a more stable output. These results affect the sealing performance of the device and give insight on fatigue life and possible stent failure.

Conclusions

Challenging the circularity assumption lead to an extensive study of the neck’s of AAAs, the landing zone of stenting. The lack of correlation between the shape of the cross section and other geometric variables suggest that attention must be taken on a patient to patient basis. Furthermore, the deployment of ring stents into various shaped vessels allowed mechanical implications, useful in the clinical practice, to be revealed.

The incidence of in-stent restenosis (ISR) after carotid artery stenting (CAS) has been reported to range from 5.0 to 17.3% [1, 2]. Differences in stent design may contribute to disparity in wall shear stress (WSS) distribution in the stented area and subsequent neointimal hyperplasia and ISR. Previous studies have employed computational models to investigate haemodynamic alterations induced by carotid stenting [3,4], however, patient-specific models incorporating detailed stent design are lacking [5]. This shortfall is possibly due to the difficulties in visualizing the stent geometry from medical images and in creating a good quality mesh that is able to capture the protrusion of slender stent struts into the blood volume.

The present study proposes a robust computational method for reconstruction of post-CAS geometry by virtually implanting a given stent in a patient-specific carotid bifurcation. This allows detailed assessment of the haemodynamic conditions and how these are influenced by different stent designs.

An efficient approach for reconstruction of post-CAS geometry with a deployed stent was developed by making use of in-vivo images of the carotid bifurcation and stent-specific properties. The procedure consists of (i) parametric sketching of the stent including stent geometric design, and (ii) Boolean operations to virtually implant the stent.

By using this method, a patient-specific model of post-stent carotid bifurcation was reconstructed from CTA images, together with two types of stent design resembling open- and closed-cell stent. A healthy and a pre-stent model were also reconstructed for comparison. Computational fluid dynamics (CFD) simulations were performed on all the geometries, with blood being treated as a non-Newtonian fluid. Patient-specific pulsatile flow waveforms were applied at the inlet and 3-element Windkessel models were prescribed at the two outlets.

Our results showed that the carotid bifurcation model implanted with a closed-cell stent presented more atheroprone and procoagulant flow conditions than the one with an open-cell design due to its smaller free cell area, which leads to larger regions of low WSS (TAWSS <0.4 Pa), elevated OSI (>0.1) and high RRT (>10 Pa^-1).

The methodology presented in this study allows us to replicate the stent apposition from the post-CAS patient-specific geometry, maintaining the complex shape of stent strut such as floating strut at the ECA entrance and a small constriction downstream of the bifurcation; all of these influence the local hemodynamics. This method should be applied to multiple patient cases in order to further elucidate the role of stent design on the development of ISR after CAS.  

References

1. Moon, K. et al., J. Neurointerv. Surg. 2015; 0:1-5
2. Bonati, L. et al., Lancet. 2015;385:529-538
3. De Santis, G. et al., Artif. Organs. 2013;37:96-106
4. Uemiya, N. et al., J. Clin. Neurosci. 2013;20:1582-1587
5. Conti, M. et al., Comput. Fluids. 2016;141:62-74

For the treatment planning of coronary artery bifurcations virtual stenting techniques become more and more popular, since they are risk-free for the patient and an arbitrary number of treatment scenarios can be considered. However, numerous model assumptions lead to severe limitations, which limit the acceptance of those methods among physicians.                  

Within this study, CT angiography as well as intravascular optical coherence tomography are used to reconstruct a patient-specific coronary artery bifurcation before and after treatment. Afterwards, two approaches for the virtual deployment of a coronary stent are applied. The first one is based on a physical model using finite element analysis (FEA) [1]. It considers the complete stent deployment process and allows for vessel wall deformations. The second one, namely fast virtual stenting (FVS), enables a geometrical placement due to free-form deformations [2]. After the comparison of both methodologies with respect to their deployment results, hemodynamic simulations are carried out. Here, time-dependent flow phenomena are considered based on an explicit description of the individual stent struts. This enables the quantification of effects of each treatment planning strategy on the local hemodynamics.

Qualitatively, the application of different stent deployment techniques for the virtual treatment of a patient-specific coronary artery bifurcation shows a good agreement, although their model assumptions strongly differ (see Figure 1, left). However, quantitative differences are mainly present due to the vessel wall deformation that occurs during the intervention. Such phenomena can only be reproduced by the FEA approach, while the second FVS method, although it is clearly faster, only considers rigid vessel walls. Further, minor differences with respect to the wall apposition can be observed. While the struts of the first approach slightly deform the coronary artery, the second method rather attaches to it. This situation leads to deviations in the prediction of representative blood flow parameters (e.g., wall shear stresses or oscillatory shear indices). Specifically, although the velocity distribution appears to be qualitatively similar (see Figure 1, middle), differences of the spatially-averaged wall shear stresses along the stented vessel lumen are approximately 20 %.

Comparison of both virtual stenting techniques allows for a quantification of geometric as well as hemodynamic predictions. Hence, limitations due to the existing model assumptions can be identified in order to improve the clinical applicability of those methods.



Figure 1: Left - Comparison of different deployment strategies: FVS (green), FEA (blue). Middle - Velocity cut-plane; Right - WSS distribution between the stent struts.

Acknowledgements

The authors thank prof. J. F. LaDisa Jr. for his contribution on FEA. The work was funded by the Federal Ministry of Education and Research in Germany within the Research Campus STIMULATE under Grant No. 13GW0095A.

References

[1]            Chiastra et al., J Biomech, 49 (2016)

[2]            Janiga et al., J Biomechs, 48 (2015)

Introduction

Aortic regurgitation and new conduction abnormalities (with need of pacemaker implantation) are the major complications related to transcatheter aortic valve replacement (TAVR). Although new device generations are designed to address those complications, the design phase of a device is a tedious and time consuming process, based on benchmark tests and animal experiments. It therefore remains unclear how the final device will perform in actual patients. Patient-specific computer simulations can predict complications [1] and can be used to optimize device design to improve clinical outcomes. In this preliminary study, we investigated the effect of geometrical parameters on the pressure generated by TAVR valves on the aortic wall at the level of the conduction system.

Methods

Four nitinol self-expanding TAVR valves resembling the shape of existing devices were created [2] and virtually implanted in 3 patient-specific 3D aortic models (using Abaqus v6.12), derived from pre-operative MSCT. Each device consisted of sets of diamond shaped cells (15 along the circumference and 5 along the length). The height (40mm), strut thickness (450µm) and width (300µm), (ventricular) inflow and (aortic) outflow diameters (29mm and 38mm respectively) were identical for each device, while the diameters at the level of half of the first cell (d₁, about 4mm from the inflow) and at the first diamond from the inflow (d₂, about 8mm from the inflow) differed among the devices [25.5-29.0mm] (fig. 1A).

The inferior border of the membranous septum was selected from pre-operative MSCT and used to identify the region of interest where the atrioventricular conduction system is located. Maximum contact pressure generated by the interaction of the device with the surrounding structures and area of contact within the region of interest were extracted from each simulation and analysed.

Results

All devices were virtually implanted at comparable implantation depth (7mm from the annular plane, fig 1B). Independently from d₁, the larger d₂ the higher was the maximum contact pressure measured in the region of interest. Also, we noticed that enlarging d₁ significantly decreased the maximum contact pressure (up to 58%), probably because of a more homogeneous contact distribution. However, the observed area of contact was not significantly affected by different device designs.

Discussion

TAVR devices can be geometrically optimized in order to improve clinical outcomes. Our results suggest that larger diameters at about 4mm from the inflow can reduce the pressure generated by the frame on the aortic region near the atrioventricular conduction system, potentially avoiding new conduction abnormalities.

Acknowledgements

This work was supported by the European Commission within the Horizon 2020 Framework through the MSCA-ITN European Training Networks (project number 642458).

References

1. de Jaegere P, et al. (2016). EuroIntervention, 11:1044-52.

2. Gessat M, et al. (2011). IEEE Eng Med Biol Soc Conf, 2667-70.

Significant advances in computing power with decreasing computational overhead have allowed us to analyze in details the complex biomechanical interaction between the stent and the surrounding tissue. Stent implantation is one of the most popular and effective treatment of stenotic artery in surgery due to its high success rate and minimal invasive advantage, however the implanted stent can potentially influence the risk of plaque rapturing.

Stent design is an important issue to in-stent restenosis, where it can plays a critical role in inducing various degrees of stresses to stent, plaque and artery. In this presentation, we compare the effects of different stent designs on plaque, artery layers (intima, media, and adventitia) and stent stresses during stent implantation to optimize the stent design. The stenotic artery is assumed to comprise of an ideally symmetric plaque inside and normal arterial layers outside, with nonlinear finite element simulations of the stents, plaque and the artery model. Available fracture data for the plaque is used to estimate the percentage of plaque tissue that might rupture during implantation.

These different stent designs can induce high stresses on plaque and artery tissue in terms of both distribution and magnitude during implantation over the plaque and the artery tissue, and thereby potentially impact the plaque’s vulnerability. The success of a particular stent design will depend on its ability to moderate the stress distributions across the plaque surface without compromising the plaque’s vulnerability to rupturing. This work will provide invaluable knowledge for plaque vulnerability assessment to minimize injury to plaque and artery tissue. In addition, the study could enhance our understanding of the mechanics of stent-plaque-artery interaction to improve the stent’s long-term clinical outcome.

Introduction

Biodegradable stent symbolize a promising technological development in the field of cardiovascular intervention due to its ability to avoid the long-term adverse events of permanent stent such as in-stent restenosis, late thrombosis, and hypersensitivity reaction. However, the poor scaffolding is the main limitation of the biodegradable stent in clinical therapy. Therefore, in this study, we have proposed a novel structure design with strong support of biodegradable zinc alloy stent and investigated its mechanical behavior via finite element analysis.

Methods

The new structure design of the stent is shown in Fig.1. The highlight of this design is that two semi-circle rings are inserted into the links at the middle of stent, which allow the stent to expand but not contract, as the result of the wedge parts carved on the inner surface of rings (Fig.1b). In order to verify the mechanical performance of the new stent, the simulations of deployment with balloon in a 40% stenotic artery are performed with a general stent as the control stent.

Results

Figure 2 shows the radial displacement against the pressure, applied inside the balloon, for the new stent and the control stent during the process of deployment inside a diseased artery. The radial contraction rate and the Dog-boning rate after expansion of the new stent is decreased by 75% and 80% respectively compared with the control stent, which indicates that the support performance is significantly enhanced in the new stent. Moreover, the contour of the stress distribution (Fig.3) demonstrates that the new stent is more effective and safer than the control stent.

          (a)                                                           (b)

Fig.1 The new design of the biodegradable zinc alloy stent.

(a)The whole structure of the new design stent; (b)The sectional view of the links and strutting rings.

Fig.2 The radial displacement of two stents

Fig.3 The contour of stress in the new stent

Conclusion

The finite element analysis results imply that new stent with strong supporting performance proposed in this study can be the new choice in clinical therapy. In the future, some vitro experiments and optimization will be performed to promote the application of the new stent.

Acknowledgement: Major Project of Science and Technology of Beijing Municipal Education Commission and Type B Project of Natural Science Foundation of Beijing (KZ201710005007).

*Corresponding author: Aike Qiao, qak@bjut.edu.cn

Introduction

In-stent restenosis (ISR) well known as the harmful effect of stent treatment for endovascular diseases [1], [2]. Both hemodynamics and mechanical stress are believed to be the most reason behind the ISR. Therefore, this research aims to find an optimized stent design to improve stent performance for avoiding ISR. We performed design optimization with Kriging surrogate method based on computational simulation of stent deployment system with vessel wall's deformation.

Method

Simple 3D stent geometry with 5.5% wall deformation[3] has been built for this simulation. Stent’s strut side length and the distance between struts were used as design variables. The Objective functions are the minimization of both the average mechanical stress and the ratio of appearance of low wall shear stress along the deployment area. In total 50 simulations have been performed for this optimization work. All simulations were performed with simulation conditions as explained in [2]. The process of sample generation, kriging and surrogate model construction, and multiobjective optimization performed by an in-house program.

Results & Discussions

Surrogate models have been successfully constructed for both objectives. So that, the broad behavior of design variables’ combinations can be observed (Fig 1a and b). Focusing on the stent first objectives (Fig 1a), we can observe that the stent size has more influence on the stress behavior. Smaller stent side length will produce much higher average stress along the vessel wall. In the other hand, the second surrogate model (Fig 1b) shows the balance of both design configuration’s influence. However, it shows the side length size much influencing the stent’s behavior towards the objective.

Both surrogate models show a different candidate for optimized design configuration: bigger size is preferable for the first objective meanwhile smaller size for the second objective. Thus, to reach both objectives, we need to perform a multiobjective optimization algorithm based on EHVI method. Seven optimized design configuration points candidate has been obtained from this research.  

Summary

Multiobjective optimization has been performed to minimize average stress and ratio of low WSS along the deployed vessel area. We constructed two surrogate model map for each objective functions which is useful for design exploration process of the stent. Seven candidates for optimized design variable configuration have been obtained to improve the stent performance.

Acknowledgement

This research supported by LPDP scholarship and ImPACT program.

References

[1]      J. B. Elmore, et al., Interv. Cardiol. Clin., vol. 5, no. 3, pp. 281–293, 2016.

[2]      N. K. Putra, et al., Advances in Structural and Multidisciplinary Optimization: Proceedings of WCSMO12, Springer International Publishing, 2018, pp. 2097–2109.

[3]      J. W. Freeman, P. B. Snowhill, and J. L. Nosher, Connect. Tissue Res., vol. 51, no. 4, pp. 314–26, 2010.

Introduction: Poor durability of femoropopliteal artery (FPA) stenting is multifactorial, but severe mechanical deformations that occur with limb flexion are likely involved. Stent design may profoundly influence stent-artery interaction, but the effects of different stent types on arterial deformations are insufficiently understood. Our goal was to determine the influence of seven commercially available stents on limb flexion-induced FPA deformations.

Methods: Retrievable nitinol markers were deployed into n=28 FPAs of 15 human cadavers (average age 81±9 years). Bodies were perfused and thin-section CT images were acquired with limbs in the standing (180°), walking (110°), sitting (90°) and gardening (60°) postures. Image segmentation and analysis allowed measurement of baseline FPA foreshortening, bending, and twisting associated with each posture. Markers were retrieved and AbsolutePro, Supera, Innova, Zilver, SmartControl, SmartFlex stents, and Viabahn stent-grafts were deployed across the adductor hiatus in the same limbs. Markers were then redeployed in the stented FPAs, and limbs were re-imaged in each posture. Baseline and stented FPA deformations were compared to determine the influence of each stent design on FPA foreshortening, bending, twisting and cross-sectional pinching during limb flexion.

Results: Proximal to the stented segment, Innova, Supera, and SmartFlex exacerbated foreshortening by 15-71%, SmartFlex exacerbated twisting by 57%, and SmartControl restricted bending of the FPA by 96%. Within the stented segment, all stents except Viabahn restricted foreshortening by 15-54%. Supera, SmartControl, and AbsolutePro restricted twisting by 30-47%, while SmartFlex and Innova exacerbated twisting by 113 and 83%. Supera and Viabahn restricted bending by 74% within the stented segment. Distal to the stents, all devices except AbsolutePro and Innova exacerbated foreshortening by 21-49%. Viabahn, Supera, Zilver, and SmartControl exacerbated distal twisting by 31-77%. All stents except Supera were pinched in flexed limb postures.

Discussion: All stents significantly affect limb flexion-induced FPA deformations, but in different ways. No device was able to accommodate all deformation modes without either restricting or exacerbating baseline natural FPA deformations, either within or outside the stented segments. Stents that better match natural limb flexion-induced FPA deformations may mitigate arterial injury and improve flow characteristics, which could potentially improve clinical outcomes of FPA stenting.

Introduction

Motivated by the incidence of clinical stent fracture and correlated higher rates of in stent restenosis, this study predicts the number of cycles to failure of three balloon-deployed coronary stent designs subjected to representative complex cardiac loading. Coronary stents are typically metallic, with the most common alloys including 316L stainless steel, cobalt chromium, nitinol, and platinum chromium and given their classical metallic compositions the device lifetimes can be determined and predicted through failure analysis.

Methods

Finite element analysis (FEA) was used to investigate coronary stents response to complex cyclical loading representative of sequential cardiac filling and ejection, and included axially compressive, torsional and bending loads. The implantation and in vivo deformations of three 316L stainless steel stent designs were investigated using commercial FEA code Abaqus/Explicit (Fig 1). A sensitivity analysis was also performed to examine response to 25% variation in complex loadings for one stent design. Each design was post-processed to extract cycles to failure predictions using the modified Smith-Watson-Topper fatigue life model [1].

Results

FEA predictions captured the realistic deployment phenomenon of ‘dogboning’ for all three stent designs. Following four cycles of complex multimodal loading the stress-strain states were analysed as per the methodology documented in Liu [1] to extract predicted number of cycles to failure. The sensitivity analysis found that the trend in deformations remained similar with the primary bending deformation in the plane normal to the applied bend. Finally, for each stent design, under the given loading conditions, failure was predicted with the fatigue methodology implemented.

Discussion

Stents are tested for ten year survival using standardized methods before regulatory approval and yet clinical stent fracture is still evident, suggesting that such fatigue events are more complex than perhaps appreciated. This study not only demonstrates failure spatially but temporally, via an explicit calculation of number of cycles to failure based on an energy criterion. The predictions indicate fatigue lives that are significantly lower than ten years, highlighting the crucial need for continued refinement of bench-top testing and in silico exploration as new device designs emerge and understanding of clinical use conditions evolve.

Figure 1    Stent (insert illustrates repeating geometric units for three stent designs) subjected tri-folded balloon deployment, complex multimodal loading comprised of axial compression (C), torsion (T) and bending (B), and resulting fatigue life predictions. Sensitivity to loading was examined for one stent design – 25% changes in parameters listed.

Acknowledements

ORISE fellowship (CC) & NIH R01 GM49039 (ERE).

References

[1]  Liu, K.C., Advances in Multiaxial Fatigue ASTM STP 1191, McDowell & Ellis, ASTM Philadelphia (1993), pp 67-84.

Introduction

Understanding the relationship between stent design and acute arterial injury has become ever more vital as the number of patients receiving these endovascular implants increases (>1 million in the US annually). Endovascular stents have transformed the clinical outcomes for atherosclerosis treatment and the utilization of animal models can provide significant insight into the biological response to stent implantations.  However, knowledge is lacking of the precise interaction of acute vascular injury and stent geometrical features that may control this biological response. This study aims to explain stent induced arterial injury in terms of a numerically derived mechanical predictor.

Methods

12 stents (nominal diameter 3 mm), randomly split, between slotted tube (ST), corrugated ring (CR), and delta wing (DW) designs were implanted in the iliac arteries of six New Zealand white rabbits (3-4 kg, Millbrook Farm Breeding Laboratories, MA, USA).  The arteries were harvested 5-15 minutes after stenting, and stained in situ with silver nitrate to identify regions of intact (light grey) and denuded (dark grey) endothelium.

The predicted arterial state due to tri-folded balloon deployment of the ST, CR and DW designs was examined using finite element analysis (FEA) techniques (Abaqus/Explicit v6.14, Dassault Systemes, RI USA). Stent deployment methods and stent-balloon assembly material properties have been reported previously [1] and vessel tissue properties were extracted from literature [2].

Results

The resulting FEA predictions for arterial hoop stress distributions are compared to stained ex vivo stented vessel sections, for the ST, CR, and DW designs respectively, in Figure 1. FEA predictions of high tensile hoop stress concentrate in the interstrut region for all three designs and this corresponds to denuded endothelial areas in the ex vivo stented segments. Contrastingly, struts in contact with the arterial vessel are predicted to have minimal tensile hoop stress in the immediate strut region and this corresponds to intact endothelial area in the ex vivo segments.

Discussion

This study illustrates that FEA predicted tissue hoop stress, in the interstrut region, is a mechanical predictor of stent induced vascular injury, indicated by denuded endothelium. This has significant implications for optimising the effectiveness of these implants. Stent geometrical designs that achieve a minimal distribution of high interstrut arterial hoop stresses may promise less acute injury and reduced risk of adverse biological events.

Figure 1   Top: FEA predicted arterial hoop stress (MPa) contour plots (deformed stent geometry overlaid on each plot). Bottom: Corresponding silver nitrate staining of ex vivo stented rabbit iliac arterial sections for three designs,  SL – Slotted Tube, ML – Multilink, DW – Delta Wing.

Acknowledements

ORISE fellowship (CC) & NIH R01 GM49039 (ERE).

References

[1] C. Conway et al., (2012). Cardiovasc. Eng. Tech., 3(4) p374-387.

[2] C. Lally et al., (2004). Ann. Biomed. Eng., 32 p1355–1364.

Quantitative assessment of the hemodynamic impact of carotid artery stenting (CAS) represents an important ingredient to understand the CAS long-term outcomes, which are still a matter of clinical debate. Although computational fluid dynamics (CFD) has been extensively used for the biomechanical analysis of carotid arteries in healthy and stenotic conditions, there are few contributions in the literature dealing with post-stenting conditions because (i) current medical images do not have sufficient resolution to accurately reconstruct the stent geometry, and (ii) the generation of a body-fitted mesh of the stent-artery configuration is cumbersome.

Motivated by these considerations, we propose a numerical approach able to implicitly account for the stent geometry in the post-operative patient-specific hemodynamic CFD analysis by introducing velocity penalty terms in the weak form of the equations governing blood flow in arteries. Local mesh refining is also included to improve the performance of the numerical tool near the stent struts.

Such an approach allows us to compare different patient-specific CAS scenarios  paving the way for rapid CFD-based investigations of the hemodynamic impact of novel endovascular devices with complex grid designs. Test computations as well as patient-specific CAS simulations are presented.

M. Conti, C. Long, M. Marconi, R. Berchiolli, Y. Bazilevs, A. Reali. Carotid artery hemody- namics before and after stenting: a patient specific CFD study. Computers and Fluids 141, 62-74 (2016).

Introduction

Femoropopliteal stents-grafts are small diameter metallic stents wrapped with a polymeric cover, used in endovascular aneurysm treatment. High failure rates associated with femoropopliteal stenting can be attributed to the large loading deformations which the implanted artery must accommodate during knee flexion, such as axial compression and tension, bending, and torsion. To date, there has been considerable computational analysis research on covered stents in a range of applications in the body [1-2], however, the peripheral system has been somewhat neglected. Experimental benchtop testing of femoropopliteal stents has been carried out on twelve commercial stents [3], a useful validation for the computational simulations in this study. The aim of this study is to establish a framework to evaluate stent performance through simulated bench tests of physiological loading conditions to make predictions about stent geometry and the influence of a polymeric cover.

Methods

Simulated bench testing was performed using Abaqus/Standard and Abaqus/Explicit (V6.14). The process is automatized with a python code where the user specifies stent geometry, type of test, and if the stent is covered or bare. The stent geometry is imported as a 2D sketch, extruded to a thickness of 0.25mm and wrapped to a 3D cylindrical system. The stent geometries each have an outer diameter of 8mm and length 40mm. They are meshed with approximately 60,000 eight-noded linear brick (C3D8R) elements with enhanced hourglass control. The cover is 30µm thick and modelled with 19,000 shell (S4R) elements. A tie constraint is implemented to represent bonding between the stent frame and the cover, and general frictionless contact is used. To simulate the loading deformations, displacement constraints are applied to the stent. Initially, simulations have been carried out on two generic open- and closed- cell stent configurations.

Results

Figure 1 shows stent behaviour under radial compression for two different stent geometries.

Discussion

Variations in stent geometries, as well as the addition of a cover significantly alters the mechanical performance of a stent. The ideal stent-graft should have a high radial stiffness, yet be flexible in torsion and bending, and allow axial compression and tension. In general, covered stents are stiffer compared to their bare equivalent. However, use of a cover allows the stent to have a more flexible geometry, as the cover can link an open cell geometry with fewer interconnecting struts. The results will be expanded to include commercial stents, and this study will provide a comparison platform to indicate superior stent designs.

References

1. Chuang et al., (2009). Biomaterials
2. Kleinstreuer et al.,  (2008). J.Biomech.
3. Maleckis et al., (2017). J. Mech. Behav. Biomed. Mater.

The radial deformation of a stent, a tiny wire mesh tube-like scaffold, imposed by the inflation of a balloon within it, allows the compression of the lipidic plaque and the treatment of the stenosis, being one of the most popular applied techniques for this purpose.

As the promoter of the stent’s expansion, the balloon plays a very important role, offering a strong influence on its performance, mainly during the deployment process. Hence, the design of this element must be considered as a core task to achieve the best outcomes from the procedure, once it is expected that balloon’s geometry has impact in stent’s metrics such as dogboning, foreshortening and recoil.

In this scope, the use of finite element analysis arises as an advantageous tool to support the creation and development process once it allows to predict the behavior of the system and, therefore, to find the optimal geometric parameters to ensure the desirable performance.

In this study, a methodology for the simulation of the stenting procedure accounting with the presence of the stent and the balloon is proposed regarding not only their material constitutive modelling, but also the adequate boundary and contact conditions. Furthermore, the impact of the design of the inflation balloon is studied and a model to the optimization of its geometry is suggested to provide better outcomes from the deployment of an existent biodegradable stent geometry built in Magnesium alloy.

The presented balloon’s shape optimization methodology is combined with a previous developed method to the optimization of the stent, which is in agreement with a set of optimization variables and limiting values of the stent’s design parameters, imposed by a new manufacturing process for such device—ultrasonic-microcasting—, ensuring an advantageous compromise between the considered performance metrics.

The obtained results show that the decrease of the thickness of the balloon’s membrane leads to higher stent expanded diameters while its length is suggested not to have significant impact on it. Regarding the dogboning and foreshortening phenomena, both thickness and length of the balloon appears to have influence, as shorter lengths appear to lead to higher values of both metrics while the increase of its thickness is associated with lower values.

Introduction

Atherosclerotic cardiovascular disease results in millions of sudden deaths annually. Coronary artery disease (CAD) is one of the most relevant cardiovascular diseases. An important advance in the treatment of CAD is the development of Drug Eluting Stents (DES). Stents are mechanically expanded recovering the original lumen of the vessel but one of the main drawbacks is the restenosis posterior to stent deployment. The controlled delivery of anti-proliferative drugs limits this restenosis phenomenon avoiding the migration and proliferation of smooth muscle cells (SMC) from the media to the intima-media interface. A key point in the success of a DES system is a correct release of the drug maintaining a therapeutic concentration on the artery avoiding toxic levels that limits re-endothelialization.

Numerical models for drug diffusion phenomena have been extensively developed last years, but they usually represent simplified straight geometries. However actual arteries and specially coronary arteries are highly curved vessel. The objective of this contribution is to analyse of the influence of curvature and complex geometries in the drug transport phenomenon in curved vessels.

Drug transport is modelled using convection-diffusion-reaction equations.  The arterial wall is considered a multilayer model distinguishing different regions: subendothelial space and media and adventitia layers and several porous membranes separating them, endothelium, internal elastic and external lamina. Blood flow is modeled by Navier-Stokes equations in the arterial lumen domain, while Darcy’s law is used to calculate filtration velocity through porous layers. Kedem-Katchalsky equations are considered to include the drug transport through the membranes. A simplified 2D geometry of the curved vessels is modelled and the effect of distinct curvatures of the artery is compared. The embedment of the strut is also studied.

Results

As an example of the obtained results the drug concentration for 1 h after implantation is presented in Fig. 1 and the variation of normalized concentration versus time allows observing the different concentration in the internal and external sides of the artery.

Figure 1: Normalized drug concentration in the media and adventitia layers for a curvature radio r=3mm, 1 h after implantation

Conclusions

The arterial curvature affects the drug concentration few hours after stent deployment, increasing the drug level in the internal side of the vessel and decreasing in the external one. The internal concentration can be increased two or three times the external one and this ratio is magnified for higher curvatures.  The malappositioning of the struts usually associated with tortuous vessels also modified the drug level and can affect the effectiveness of the treatment.

References

Bozsak, F, Chomaz, JM, Barakat, AI (2014). Biomech. Model. Mechanobiol. 13(2), 327–347.

Acknowledgements

Support from the Spanish Ministry of Economy and Competitiveness through the research projects DPI2016-76630-C2-1-R.

Introduction

The implantation of perpetual mental stents will change the mechanic environment of the host vascular wall enormously [1]. With the degradation even the disappear of the polymer struts biodegradable polymer stents have advantages for the treatment of coronary heart disease, especially eliminating the need for stent removal [2]. Our goal is to understand mechanisms of action of endogenous mechano-growth factor (MGF), a splice variant of insulin-like growth factor 1 (IGF-1), and calcification in the vascular wall after poly (l-lactic acid) (PLLA) stents implanted for one year.

Methods

PLLA and 316L stainless steel (316L SS) stents were implanted to the abdominal aortas of Sprague Dawley rats at 1.2 : 1 (stent : artery) ratio. After 1 week, 4 weeks, 12 weeks, 6 mouths and 1 year, the stented arteries were cleaned by PBS and some of them were cut open longitudinally and fixed with 2.5 % glutaraldehyde. Human Umbilical Vein Endothelial Cells (HUVEC) were cultured in standard growth media with oscillatory fluid shear stress (5, 15, 25 dynes/cm²) supplements. The expression of MGF was quantified by western blot and qRT-PCR. Cytokine and signal pathway related to calcification were quantified by immunofluorescence. In-stent stenosis and inflammatory response were also detected by SEM and HE stain of the fixed vascular samples.

Results and Discussion

The expression of MGF after PLLA stents implantation was started to increase obviously at the beginning then decreases with the degradation of the PLLA struts. While the MGF of 316L SS group was much more stable compared to the PLLA group. When we applied shear stress to HUVEC, we obtained the similar results. Calcification of the stented arteries was much more serious in 316L SS group, many calcification cytokine and signals changed, such as BMP2, Runx2, TSP1, pSmad2. Because of obverse inflammatory response, the in-stent stenosis of PLLA stents was much more serious than 316L SS stents. But one year later with the degradation of the PLLA struts and its’ lose of the main mechanic structure, the thickness of neointima became smaller.

Conclution

This study has revealed that combined with the degradation of the PLLA stents and its’ lose of the main mechanic structure, the mechanic environment of the stented vascular wall changed dramatically which induce the changes of MGF and calcification of endothelia cells and vascular smooth muscle cells. 

Acknowledgements

NSFC (31370949, 11332003, 81400329), the National Key Technology R&D Program of China (2016YFC1102305), and the Visiting Scholar Foundation of Key Laboratory of Biorheological Science and Technology (Chongqing University) (CQKLBST-2016-003).

References

1. Yang, H., et al., (2017). Cardiovasc Eng Technol, 8(1) p81
2. Iglesias, J.F., et al., (2017). EuroIntervention. pii: EIJ-D-17-00734

Introduction: Highly-refined patient-specific finite element (FE) modeling can accurately reproduce percutaneous procedures, mimicking the complex coupling between the stent and the dysfunctional anatomy [1]; however, this strategy can be excessively time-consuming to effectively support clinical decision making. Accordingly, through adequate FE model simplifications, we sought to tackle the computational cost of FE stenting analysis, focusing on two patient-specific clinically-relevant scenarios: right ventricular outflow tract (RVOT) calcific obstruction and aortic coarctation (CoA).

Methods: Patient-specific anatomies were acquired through CT acquisitions (Siemens, Somatom AS) and segmented in Mimics Medical v.19 (Materialise, Belgium). RVOT models included calcific deposits and was complemented by the aortic root (AR) and the proximal section of the left coronary artery (CA); CoA models consisted of the ascending aorta, sovra-aortic branches and the descending aorta.

The widely used CP-Stent^TM (Numed, Hopkinton, USA) geometry was reconstructed in Gambit (Ansys Inc., Canonsbourg, USA). Subsequently, the device was discretized, following two different approaches, in: i) refined structured hexahedral grid of about 800k brick elements [2]; ii) simplified grid of about 3k linear beam elements (Fig. 1A). Also, a balloon-in-balloon delivery system (BiB®, Numed, Hopkinton, NY USA) was reproduced. The stenting procedure was numerically assessed in both the patient-specific scenarios, comparing the two FE strategies; all the simulations were run in the explicit solver Abaqus© v. 6.10 (Dassault Systèmes, France).

Results: Both the FE strategies reliably reproduced the stenting procedure in calcific RVOT (n=3) and CoA (n=2) patients, capturing the mutual interaction between the stent and the surrounding site of implantation; when using the beam model, the computational cost largely reduced by about 60%.

The FE models reported a comparable mechanical interplay between the stent and RVOT calcifications and between the device and the aortic wall, not negligible differences were visible in terms of pattern and entity of principal mechanical stress (σ_I) on the conduit wall (Fig. 1B-C). The minimum diameter of the deployed stent was localized in the same region; however, if compared with fluoroscopic data, the solid and the beam models overestimated the fluoroscopic diameter by about 8.6% and 18%, respectively. Differences arose between the two strategies comparing the mechanical stent response after implantation: the solid model revealed a Von Mises stress peak value of 644 MPa, 17% higher than the beam model.

Discussion: Our preliminary data suggest that beam-based stenting may be a valid strategy to investigate the impact of the procedure in patient-specific scenarios. Despite 3D FE stenting is still mandatory to pinpoint the biomechanical response of the implant, simplified FE models may supply practical improvements in elucidating procedural complications, in particular in complex clinical scenarios.

References

1. Biglino et al. (2017). Heart,103: 98–103.

2. Cosentino et al. (2014). Circ Cardiovas Interv, 7: 510-7.

Introduction

The long-term outcomes of endovascular therapy in patients with Peripheral Arterial Disease (PAD) show high rates of restenosis, which may be explained by the demanding mechanical environment related to leg movements. Stents that are implanted towards the popliteal artery (PA) can lead to unphysiological arterial deformations during leg flexion, such as arterial kinking¹. Although the presence of these extreme deformations has been qualitatively linked with restenosis, their effects on the flow behaviors of the Femoropopliteal (FP) arteries have not been investigated. Therefore, the objective of this work was to perform Computational Fluid Dynamics (CFD) analyses on patient-specific models of arteries in straight and flexed positions in order to investigate the changes in hemodynamic properties of the FP arteries with leg flexion.

Methods

Among a cohort of 35 patients that have undergone routine endovascular therapy, 7 of them had exhibited clinically relevant restenosis within 6 months of the procedure. Furthermore, 4 out of the aforementioned 7 patients were observed to produce a kink in the artery (max. curvature change: 0.51 cm^-1 ± 0.04 cm^-1) with leg flexion following Nitinol stent implantation¹. As such, the 3D arterial geometries of these 4 patients in straight and flexed positions (hip/knee flexion of 20°/70°) were reconstructed from 2D angiographic images. The arteries were assumed to have a constant lumen diameter of 5 mm and were discretized with approximately 500,000 tetrahedral elements. The CFD analyses were conducted with ANSYS CFX 18.0. Blood was modeled as a Newtonian fluid and the flow was simulated by applying an MRI measured volumetric flow rate at the inlet of the artery. A zero-pressure condition was applied at the outlet and a no-slip condition was prescribed for the artery wall. Three cardiac cycles were simulated and the wall shear stress (WSS), time-averaged WSS (TAWSS) and Oscillatory Shear Index (OSI) values for the last cycle were investigated.   

Results & Discussion

The CFD analyses showed that the presence of a kink affected the location of the areas that were exposed to low WSS (i.e. <1.5 Pa). Although the percentage of the nodes that were below this threshold was similar for both straight and flexed arteries (9% and 12%, respectively), the areas that were affected downstream to the kink was 2% for the straight vs. 7% for the flexed configurations (Fig. 1). This suggests that the arterial kink may negatively affect the hemodynamic properties of the arteries and may trigger arterial remodeling in the vicinity of unphysiological curvature changes.  

Fig. 1: The TAWSS (Pa) distribution for a post-stent patient-specific artery in straight and flexed positions

Acknowledgements

The work was supported by the Swiss National Science Foundation.

References

¹Gökgöl et al, Journal Endovascular Therapy, 2017

Introduction

Vascular stent implantation is widely used in clinical treatment due to its advantages such as less trauma, shorter hospital staying and faster postoperative recovery. It is one of the most effective cardiovascular disease diagnosis and treatment techniques. The blood flow of healthy coronary arteries is stable laminar flow. Vascular endothelial cells (VECs) are also subject to a relatively constant shear stress that directly regulate a variety of gene expression in VECs, involving cell proliferation, differentiation, vasomotor maintenance, thrombosis, regulation of the inflammation or immune system and many other aspects. Vascular stent implantation will cause intravascular hemodynamic changes and VECs will be damaged during the process of stent expansion. The purpose of this study is to investigate changes in morphology and function of VECs after vascular stents implantation that can affect shear stress on the surface of stents.

Methods

The stents were implanted to the coronary arteries of pigs and removed after 1 month, 3 months, 6 months, 12 months and 24 months. Then, the morphology of neointima with stents was observed by scanning electron microscopyat accelerating voltages ranging from 15 to 30 kV. According to the results of the mechanical changes computer simulation, the VECs were cultured under different mechanical conditions to observe the morphology and arrangement of the cells. To verify the changes of intercellular junctions after stent implantation the expression of F-actin, VE-cadherin and Rac1 were investigated.

Results and Discussion

Stents implantation will change the microenvironment of VECs ^[1], which will affect its arrangement and intercellular junctions. 4 weeks after implantation, rounded neonatal VECs had completely covered the surface of the scaffolds. As the time prolonged, the cell morphology gradually changed to the spindle shape. Studies have shown that round-like VECs were unhealthy, and long spindle-shaped VECs could play a normal function to maintain the stability of the blood vessel environment. In the strut and V-shaped of the stent, the VECs were paving along the strut. The connecting rod (“S” position) was structurally complex, and the arrangement of the VECs in these parts was also complicated（Fig 1）. The phenomena in animal experiments were basically consistent with the computer simulation. In vitro, cell experiments found the cytokines related to arrangement and intercellular connection, and obtained similar results.

Conclusion

This study has revealed that the changes of shear stress on the surface of stents induced by vascular stents implantation will affect the morphology and function of VECs.

Acknowledgements

NSFC (31370949, 11332003, 81400329), the National Key Technology R&D Program of China (2016YFC1102305), and the Visiting Scholar Foundation of Key Laboratory of Biorheological Science and Technology (Chongqing University) (CQKLBST-2016-003).

References

1. Wang, J., et al., (2016). Acta Biomaterialia, 50p546

Introduction. Despite their wide clinical usage, stent functionality may be compromised by the occurrence of complications, including early/late stent thrombosis and occlusion, which may have significant impact on the treatment efficacy and cost of patient care. Although several studies have described the effect of fluid-structure interaction on local hemodynamics, there is yet limited information on the effect of the stent presence on hemorheological parameters. The current work aims to investigate how red blood cell (RBC) mechanical behavior and physiology changes for flows in stented vessels and what the short/long term effects might be.

Methods. Blood samples from healthy volunteers were prepared as RBC suspensions in serum and in phosphate buffer saline at 45% haematocrit (Bioethics reference: EEBK/ΕΠ/2016/18). Self-expanding nitinol stents (1.0x3.3mm, closed-cell design with diamond-shaped pattern) were inserted in clear perfluoroalkoxy alkane tubing of I.D. 0.75mm. The tubing was connected to a syringe, integrated in a syringe pump. The samples were tested at flow rates of 17.5, 35 (physiological) and 70ml/min, and control tests were performed in non-stented vessels. For each flow rate, the sample viscosity was calculated (for shear rates between 251.2 and 1.4 s^-1), and RBC aggregation (Rheoscan-A200) and deformability (LINKAM-CSS450) were estimated. RBC membrane lysis was also examined by measuring the serum free haemoglobin concentration (Cayman Hemoglobin Assay).

Results. Our results indicate that the presence of a stent in a vessel indeed affects the hemorheological characteristics. Figure 1(a) shows that the relative viscosity η_rel of all samples (n=7) increases slightly with the increase of the flow rate (η_rel = sample viscosity / suspending medium viscosity, normalised with the high-shear viscosity of the sample). η_rel further increases when tripling the exposure (number of fluid sample passages through the device) at the 70 ml/min flow rate (square-brown data points). RBC aggregation (AI) and elongation index (EI) decrease as the flow rate, and exposure, increases (Figure 1(b) and (c)). Figure 1(d) illustrates that the levels of serum free haemoglobin (SFH) are higher for the case of triple exposure.

Discussion. The preliminary results in this study agree with the current understanding in hemorheology that the reduction in RBC deformability results in higher viscosity and reduced RBC aggregation¹. The results also indicate that the stresses developed in the stent area for the extreme condition could be sufficiently high to cause destruction of the RBC membrane. Further work on additional flow, geometry and device parameters is warranted in order to establish the clinical significance of the trends apparent in the present study.

Acknowledgements. Financial support received through the Start-Up Scheme of the Cyprus University of Technology.

References. [1] Baskurt, O., and Meiselman, H. Seminars In Thrombosis and Hemostasis, 2003, 29(5), p435-450.

Introduction

Thoracic aortic endovascular repair (TEVAR) has been accepted as an alternative therapy for aortic dissections, due to its less invasive nature when compared to traditional open surgery. Nevertheless, TEVAR has been reported with several biomechanical related complications, one of which is stent induced new entry (SINE), defined as a new dissection caused by stent-graft (SG) itself. Recent studies suggest a strong correlation between the incidence of SINE and SG oversizing ratio [1]. This study presents the development of a new computational model for virtual SG deployment in Type B aortic dissection (TBAD) reconstructed from computed tomography angiogram (CTA) images, and finite element analyse of biomechanical conditions in post-TEVAR TBAD. The aim is to identify morphological and biomechanical predictors of SINE.

Method

Two TBAD patients from Zhongshan Hospital (Fudan University, Shanghai, China) treated with Valiant Thoracic Stent Graft (Medtronic Vascular, California) are included in this study. The Valiant SG used in both patients were Proximal Freeflo Straight type but with different geometry codes, 32-300mm for Patient A and 28-150mm for Patient B. Patient A was found with SINE at the distal end of SG at 14-month follow-up, while the condition of Patient B remained stable.

CTA images of the two patients at three separate time phases, i.e. pre-TEVAR (Phase I), post-TEVAR (Phase II) and 14-months after TEVAR (Phase III) have been processed for 3D reconstruction. The adopted SG geometry has also been developed for the simulation. The aortic wall is modelled as a Mooney-Rivlin hyperelastic material, and its thickness is assumed to 2 mm. Virtual SG deployment is performed with the ABAQUS explicit solver (Dassault Systemes, France) following the steps described below:

Step 1: SG is compressed by a rigid cylinder from its diameter in zero stress condition (32 mm for Patient A and 28 mm for Patient B) to catheter diameter 22 mm.

Step 2: Compressed device is delivered into the pre-TEVAR aorta geometry in Phase I. The position of SG device is determined based on images acquired in Phase II.

Step 3: The sheath is removed and contact between the aorta and SG is activated. Finite element analysis is then performed on the patient-specific post-TEVAR aorta.

Results and Discussion

Detailed stress analysis will be performed on both patients. Biomechanical factors, including von Mises stress and shear load will be obtained and compared between the two patients. Regions of high stress concentrations will be identified and compared with the location of SINE observed in Phase III.

Acknowledgements

The authors would like to acknowledge the PhD Scholarship (grant for X.K.) from the China Scholarship Council (CSC) and the International Exchange Award from The Royal Society (IE161052), UK.

Atheroscerosis is the leading cause of deaths worldwide.Stents are medical devices used to treat such diseases. Stents are expanded slightly oversized to artery diameter to prevent flow-induced or pressure induced stent migration.This causes injury to the artery and altered hemodynamics in the vicinity of the stent struts.  Stent must have enough radial strength to act as a scaffold and allow linear flow of blood. Abnormally high stresses imparted to the arterial wall during the expansion are important factor in restenosis. Previous studies indicate a correlation between altered near-wall haemodynamics, specifically low wall shear-stress (WSS), and neointimal hyperplasia that leads to restenosis. The objective of this study is to investigate the effect of stent design parameters on radial strength,radial stiffness. stresses on the artery wall and wall shear stresses.

A parametric model of stent is made in CAD software and imported for CAE simulations.The stent is first crimp to 1 mm diameter and then expanded to 3.5 mm diameter. An elastic-plastic material property is used for the stent and hyper elastic material for the artery as reported in literature. Several stent geometries are generated by varying strut-spacing, ring amplitude and crown radii. Stress field on arterial wall is studied. Same solid model is also imported for CFD analysis. In this work a framework is developed to couple CFD and structural analysis to achieve an optimized stent design.  

Preliminary studies show strut spacing, amplitude and crown radius influences the wall shear stress on artery wall during the deployment and increases with increase in these parameters and vice versa. So an optimum value of these parameters is chosen to achieve adequate radial strength,radial stiffness and minimum crimping profile.


                             (a)                                               (b)

Fig. 1: Stent-Artery interaction study (a) Von mises stress and  (b)Wall shear stress

References

1. Gundert TJ1, Marsden AL, Yang W, LaDisa JF Jr., Optimization of cardiovascular stent design using computational fluid dynamics, J Biomech Eng. 2012 Jan;134(1):011002
2. Sanjay Pant, Georges Limbert, Nick P. Curzen, Neil W. Bressloff, Multiobjective design optimisation of coronary stents, Biomaterials,Volume 32, Issue 31, November 2011, Pages 7755–7773

Percutaneous coronary intervention with drug eluting stents (DES) has displaced coronary artery bypass surgery as the most widely used method for revascularisation in patients with coronary heart disease. In 2018, 3^rd generation metallic drug eluting stents (DES) have become a commodity. Irrespective of the manufacturer, they share very similar design features in terms of their metallic platforms, polymer coatings and limus based drug elution characteristics. The use of biodegradable polymers does not seem to have improved clinical outcomes and 1^st generation coronary bioresorbable vascular scaffolds have proved to be inferior to metallic DES. Specifically, no manufacturer currently offers any variation in either drug dose or elution kinetics. In mixed populations of patients treated with DES, stent failure rates at 1 year are < 5% and it has become increasingly difficult to demonstrate any differences in clinical outcomes between devices. However, in specific patient sub-groups such as those with diabetes, renal failure and vascular calcification there is clear evidence that outcomes are less good and such patients represent target populations for new devices with different limus doses and elution, novel drugs or perhaps drug combinations. In addition, certain lesion subsets such as ostial stenoses and complex left main coronary bifurcations also have inferior outcomes and may require lesion-specific stent designs. I will present data from clinical trials and case examples to support the contention that in the field of DES, one “size” does not fit all and it is time for some personalised medicine.

Introduction

For biodegradable stents, what is lacking in computational modelling literature is the representation of the active response of the arterial tissue in the weeks and months following stent implantation, i.e. neointimal remodelling. The main objective of this research is to build on the previous investigations (Boland et al. [1,2]) to develop a computational modelling framework that accounts for several major physiological stimuli responsible for neointimal remodelling and to combine this with a stent corrosion model that is capable of simulating both uniform and localised pitting corrosion. Using the framework, a second objective is to investigate the effects of the presence of the neointima on the mechanical performance (scaffolding support) of the biodegrading stent.

Methods

A continuum damage mechanics (CDM) approach is taken for the neointimal remodelling. A damage variable due to the Von-Mises stress in the artery (D_VM) is calculated through a stent deployment and recoil simulation. Separately, a computational fluid dynamics (CFD) model for blood flow is used to compute transient arterial wall shear stresses around the deployed stent. The wall shear stresses damage variable (D_WSS) is calculated using the time-averaged wall shear stresses (TAWSS) during a cardiac cycle. These two damage variables (D_VM and D_WSS) are then used as the stimuli in a neointimal remodelling simulation; this involves a three-layer artery modelled using an anisotropic HGO model, radially supported by an expanded magnesium stent. The arterial lumen is filled with mesh of deactivated finite elements, namely the “ghost mesh”, with negligible initial mechanical properties. The elements in the ghost mesh become activated according to the CDM neointimal remodelling. The activation of elements in the ghost mesh is representative of the remodelling of the artery through the growth of new neointimal tissue and involves a change in the material properties assigned to those elements. This neointimal growth model is subsequently combined with the both uniform (best case) and localised pitting (realistic) magnesium stent corrosion models.

Results and Discussion

Sumulation results show that non-uniform neointimal growth patterns occur, which are accelerated in regions of TAWSS and/or high Von-Mises stresses; this is most clearly demonstrated for the D_WSS only stimulus case in Fig. 1, with neointimal remodelling concentrated in regions of low TAWSS (high D_WSS) around the biodegrading stent connectors and struts. Crucially, this aligns with clinical observations. Hence, this new multi-physics modelling platform that for the first time combines multi-stimulus arterial remodelling around a biodegrading stent, allows for insight to be gained into the interplay between remodelling and degradation, to aid the development of the next generation of biodegradable stenting technology.

Acknowledgements

Irish Research Council

References

1. E. L. Boland, et al. (2016) JOM, 68(4),1198–1203.
2. E. L. Boland et al. (2017), “J. Med. Device., 11.
   
   

Introduction

Bioresorbable scaffolds (BRS) are considered as the next vascular interventional revolution. They reopen narrowed vessels while degrading with time under the presumption that in doing so they release vessels from caging, allow them to remodel, and eliminate long-term complications associated with metal stents. Yet, results have been disappointing with substantially higher incidence of clinical failures, especially thrombosis, compared to metal stents. Current research strategies inherited from metal stents lack the consideration of polymer micro-structures and dynamics, and rely heavily on clinical studies to identify failures. The objective of the study is to identify potential failure modes associated with 1^st generation BRS, and redefine research strategies for future BRS development, fabrication and evaluation processes.

Methods

The micro-structure and macro-performance of clinical-grade BRS were examined. Raman spectroscopy was used to track changes in molecular orientations and crystallinity at different stages of BRS during fabrication and implantation. A high-throughput multi-modal fatigue tester was designed to conduct durability tests on BRS. Cyclic multiaxial loadings included 16^o bending, 15^o torsion, and 4% axial compression which were applied in different combinations onto the scaffolds. Numbers and locations of fractures were compared to results acquired from 30-day animal studies in porcine coronary arteries.

Result

Microstructural heterogeneities were found within scaffolds arising from thermal and strain history during fabrication and were amplified after scaffold crimping on the balloon catheter and then balloon inflation (Fig 1A). Differences in crystallinity and polymer alignment across scaffolds produce structural irregularities even before chemical degradation, and lead to faster degradation in scaffold cores than on the surface (Fig 1B), which further enlarge the localized deformation (Fig. 1C). Macro-performance, in terms of fracture resistance, can be well predicted using multi-modal fatigue tester where the number of fractures on BRS and their location patterns matches with the results from animal studies (Fig. 1D).

Figure 1: A) Spatial heterogeneity in crystallinity exists between the surface and the core. B) Non-uniform degradation due to material heterogeneities. C) Extreme structure irregularity due to non-uniform degradation. D) Durability test with multi-modal fatigue tester can predict fractures of BRS in specific designs.

Discussion

We postulate that observed structural irregularities and asymmetric material degradation are responsible for unpredictable and variable clinical performance. Polymer material microstructures should be considered in earliest design stages of resorbable devices, and fabrication processes must be well-designed with microscopic perspective. In addition, multi-modal loading environments are critical during early development phase to identify potential failures associate with certain designs. With these micro- and macroscopic perspectives, we can now provide a more scientifically-driven design and testing paradigm in the development of next generation BRS specifically and devices of degradable materials in general.

Introduction: Coronary artery disease (CAD) remains a leading cause of death worldwide, with ischemic symptoms associated with atherosclerosis, often leading to fatal cardiac events. Percutaneous stent implantation is an established treatment for atherosclerosis. Unfortunately, stent deployment via balloon angioplasty can remove endothelial cells (ECs), which are critical in vasodilation and as an anti-thrombogenic surface. Lack of post-stenting re-endothelialization is associated with restenosis and late-stent thrombosis, with a high mortality rate. Given the endothelium’s pivotal role, promoting re-endothelialization of stented blood vessels can improve clinical outcome in stent therapy. We have demonstrated the effects of stent strut geometry on the local flow field and regulation of EC wound healing in vitro.

 Methods: Human umbilical vein endothelial cells (HUVECs) were grown to confluence on polydimethylsiloxane (PDMS) substrates coated with fibronectin, except for a narrow band that included the stent strut models. Equally spaced steps on the PDMS surface, either rectangular (RT) or circular arc (CA) and ranging from 50 to 150 µm high with a constant width of 200 µm, served as stent strut models with a control substrate void of stent struts (Blank). Once confluent, ECs were exposed to static (no flow, ST) or pulsatile, disturbed flow (DF) or undisturbed flow (UF) waveforms. Pulsatile wall shear stress (WSS) for DF ranged from -0.06 to 0.17 Pa with a mean of 0.06 Pa and an oscillatory shear index (OSI) of 0.15, while UF WSS ranged from 0.15 to 1.05 Pa with a mean WSS of 0.54 Pa and OSI of 0. Cell migration was tracked live using a microscope and camera.

Results: Figure 1a shows the EC area coverage of the initial wound with respect to time. A value of zero represents the initial gap while complete endothelialization of the gap, up to the stent edge and excluding the stent surface, is represented by a value of 1. Under static and DF conditions, ECs successfully migrate and close the initial wound for all strut geometries and heights (Figure 1a). However, under UF conditions the gap is completely endothelialized for the Blank, CA050 and RT150 geometries. Figure 1b only shows the stent surface area coverage. For ST, the stent surface of all stent strut geometries is endothelialized, while for DF only Blank, CA050 and RT050 are covered by ECs. Under UF conditions only Blank and CA050 stent surface are endothelialized.

Discussion: The CA050 stent strut was the only stent geometry completely covered by ECs regardless of flow conditions. Two variables, streamlining the stent strut geometry and reducing the strut height, promote endothelialization of the strut vicinity and surface. These results highlight the importance of the stent strut geometry and hemodynamics as important stent design parameters for improved wound healing and optimal clinical outcomes.

Introduction

Many mathematical models have been developed to try to understand drug release from stents and subsequent redistribution in the arterial wall¹.  Models have highlighted the importance of accounting for specific and non-specific binding², concluding that for sirolimus-eluting stents it is more important to sustain release than to increase dose.  Modelling has also been used to explain how differences in the binding properties of paclitaxel and sirolimus lead to different retention, suggesting that the optimal delivery strategy is drug-dependent³.  However, these conclusions have been made based on the assumption that the density of binding sites is uniform across the arterial wall.  This is despite experimental evidence to the contrary, suggesting a variation across the wall thickness, with noticeable differences between and within the media and adventitia^4-5. Target receptor densities for paclitaxel and sirolimus do not follow the same spatial pattern⁵ and when components of disease are present, the picture is further complicated⁶.  The aim of this study is therefore to investigate the role of non-uniform binding site density in determining arterial drug distribution following stent-based delivery.

Methods

We develop a 2D axisymmetric model of coupled stent drug release and redistribution in the arterial wall, similar to that employed by Bozsak et al.³ but with two important differences:  we model binding to both specific and non-specific binding sites, and we consider the density of binding sites to be a function of radial distance.  The form of the function is derived from experimental data, by relating binding site density to the partition coefficient. We simulate a number of cases including different drugs and initial drug loadings.

Results

Whilst plots of time-varying normalised mean concentration are similar for the range of cases considered, our results highlight clear differences in spatially-varying concentration of bound and free drug in the arterial wall.

Discussion

By assuming a uniform density of binding sites across the wall, current models may be misguiding the optimal drug delivery strategy. Accounting for spatial variation in binding site density can have a significant influence on local drug concentrations and binding site saturation levels. Since the distribution of binding sites will vary from artery-to-artery and from patient-to-patient, our results may help in the development of optimal drug delivery strategies and open up opportunities for personalised stent treatments.

Acknowledgements

We acknowledge funding provided by the Spanish Ministry of Economy, Industry and Competitiveness (Project DPI2016-76630-C2-1-R and grant BES-2014-069737).

¹McGinty, S. Math. Biosci. 2014,257:80-90.

²Tzafriri, A.R., et al. J. Control. Release. 2012,161(3):918-926.

³Bozsak, F., et al. Biomech. Model. Mechanobiol. 2014,13(2):327-47.

⁴Creel, C.J., et al. Circ. Res. 2000,86:879-884.

⁵Levin, A.D., et al. Circ. Res. 2004,101(25):9463-7

⁶Tzafriri, A.R, et al. J. Control. Release. 2010,142(3):332-338.

Introduction

Vascular tissue responses to percutaneous coronary intervention, such as in-stent restenosis, are influenced by alterations of local flow pattern due to stent insertion. In this context, computational tools alongside the application of medical imaging techniques enable the estimation of hemodynamic quantities that are known as trigger for events of restenosis, but not routinely measured in clinical practice.

A previously validated methodology was used to reconstruct patient-specific coronary artery where a bioresorbable polymeric stent was implanted to treat a bifurcation disease. The achieved high fidelity geometry was used to simulate the hemodynamics under the effect of the device. Seeking to investigate any correlation between endothelial wall shear stress at baseline and neointimal thickening at follow-up, we compared results from computational fluid dynamics (CFD) simulations with the neo-intimal coverage.

Methods

The treated blood vessel was reconstructed from angiograms and optical coherence tomography (OCT) images with a methodology described in (1). In particular, angiograms were elaborated with a commercial software (CAAS, PIE Medical Imaging) to obtain the centreline of the treated bifurcated vessel. The centreline was then used to reconstruct the 3D main branch with the implanted stent from the processed OCT images.

CFD simulations were performed with ANSYS Fluent v18 (Ansys Inc., USA). The inlet pulsatile flow-rate resembled a typical left anterior descending coronary artery flow-rate waveform. The mean flow-rate value was computed from the flow of the radiopaque liquid injected during angiography. The diameters of the branches were used to derive the flow-split conditions at the outlets (2).

Results

Blood flow into the bifurcation regions was disrupted by the presence of the scaffold over the orifice leading, eventually, to micro-recirculation and flow stasis (Fig.1A). The region of low values of time-averaged wall shear stress (TAWSS) reflected the flow recirculation at the proximal side branch (Fig.1B) and the 39% of the scaffolded main branch segments underwent TAWSS ≤ 0.4 Pa, this value is considered as potential trigger for complications (3). The distribution of neointimal growth will be computed in the scaffolded segment and compared with baseline TAWSS. 

Figure 1: Streamlines of the velocity coloured by the velocity magnitude at peak of flow-rate (A) and contour map of the time-averaged wall shear stress (TAWSS) (B).

Conclusion

This work illustrates the use of a patient-specific technique to model blood flow in stented bifurcation segments. This approach provides valuable complementary information that is expected to improve the clinical outcomes in this subset of coronary disease.

Acknowledgments

Susanna Migliori is supported by the EU-MSCA. GA No 642612 (www.vph-case.eu).

References

1.        Migliori S., et al. Med Eng Phys 2017;10(0):40–0.

2.        van der Giessen AG., et al. J Biomech 2011;44(6):1089–95.

3.        Malek AM., et al. JAMA 1999;282(21):2035.

Introduction

Biodegradable stents show promise to revolutionize coronary artery disease treatment. Cold gas dynamic spraying (CGDS) has been proposed as a novel manufacturing approach to fabricate biodegradable stents with reduced strut thickness, where iron and stainless steel 316L are combined in a 4:1 ratio¹. The objective of this study is to develop an experimental protocol to assess the effect of degradation on the micromechanical properties of Fe-316L, as well as the influence of size effects. Such an investigation is critical to the successful implementation of Fe-316L stents in a clinical setting.

Methods

Uniaxial micro-tensile samples were laser-cut from 150 µm and 250 µm thick Fe-316L foils, where the gauge section was comparable to a stent strut. In situ uniaxial tensile tests were conducted with a micro-tensile stage underneath a confocal 3D laser microscope to assess crack propagation and fracture location. Prior to tensile testing, samples were exposed to degradation by means of a dynamic flow chamber, which replicated peak in vivo shear stress rates observed at the strut surface (6-8 Pa), as flow-induced shear stress can accelerate stent corrosion². Samples were exposed to degradation for 24 and 48 hours.  

Results

The results of the uniaxial tensile tests are presented in Figure 1. For the non-corroded test specimens, similar stress-displacement curves were observed between the 150 µm and 250 µm samples. Furthermore, the ultimate tensile strengths (UTS) of the 150 µm and 250 µm samples were not significantly different from that of bulk Fe-316L. An overall decrease in UTS was observed for both sample thickness’ after 24 hours of corrosion, though a more significant loss was found for the 250 µm sample. Further loss of strength did not occur after an additional 24 hours of corrosion time. The greater loss in strength of the 250 µm sample may be attributed to a weakening effect of increased flow-induced corrosion due to higher shear stress on the sample surface.  

Discussion

The results of the investigation illustrate that CGDS is an advantageous fabrication method for biodegradable stent application. The influence of size effects was not found to be significant, due to the decreased grain size observed with CGDS. Further reduction of strut thickness may be possible without compromising scaffolding capability. While a reduction in strength was observed as the samples were exposed to corrosive media, this was restricted to the first 24 hours. Further investigations are ongoing to assess the loss over longer degradation periods.

Acknowledgements

The authors acknowledge the Natural Sciences and Engineering Research Council of Canada (NSERC) and the McGill Engineering Doctoral Award (MEDA).

References

1. Frattolin J. et al., Ann. Biomed. Eng. 44(2):404-418, 2016.
2. J. Wang, et al., Acta Biomater. 10(12): 5213-5223, 2014.