Fluoroscopy, CT, and MR imaging characteristics of a novel primarily bioresorbable flow-diverting stent for aneurysms

Abstract

Background

Five to ten percent of the global population have unruptured intracranial aneurysms, and ruptured brain aneurysms cause approximately 500,000 deaths a year. Flow-diverting stent treatment is a less invasive intracranial aneurysm treatment that induces aneurysm thrombosis. The imaging characteristics of a novel primarily bioresorbable flow-diverting stent (BFDS) are assessed in comparison to the leading metal stent using fluoroscopy, CT, and MRI.

Methods

X-ray/fluoroscopic images of stents were taken using a human cadaveric skull model. CT and MRI were acquired using silicone flow models of residual aneurysms. Images were analyzed with Likert scales in anonymous surveys by neurointerventionalists. Quantitative measurements of radiographic density (CT) and artifact boundary size (CT & MRI) were also obtained.

Results

Visibility of the BFDS on X-ray was less than the metal stent but deemed adequate for deployment and intraprocedural assessment. The metal stent was more radiopaque than the BFDS on CT, but qualitative assessment was not significantly different for the two stents. MRI imaging was significantly better using the BFDS in terms of overall artifact and intraluminal assessment.

Conclusions

The BFDS has adequate visualization on X-ray/fluoroscopy and should be clinically acceptable for fluoroscopic deployment. On MRI, there is less quantitative artifact as well as overall improved qualitative assessment that will allow for more detailed non-invasive imaging follow-up of treated aneurysms, potentially reducing the need for digital subtraction catheter angiography.

Introduction

It is estimated that 5% to 10% of the global population have unruptured intracranial aneurysms. ¹ Intracranial aneurysm rupture can be imminently life threatening, but rupture is preventable with prophylactic treatment. ² Current preventative measures include open surgical clipping, and endovascular treatments such as aneurysm coiling and flow diversion. Flow-diverting stents are a preferred treatment for wide-necked proximal internal carotid aneurysms and have high cure and low complication rates. ³

Flow-diverting stents redirect blood flow away from the neck of an aneurysm, allowing it to thrombose, and reducing subsequent rupture risk.^1,4 Limitations of this treatment include incomplete aneurysm thrombosis, vessel wall rigidity and degradation, stent thrombosis, intimal hyperplasia with luminal narrowing, and the extensive metallic artifact which can limit non-invasive imaging follow up. ⁴ Bioabsorbable stents have been proposed as an alternative to metal flow-diverting stents in order to reduce some of these problems. Potential benefits of a biodegradable stent would include reduced or eliminated delayed stent thrombosis, restoration of vasomotion, reduced restenosis, and improved follow-up imaging with CT or MR. ⁵

A novel self-expanding primarily bioresorbable intracranial flow-diverting stent (BFDS) is currently involved in preclinical studies. ⁶ Radiopacity is an important consideration of intracranial stent development for both deployment in clinical situations and imaging follow-up.^7,8 This study is an imaging evaluation of novel BFDS compared to the leading metal stent using fluoroscopy, CT, and MRI.

Methods

The primarily bioresorbable stent utilized in these studies is a braided self-expanding polymer/metal hybrid device. The resorbable component is a poly-l-lactic acid polymer, while the metal component is a radio-opaque composite wire. There are terminal radio-opaque marker bands at each end of the stent. The comparison metal stent (PED; Medtronic Inc) has a braided mesh design composed of cobalt chromium and platinum tungsten wires. ⁹

Fluoroscopy

Each size offering of the stent samples were placed in a relaxed (non-loaded/compressed) state in the cavernous sinus of a human cadaveric skull. The samples were imaged using the x-ray parameters shown in Supplementary Table 1. Rotational 3D reconstructed cross-sectional images were also acquired. The various size offerings of the BFDS (part A) (n = 6) and the BFDS in comparison to a metal stent (part B) (n = 1 for each stent type) were evaluated using an anonymous survey involving Likert scales by eight neuroradiologists for part A and by nine neuroradiologists for part B (Supplementary File).

Computed tomography

CT images were acquired using a straight tubed silicone aneurysm flow model with a singular channel in which the BFDS (4.0×26 mm) and the metal stent (4.25 × 16 mm) were separately deployed, imaged, and removed. A diluted contrast solution (20 mL 350 mg I/mL Omnipaque per litre of normal saline) was circulated via a Masterflex LS pump in ambient conditions at 240 mL/min through a flow loop. ¹⁰ The silicone model was determined by pin gage to have an inner diameter of 4.00 mm. Images were acquired with a Siemens Symbia Intevo Bold scanner using the parameters shown in Supplementary Table 2. Select images from the CT were assessed using an anonymous survey by eight neuroradiologists involving Likert scales (Supplementary File). Quantitative CT measurements of radiodensity (pixel value) were also measured from initial static images according to ASTM F640-12 standards, and CT artifact boundary sizes were measured manually based on ASTM F2119 − 07 standards.^11,12

Magnetic resonance imaging

MRI images were acquired in two trials, using straight tubing then using silicone aneurysm models, both with pulsatile flow using the same pump system as the CT model. The straight tubing model diverted flow into two parallel channels where the BFDS (4.0 × 22 mm) and the metal stents (3.75 × 15 mm) were deployed. The polyvinyl chloride polymer (PVC) tubing encompassing the stents had an inner diameter of 4.06 mm, verified by a pin gage (Accusize Equipment Serial #1064058). A total of 60 ft of clear PVC ¼” mm ID tubing was used to accommodate the metal restrictions of the scanner. The model was filled with 1000 mL of distilled water and circulated with a total flow rate of 480 mL/min (240 mL/min for each channel) simulating physiological flow rates. ¹⁰ Pre and post contrast images were obtained, the latter using a gadolinium (Magnevist 469 mg/mL) concentration of 2.35 g/L. ¹³ Straight and curved tubes were utilized for the aneurysm model, both having a 4.0 ± 0.3 mm saccular aneurysm, a 4.0 ± 0.3 mm inner diameter, a 150 mm length, and a 2.0 ± 0.3 mm wall thickness. The BFDS was deployed in the curved model and the metal stent in the straight. Both models were imaged in a 1.5 T MR system (Model GE Model Optima 360) following the protocols shown in Supplementary Table 3. Select images from the MR were assessed using an anonymous survey by nine neuroradiologists involving Likert scales (Supplementary File). MR artifact boundary sizes were also measured manually, in the area of highest artifact severity, based on ASTM F2119 − 07 standards. ¹²

Statistical analysis

Qualitative results were compared using Mann–Whitney and Wilcoxon paired tests, with p values <0.05 considered statistically significant.

Results

Fluoroscopy

Generally acceptable Likert scale ratings were obtained for each of the available size offerings of the BFDS (Figure 1). Compared to the BFDS, the metal stent was significantly better visualized on static fluoroscopic images. There was no significant difference, however, in the anticipated stent visibility for a delivery procedure using either x-ray fluoroscopy or 3D reconstructions (Figures 2 and 3).

Figure 1.

Average Likert scores given to six BFDS stents of different sizes assessed with X-ray fluoroscopy.

Figure 2.

X-ray images used in the fluoroscopy survey part b. The BFDS is shown in the static image (a) and axial reconstruction from the rotational acquisition (c, blue arrow). The equivalent metal stent images are shown in images b and c (red arrow).

Figure 3.

Average Likert scores of the BFDS and metal stents. The metal stent was rated to have the highest visibility and adequacy of visibility overall, but differences were only statistically significant using both the Mann–Whitney test (p = 0.007)* and the Wilcoxon paired test (p = 0.02)* for fluoroscopic visibility of stent in the resting state (column 1).

Computed tomography

For the qualitative CT studies, there were no statistically significant differences using the Mann–Whitney and Wilcoxon paired tests for any of the categories assessed (Figures 4 and 5). Quantitative CT results showed that the BFDS stent body artifact is on average approximately 67-87% of the metal stent artifact (Table 1). Marker band artifact was only present on the BFDS (as expected) but this was still less than the body artifact from the metal stent. Radiodensity measurements showed that the pixel values of the BFDS marker bands are around 2.11 to 8.43% greater than the ends of the metal stent, and that the radiodensity of the BFDS is 27.59% less than the metal stent at the neck of the aneurysm (Table 2).

Figure 4.

CT images used in survey. BFDS (A₁) and metal stent (A2).

Figure 5.

Results of the CT imaging survey comparing the BFDS to the metal stent. There was no statistical significance between the stents for any survey question.

Table 1.

Quantitative estimates of CT artifact boundary size.

Stent type	Average marker band region artifact (mm) with tubing ID ref scale	Average stent body artifact (mm) with tubing ID ref scale	Average marker band region artifact (mm) with CT internal ref scale	Average stent body artifact (mm) with CT internal ref scale
BFDS	1.31	0.97	1.28	1.18
Metal	NA	1.43	NA	1.36

Table 2.

Quantitative estimates of CT radiodensity.

	Average pixel value at distal end of stent	Average pixel value at the neck of the aneurysm	Average pixel value at the proximal end of stent
BFDS	1567	1404	1648
Metal	1534	1939	1509
Percent difference between the BFDS and Metal (%)	+2.11	−27.59	+8.43

MRI

For the qualitative MRI results, the BFDS was consistently rated higher than the metal stent in all categories (Figures 6 and 7). Statistically significant differences (by Mann–Whitney and Wilcoxon paired tests, respectively) were seen in the assessment of stent artifact (p = 0.001, p = 0.01), ability to evaluate the status of parent vessel lumen (p = 0.008, p = 0.01), diagnostic confidence in assessing in-stent stenosis (p = 0.004, p = 0.008), and diagnostic confidence in assessing intraluminal thrombus (p = 0.03, p = 0.02). The assessment of residual aneurysm filling (p = 0.05, p = 0.07) and any residual aneurysm (p = 0.13, p = 0.16) were also rated higher but were not significantly different. For the quantitative results, MRI boundary artifact was consistently higher for the metal stent compared to the BFDS (Table 3). The estimated artifact boundary size for the BFDS in the straight tube model was 34% of the width of the vessel, and was 45% for the metal stent. The estimated artifact boundary sizes in the aneurysm model for the BFDS and the metal stent were 48.5% and 77.6% at the aneurysm site, respectively. The estimated artifact boundary size at the extremities of the stents was not measurable for the BFDS and was 39.3% for the metal stent.

Figure 6.

MRI scans used in the survey, with the metal stent deployed on the left and the BFDS deployed on the right.

Figure 7.

MRI average survey results comparing the BFDS to the metal stent. *Categories with statistically significant differences between stents (Mann–Whitney test and Wilcoxon paired test).

Table 3.

Estimated quantification of MRI artifact boundary size.

Artifact boundary size	BFDS	Pipeline™
Straight tube model average measurement (mm)	1.4	1.85
Aneurysm model average measurement at neck of aneurysm site (mm)	1.95	0.24
Aneurysm model average measurement outside aneurysm (mm)	None	2.89
Aneurysm model average measurement at extremities of stents	NA	1.58

Discussion

Endovascular approaches have become a mainstay for treating ruptured and unruptured intracranial aneurysms. Flow-diverting stents have become an important tool within this, used mostly for wide-necked unruptured internal carotid artery aneurysms. Flow diversion using stents minimizes aneurysm rupture risk because a stent is deployed across the neck of an aneurysm to redirect blood flow, leading to aneurysm occlusion without requiring device placement inside the aneurysm itself. Currently available flow-diverting stents are metallic, made with a variety of metals including platinum, nitinol, tungsten, and nickel titanium. Although they have been overall successful in treating intracranial aneurysms, they still have significant limitations including: thromboembolic events, technical issues related to device placement, vessel rupture, occlusion of side branches, delayed hemorrhage and aneurysm rupture due to incomplete aneurysm thrombosis. ¹⁴ There is also the issue of metallic artifact from these stents which limits post-procedure imaging follow-up. Bioresorbable flow-diverting stents have the potential to impact aneurysm treatment by reducing the need for anticoagulation, minimizing long term complications associated with metal stents, and by improving post-procedure follow-up and assessment. ⁶

Bioresorbable vascular stents have been in development over the past 30 years, with the majority of these stents having been developed for coronary use. The Igaki-Tamai bioabsorbable coronary stent was the first to be implanted in humans and was made from poly-l-lactic acid. ⁵ Poly-l-lactic acid is absorbed through the hydrolysis of bonds between lactide units which produces lactic acid that enters the Krebs cycle to be metabolized into carbon dioxide and water. ⁵ The Igaki-Tamai stent had radio-opaque gold markers at each end, but was otherwise radiolucent. ⁵ The Absorb BVS (Abbott Laboratories, Chicago, IL) coronary stent, having markers at each end for radio-opacity, was recently approved then subsequently withdrawn from the market due to low sales. Other biodegradable coronary stents, which vary in materials and characteristics, have also been developed. The Magmaris BRS (Biotronik AG, Bülach, Switzerland) is the first bioabsorbable metallic coronary stent, made from a magnesium alloy, and has radio-opaque markers at each end. ⁵ In 2007, the REMEDY^TM biodegradable peripheral stent, based on the Igaki-Tamai Stent was commercialized for peripheral vascular use. It is the only CE Mark certified biodegradable peripheral stent to date, and contains two gold markers for fluoroscopic placement. ¹⁵ There are no biodegradable stents, however, that are currently approved for neurovascular use. The novel BFDS stent whose imaging characteristics are reported here is the first primarily bioresorbable self-expanding braided stent for aneurysm treatment.

Device characteristics are all important because in the case of intracranial aneurysm treatment, the importance of long-term follow up imaging and suitability of materials is emphasized. ¹⁶ Flow-diverting stents are deployed with fluoroscopic guidance, and for stent placement to be successful, the stent needs to be easily visualized during deployment. Our results demonstrated that the visibility of the BFDS has overall acceptable imaging on fluoroscopy, and despite the metal stent being significantly more radiopaque, there was no anticipated difference between the stents during deployment using fluoroscopy or rotational 3D imaging. Of course, these results are limited by the inability of surveyed neuroradiologists to evaluate the BFDS characteristics during actual human use.

CT imaging following stent placement is often used to determine stent wall apposition, persistence of large or giant aneurysms, or the presence of any bleeding associated with the aneurysm. ¹⁷ The presence of significant metal artifacts from a flow-diverting stent, however, limits the use of CT for more detailed assessments of small aneurysm persistence, parent vessel thrombosis and in-stent stenosis. In this study, we attempted to determine if the BFDS has the potential to yield more useful information on CT compared to metal stents. Our results showed that there is quantitatively more CT artifact with the metal stent compared to the BFDS, which is not surprising. Quantification of artifact, however, is an estimate with many contributing factors including image quality, definition of boundary (pixel density/colour), subjectiveness in the selection of start and end points with image processing tools, small sample size, orientation in scanner field, differences in sample mass/volume/dimensions, etc. Interestingly, although quantitatively there was increased CT artifact with the metal stent, there was no statistically significant difference between the two stents when assessed qualitatively.

With MRI imaging there was less boundary artifact associated with the BFDS as compared to the metal stent, which is also not surprising. The BFDS stent displayed less artifact than the metal stent in both a straight tube as well as in the presence of a residual aneurysm model. When evaluated by clinicians, the BFDS stent was rated on average 44% superior to the metal stent in the degree of artifact, impact on ability to evaluate parent vessel lumen, residual aneurysm filling, and diagnostic confidence in assessing in-stent stenosis, intraluminal thrombus, and residual aneurysm. Results were statistically significant (p < 0.05) aside from assessment of residual aneurysm. For MRI, the BFDS was significantly better than the metal stent in most respects. Therefore, it may be that the largest imaging benefit of the BFDS is related to the ability to perform MRI follow-up to obtain detailed information on the status of the parent vessel and any residual aneurysm. This could potentially lessen the need for conventional digital subtraction catheter angiography, which is more costly, requires a longer hospital stay, and has associated risks of puncture site complications and stroke. Further studies, however, would be required to confirm this.

There are several important limitations of this study. Boundary measurements for CT and MR were not normalized for the small size differences between the two devices. There are also manual measurement and nominal dimension limitations when measuring artifact. Furthermore, this in vitro assessment may not reflect the imaging characteristics experienced in vivo. During our fluoroscopic studies, for example, the stent was in its relaxed configuration and simply placed into the location of the cavernous carotid artery in a human cadaveric skull. A braided stent implanted into a blood vessel, on the other hand, would be in a constrained and more compressed state. Because braided stents are typically constructed oversized compared to their nominal diameters, in order to ensure enough radial force on the vessel wall to resist migration, imaging these stents in their unconstrained state leads to artificially increased radio-opacity, particularly on the outer edge of the two-dimensional image. This can potentially exaggerate the visibility of the stent on imaging compared with what would otherwise be encountered during actual stent placement in a human blood vessel. Flow models used in the CT and MR studies were also limited by the quantity and distribution of contrast, which may not be completely representative of the human condition. Evaluations of in vivo imaging characteristics using animal models, as well as studies of flow diversion and bioresorption, are in progress and will be the topic of future manuscripts.

Conclusions

Bioresorbable flow-diverting stents are an exciting and promising technology with the potential to improve intracranial aneurysm treatment and patient outcomes. Imaging characteristics of a novel flow-diverting, primarily bioresorbable stent have been presented. From this data, there is adequate fluoroscopic visualization and there would be no expected detriment for fluoroscopic deployment or intraprocedural assessment. The primary imaging benefit appears to be with MRI follow-up due to less associated metal artifact. This has the potential to improve non-invasive post-procedural aneurysm assessment and reduce the need for catheter angiography.