WEB Device Shape Changes in Elastase-Induced Aneurysms in Rabbits

BACKGROUND AND PURPOSE: While WEB devices have been shown to be safe and effective for aneurysm treatment, WEB-shape modi ﬁ cation compression has been associated with incomplete aneurysm occlusion. We explored the relationship between occlusion rates and WEB-shape modi ﬁ cation in different WEB device types in an experimental aneurysm model. MATERIALS AND METHODS: Elastase-induced aneurysms were created in rabbits and treated with dual-layer ( n ¼ 12), single-layer ( n ¼ 12), or single-layer sphere ( n ¼ 12) WEB devices. Aneurysms were followed up either at 3 or 12months. Angiographic occlusion was graded using the WEB Occlusion Scale: grade I, complete; grade II, complete but recess ﬁ lling; grade III, residual neck; or grade IV, residual aneurysm. WEB-shape modi ﬁ cation and histologic features were also analyzed. RESULTS: Grade I or II occlusion was seen in 16 (44%) aneurysms, and grade I, II, or III ( “ adequate ” ) occlusion was observed in 22 (61.1%) aneurysms at follow-up. WEB-shape modi ﬁ cation was observed in 22 (61.1%) aneurysms. WEB-shape modi ﬁ cation was higher in single-layer (9/12) and dual-layer (10/12) devices compared with single-layer sphere devices (3/12). Aneurysms with WEB-shape modi ﬁ cation had a higher level of thrombus organization in the dome compared with those without WEB-shape modi ﬁ cation (68% [15/22] versus 50% [7/14]). WEB-shape modi ﬁ cation was not correlated with angiographic or histologic outcomes but was signi ﬁ cantly correlated with levels of ﬁ brosis and smooth muscle cells in the aneurysm. CONCLUSIONS: WEB-shape modi ﬁ cation is not associated with incomplete aneurysm occlusion of WEB devices in the rabbit model but may be related to connective tissue formation and the healing response to WEB device implantation.

D uring the past 2 decades, the treatment of intracranial aneurysms has rapidly evolved from the single option of open skull surgical clipping to the increasingly common use of minimally invasive endovascular techniques. While endovascular coiling has been shown to offer significant advantages over an open surgical approach, it is still associated with significant limitations. These include treatment-associated morbidity of at least 5%-7%, a high aneurysm recurrence rate (as high as 20%-30% in some series), and low aneurysm occlusion rates. 1 Flow diverters (FDs) are innovative and promising devices used to treat complex and/or wide-neck aneurysms. 2,3 They divert blood away from aneurysms, and their use has expanded since inception. 4,5 Use of endoluminal FDs, however, necessitates the use of antiplatelet therapy, which can be problematic for ruptured aneurysms. Unfortunately, the use of FDs does not eliminate the possibility of delayed aneurysm rupture, a thromboembolic event, and parent artery stenosis. 6 The Woven EndoBridge (WEB) device (MicroVention) is a completely intrasaccular FD device, which was designed to treat challenging, wide-neck bifurcation aneurysms that are difficult to embolize safely and effectively with other existing treatment approaches. 7,8 From the original dual-layer (DL) to current single-layer (SL) design, WEB devices have shown a good safety profile and a high rate of adequate occlusion both in experimental and clinical studies. [9][10][11][12][13][14][15][16] In the meantime, WEB-shape modification (WSM) has been reported in the literature and could potentially be associated with the approximately 15% rate of incomplete/inadequate aneurysm occlusion observed in the published WEB Intrasaccular Therapy (WEB-IT) trial. 17 However, this phenomenon has not been well-understood until now because other researchers have reported that this shape change did not impact the anatomic outcome after WEB deployment. 18 In this study, we compared the occlusion rates of DL, SL, and single-layer sphere (SLS) WEBs in rabbit elastase (Worthington Biochemical)-induced aneurysm models and evaluated the relationship between WSM using histologic results and angiographic occlusion rates.

Aneurysm Creation
Elastase-induced aneurysms were created in 36 New Zealand white rabbits. Animal procedures were approved by the Institutional Animal Care and Use Committee at Mayo Clinic. Aneurysm-creation procedures were performed using an elastase-induction model as previously described. 19 Endovascular treatment of each aneurysm was undertaken at least 3 weeks after aneurysm creation. 20 Devices WEB devices are classified according to shape and the number of mesh layers each one contains. DL devices consist of 2 layers of braided, nitinol wire mesh. SL and SLS devices consist of a single layer of braided, nitinol/platinum wire mesh; however, SLS implants have a more rounded 3D shape than the "barrel-like" SL and DL devices. The specific qualities of the WEB devices have been previously described. 21,22 Device Deployment The WEB device deployment procedure has been reported previously. 7,8 The right femoral artery was briefly exposed. A 5F sheath (Envoy; Cordis) was inserted, 500 U of heparin was injected, and a 5F catheter was then advanced into the brachiocephalic trunk from the aortic arch. DSA was performed through the guide catheter. A 0.027-inch ID (interior diameter) microcatheter (VIA-27; Sequent Medical) was advanced into the aneurysm lumen over a microguidewire (Transend-14; Stryker). Appropriately sized devices were placed in the aneurysm cavity for each device (n ¼ 12 [DL], n ¼ 12 [SL], n ¼ 12 [SLS]). DSA was performed immediately following device placement. No animals received antiplatelet therapy during the course of the study.

Follow-Up and Euthanasia
Angiographic evaluation was completed immediately after device implantation and again at follow-up. The cohort was divided into 2 subgroups preselected for euthanasia at 3 months (n ¼ 6 [DL], n ¼ 6 [SL], n ¼ 6 [SLS]) and 12 months (n ¼ 6 [DL], n ¼ 6 [SL], n ¼ 6 [SLS]), respectively, by intravenous injection with a lethal dose of pentobarbital through the ear vein. Following euthanasia, aneurysm tissue was harvested and placed in a 10% formalin solution.

Aneurysm Grading and WSM Assessment
The degree of angiographic aneurysm occlusion at follow-up was graded as follows: grade I, complete; grade II, complete but with recess filling; grade III, residual neck; or grade IV, residual aneurysm. 23,24 Two reviewers independently evaluated the angiographic occlusion, and disagreements were resolved by a third reviewer. Angiographic occlusion outcome was dichotomized into either complete occlusion (grade I or II) or incomplete occlusion (grade III or IV). Grades I, II, and III were considered adequate. The distance between proximal and distal device markers was measured on unsubtracted angiographic images. WSM was defined as a change in distance (#À10% to $10%) between markers at follow-up compared with immediate postdevice deployment.

Histopathologic Processing and Analysis
A histopathologist who was blinded to the angiographic results did the processing and analysis for healing evaluation. Aneurysm samples were processed at 1000-mm intervals in a coronal orientation, permitting long-axis sectioning of the aneurysm neck, with use of an IsoMet Low Speed Saw (Buehler). After the device segments were removed under a dissecting microscope, the samples were then re-embedded in paraffin, sectioned at 4 mm, and stained with hematoxylin-eosin.
Histologic healing of aneurysms was assessed using an ordinal scale. 25 The extent of blood clot organization within the aneurysm and the neointimal coverage of the aneurysm neck were evaluated. Tissue compaction, seen as a concave surface of clot tissue toward the aneurysm dome, was also analyzed. The degree of inflammation within each aneurysm was scored as 0 (no inflammatory cell infiltration); 1 (minimal or mild: scant, scattered inflammatory cell infiltration); 2 (moderate: patchy-butlocalized or limited inflammatory cell infiltration); or 3 (marked: attenuated, diffuse inflammatory cell infiltration).
Masson trichrome staining was performed to evaluate collagen deposition within the aneurysm; collagen deposition within the aneurysm dome was segmented and quantified using the previously described method. 26 The fibrosis ratio (total area of fibrosis within the aneurysmal cavity divided by the total area of the aneurysmal cavity) was calculated for each aneurysm.
Immunohistochemical staining of smooth muscle cells in tissue sections was performed with smooth muscle actin (SMA). 27 The SMA-positive area was segmented and quantified with the same method used for collagen deposition as described above.

Statistical Analysis
Continuous variables were described as mean [SD] and compared using a Student t test. Categoric variables were presented as number (percentage) and compared using the Fisher exact test. The correlation between WSM and aneurysm occlusion was assessed by the Spearman rank correlation. The correlation between WSM and aneurysm geometries, histologic healing, fibrosis, and smooth muscle actin levels was evaluated by simple linear regression. A P value , .05 was considered statistically significant.

Angiographic Findings
All aneurysms were implanted with appropriately sized devices. No morbidity or mortality was observed throughout the study.

WEB-Shape Modification
WSM was observed in 22 (61%) aneurysms (Fig 1), of which half (5 with complete occlusion, 6 with incomplete occlusion) were in the 3-month group and the remaining half (4 with complete occlusion, 7 with incomplete occlusion) were in the 12-month group ( Table 2). WSM was not statistically associated with aneurysm neck width or height (Fig 2). There were no significant differences in occlusion grades in aneurysms with or without WSM (9 with complete occlusion, 13 with incomplete occlusion versus 7 with complete occlusion and 7 with incomplete occlusion; P ¼ .73) (Fig  1). Furthermore, no correlation was found between the percentage of WSM and the occlusion grade. Most interesting, SLS had significantly less WSM (25%, 3/12) compared with the DL (83%, 10/12; P ¼ .012) and SL (75%, 9/12; P ¼ .039) devices.

Histologic Findings
The mean histologic healing score was not significantly different between the 3-and 12-month groups (5.6 [SD, 2.6] versus 4.2 [SD, 2.5]; P ¼ .11), but it was significantly higher in aneurysms with complete occlusion compared with those with incomplete occlusion (6.39 [SD, 2.07] versus 2.80 [SD, 1.75]). Histologic evaluation of aneurysm sacs showed a combination of loose connective tissue and thrombus (unorganized, organized, and poorly organized).    (Fig 3). The mean histologic healing score, inflammation score, fibrosis percentage, and SMA percentage , respectively, in aneurysms without WSM. There were no statistically significant differences in histologic healing, inflammation, fibrosis, and SMA between aneurysms with and without WSM. However, the percentage of WSM was moderately correlated with both fibrosis (r ¼ 0.37, P ¼ .02) and SMA levels (r ¼ 0.36, P ¼ .032) (Fig 4). WSM was not significantly correlated with either total histologic healing or inflammation.

DISCUSSION
Our study, which aimed to delineate the underlying mechanisms of WSM in WEB devices, demonstrates that WSM does not correlate with angiographic aneurysm occlusion or total histologic healing outcomes in the rabbit aneurysm model. However, WSM is positively associated with collagen and SMA levels, supporting the hypothesis that WSM is likely more related to aneurysm healing than external (eg, hemodynamic) compression alone.
The underlying mechanisms of WSM or compression could be multifaceted (eg, device size and construction, exact anatomic  configuration of aneurysm geometry and surrounding vessels, parent artery, aneurysm neck). Computational fluid dynamics simulations have shown that WEB device compression was positively correlated with computational fluid dynamics-derived inflow into the aneurysm. 28 In endoluminal FDs, pore density has been positively associated with aneurysm occlusion. Pore density at the proximal WEB device marker is relatively high compared with that in endoluminal devices, 29,30 and it is highly unlikely that device compression would lead to reduced pore density at the neck. Rouchaud et al 31 demonstrated, in coiled aneurysms, that a higher level of collagen in the aneurysm dome could trigger coil retraction from the neck orifice into the aneurysm cavity, resulting in aneurysm recurrence. Our findings suggest that WEB devices could behave like coils in the vascular microenvironment and result in device compression, and the contractile properties of connective tissue components in WEB-implanted aneurysms could lead to WSM. Proper endothelial cell growth across the pores of FDs in the aneurysm neck is also considered critical for aneurysm occlusion, in addition to the change in blood hemodynamics, for the mechanistic action of FDs. Endoluminal FDs placed in the healthy parent artery act as scaffolds for endothelial cell migration and neointima formation. In contrast, intrasaccular FDs are implanted in the aneurysm cavity, which represent a nonfunctional endothelium and a smooth-muscle layer and could explain the lack of neointimal coverage at the neck.
The observed WSM rate of 62% in the rabbit model is similar to that reported in clinical studies for the WEB device. [32][33][34][35][36] Although DL WEB devices had higher adequate occlusion rates compared with SL devices in this study, the DL devices also demonstrated higher WSM. Given the small sample sizes and the difference of a single result providing statistical significance, taken together, these results suggest that WSM is not well-correlated with either device type.
Our study has several limitations. First, the variation in the angiographic working projection angle resulted in inaccurate measurements of the distance between device markers. Second, we arbitrarily defined 10% of the device as a threshold for WSM.
A high stringent limit may provide different results. Third, we did not evaluate numerous other factors, including mechanical and hemodynamic factors, which could influence the device shape changes.

CONCLUSIONS
WSM is not associated with incomplete aneurysm occlusion in the rabbit model but may be related to connective tissue formation and collagen deposition after WEB implantation.