Annals of Biomedical Engineering ( 2014) DOI: 10.1007/s10439-013-0967-3

Design and Mechanical Properties of a Novel Cerebral Flow Diverter Stent JIAYAO MA,1 ZHONG YOU,1 JAMES BYRNE,2 and RAFIK R. RIZKALLAH2 1

Department of Engineering Science, University of Oxford, Oxford, UK; and 2Nuffield Department of Surgical Sciences, University of Oxford, Oxford, UK (Received 9 September 2013; accepted 23 December 2013) Associate Editor Nathalie Virag oversaw the review of this article.

Abstract—Brain arterial aneurysms are localised dilatation in the wall of cerebral arteries that are common among adult population and associated with high incidence of morbidity and mortality. Using flow diverter stent alone to treat cerebral aneurysm is recognized as a safe and effective method. However, flow diverter stents currently available have limitations due to their braided structures. In this paper a novel flow diverter stent is proposed. It is made out of nitinol tubes that provide adequate radial stiffness and tailored surface coverage to overcome problems of currently available braided stents while retaining low porosity and excellent longitudinal flexibility. Finite element analysis using Abaqus has been conducted to investigate radial stiffness, longitudinal flexibility, and maximum strain during packaging of a series of novel stent designs with varied geometric parameters. Results show that porosity below 70% can be achieved and provides radial stiffness and longitudinal flexibility comparable to those of the Neuroform stent that is commonly used for stent assisted coiling. The novel flow diverter has showed great potential for direct treatment of cerebral aneurysms. Keywords—Cerebral aneurysm, Flow diverter stent, Finite element analysis, Radial stiffness, Longitudinal flexibility.

INTRODUCTION Cerebral arterial aneurysm is a localised dilatation within the wall of a brain vasculature which can cause it to swell out like a balloon. If untreated, it will continue to grow and eventually tear or rupture resulting in stroke, and in around 25% of cases, death.3 Cerebral aneurysms affect up to 5% of adult population with 1% of detected aneurysms rupture every year.18

Address correspondence to Zhong You, Department of Engineering Science, University of Oxford, Oxford, UK. Electronic mail: [email protected]

Current treatment methods of cerebral aneurysms include surgical clipping and endovascular coiling.4,7 In surgical clipping, the skull is opened and a clip is placed across the neck of the aneurysm to prevent the blood vessel from bleeding. This method carries a high risk of post-operative complications (e.g., infection, neurological deficit, etc.). Endovascular coiling involves the insertion of soft metallic coils into the aneurysm sac to form a thrombus in the aneurysm to prevent it from rupture. This method allows minimally invasive aneurysm treatment, and is therefore widely accepted as a safer method than surgical clipping. However, it also has its limitations. First, coils are likely to fall out of the aneurysmal sac or protrude into the parent vessel in the case of wide-necked aneurysms leading to treatment failure.9 Second, the aneurysm that has been treated with coils retains roughly its original size since coils exist permanently within the aneurysm. Consequently, the mass effect of the aneurysm persists. Third, aneurysms that have been treated with coils may be refilled with blood, a process referred to as recanalization because of coil compaction.17 To overcome the limitations of endovascular coiling, stent-assisted coiling, in which a high porosity (>90%) stent is placed in the parent vessel to keep the coils in the aneurysmal sac, has also been utilized. This procedure can effectively prevent coils from prolapse, but the mass effect remains as the aneurysmal sac is still filled with coils. Moreover, this procedure still leaves many wide-neck and delicate aneurysms, such as pseudoaneurysm where no fully-formed aneurysm sac can be identified, untreatable.6 Being aware of the limitations of the treatment methods outlined above, therefore, there is a need for a new generation of devices and procedures that can effectively overcome the drawback of those methods.  2014 Biomedical Engineering Society

MA et al.

Among them, the using of a low porosity stent alone as a flow diverter to treat cerebral aneurysms is a more promising method. Due to the low porosity of the stent, blood is restrained from flowing into the cerebral aneurysm, thus encouraging thrombus (clot) formation in the aneurysm to stabilize, halting the growth of the aneurysm and reducing rupture risk. For this reason, the stent acts as a blood flow diverter. According to preliminary series published in the literature, treatment of aneurysms with flow diverters is highly efficacious with acceptable morbidity and mortality.8,12 Endovascular stents have three potential actions in this situation; (a) they improve anatomical reconstruction of the parent artery, (b) they modify the flow across the neck of the aneurysm, and (c) they allow biological repair of the aneurysm neck by endothelial overgrowth of the stent mesh.16 To exploit the last two effects, flexible self-expanding microcatheter delivered flow diverter stents with low porosity have been designed to induce aneurysm occlusion and endoluminal reconstruction of the diseased segment of the parent artery. A number of flow diverter stents, e.g., SILK from Balt and PED from Coviden, are currently available in the market. However, current designs are associated with high complication rates and monetary cost.5 The former are partially due to a low radial stiffness that makes endovascular deployment difficult, and results in poor stent-wall opposition, causing extra-aneurysmal thrombosis and stroke. As a result, we propose a novel flow diverter stent design that can be manufactured out of nitinol tubes to achieve low porosity and adequate radial stiffness while retaining excellent longitudinal flexibility to conform to the often tortuous brain arteries. The paper outlines the design of the novel device and computer structural analysis to study the maximum strain during packaging, radial stiffness, and longitudinal flexibility of the novel stent design. The layout of the paper is as follows. The design of the novel flow diverter stent is introduced in ‘‘Design of the Flow Diverter Stent’’ section, while ‘‘Geometric Parameters Analysis and Numerical Simulation’’ section deals with the finite element modelling of the stent. The numerical results of the stents are given in ‘‘Results’’ section. Finally, ‘‘Discussion’’ section sums up the main conclusions drawn from the work, which ends the paper.

DESIGN OF THE FLOW DIVERTER STENT A prototype of the novel flow diverter stent is shown in Fig. 1a. It is engraved from a nitinol tube. The developed pattern for the stent is presented in Fig. 1c. The stent comprises a low porosity region in the middle that to be positioned across the neck of the aneurysm,

and high porosity regions at both ends. The high porosity ends are designed to realize a good conformity of the device to the parent artery. Different from existing flow diverters which have little radial retention force so that they need expanded ends to retain them in place, the new stent has excellent radial retention force and therefore no expanded ends are necessary. The new flow diverter will be usually chosen to have a slightly larger diameter than that of the parent artery, and therefore the parent artery will be slightly expanded by the flow diverter along its entire length, which gives good anchorage against migration. The ends designed to have a lower radial force than that in the middle section provide a smooth transition from the undeformed part of the parent artery to the deformed part. Meanwhile, the ends can also be geometrically adjusted not to be overly soft so that a complete opening of the flow diverter is ensured. The low porosity region is formed by combining a number of identical typical modules. Two types of typical modules, referred to as type I and type II module, respectively, are presented in Figs. 1d and 1e. Type II module is a slightly modified version of type I module by replacing the edges of the straight portion with curved ones in order to reduce its porosity. Here the term porosity, denoted by q, refers to the ratio of the surface area of open regions to the total external surface area occupied by the stent.2 The total external surface area is the sum of the surface area of the open regions and the surface area of the regions occupied by the material of the stent. The stent can reach a radially contracted configuration that has a much smaller diameter than the original fully expanded configuration through a process involving elongation of the stent longitudinally. Figure 1b shows a portion of the prototype stent in a radially contracted configuration. As a result, the stent can be manufactured in the fully expanded configuration at room temperature, and then elongated longitudinally to reach a much smaller lateral profile. The stent can be elongated by over 300% of its original length to reduce the diameter by over 50% should a suitable design is chosen, while the stent material remains elastic. The stent can be manufactured by laser cutting a nitinol tube, unlike the braided stents made from interwoven nitinol wires, which has several important features. First, this design allows the radial contraction process to be achieved using a simple elongating and pushing action, even when the stent is in an elastic state (e.g., when a nitinol stent at room temperature is used). This method facilitates reliable and efficient insertion into a delivery catheter without the need for cooling the stent. Upon deployment, the stent will self-expand to restore its original configuration. Second, this design

Novel Cerebral Flow Diverter Stent

FIGURE 1. (a) The novel flow diverter stent in its radially expanded configuration, (b) a portion of the stent in its radially contracted configuration, (c) the developed pattern for the novel flow diverter stent, (d) module type I, and (e) module type II.

can be made out of a circular tube. As compared to current braided flow diverter stents, e.g., PED and SILK stents, this design can provide higher radial stiffness. Finally, this design is composed of both low porosity regions in the middle and high porosity regions at the ends, and it is, therefore, possible for the porosity to vary gradually, rather than suddenly change at the boundary of the low porosity region, through a proper design of the end regions. Such gradual change may produce a smooth change in the shape of the stented vessel. GEOMETRIC PARAMETERS ANALYSIS AND NUMERICAL SIMULATION Geometric Parameters Analysis As previously mentioned, the novel stent comprises of repetitive modules with low porosity in the middle

and high porosity regions at the ends. Here the focus is put on the modules which are critical for the stent. For type I module, five parameters define its geometry, i.e., the fully expanded radius of the stent R0, which determines its circumference, inner radius of the curved portion r, strut width w, thickness t, and minimum distance between adjacent struts d. For type II module, an additional parameter, i.e., the radius of curvature of the curved edges r¢, is needed to completely define its geometry. To investigate the structural performance of the stent, 12 models, including eight type I modules and four type II modules, were built for structural analysis. The R0 of all models was selected as 1.9 mm, which was typical of a cerebral stent, and d was taken as 0.05 mm. Parameters r, w, and t, on the other hand, were varied systematically to investigate the effect of geometric parameters on the mechanical properties of the stent. The geometries of all of the models, namely

MA et al.

A1–A8 and B1–B4, are listed in Table 1. In addition, the porosity of the models is presented in Table 1. A module of A1 is shown in Fig. 2a as an example. It can be seen in Table 1 that the porosity of the new stent ranges from 87.2 to 64.6%. This wide range in porosity is provided because there is still no conclusive standard on the optimum porosity of flow diverter stents. It is a common practice in industry to control the porosity at 80% or above to reduce the risk of restenosis. However, flow diverters with porosity below 70% have also been adopted in recently clinical study.10,15 Therefore, it is advantageous for the stent design to cover a wide range of porosity so that a proper porosity can be selected by clinicians. Three Neuroform stent structural models (Boston Scientific, USA), namely N1–N3, were also used to set the benchmark for the performance of the novel stent. We chose Neuroform stent as the benchmark for two reasons. First, both the new flow diverter and Neuroform stent are cut from a nitinol tube, which makes them structurally comparable. Second, the Neuroform stent is widely used in cerebral blood vessels so that it is deemed to have proper radial stiffness and longitudinal flexibility, which makes it a good candidate for comparison. The geometry of the three models is listed in Table 2. A ring of N1 is also shown in Fig. 2b as an example.

analysis through longitudinally elongating stent model A1. Four different mesh densities, i.e., 2 9 2, 3 9 3, 4 9 4, and 5 9 5 elements in section, were assigned to A1, and the maximum equivalent uniaxial tensile stress and strain were calculated for each density and normalized against the results of the finest mesh 5 9 5. A clear trend of convergence was observed from the data, and both the maximum stress and strain for the 4 9 4 mesh was within 5% of those of the finest mesh 5 9 5. Therefore we chose the mesh with 4 9 4 elements in section for the analysis. The first objective of the analysis was to examine the maximum equivalent uniaxial strain, emax, of the novel stent when inserted inside a 6F catheter of internal radius 0.9 mm. One module for each model was

Numerical Simulation Finite element analysis software package ABAQUS/ Standard1 was applied to investigate the maximum strain during radial contraction, radial stiffness, and longitudinal flexibility of the stent models. Superelastic nitinol was chosen as the material due to its large elastic range at room temperature (20 C). The material properties of nitinol are listed in Table 3.13 Each stent model was meshed with C3D8R, an 8-node linear brick element with reduced integration and hourglass control. Mesh density tests were conducted prior to

FIGURE 2. (a) A module of A1, (b) a ring of N1, (c) half a module of A1, (d) half a ring of N1, (e) stent I, and (f) stent III.

TABLE 1. Geometries of the novel stent models. Model A1 A2 A3 A4 A5 A6 A7 A8 B1 B2 B3 B4

type

r (mm)

w (mm)

t (mm)

r¢ (mm)

q (%)

emax (%)

I I I I I I I I II II II II

0.4 0.4 0.4 0.4 0.5 0.5 0.3 0.3 0.5 0.5 0.3 0.3

0.075 0.075 0.05 0.05 0.075 0.075 0.05 0.05 0.075 0.075 0.05 0.05

0.075 0.05 0.075 0.05 0.075 0.05 0.075 0.05 0.075 0.05 0.075 0.05

– – – – – – – – 2 2 7 7

82.8 82.8 87.2 87.2 85.1 85.1 85.3 85.3 68.5 68.5 64.6 64.6

7.97 7.80 5.95 6.04 7.30 7.30 7.28 7.30 7.39 7.34 7.63 7.59

Novel Cerebral Flow Diverter Stent TABLE 2. Geometries of the Neuroform stent models. Model N1 N2 N3

w (mm)

t (mm)

0.075 0.05 0.05

0.075 0.075 0.05

TABLE 3. Material properties of nitinol. Property EA EM rsM rfM rsA rfA eL

Value

Definition

50 GPa 37 GPa 400 MPa 650 MPa 350 MPa 80 MPa 0.055

Austenite elasticity Martensite elasticity Starting transformation stress of loading End transformation stress of loading Starting transformation stress of unloading End transformation stress of unloading Maximum residual strain

analyzed. During simulation, each model was longitudinally elongated until its radius reached 0.9 mm. This process was realised by assigning displacement in the longitudinal (L) direction as shown in Fig. 3a to the nodes at the extreme position at each longitudinal end of the model. Rigid body motion was eliminated by constraining two nodes in the middle of the model in both longitudinal (L) and circumferential (C) directions. Longitudinal elongation of both a module of a stent and an entire stent composed of 12 modules were conducted and it was found that both the standalone module and the module of the entire stent deformed in the same manner. Therefore, it was concluded that the stand-alone module was capable of representing the module of an entire stent. The emax for all of the 12 novel stent models were calculated and listed in Table 1. The second objective of the analysis was to investigate the radial stiffness of the novel stent. This was achieved by first longitudinally stretching the stent from the original fully expanded configuration to a radially contracted configuration using the same approach as described above, and then deploying it inside a rigid cylinder with a radius R that is smaller than the initial radius of the stent R0. Only half of a module, shown in Fig. 2c, was analyzed for the convenience of calculating hoop force. Symmetric boundary conditions were applied to the two circumferential ends of the stent. Rigid body motion was eliminated by constraining a node in the middle of the model in the longitudinal (L) direction. In addition, R3D4, a 4-node 3-D quadrilateral element was employed to mesh the rigid cylinder inside which the stent was deployed. The cylinder was completely fixed in space by six degrees of freedom. Surface-to-surface contact was assigned between the stent and the rigid cylinder while the frictionless contact interaction properties was utilised. The hoop force

Fh was calculated by summing up the reaction forces in the circumferential (C) direction of all the nodes at one circumferential end. Models A1, A2, A3, A4, A7, A8, B3, and B4 were analyzed, each of which being deployed in rigid cylinders of radius 1.8, 1.7, 1.6, and 1.5 mm, respectively. The hoop force Fh and corresponding longitudinal length l of each model at the four radii are listed in Table 4. In addition, half a ring of Neuroform stent N1 as shown in Fig. 2d was analyzed, and the result was compared with those of the novel stent models. The longitudinal flexibility was compared among four entire stent models. Stent I comprised 12 modules of B3 and two ends. The end geometry is shown in Fig. 1c. Stent II comprised 12 modules of B4 and two ends identical to those for Stent I. Stent III consisted of eight rings of the Neuroform stent N2, and stent IV consisted of eight rings of the Neuroform stent N3. Stents I and III are shown in Figs. 2e and 2f, respectively, as examples. To simulate the bending of the stent, the nodes at the extreme longitudinal positions at each longitudinal end of the stent were coupled to a reference point which was located on the central axis of the stent with a single translational degree of freedom along the axis. Rotational displacements in opposite directions were applied to the reference points to bend the stent until its curvature reached 0.1 mm21. Results are expressed in terms of bending moment vs. curvature.

RESULTS Strain Analysis For the novel self-expanding stent, it is crucial that the maximum equivalent uniaxial strain in the stent does not exceed 8%, the maximum recoverable strain of nitinol11 during the process of radial contraction to ensure that the stent can completely restore its original configuration upon deployment. The maximum equivalent uniaxial strains emax of the 12 novel stent models at the radius of 0.9 mm are shown in Table 1. The strain contour maps of the selected models are presented in Fig. 3. It can be seen that no strain is larger than 8%, indicating that no plastic deformation has occurred in the models during radial contraction. In addition, the following observations are made from the results. First, the strain contour maps in Fig. 3 clearly show that large strains occur in the curved portions of the model, while the rest undergoes only very small deformation. This is because when a stent is longitudinally stretched, the curved portions are bent severely in order to allow radial contraction. Therefore, the curved portions are critical in the design of the stent to ensure full elasticity during the process of

MA et al.

FIGURE 3. Strain contour maps of (a) A1, (b) A2, (c) A3, (d) A5, (e) A7, and (f) B1.

radial contraction. In addition, this deformation mechanism of the stent helps to explain the variation in emax with respect to stent geometric parameters. Second, comparison of the emax of A1 and A5, A2 and A6, A3 and A7, and A4 and A8, respectively, shows that when t and w are fixed, increasing r results in a reduction in emax. This phenomenon is understandable because the main deformation of a longitudinally elongated stent is in the form of bending of the curved portions as shown before. As a result, the larger r, the less bending deformation needed to straighten the curved portions, and the lower emax. Third, it is shown by comparing the emax of A1 and A3, and A2 and A4, respectively, that when t and r are fixed, reducing w leads to a reduction in emax. The reason for this is that when the bending deformation in the curved portions of the stent occurs, the direction of the bending moment is roughly perpendicular to the direction of w but parallel to the direction of t. Based on the elementary beam theory, when the change in curvature of the curved portions is identical, the largest strain is proportional to the height of the cross section, i.e.,

the strut width w. As a result, the smaller w, the lower emax is. Fourth, comparing the emax of A1 and A2, A3 and A4, A5 and A6, and A7 and A8, respectively, shows that when r and w are fixed, changing t has negligible effect on the emax. This is reasonable because as mentioned above, the direction of bending moment in the curved portions is roughly parallel to the direction of t. As a result, the emax is not significantly affected by t. Finally, when r, w, and t are all fixed, changing the module type from type I to type II leads to an increase in the emax. However, the amount of increase is invariably below 0.4%. This conclusion can be easily drawn by comparing the emax of A5 and B1, A6 and B2, A7 and B3, and A8 and B4, respectively. The reason for this has also been shown above, i.e., the main deformation occurs in the stent is due to the bending of the curved portions. As a result, change in the shape of straight portions has little effect on the deformation, and thus the emax of the stent. This feature is very useful for using type II module and can significantly reduce the porosity of the stent. Results in

Novel Cerebral Flow Diverter Stent TABLE 4. Hoop force per unit length of the stent models. Model A1

A2

A3

A4

A7

A8

B3

B4

N1

R (mm)

Fh (N)

l (mm)

fh (N/mm)

1.8 1.7 1.6 1.5 1.8 1.7 1.6 1.5 1.8 1.7 1.6 1.5 1.8 1.7 1.6 1.5 1.8 1.7 1.6 1.5 1.8 1.7 1.6 1.5 1.8 1.7 1.6 1.5 1.8 1.7 1.6 1.5 1.8 1.7 1.6 1.5

0.148 0.117 0.109 0.110 0.090 0.072 0.066 0.066 0.062 0.050 0.049 0.050 0.037 0.031 0.030 0.030 0.060 0.053 0.053 0.055 0.037 0.031 0.031 0.031 0.072 0.063 0.061 0.060 0.041 0.036 0.035 0.035 0.030 0.054 0.064 0.070

2.90 3.70 4.23 4.64 2.95 3.73 4.25 4.66 2.88 3.68 4.21 4.63 2.90 3.68 4.21 4.63 2.74 3.46 3.98 4.39 2.75 3.47 3.98 4.40 2.78 3.53 4.08 4.51 2.76 3.50 4.05 4.48 1.21 1.23 1.26 1.27

0.0510 0.0316 0.0258 0.0237 0.0305 0.0193 0.0155 0.0142 0.0215 0.0136 0.0116 0.0108 0.0128 0.0084 0.0071 0.0065 0.0219 0.0153 0.0133 0.0125 0.0135 0.0089 0.0078 0.0070 0.0259 0.0178 0.0150 0.0133 0.0149 0.0103 0.0086 0.0078 0.0248 0.0439 0.0508 0.0551

Table 1 show that, a porosity of 64.6% can be reached in B3 and B4 by properly selecting the geometric parameters while keeping the emax below 8%. To sum up, increasing r and reducing w are effective to reduce emax in the design of the novel stent, whereas employing type II module can significantly lower the porosity while not significantly affecting the emax of the stent.

treatment of cerebral aneurysms. To evaluate the radial stiffness of the novel stent, the Neuroform stent, which has been successfully used for treating intracranial aneurysms, is used here as a benchmark. The hoop forces per unit length, fh, of novel stent model A1 and Neuroform stent model N1, which have identical w and t, are plotted against radius R in Fig. 4a. It can be seen that when the radius varies from 1.5 to 1.8 mm, fh of the two stents can be seen as within the same range since the minimum and maximum fh of the new stent, 0.0237 and 0.0510 N/mm, are respectively very close to those of the Neuroform stent, 0.0248 and 0.0551 N/mm, suggesting that the new stent has a reasonable stiffness. However, the curves show opposite trends, i.e., fh of N1 increases with decreasing R whereas that of A1 reduces as R reduces. This result indicates that the new stent tends to be weaker that the Neuroform stent in the radial direction when the mismatch in size between stent and parent artery is large. For instance, in the case of 1.5 mm when stent over-sizing is 27%, fh of the new stent is only 43.0% of that of the Neuroform stent. Therefore, a good match in size between the new stent and parent artery is desired if a large radial stiffness is required. The fh of A1, A2, A3, A4, A7, A8, B3, and B4 at varying R are listed in Table 4 and plotted in Figs. 4b and 4c, and four observations are made. First, it can be seen by comparing the fh of A1 and A2, A3 and A4, and A7 and A8, respectively, that when r and w are fixed, reducing t results in a reduction in fh. Second, comparing the fh of A2 and A4, it is seen that when r and t are fixed, reducing w also leads to a reduction in fh. Third, when t and w are fixed, changing r has little effect on fh. This can be seen by comparing the fh of A3 and A7, and A4 and A8, respectively. Finally, when r, t, and w are all fixed, changing the module type from type I to type II leads only to minor increase in fh. In addition, comparing the fh of A7 and B3, and A8 and B4, respectively, clearly shows this trend. To sum up, increasing the cross section of the strut, i.e., increasing w or t, or both, helps to improve the radial stiffness of the novel stent. Longitudinal Flexibility

Radial Stiffness Radial stiffness is another important mechanical property that affects the performance of a flow diverter stent. As it is mentioned earlier, a low radial stiffness makes the endovascular deployment of the stent difficult and results in poor wall opposition. High radial stiffness, on the other hand, may severely deform the brain vessel in which the stent is to be deployed and cause potential damage to it. Ideally, a moderate radial stiffness that is just large enough to prevent the stent from migrating in the vessel is desirable for direct

The deformed configuration of stents I–IV is shown in Fig. 5. Large gaps are observed between adjacent rings of the Neuroform stents III and IV because of their open-cell design, resulting in a non-uniform profile. The deformed configuration of the novel stents I and II, in contrast, is quite uniform, thus providing an even coverage across the aneurysmal sac. The bending moment vs. curvature curves of stents II and IV, which have identical strut width and wall thickness, are plotted in Fig. 4d. It can be seen from the curves that the novel stent II requires a much lower bending moment than

MA et al. A1

(b) 0.06

N1

0.05

0.05

0.04

0.04

f (N/mm) θ

f (N/mm) θ

(a) 0.06

0.03 0.02

A1 A2 A3 A4

0.03 0.02 0.01

0.01 0 1.5

1.6

0 1.5

1.7

1.6

1.7

R (mm)

(c)

1.9

0.06

0.08

(d) 0.016

0.03

0.02

-1

Bending moment(N*mm )

A7 A8 B3 B4

0.025

f (N/mm) θ

1.8

R (mm)

0.015 0.01 0.005 0 1.5

0.012 0.01 0.008 0.006 0.004 0.002 0

1.6

1.7

1.8

1.9

Stent II Stent IV

0.014

0

0.02

0.04 -1

Curvature(mm )

R (mm)

Stent II Stent IV

-1

Bending moment(N*mm )

(e) 0.05 0.04 0.03 0.02 0.01 0

0

0.02

0.04

0.06

0.08

-1

Curvature(mm )

FIGURE 4. (a) fh vs. R curves of A1 and N1, (b) fh vs. R curves of A1, A2, A3, and A4, (c) fh vs. R curves of A7, A8, B3, and B4, (d) bending moment vs. curvature curves of stents II and IV at expanded state, and (e) bending moment vs. curvature curves of stents II and IV at crimp state.

that of the Neuroform stent IV when bent to an identical curvature, indicating that the longitudinal flexibility of the novel stent design is superior over the Neuroform stent. Moreover, it is deemed that the longitudinal flexibility of a stent at the crimp stage is also important because it affects the deliverability of a stent. And therefore the bending moment vs. curvature curves of stents II and IV when crimped to a radius of 0.9 mm are plotted in Fig. 4e. It can be seen that the new stent is about 80% stiffer than the Neuroform stent, which is not desirable from the perspective of stent deployment. However, as will be shown in the in vivo experiment in ‘‘Preliminary In Vivo Animal Trial’’ section, the stent has no problem of being guided through very torturous blood vessels. Therefore, this drawback is not deemed to

significantly affect the applicability of the new stent. At the same time, further design optimization is desirable to increase the longitudinal flexibility of the new stent at crimp stage. DISCUSSION Manufacturing Prototypes of the novel stent were cut out of nitinol tubes using our in-house laser cutting machine A3200 (Rofin Bassel, UK). Each stent then underwent acidpickling to remove the burrs and oxide films covering its surfaces, and electro-polishing to achieve a smooth surface. The liquor used for acid-pickling is a mixture

Novel Cerebral Flow Diverter Stent

FIGURE 5. Deformed configurations of (a) stent I, (b) stent II, (c) stent III, and (d) stent IV subjected to bending.

of 1% hydrofluoric acid, 9% nitric acid and 90% deionized water. Stent specimens are pickled in a plastic beaker with ultrasonic bath at 45 C for 10 min. Regarding electro-polishing, a stent specimen mounted on a titanium wire is immersed in the electrolyte in a glass beaker and forms the anode, while a stainless steel ring is selected as the cathode. The electrolyte contains 4.8% perchloric acid and 95.2% acetic acid. Each stent specimen is polished at 20 V for 100 s. Radiopaque markers made of platinum were installed at both ends of the stent to increase its visibility under X-ray. Delivery System The unique design of the novel stent allows it to be inserted into a delivery catheter in a single elongating and pushing action. Figure 6a shows the sketch of one delivery system design. In the delivery system the stent is loaded near the distal tip of the delivery catheter, and a stopper is placed coaxially and behind it. When the stent is deployed, the stopper is held stable in place, and the delivery catheter is pulled back in the proximal direction to unsheath the stent. A prototype of the delivery system with a stent loaded inside is shown in Fig. 6b. The device has been successfully loaded into a 6F delivery catheter; however, efforts are under way to optimise the delivery system into 5/4F catheter. The friction between the device and the delivery catheter was comfortably acceptable. Preliminary In Vivo Animal Trial With the recent advances of endovascular treatment of cerebral aneurysms,14 experimental models of saccular aneurysm are encouraged to adapt this novel therapeutic modality. The technique used in the surgical construction of experimental aneurysm is based on grafting a venous pouch (usually taken from the external jugular vein) onto the common carotid artery. To assess the performance of the novel stent design, a

FIGURE 6. (a) Sketch of the delivery system and (b) a prototype of the delivery system.

preliminary in vivo animal trial in which a stent was deployed across an experimental aneurysm in a carotid vessel of a swine model was conducted. The delivery system loaded with the stent was found to be quite flexible to be guided through to reach carotid. Figure 7 shows two fluoroscopy images of the experimental aneurysm before and after stenting. The results indicate that the aneurysm is successfully isolated by the flow diverter stent. Summary A novel flow diverter stent design for direct treatment of cerebral arterial aneurysm has been presented. The stent is made out of a nitinol tube through laser cutting technique. It comprises a low porosity region in the middle which is to be located across the neck of the aneurysm to reduce the blood flow in the aneurysmal sac, and high porosity regions at both ends. The radial contraction of the stent is realized by a simple longitudinal elongation action at room temperature. The mechanical properties, including maximum equivalent uniaxial strain during radial contraction,

MA et al.

FIGURE 7. Animal trial results: (a) without diverter and (b) with diverter.

radial stiffness, and longitudinal flexibility of the novel stent were investigated through finite element analysis of 12 models. The main conclusions drawn from the numerical results are as follows:

 The maximum equivalent uniaxial strain, emax, is lower than the maximum elastic strain of nitinol for all of the novel stent models when they are radially contracted to 0.9 mm in radius from the original radius of 1.9 mm. By properly selecting stent geometric parameters, a porosity of 64.6% is achieved.  The hoop force per unit length, fh, of the novel stent is reduced when being deployed in cylinders with reducing radii. With identical strut width and wall thickness, the fh of the novel stent is comparable to that of the Neuroform stent which is commonly used for stent-assisted coiling treatment of cerebral aneurysms.  When bent to an identical curvature, the novel stent requires a much less bending moment than that of the Neuroform stent with identical strut width and wall thickness. In addition, the novel stent deforms in a quite uniform manner, thus providing an even coverage for the aneurysm. Quantitative in vitro flow dynamics tests utilizing particle image velocimetry (PIV) are being prepared and the results will be reported in due course. Further in vivo experiments in which a swine model is stented and kept for 3–6 months are also underway.

ACKNOWLEDGMENTS The authors would like to thank Wellcome Trust and EPSRC for their financial support under Grant Number WT 088877/Z/09Z.

REFERENCES 1

Abaqus. Version 6.7 Edition. Providence, RI: SIMULIA Corp. 2 Augsburger, L., M. Farhat, P. Reymond, E. Fonck, Z. Kulcsar, N. Stergiopulos, and D. A. Ru¨fenacht. Effect of flow diverter porosity on intraaneurysmal blood flow. Clin. Neuroradiol. 19(3):204–214, 2009. 3 Bederson, J. B., E. S. Connolly, H. H. Batjer, R. G. Dacey, J. E. Dion, M. N. Diringer, J. E. Duldner, R. E. Harbaugh, A. B. Patel, and R. H. Rosenwasser. Guidelines for the management of aneurysmal subarachnoid hemorrhage a statement for healthcare professionals from a special Writing Group of the Stroke Council, American Heart Association. Stroke 40(3):994–1025, 2009. 4 Brisman, J. L., J. K. Song, and D. W. Newell. Cerebral aneurysms. N. Engl. J. Med. 355:928–939, 2006. 5 Byrne, J. V., R. Beltechi, J. A. Yarnold, J. Birks, and M. Kamran. Early experience in the treatment of intra-cranial aneurysms by endovascular flow diversion: a multicentre prospective study. PLoS ONE 5(9):e12492, 2010. 6 Fiorella, D., F. C. Albuquerque, V. R. Deshmukh, H. H. Woo, P. A. Rasmussen, T. J. Masaryk, and C. G. McDougall. Endovascular reconstruction with the Neuroform stent as monotherapy for the treatment of uncoilable intradural pseudoaneurysms. Neurosurgery 59(2):291–300, 2006. 7 Linfante, I., and A. K. Wakhloo. Brain aneurysms and arteriovenous malformations advancements and emerging treatments in endovascular embolization. Stroke 38(4): 1411–1417, 2007. 8 Lubicz, B., L. Collignon, G. Raphaeli, J.-P. Pruvo, M. Bruneau, O. De Witte, and X. Leclerc. Flow-diverter stent for the endovascular treatment of intracranial aneurysms a prospective study in 29 patients with 34 aneurysms. Stroke 41(10):2247–2253, 2010. 9 Lubicz, B., X. Leclerc, M. Levivier, J. Brotchi, J.-P. Pruvo, J.-P. Lejeune, and D. Bale´riaux. Retractable self-expandable stent for endovascular treatment of wide-necked intracranial aneurysms: preliminary experience. Neurosurgery 58(3):451–457, 2006. 10 Lylyk, P., C. Miranda, R. Ceratto, A. Ferrario, E. Scrivano, H. R. Luna, A. L. Berez, Q. Tran, P. K. Nelson, and D. Fiorella. Curative endovascular reconstruction of

Novel Cerebral Flow Diverter Stent cerebral aneurysms with the pipeline embolization device: the Buenos Aires experience. Neurosurgery 64(4):632–643, 2009. 11 McKelvey, A., and R. Ritchie. Fatigue-crack propagation in Nitinol, a shape-memory and superelastic endovascular stent material. J. Biomed. Mater. Res. 47(3):301–308, 1999. 12 Pierot, L. Flow diverter stents in the treatment of intracranial aneurysms: where are we? J. Neuroradiol. 38(1): 40–46, 2011. 13 Rebelo, N., X.-Y. Gong, A. Hall, A. R. Pelton, and T. W. Duerig. Finite element analysis on the cyclic properties of superelastic nitinol. In: Proceedings of the International Conference on Shape Memory and Superelastic Technologies, 2006, pp. 157–163. 14 Strother, C. M. Electrothrombosis of saccular aneurysms via endovascular approach: part 1 and part 2. AJNR Am. J. Neuroradiol. 22(5):1011–1012, 2001.

15

Szikora, I., Z. Berentei, Z. Kulcsar, M. Marosfoi, Z. Vajda, W. Lee, A. Berez, and P. Nelson. Treatment of intracranial aneurysms by functional reconstruction of the parent artery: the Budapest experience with the pipeline embolization device. AJNR Am. J. Neuroradiol. 31(6):1139–1147, 2010. 16 Szikora, I., L. R. Guterman, K. M. Wells, and L. N. Hopkins. Combined use of stents and coils to treat experimental wide-necked carotid aneurysms: preliminary results. AJNR Am. J. Neuroradiol. 15(6):1091–1102, 1994. 17 Van Rooij, W., and M. Sluzewski. Endovascular treatment of large and giant aneurysms. AJNR Am. J. Neuroradiol. 30(1):12–18, 2009. 18 Vernooij, M. W., M. A. Ikram, H. L. Tanghe, A. J. P. E. Vincent, A. Hofman, G. P. Krestin, W. J. Niessen, M. M. B. Breteler, and A. van der Lugt. Incidental findings on brain MRI in the general population. N. Engl. J. Med. 357(18):1821–1828, 2007.