Evaluation of FeMnN alloy bioresorbable flow diverting stents in the rabbit abdominal aorta

Abstract

Flow diverting stents are braided, metallic endoluminal devices widely used to treat intracranial aneurysms. Bioresorbable flow diverters (BRFDs) are gaining traction as the next generation of flow diverter technology. BRFDs aim to occlude and heal the aneurysm before safely dissolving into the body, mitigating or eliminating complications associated with the permanent presence of conventional flow diverters such as thromboembolism and stenosis. Additional putative advantages of a BRFD include a reduction in metal induced medical imaging artifacts, a restoration of physiological vasoreactivity, and allowing physicians to re-access the aneurysm if an additional procedure is required. In this current study, iron-manganese-nitrogen (FeMnN) alloy BRFDs and permanent control FDs composed of an industry standard Cobalt-Nickel-Chromium alloy were deployed in the rabbit aorta. Micr°CT and SEM corrosion analysis determined the FeMnN wire volumes and cross-sectional areas had reduced approximately 85 % and 95 % after 3- and 6-months implantation duration, respectively. Histological analysis demonstrated the BRFDs exhibited suitable biocompatibility, with no cases of in-stent thrombosis, clinically significant stenosis, or adverse tissue responses observed. Immunohistochemistry revealed the neointimas surrounding the BRFDs featured a confluent endothelium covering several layers of smooth muscle cells, with macrophages adjacent to the device wires. The macrophages were able to penetrate the corrosion product and were observed transporting corrosion products away from the implant site. This current work provides primary in vivo corrosion and biocompatibility data to the field for FeMn alloys, which we feel will stimulate and inform the design of next-generation bioresorbable endovascular devices.

Headnote

ARTICLE INFO

Keywords:

Bioresorbable

Biodegradable

Flow diverter

Stent

Rabbit aorta

ABSTRACT

Flow diverting stents are braided, metallic endoluminal devices widely used to treat intracranial aneurysms. Bioresorbable flow diverters (BRFDs) are gaining traction as the next generation of flow diverter technology. BRFDs aim to occlude and heal the aneurysm before safely dissolving into the body, mitigating or eliminating complications associated with the permanent presence of conventional flow diverters such as thromboembolism and stenosis. Additional putative advantages of a BRFD include a reduction in metal induced medical imaging artifacts, a restoration of physiological vasoreactivity, and allowing physicians to re-access the aneurysm if an additional procedure is required. In this current study, iron-manganese-nitrogen (FeMnN) alloy BRFDs and permanent control FDs composed of an industry standard Cobalt-Nickel-Chromium alloy were deployed in the rabbit aorta. Micr°CT and SEM corrosion analysis determined the FeMnN wire volumes and cross-sectional areas had reduced approximately 85 % and 95 % after 3- and 6-months implantation duration, respectively. Histological analysis demonstrated the BRFDs exhibited suitable biocompatibility, with no cases of in-stent thrombosis, clinically significant stenosis, or adverse tissue responses observed. Immunohistochemistry revealed the neointimas surrounding the BRFDs featured a confluent endothelium covering several layers of smooth muscle cells, with macrophages adjacent to the device wires. The macrophages were able to penetrate the corrosion product and were observed transporting corrosion products away from the implant site. This current work provides primary in vivo corrosion and biocompatibility data to the field for FeMn alloys, which we feel will stimulate and inform the design of next-generation bioresorbable endovascular devices.

Abbreviations: BRFD, bioresorbable flow diverter; FD, flow diverter; DSA, digital subtraction angiography; EDS, energy dispersive X-ray spectroscopy; a-SMA, alpha-smooth muscle actin; CSA, cross sectional area; IPR, interpercentile range; BSE, back-scattering electron microscopy.

(ProQuest: ... denotes formulae omitted.)

1. Introduction

Metallic bioresorbable endovascular stents were first proposed in the literature over two decades ago [1]. These devices aim to serve their temporary mechanical function long enough to treat their intended vascular condition and allow their host vessel to remodel and heal. They then aim to completely dissolve into the body, mitigating or eliminating long term complications associated with the permanent presence of conventional stents such as device induced thromboembolism and the foreign body response: associated chronic inflammation, neointimal hyperplasia, and ultimately stenosis [2,3]. Furthermore, a bioresorbable stent may reduce metal induced medical imaging artifacts [4], allow for the restoration of physiological vasomotor function in the stented vessel [5], provide endovascular solutions to pediatric applications where the vessel is growing [6-8], and may open up more options to clinicians if retreatment of the stented vessel is required. Magnesium (Mg)-based bioresorbable coronary stents now have achieved market approval in Europe with promising results in their ongoing post market trial [9]. Iron (Fe)-based bioresorbable stents are currently in clinical trials for pediatric, peripheral, and coronary artery applications and so far have demonstrated promising safety profiles out to 26-month follow ups [10].

Bioresorbable flow diverting stents (BRFDs) are starting to gain traction in the preclinical literature as another application where a bioresorbable endovascular device may be advantageous [11]. Flow diverters (FDs) are specialized stents composed of a high density of braided wires used to treat intracranial aneurysms [12]. FDs are deployed in the parent artery over the aneurysm neck and function by diverting some blood flow away from the aneurysm sac. This results in the slowing or stagnation of blood flow and the formation of a thrombus plug within the aneurysm sac. The device then also acts as a bridge for tissue to grow over the aneurysm neck, ultimately occluding the aneurysm from blood flow [13]. FDs have demonstrated high aneurysm occlusion rates in the clinic and consequently are growing in popularity [14]. In patients, aneurysms treated with FDs typically take ~6-12 months to occlude and heal [15]. However, all market-approved FDs are composed of permanent metals such as cobalt-nickel-chromium (CoNiCr) and nitinol alloys that will remain in the patient for the duration of their life [16]. Similar to permanent cardiac stents, the permanent presence of FDs exacerbates complications such as device induced thromboembolism [15,17], stenosis [18-20], occlusion of adjacent branching arteries [14], and metal induced imaging artifacts [21,22]. The ideal BRFD would remain long enough for the aneurysm to heal, prior to safely dissolving into the body, ameliorating these complications.

Our group recently prototyped BRFDs composed primarily of Mg and Fe-manganese(Mn)-nitrogen(N) alloys. Benchtop tests concluded that the higher strength and slower corrosion rate of FeMnN relative to Mg alloys makes them more suitable for the BRFD application [23]. The Fe-based BRFD prototypes had a total wire count, individual wire diameter, and overall device radial strength that was comparable to FDA approved FDs [23]. FeMn alloys were first proposed as a candidate bioresorbable stent material in 2008 [24]. They offer a comparable yield strength to medical grade 316L stainless steel and exhibit an austenitic crystal structure, resulting in non-ferromagnetic properties that are critical for MRI safety considerations [24,25]. We previously demonstrated the MRI compatibility of our FeMnN based BRFDs [26]. Furthermore, FeMn alloys have demonstrated suitable biocompatibility to vascular cell types in vitro [27]. In the current work, FeMnN based BRFDs and control permanent FDs were deployed in the rabbit aorta for 3 or 6 months. The in vivo corrosion behavior of the BRFDs was characterized over time. The biological response to BRFD corrosion was investigated and compared to the control FDs.

2. Methods

2.1. Devices

The control FDs and BRFDs used in this study were both composed of 48 wires with an individual wire diameter of ~25 pm. The wires were braided at ~60° around a 4.75 mm diameter mandrel and heat treated. All devices used in the study were cut to 7 mm length. The control FDs were constructed from drawn filled tubing wires, with an outer shell of permanent industry standard CoNiCr alloy and an inner filament of Pt to impart radiopacity. The BRFDs used in this study have been previously described in detail [23]. The BRFDs contained 36 bioresorbable FeMn-nitrogen (№ alloy wires (35 % Mn, 0.15 % N, balance Fe, by wt%). This specific alloy was selected because the 35 % Mn alloying addition to Fe has been previously shown to produce a material with an elastic modulus and yield strength comparable to industry standard 316L stainless steel, a non-ferromagnetic austenitic crystal structure, and suitable biocompatibility to vascular cell types in vitro [25,27]. The N addition was added in attempt to increase the uniformity of corrosion [28]. The remaining 12 wires were permanent tantalum (Ta), included to provide radiopacity, coated by a ~5 pm layer of polyimide to prevent galvanic corrosion at FeMnN/Ta wire interfaces. All wire components for both devices were manufactured by Fort Wayne Metals. The BRFDs were sonicated in 15-20 % citric acid for 5 min to remove the thin surface oxide layer and stored in high purity ethanol until use, as previously described [27].

2.2. Rabbit procedures

All animal experiments were approved by our local Institutional Animal Care and Use Committee. Thirteen female New Zealand white rabbits, weighing approximately 2.5-3 kg, were used in this study. The rabbits received a daily dual antiplatelet therapy of 10 mg/kg of aspirin and clopidogrel, starting 2 days before and ending 30 days after the FD deployment procedure. On the day of the procedure, anesthesia was induced with an intramuscular injection of ketamine and xylazine. The rabbit was intubated, and anesthesia was maintained with 2-3% isoflurane.

A cut down was made to the right femoral artery and vascular access was achieved with a 5F introducer sheath. A guidewire and catheter were advanced into the aorta under fluoroscopic guidance. Digital subtraction angiography (DSA) was performed to assess the anatomy of the aorta before device deployment. A control FD and BRFD were then deployed in the abdominal aorta of each rabbit using a custom delivery catheter system. The cranial/caudal order of the control FD and BRFD was alternated for each rabbit. Balloon angioplasty was used for both devices when necessary to ensure suitable wall apposition. A DSA was performed to assess the device placement and patency immediately following deployment.

The rabbits survived for either 3 (n = 7 rabbits) or 6 (n = 6 rabbits) months after the FD deployment procedure. A DSA was performed immediately prior to sacrifice to assess device patency. The rabbits were then injected with a lethal dose of pentobarbital. The device containing segments of the aorta were immediately excised and rinsed with heparinized saline to remove postmortem clotting. The BRFDs were then cut in half. The distal half of the BRFD and the control FD were stored in formalin until processing for histological analysis. The proximal half of the BRFD was dehydrated using one 5-min wash of 70 % ethanol, three 5-min washes with 100 % ethanol, and then was air dried overnight before storing in a desiccating environment until micr°CT analysis.

2.3. Angiographic outcomes

Angiograms of the control FDs and BRFDs immediately after deployment and at the 3 or 6 month follow up were compared to assess device migration. Devices that were no longer located where they were initially deployed by the follow up were considered to have migrated and the device migration rate was reported for both device types. Migrated devices were excluded from the angiographic stenosis and histological analysis. Angiographic % stenosis of the control FDs and BRFDs was assessed using the 3 or 6 month follow up DSAs. Two blinded reviewers measured the diameter at the proximal, middle, and distal segment of each device, as well as the native aorta proximal and distal to each device. The device measurements and native aorta measurements were averaged, and % stenosis was calculated using Equation (1) below. The % stenosis calculations from the two blinded reviewers were averaged for each device.

...Equation 1

2.4. Micr°CT

Corrosion of the BRFDs was first characterized with a Bruker SkyScan 1276 micr°CT scanner using methods previously described [23]. The scans were taken with a source voltage of 55 kV and current of 200 pA with a 0.25 mm aluminum filter. Projection images were taken with a 0.48 s exposure time every 0.2° for 360°. The image data were reconstructed using a modified Feldkamp conebeam algorithm (NRecon, Bruker) to produce image data with a voxel resolution of 6.5 pm. Every scan of a BRFD implanted for 3 or 6 months included a scan of a non-implanted baseline BRFD. Thresholding values were determined using the baseline BRFD based on the manufacturer's specifications of wire diameter. The micr°CT scans were used to generate 3D renderings to qualitatively evaluate changes to the BRFD structures and quantify BRFD volume reductions over time. To account for differences in device lengths, FeMnN volume reduction was presented as a ratio of FeMnN/Ta volume, normalized such that the average of the baseline BRFDs started at 100 %.

2.5. Scanning electron microscopy

After micr°CT, the samples were processed for scanning electron microscopy. Samples were stored in absolute ethanol vials and desiccated at room temperature. Each sample was affixed with super glue to the bottom of a plastic mold then embedded with two-part epoxy resin. Cross-sections were prepared by standard metallographic procedures which included an initial grind with coarse silicon carbide paper and polished with intermediate decreasing steps of glycol-based polycrystalline diamond suspension and glycol-based lubricant on microfiber cloths. The polished cross-sections were then sputter coated with a thin layer of gold to increase conductivity. Cross sections from regions of wire adjacent to where the device was cut in half for histology were excluded from analysis. Samples were then imaged with a JEOL (Peabody, MA) JSM-6510V scanning electron microscope (SEM), utilizing the back-scattering electron microscopy (BSE) detector. SEM was carried out with a tungsten filament, at an accelerating voltage of 15 kV acceleration voltage, working distance of 15 mm, and a 500-800x nominal magnification. In addition, energy dispersive X-ray spectroscopy (EDS) was taken for representative cross-sections. The crosssection analysis of the remaining metal of individual braids was performed in Image J software. Area of the remaining metal for each sample was estimated using the threshold adjustment and magic wand tool on the back scattered electron images.

2.6. Histology staining

After at least 24 h of formalin fixation, the distal segments of the BRFD and control FD containing aortas were paraffin embedded and cut into ~1 mm thick segments using an Isomet low-speed saw with a diamond blade. The device wires were then removed from the segments with microforceps under a dissection microscope using methods previously described [29]. The segments were then re-embedded in paraffin and cut into 4 ит thick sections using a standard microtome and adhered to glass microscope slides. Hematoxylin and Eosin (H&E) staining was performed using a conventional approach. Briefly, the sections were deparaffinized and hydrated. They were then stained with hematoxylin for 5 min, differentiated in 1 % acid-alcohol, and rinsed in tap water. The sections were then stained with eosin for 1 min, dehydrated in grated alcohol solutions, and cleared in xylene before mounting a coverslip.

Immunohistochemistry staining of the 4 um thick sections was performed with an automated slide stainer (Bond Max, Leica). The slides were stained with a-smooth muscle actin (x-SMA)(Dako, Monoclonal mouse anti-Human Smooth Muscle Actin, clone 1A4, diluted 1:200), RAM11(Dako, Monoclonal mouse anti-Rabbit macrophages Clone RAM11, diluted 1:200), or CD31 (Dako, Monoclonal mouse anti-human CD31, Endothelial cells clone JC70A diluted 1:30) primary antibodies to identify smooth muscle cells, macrophages, and endothelial cells, respectively. A citrate-based epitope retrieval solution (Leica Biosystems AR9961) associated with tissue heating was applied for the RAM11 and CD31 samples for 2 min. No epitope retrieval was used for the a-SMA samples. The BOND Polymer Refine Red detection system (Leica Biosystems DS9390) was used to mark the primary antibodies with a red stain and apply a hematoxylin counterstain to background tissue. The samples were then rinsed in water, dehydrated in grated alcohol solutions, and cleared in xylene before mounting a coverslip. All stained histology samples were imaged at 20x normal magnification using a slide scanner (Miotic Easy Scan Pro, Miotic Digital Pathology, San Francisco, CA).

2.7. Neointimal morphometrics

Neointimal morphometrics were taken from the H&E stained samples using methods previously described in detail [30]. Three H&E-stained cross sections, all from a different region along the device length, were used for measurements and averaged together to represent each device containing sample. Neointimal thickness was defined as the distance between the luminal facing surface of each wire to the luminal border of the neointima. For every tissue section, the neointimal thickness for every wire around the circumference of the device was measured and averaged together to represent the section. Neointimal area was defined as the area of tissue that had developed on the luminal side of the internal elastic lamina. Histological % stenosis was defined as the % of area luminal to the internal elastic lamina occupied by neointimal area.

2.8. Serum analysis

Approximately 3 mL of blood were collected from the ear artery of each rabbit immediately before the device deployment and end study procedures. The blood was collected into trace element serum tubes (BD 368380) and was allowed to clot for 30 min at room temperature. The tubes were then centrifuged at 2000 xG for 10 min at 4 °C. The serum was collected and stored in trace metal free tubes (Sarstedt 63.550.121) at -80 °C until analysis.

The concentration of Fe and Mn in the serum was measured using triple quadrupole inductively coupled plasma mass-spectrometry and validated for Fe and Mn, respectively, using a standard curve. The lower limit of quantitation for Fe was 6.25 ng/mL and the analytical measurement range was 6.25-40,000 ng/mL. The lower limit of quantitation for Mn was 0.2 ng/mL and the analytical measurement range was 0.2-100 ng/mL. For both methods, 100uL of sample was diluted 1x25 with a slightly basic diluent. Quality control materials were analyzed with the run to ensure accuracy.

2.9. Statistics

Angiographic % stenosis, micr°CT FeMnN volume reduction, neointimal morphometrics, and Fe and Mn serum concentrations are presented as the mean + the standard deviation. For angiographic % stenosis and the neointimal morphometrics, a one-way anova with a Sidak's post hoc test was used to make comparisons between control FDs and BRFDs at 3 and 6 months, as well as comparisons of the control FD and BRFD groups between 3 and 6 months. For micr°CT FeMnN volume reduction, a one-way anova with a Sidak's post hoc test was used to make comparisons between the 0- and 3-months, and 3- and 6-month time points. To evaluate SEM-based cross-sectional area (CSA) reduction over time, we used a linear quantile mixed model based on the asymmetric Laplace distribution. Predicted results for the 50th percentile and the 90th interpercentile range (IPR), representing the range from the 5th to 95th percentile, were extracted from the model results. To account for dependency between measurements from the same device within the same subject, random intercepts were estimated for each subject. Time was transformed as a factor and included as fixed covariate. Quantile regression was employed due to the irregular distribution of degradation for individual wires within a given device. For the serum analysis, paired t-tests were used to compare Fe or Mn concentrations in the serum before device deployment and at the 3 or 6 month follow up. For all statistical tests, p < 0.05 was considered statistically significant. The SEM CSA reduction analysis was performed using the "lgmm" package in R (version 1.5.8). All other statistics were performed in GraphPad Prism 10.

3. Results

3.1. Angiographic outcomes

All devices were successfully deployed in the abdominal aorta. For the control devices, balloon angioplasty was used to ensure proper wall apposition after deployment in 2/7 rabbits in the 3-month group and 1/ 6 rabbits in the 6-month group. For the BRFDs, balloon angioplasty was used for 3/7 rabbits in the 3-month group and 1/6 rabbits in the 6month group. Fig. 1A and B demonstrate a representative radiograph and associated DSA, respectively, of a control FD and BRFD deployed within the abdominal aorta. Fig. 1C demonstrates a representative DSA of the devices within the aorta at the 6-month follow up. In the 3-month group, 2/7 control FDs and 1/7 BRFDs had migrated by the follow up. The BRFD was never recovered and was consequently not included in the corrosion analysis. In the 6-month group, 0/6 control FDs and 2/6 BRFDs had migrated within the aorta. We attribute device migration to poor wall apposition after deployment. We speculate that device migration occurred soon after deployment, before the device became constrained and encapsulated in neointimal tissue, and before notable corrosion. The angiographic % stenosis is presented in Fig. 1D. The follow up DSA for one rabbit in the 3-month group was unavailable for analysis. There was no statistical significance between the control FDs and BRFDs at 3 months (18.0 + 5.3 % vs 13.9 + 5.2 %, р = 0.849) or 6 months (11.0 + 8.1 % vs 13.3 + 7.6 %, р = 0.978). Furthermore, there was no statistical significance in the control FDs (18.0 + 5.3 % vs 11.0 + 8.1 %, р = 0.434) or BRFDs (13.9 + 5.2 % vs 13.3 + 7.6 %, р > 0.999) between 3 and 6 months. № device had an angiographic % stenosis greater than 25 % at either time point.

3.2. Micr°CT

Representative micr°CT 3D renderings of BRFDs corroding over time are presented in Fig. 2A. Qualitatively, by 3 months, a majority of the FeMnN volume has been reduced. The FeMnN has almost entirely disappeared by 6 months, leaving only the permanent Ta wires behind. The implanted 3- and 6-month devices are approximately half the length of the non-implanted baseline control because the implants were cut in half, with the remaining half being used for histological analysis. Quantitative normalized FeMnN volume reduction over time is presented in Fig. 2B. The FeMnN volume had reduced by 81.2 + 10.7 % of its initial volume after 3 months and 93.5 + 7.2 % after 6 months. There was a statistically significant reduction in volume between baseline and 3 months (p < 0.0001) but not between 3 and 6 months (p = 0.133).

3.3. Scanning electron microscopy

A representative SEM micrograph of a BRFD deployed in the rabbit aorta for 3 months is presented in Fig. 3A. The polyimide coating was intact, as shown in Fig. 3A, in every analyzed Ta wire cross section. Three FeMnN wire cross sections are contained within the field of view that have corroded to varying degrees. One FeMnN wire is minimally degraded, the second partially degraded, and the third completely degraded and converted into a less electron dense oxide substance. This qualitatively demonstrates the high degree of non-uniformity in corrosion along the length of the FeMnN wires. The quantified FeMnN wire CSA reduction is presented in Fig. 3B and C. Fig. 3B presents the groupaveraged plot of normalized cross-sectional areas over time. The observed median value for normalized CSA reduction for a given wire was 93.9 % at 3 months and 100 % at 6 months. On a mixed effects quantile regression that accounted for subject-level dependencies and over- or under-sampling for a given subject, the predicted median value for normalized CSA reduction for a given wire was 85.0 % (90th IPR: 67.8 %-92.9 %) at 3 months and 96.7 % (90th IPR: 87.4 %-100 %) at 6 months. In other words, based on the predicted median values, approximately 15 % of the FeMnN wire CSA remained after 3 months, and 3.3 % remained after 6-months of implantation in the rabbit abdominal aorta. On average, the percentage of wires completely degraded for a given device was approximately 40 % for every 3-month interval (43.7 % at 3-months, 81.8 % at 6-months). Fig. 3C presents the plot of normalized cross-sectional areas over time for individual rabbits. Within a given device, there was considerable variability in degradation, with a heavily right-skewed distribution. At 3 months, between individual rabbits, the minimum wire CSA reduction ranged from 32.1 % to 69.7 %. By 6 months, between individual rabbits, the minimum wire CSA reduction ranged from 41.1 % to 90.7 %. Between individual rabbits, the percentage of analyzed wire cross sections completely degraded (i.e., without a metal core) ranged from 23.5 % to 55.8 % at 3 months compared to 64.0 %-97.0 % at 6 months.

SEM-EDS results for representative cross sections at 3 and 6 months are presented in Fig. 4. BSE images are shown for partially and fully degraded struts. The corrosion attack of individual struts is localized and heterogenous, propagating from the surface detailed by the BSE images of partially degraded struts. EDS spectra analysis detected Fe, Mn, O, P, Cl, Ca, and C. Comparatively, partially degraded struts at 3 and 6 months do not show distinct differences in element localization, yet fully degraded struts demonstrate some differences in element localization as time progresses. Fe signal is seen throughout the corrosion product footprint in fully degraded wire segments, with varying counts contingent on localization, indicating possibly a variety of stoichiometric Fe based species. Mn localization in the central converted product is remnant of the original wire footprint and is seemingly reduced in quantity from 3 to 6 months with and accompanied by an increase in Cl presence within the region at 6 months. Ca consistently localizes in the outermost portion of the product, while P remains more diffuse, with P avoiding co-localization with the Mn rich center of the fully degraded struts at any timepoint.

3.4. Hematoxylin and Eosin staining

Representative low magnification images of H&E-stained cross sections of control FDs and BRFDs deployed in the rabbit abdominal aorta for 3 or 6 months are presented in Fig. БА. The selected images in Fig. 5A represent the median cross sections in the histological % stenosis assessment for each device type and timepoint. In general, no excessive inflammation, neointimal hyperplasia, or local adverse tissue reactions were observed in response to the deployment of control FDs or BRFDs for 3 or 6 months. Fig. 5B demonstrates representative high magnification images of BRFDs deployed in the aorta for 3 or 6 months. At both time points: 1) Inflammatory cells were observed infiltrating the corrosion products at the tissue/implant interface; 2) Foreign body giant cells formed at the interface of many of the wires; and 3) Inflammatory cells were able to consume corrosion products and were observed transporting them away from the implant site. BRFD corrosion and the biological response appeared to be more developed in the 6-month group. The 6-month samples exhibited a higher volume of corrosion products and had more Fe-laden cells that were dispersed further away from the wire location. Supplemental Fig. 1 presents high magnification images of H&E-stained cross sections of control FDs and BRFDs deployed in the rabbit abdominal aorta for 3 or 6 months.

Supplemental Fig. 2A demonstrates low magnification images of H&E-stained cross sections of aortas implanted with control FDs and BRFDs for 3 and 6 months that represent the worst-case biological responses. The selected images correspond to the highest histological % stenosis for each device type and timepoint. Supplemental Fig. 2B demonstrates high magnification images of the worst case BRFD implants at 3 and 6 months. Qualitatively, the worst-case sections appeared to have a comparable degree of inflammation and a slightly higher degree of smooth muscle cell proliferation into the neointima relative to the representative sections in Fig. 5.

3.5. Neointimal morphometrics

The neointimal morphometrics, measured from the H&E-stained cross sections, are presented in Fig. 6. The general trend for the three neointimal morphometrics was that the BRFDs were significantly higher than the control FDs at both 3 and 6 months, however there was no difference for either device type over time. The control FDs had a significantly thinner neointimal thickness than the BRFDs at 3 (0.09 + 0.01 mm vs 0.16 + 0.03 mm, p = 0.037) and 6 (0.07 + 0.01 mm vs 0.21 + 0.08 mm, p < 0.001) months. The neointimal thickness was not significantly different for the control FDs (0.09 + 0.01 mm vs 0.07 + 0.01 mm, p = 0.810) or BRFDs (0.16 + 0.03 mm vs 0.21 + 0.08 mm, p = 0.251) between 3 and 6 months. The control FDs had smaller neointimal areas than the BRFDs that was nearly significantly different at 3 months (0.71 + 0.15 mm? vs 1.31 + 0.42 mm", p = 0.054) and significant at 6 months (0.48 + 0.13 mm? vs 1.63 + 0.62 mm", р < 0.001). The neointimal area was not significantly different for the control FDs (0.71 + 0.15 mm? vs 0.48 + 0.13 mm", р = 0.763) or BRFDs (1.31 + 0.42 mm? vs 1.63 + 0.62 mm", p = 0.563) between 3 and 6 months. The control FDs had a significantly lower histological % stenosis than the BRFDs at 3 (15.6 + 1.5 % vs 28.5 + 8.6 %, р = 0.037) and 6 (9.8 + 2.8 % vs 31.6 + 12.7 %, р < 0.001) months. The histological % stenosis was not significantly different for the control FDs (15.6 + 1.5 % vs 9.8 + 2.8 %, р = 0.611) or BRFDs (28.5 + 8.6 % vs 31.6 + 12.7 %, р = 0.945) between 3 and 6 months. Balloon angioplasty did not appear to have a notable effect on neointimal morphometrics when comparing ballooned versus non-ballooned devices within the same device type and endpoint groups. However, these comparisons were underpowered. The raw data is presented in Supplemental Table 1.

3.6. Immunohistochemistry

Representative high magnification images of aortic cross sections implanted with control FDs or BRFDs for 3 or 6 months stained with immunohistochemistry are presented in Fig. 7. For both device types and time points, a-SMA + smooth muscle cells made up the majority of the neointima. In general, there was slightly more smooth muscle cell proliferation into the neointima in the control FDs relative to the BRFDs, which likely contributed to the larger neointimal morphometrics in the BRFD groups. There was no observable difference in smooth muscle cell proliferation for both device types between the time points. At both 3 and 6 months, a few RAM11+ macrophages were observed at the wire/ tissue interfaces for the control FDs. There appeared to be more RAM11+ macrophages at the implant interfaces in the BRFD groups. RAM11+ cells were capable of infiltrating the corrosion products. The cells that contained corrosion products, first observed in H&E staining, appeared to be RAM11+ after immunohistochemistry. The qualitatively higher number of RAM11+ cells in the neointimas for the BRFDs relative to the control FDs also likely contributed to higher neointimal morphometric values for the BRFD groups. Qualitatively, the number of RAM11+ cells appeared to slightly increase from 3 to 6 months in the BRFD group. In the BRFDs, there was a clear distinction between regions of RAM11+ and a-SMA + cells, with the RAM11+ macrophages localizing around the wires, and a-SMA + smooth muscle cells localizing either luminal in the neointima or abluminal in the medial layer of the native aorta. For both device types and time points, a confluent CD31" endothelium was observed lining the luminal surface of the neointima, suggesting that the control FDs and BRFDs did not impair the development of the endothelium. Supplemental Fig. 3 presents additional high magnification CD31 stained cross sections of control FDs and BRFDs implanted in the aorta for 3 months, emphasizing the CD31" endothelium on the luminal surface of neointimas for both device types.

3.7. Serum analysis

The Fe and Mn serum concentrations are presented in Fig. 8. The first three rabbits in the study, all belonging to the 3-month group, were not included in the serum analysis. There was no statistically significant difference in the Fe serum concentration between before BRFD deployment and the 3-month endpoint (1804 + 302 ng/mL vs 2095 + 25 ng/mL, p = 0.153), the 6-month endpoint (2313 + 254 ng/mL vs 2204 + 155 ng/mL, p = 0.331), or pooled 3- and 6-month endpoints (2109 + 368 ng/mL vs 2160 + 129 ng/mL, p = 0.638). There was no statistically significant difference in the Mn serum concentration between before BRFD deployment and the 3-month endpoint (3.6 + 0.6 ng/mL vs 3.8 + 0.8 ng/mL, p = 0.391) and the pooled 3- and 6-month endpoints (3.8 + 0.6 ng/mL vs 3.6 + 0.6 ng/mL, p = 0.139). However, interestingly, the Mn serum concentration significantly decreased from before BRFD deployment to the 6-month endpoint (4.0 + 0.6 ng/ mL vs 3.4 + 0.6 ng/mL, p = 0.021).

4. Discussion

In this study, we evaluated the corrosion and biocompatibility of FeMnN BRFDs deployed in a rabbit aorta for 3 and 6 months. Micr°CT analysis determined the FeMnN wires had reduced by approximately 85 % and 95 % of their original volumes after 3 and 6 months in the rabbit aorta, respectively. We found high variability in the degree of corrosion along the length of the FeMnN wires. Indeed, at both time points, we observed some FeMnN wire cross sections that were completely corroded and others that were over 50 % intact. The BRFDs exhibited suitable biocompatibility, and although the BRFDs had significantly greater neointimal morphometrics than the control FDs at both time points, the degree of stenosis was not clinically significant, which is typically defined as >50 % stenosis in the FD application [31]. Furthermore, there was no statistical difference in angiographic % stenosis between the BRFDs and control FDs at 3 or 6 months. No in stent thrombosis or adverse tissue responses were observed in response to any control FD or BRFD implants. Immunohistochemistry revealed the neointimas surrounding the BRFDs featured a confluent CD31" endothelium covering several layers of smooth muscle cells, with macrophages adjacent to the device wires. The macrophages were able to penetrate the corrosion product and were observed transporting corrosion products away from the implant site, as previously reported in other bioresorbable Fe stent studies [32,33]. Inflammatory cells are known to activate smooth muscle cells to proliferate into the lumen and deposit matrix. We speculate that the dynamic implant/tissue interface of the BRFDs, and consequent sustained presence of macrophages acting to clear corrosion products, resulted in a slightly higher degree of smooth muscle cells proliferating into the neointima relative to the control FDs [34]. No accumulation of Fe or Mn was detected in the rabbit serum over time. These findings provide primary in vivo corrosion and biocompatibility information critical to the design of next generation FeMn alloy bioresorbable endovascular devices.

Historically, Fe-based bioresorbable materials have exhibited much slower in vivo corrosion rates relative to other bioresorbable metals. Peuster et al. investigated pure Fe stents (120 pm strut thickness) deployed in minipigs for up to a year and found that their stents were still mostly intact [32]. Lin et al. investigated FeN alloy stents (70 pm strut thickness) in a rabbit abdominal aorta for up to 53 months and concluded that they expect their device to completely corrode after 4-5 years [33]. Consequently, the slow corrosion rate of Fe based bioresorbable materials has been considered a limitation for stenting applications and many strategies have been applied to increase it [35]. We hypothesized that the previously reported slow corrosion rate of Fe-based materials would be advantageous in the BRFD application, where 1) device wires/struts need to be ~25 pm in diameter to match market approved FDs; and 2) we believe the device lifetime is ~12 months to properly heal the aneurysm. The corrosion rate of the BRFD in this study was faster than anticipated and is likely too fast to allow the device to properly treat an aneurysm. Furthermore, the non-uniformity of corrosion along the length of the FeMnN wires may result in premature mechanical failure of the device. We believe a material with a slower, more uniform corrosion rate is more suitable for the BRFD application. Different alloying compositions, thermal treatments, or surface treatments may be applied to tune the corrosion behavior [28]. Additionally, the incorporation of slower degrading metals inside FeMnN to form composite wires may further extend the functional lifetime [36].

Lu et al. recently proposed the corrosion mechanism for bioresorbable FeMnN alloys [28]. Broadly, in the presence of oxygen and water, as Fe materials corrode Fe cations are produced at the local anode and electrons are consumed at the local cathode. Oxygen facilitates the corrosion process by accepting the electrons in the reduction reaction. Consequently, insufficient oxygen may limit the corrosion rate. Lu et al. proposed that N atoms, introduced from alloying, can also accept electrons from the anodic reaction of the Fe atoms, facilitating corrosion even if oxygen availability is insufficient, ultimately resulting in an increased corrosion rate. The hydrolysis of Fe ions produces H·, which accelerates the local acidic corrosion runway associated with non-uniform pitting corrosion [37]. N atoms introduced from alloying are expected to form ammonia during corrosion release, which can scavenge H·, mitigate the localized acidic corrosion runway, and ultimately result in more uniform corrosion. Indeed, Lu et al. reported that increasing № content from 0 to 0.6 % in a Fe30Mn (wt%) alloy increased the corrosion rate and uniformity when evaluated in vitro in Hank's solution [28]. Uniform corrosion of our FeMnN alloy was not conserved in vivo, however increasing the N content from 0.15 % may have resulted in more uniform corrosion.

Similar to other in vivo corrosion evaluations of pure Fe [38] and FeN alloy [33] bioresorbable endovascular implants, our SEM-EDS analysis revealed the corrosion products of our BRFDs were composed primarily of Fe and O (likely in the form of iron oxides and hydroxides [28,33]), as well as Ca located on the outermost layers. Ca and phosphate present in the physiological environment can result in the formation of calcium phosphate based compounds at the material-fluid interface [39]. The presence of Ca and phosphate may result in vascular calcification; however, we did not observe any cases. We observed the presence of Cl in the corrosion products of our BRFDs. Cl anions from the physiological environment may diffuse through the passivating oxide layer and can accelerate corrosion at the metal-oxide interface [39]. Mn hydroxide, the primary expected corrosion product of the Mn alloying addition, has a high solubility in physiological saline, and may have eluted away from the implant site [28]. This may explain the overall lack of Mn observed in the corrosion products of our BRFDs which continuously decreases from the fully degraded wire remnants from 3 to 6 months.

We previously evaluated corrosion of the BRFDs in vitro [23]. The BRFDs were deployed in a flow loop with Dulbecco's Modified Eagle Medium without the addition of protein, circulating at a flow rate of 0.5 mL/s at a 37 °C. The media volume and frequency of media changes were performed in accordance with the ASTM G31-72 standard. Volume reduction over time was assessed using the same micr°CT approach described in the current study. We found that the BRFD volume had reduced -35 % after 3 months, which is more than 2 times slower than the -85 % reduction after 3-months in vivo observed in the current study. This is opposite of what is typically observed in the bioresorbable metals literature, and what we have previously observed for FeMnN implants [36], where the in vitro corrosion rate is expected to be faster than the in vivo rate. The faster observed in vivo rate could be due to several factors. First, multiple in vitro studies have demonstrated that the addition of proteins increases the corrosion activity of FeMn alloys [36, 40]. While the mechanism behind this is unclear, the lack of proteins in our in vitro model and abundance of proteins in vivo may have contributed to the faster in vivo corrosion rate. Second, the flow rate in the rabbit aorta is on the order of ~3 mL/s [41]. The faster in vivo flow rate may have accelerated corrosion. Third, in vivo, the FeMnN wires were exposed to an inflammatory response, which can lower the pH of the local environment and accelerate corrosion [42]. A previous study by Hou et al. reported that the corrosion of Mg screws implanted in rat femurs was accelerated when titanium screws were implanted 5-10 mm away, with the greatest affect occurring at 5 mm [43]. They speculated that abundant blood vessels in the periosteum between the Mg and titanium screws facilitated the completion of the galvanic cell, increasing the rate of Mg corrosion. We do not believe this phenomenon occurred in the current study between control FD and BRFD deployments because the devices were placed over 15 mm apart. Furthermore, we speculate that great vessels would not have equivalent ion diffusion behavior as bone, possibly limiting effective charge transfer and subsequently shortening the distance required for contactless galvanic corrosion.

The biocompatibility of actively corroding Fe based materials has received scrutiny due to the potential production of reactive oxygen species during Fenton reactions and their impact on resident vascular cell types. In the Fenton reaction, ferrous Fe oxidizes hydrogen peroxide to form hydroxyl radicals [44]. Hydroxyl radicals are considered potent proinflammatory reactive oxygen species [45]. Furthermore, studies have shown that hydroxyl radicals, formed from Fe corrosion, are cytotoxic to endothelial and smooth muscle cells in vitro [46] and may impair endothelial function ex vivo [47]. In our in vivo model, the corrosion of the FeMnN wires did not appear to have any toxic effects on smooth muscle cells or endothelial cells. Immunohistochemistry revealed that the neointimas surrounding the FeMnN wires were composed primarily of organized concentric layers of smooth muscle cells and a confluent CD31" endothelium on the luminal surface. We did observe more macrophages at the FeMnN wire interfaces relative to the control wires, which we attribute to FeMnN corrosion and the dynamic tissue/wire interface of the FeMnN wires. It is thought that insoluble corrosion products are primarily removed by cellular uptake [3,48]. Peuster et al. first observed macrophages phagocytosing Fe corrosion products and clearing them through the lymphatic system [32]. This clearance mechanism was later corroborated by Zheng et al. [49]. In this study, we also observed macrophages that appeared to contain corrosion products, corroborating prior findings in other studies.

A limitation of this study is that we did not include a long enough time point to observe complete bioresorption of the FeMnN wires. Complete bioresorption of the FeMnN wires may have resulted in a resolution of the inflammatory response and a reduction in neointimal morphometrics over time. We did not observe any cases of thrombosis of either device type. The use of young healthy rabbits with dual antiplatelet therapy results in a low risk of device induced thrombosis, making it difficult to study differences in thrombosis between devices. Another limitation of this study was the use of balloon angioplasty for some of the devices to ensure proper wall apposition during their deployment. Excessive ballooning may induce injury to the blood vessel wall, resulting in higher degrees of inflammation and stenosis [50]. Therefore, balloon angioplasty may have been a confounding variable in our angiographic % stenosis and neointimal morphometric comparisons. However, in our study, the number of devices requiring balloon angioplasty was relatively consistent between groups, with 3/13 control FDs and 4/13 BRFDs requiring ballooning. Furthermore, there were no notable trends suggesting increased neointimal morphometrics in response to balloon angioplasty when comparing ballooned vs non-ballooned devices of the same type at the same end point. Another limitation of the study is that we did not evaluate the efficacy of the BREDs to treat aneurysms. Future work is to evaluate the BRFDs in the rabbit elastase induced aneurysm model, a standard preclinical model accepted by the FDA for evaluating FDs [51,52]. Finally, other than our serum analysis, we did not investigate any systemic toxicity of the BRFD implants. Future work will focus on determining if there is any systemic toxicity resulting from bioresorbable FeMnN implants by looking for circulating markers of toxicity and examining organs such as the kidneys and liver.

5. Conclusion

We evaluated the corrosion and biocompatibility of FeMnN BRFDs deployed in the rabbit aorta for 3 and 6 months. The FeMnN wire volume had reduced approximately 85 % and 95 % after 3- and 6-months implantation, respectively. The degree of corrosion was highly variable along the length of the FeMnN wires. The BRFDs exhibited suitable biocompatibility, with no cases of in stent thrombosis, clinically significant stenosis, or adverse tissue responses observed. This work provides primary in vivo corrosion and biocompatibility data to the field for FeMn alloys, which we feel will stimulate and inform the design of next generation bioresorbable endovascular devices.

CRediT authorship contribution statement

Alexander A. Oliver: Writing - original draft, Methodology, Funding acquisition, Formal analysis, Data curation, Conceptualization. Cem Bilgin: Methodology, Data curation. Mitchell L. Connon: Writing - review & editing, Writing - original draft, Methodology, Formal analysis, Data curation. Andrew J. Vercnocke: Writing - review ¿ editing, Methodology, Formal analysis, Data curation. Esref A. Bayraktar: Writing - review & editing, Data curation. Jonathan Cortese: Writing - review ¿ editing, Data curation. Daying Dai: Writing - review & editing, Methodology, Formal analysis, Data curation. Yong Hong Ding: Data curation. Sarah A. Erdahl: Writing - review & editing, Methodology, Formal analysis, Data curation. John Pederson: Writing - review & editing, Writing - original draft, Methodology, Formal analysis. Kent D. Carlson: Writing - review & editing. Adam J. Griebel: Writing - review & editing, Resources, Conceptualization. Jeremy E. Schaffer: Writing - review ¿ editing, Resources, Conceptualization. Dan Dragomir-Daescu: Supervision. Ramanathan Kadirvel: Writing - review & editing, Supervision, Methodology, Funding acquisition. Roger J. Guillory: Writing - review & editing, Supervision, Methodology, Data curation, Conceptualization. David Е. Kallmes: Writing - review & editing, Supervision, Methodology, Funding acquisition, Data curation, Conceptualization.

Data statement

Raw data is available upon request. Contact the corresponding author.

Ethics approval and consent to participate

The submitted manuscript does not include any clinical data or involve the use of human subjects. All animal work described in the submitted manuscript was approved by the Mayo Clinic Institutional Animal Care and Use Committee under protocol A00006455. The authors confirm compliance with all relevant ethical regulations.

Declaration of competing interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Ramanathan Kadirvel reports a relationship with Cerenovus Inc., Medtronic, Endovascular Engineering, Insera Therapeutics, Frontier Bio, Sensome Inc, Endomimetics, Ancure LLC, Neurogami Medical, MIVI Biosciences, Monarch Biosciences, Stryker Inc, Piraeus Medical, and Bionaut Labs. that includes: funding grants. David F. Kallmes reports a relationship with Cerenovus, Sensome, Neurogami Medical, Insera Therapeutics, Medtronic, Microvention, Balt, Monarch Biosciences, Brainomix, MiVi, Stryker that includes: funding grants. David F. Kallmes reports a relationship with Medtronic, Nested Knowledge, Superior Medical Experts, Marblehead Medical, Conway Medical, Monarch Biosciences, and Piraeus Medical. that includes: equity or stocks. John Pederson is currently employed by Superior Medical Experts. Adam J. Griebel and Jeremy E. Schaffer are currently employed by Fort Wayne Metals. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

The graphical abstract was created using BioRender. The authors would like to acknowledge the use of the Mayo Clinic X-ray Imaging Core. This work was partially funded by National Institutes of Health grant #R01 NS076491. Alexander Oliver is supported by American Heart Association grant 23PRE1012781.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi. org/10.1016/j.bioactmat.2025.01.039.

Sidebar

Peer review under the responsibility of KeAi Communications Co., Ltd.

Received 4 November 2024; Received in revised form 28 January 2025; Accepted 30 January 2025

Available online 12 February 2025

Footnote

* Corresponding author. Biomedical Engineering and Physiology, Mayo Clinic Graduate School of Biomedical Sciences, 200 1st St SW, Rochester, MN, 55905, USA

E-mail addresses: [email protected], [email protected] (A.A. Oliver), [email protected] (C. Bilgin), [email protected] (M.L. Connon), Vercnocke. [email protected] (A.J. Vercnocke), [email protected] (Е.А. Bayraktar), [email protected] (J. Cortese), [email protected] (D. Dai), [email protected] (Y.H. Ding), [email protected] (S.A. Erdahl), [email protected] (J. Pederson), [email protected] (K.D. Carlson), [email protected] (A.J. Griebel), jeremy [email protected] (J.E. Schaffer), [email protected] (D. Dragomir-Daescu), kadir@mayo. edu (В. Kadirvel), [email protected] (R.J. Guillory), [email protected] (D.F. Kallmes).

References

References

[1] M. Peuster, P. Wohlsein, M. Brügmann, M. Ehlerding, K. Seidler, C. Fink, H. Brauer, A. Fischer, G. Hausdorf, A novel approach to temporary stenting: degradable cardiovascular stents produced from corrodible metal-results 6-18 months after implantation into New Zealand white rabbits, Heart 86 (5) (2001) 563-569.

[2] C. Indolfi, S. De Rosa, A. Colombo, Bioresorbable vascular scaffolds-basic concepts and clinical outcome, Nat. Rev. Cardiol. 13 (12) (2016) 719.

[3] A.A. Oliver, M. Sikora-Jasinska, A.G. Demir, R.J. Guillory II, Recent advances and directions in the development of bioresorbable metallic cardiovascular stents: insights from recent human and in vivo studies, Acta Biomater 127 (2021) 1-23.

[4] D. Bian, L. Qin, W. Lin, D. Shen, H. Qi, X. Shi, G. Zhang, H. Liu, H. Yang, J. Wang, Magnetic resonance (MR) safety and compatibility of a novel iron bioresorbable scaffold, Bioact. Mater. 5 (2) (2020) 260-274.

[5] H.M. Garcia-Garcia, M. Haude, K. Kuku, A. Hideo-Kajita, H. Ince, A. Abizaid, R. Tölg, P.A. Lemos, C. von Birgelen, E.H. Christiansen, In vivo serial invasive imaging of the second-generation drug-eluting absorbable metal scaffold (Magmaris-DREAMS 2G) in de novo coronary lesions: insights from the BIOSOLVE-II first-in-man trial, Int. J. Cardiol. 255 (2018) 22-28.

[6] D.I. McLennan, J.R. Maldonado, S.R. Foerster, S.S. Handler, J.F. LaDisa Jr., T. M. Gudausky, R.J. Guillory, Absorbable metal stents for vascular use in pediatric cardiology: progress and outlook, Front. Cardiovasc. Med. 11 (2024) 1410305.

[7] L. Sun, J.-j. Li, Y.-k. Xu, Y.-m. Xie, S.-s. Wang, Z.-w. Zhang, Initial status and 3- month results relating to the use of biodegradable nitride iron stents in children and the evaluation of right ventricular function, Front. Cardiovasc. Med. 9 (2022) 914370.

[8] P.A. Zartner, D. Schranz, N. Mini, M.B. Schneider, K. Schneider, Acute treatment of critical vascular stenoses with a bioabsorbable magnesium scaffold in infants with CHDs, Cardiol. Young 30 (4) (2020) 493-499.

[9] S. Verheye, A. Wlodarczak, P. Montorsi, J. Torzewski, J. Bennett, M. Haude, G. Starmer, T. Buck, M. Wiemer, A.A. Nuruddin, BIOSOLVE-IV-registry: safety and performance of the Magmaris scaffold: 12-month outcomes of the first cohort of 1,075 patients, Cathet. Cardiovasc. Interv. 98 (1) (2020) E1-E8.

[10] R.-L. Gao, B. Xu, Z. Sun, C. Guan, L. Song, L. Gao, C. Li, J. Cui, Y. Zhang, K. Dou, First-in-human evaluation of a novel ultrathin sirolimus-eluting iron bioresorbable scaffold: 3-year outcomes of the IBS-FIM trial, EuroInterven.: J. EuroPCR Collabo. Working Group Interven. Cardiol. European Soc. Cardiol. (2023) EIJ-D.

[11] A.A. Oliver, K.D. Carlson, C. Bilgin, J.L.A. Larco, R. Kadirvel, R.J. Guillory II, D. D. Daescu, D.F. Kallmes, Bioresorbable flow diverters for the treatment of intracranial aneurysms: review of current literature and future directions, J. Neurointerventional Surg. 15 (2) (2022) 178-182.

[12] A.A. Oliver, Y.C. Senol, C. Bilgin, J.E. Schaffer, R. Kadirvel, D.F. Kallmes, J. M. Wainwright, Intraluminal flow diverter design primer for neurointerventionalists, Am. J. Neuroradiol. 45 (4) (2024) 365-370.

[13] R. Kadirvel, Y.-H. Ding, D. Dai, I. Rezek, D.A. Lewis, D.F. Kallmes, Cellular mechanisms of aneurysm occlusion after treatment with a flow diverter, Radiology 270 (2) (2014) 394.

[14] W. Brinjikji, M.H. Murad, G. Lanzino, H.J. Cloft, D.F. Kallmes, Endovascular treatment of intracranial aneurysms with flow diverters: a meta-analysis, Stroke 44 (2) (2013) 442-447.

[15] T. Becske, W. Brinjikji, M.B. Potts, D.F. Kallmes, M. Shapiro, C.J. Moran, E.I. Levy, C.G. McDougall, I. Szikora, G. Lanzino, Long-term clinical and angiographic outcomes following pipeline embolization device treatment of complex internal carotid artery aneurysms: five-year results of the pipeline for uncoilable or failed aneurysms trial, Neurosurgery 80 (1) (2017) 40-48.

[16] S. Dandapat, A. Mendez-Ruiz, M. Martínez-Galdámez, J. Macho, S. Derakhshani, G. F. Torres, V.M. Pereira, A. Arat, A.K. Wakhloo, S. Ortega-Gutierrez, Review of current intracranial aneurysm flow diversion technology and clinical use, J. Neurointerventional Surg. 13 (1) (2021) 54-62.

[17] D.F. Kallmes, R. Hanel, D. Lopes, E. Boccardi, A. Bonafé, S. Cekirge, D. Fiorella, P. Jabbour, E. Levy, C. McDougall, International retrospective study of the pipeline embolization device: a multicenter aneurysm treatment study, Am. J. Neuroradiol. 36 (1) (2015) 108-115.

[18] A. Gúedon, F. Clarençon, F. Di Maria, C. Rosso, A. Biondi, J. Gabrieli, P. Rojas, J. Chiras, N. Sourour, Very late ischemic complications in flow-diverter stents: a retrospective analysis of a single-center series, J. Neurosurg. 125 (4) (2016) 929-935.

[19] J. Caroff, T. Tamura, R.M. King, P.N. Lylyk, E.T. Langan, O.W. Brooks, F. Clarençon, J.M. Wainwright, L. Spelle, M. Marosfoi, Phosphorylcholine surface modified flow diverter associated with reduced intimal hyperplasia, J. Neurointerventional Surg. 10 (11) (2018) 1097-1101.

[20] T.F. Flood, I.M. van der Bom, L. Strittmatter, A.S. Puri, G.M. Hendricks, A. K. Wakhloo, M.J. Gounis, Quantitative analysis of high-resolution, contrastenhanced, cone-beam CT for the detection of intracranial in-stent hyperplasia, J. Neurointerventional Surg. 7 (2) (2015) 118-125.

[21] B. Halitcan, S. Bige, B. Sinan, A. Ilkay, D. Ergun, A. Fatih, A. Anil, The implications of magnetic resonance angiography artifacts caused by different types of intracranial flow diverters, J. Cardiovasc. Magn. Reson. 23 (1) (2021) 1-14.

[22] P. Bouillot, O. Brina, B.M. Delattre, R. Ouared, A. Pellaton, H. Yilmaz, P. Machi, K. O. Lovblad, M. Farhat, V.M. Pereira, Neurovascular stent artifacts in 3D-TOF and 3D-PCMRI: influence of stent design on flow measurement, Magn. Reson. Med. 81 (1) (2019) 560-572.

[23] A.A. Oliver, C. Bilgin, A.J. Vercnocke, K.D. Carlson, R. Kadirvel, R.J. Guillory, A. J. Griebel, J.E. Schaffer, D. Dragomir-Daescu, D.F. Kallmes, Benchtop proof of concept and comparison of iron-and magnesium-based bioresorbable flow diverters, J. Neurosurg. 1 (aop) (2022) 1-7.

[24] H. Hermawan, H. Alamdari, D. Mantovani, D. Dube, Iron-manganese: new class of metallic degradable biomaterials prepared by powder metallurgy, Powder Metall. 51 (1) (2008) 38-45.

[25] J.E. Schaffer, E.A. Nauman, L.A. Stanciu, Cold-drawn bioabsorbable ferrous and ferrous composite wires: an evaluation of mechanical strength and fatigue durability, Metall. Mater. Trans. B 43 (4) (2012) 984-994.

[26] A. Oliver, E. Koons, P. Trester, J. Kleinow, R. Jonsgaard, A. Vercnocke, C. Bilgin, R. Kadirvel, S. Leng, A. Lu, Medical imaging compatibility of magnesium-and ironbased bioresorbable flow diverters, Am. J. Neuroradiol. 44 (6) (2023) 668-674.

[27] J.E. Schaffer, E.A. Nauman, L.A. Stanciu, Cold drawn bioabsorbable ferrous and ferrous composite wires: an evaluation of in vitro vascular cytocompatibility, Acta Biomater. 9 (10) (2013) 8574-8584.

[28] S. Lu, Q. Wang, Y. Zhang, H. Li, H. Feng, L. Tan, K. Yang, A novel biodegradable high nitrogen iron alloy with simultaneous enhancement of corrosion rate and local corrosion resistance, J. Mater. Sci. Technol. 152 (2023) 94-99.

[29] D. Dai, Y.H. Ding, M.A. Danielson, R. Kadirvel, D.A. Lewis, H.J. Cloft, D.F. Kallmes, Histopathologic and immunohistochemical comparison of human, rabbit, and swine aneurysms embolized with platinum coils, Am. J. Neuroradiol. 26 (10) (2005) 2560-2568.

[30] D.F. Kallmes, Y. Ding, D. Dai, R. Kadirvel, D. Lewis, H. Cloft, A second-generation, endoluminal, flow-disrupting device for treatment of saccular aneurysms, Am. J. Neuroradiol. 30 (6) (2009) 1153-1158.

[31] J. Fiehler, S. Ortega-Gutierrez, V. Anagnostakou, J. Cortese, H.S. Cekirge, D. Fiorella, R. Hanel, Z. Kulcsar, S. Lamin, J. Liu, Evaluation of flow diverters for cerebral aneurysm therapy: recommendations for imaging analyses in clinical studies, endorsed by ESMINT, ESNR, OCIN, SILAN, SNIS, and WFITN, J. Neurointerventional Surg. (2024).

[32] M. Peuster, C. Hesse, T. Schloo, C. Fink, P. Beerbaum, C. von Schnakenburg, Longterm biocompatibility of a corrodible peripheral iron stent in the porcine descending aorta, Biomaterials 27 (28) (2006) 4955-4962.

[33] W. Lin, L. Qin, H. Qi, D. Zhang, G. Zhang, R. Gao, H. Qiu, Y. Xia, P. Cao, X. Wang, Long-term in vivo corrosion behavior, biocompatibility and bioresorption mechanism of a bioresorbable nitrided iron scaffold, Acta Biomater. 54 (2017) 454-468.

[34] P.K. Shah, Inflammation, neointimal hyperplasia, and restenosis: as the leukocytes roll, the arteries thicken, Am. Heart Assoc. (2003) 2175-2177.

[35] G. Gąsior, J. Szczepánski, A. Radtke, Biodegradable iron-based materials-what was done and what more can be done? Materials 14 (12) (2021) 3381.

[36] A.J. Griebel, P. Maier, H. Summers, B. Clausius, I. Kanasty, W. He, N. Peterson, C. Czerniak, A.A. Oliver, D.F. Kallmes, Radiopaque FeMnN-Mo composite drawn filled tubing wires for braided absorbable neurovascular devices, Bioact. Mater. 40 (2024) 74-87.

[37] S. Ningshen, U.K. Mudali, V. Mittal, H. Khatak, Semiconducting and passive film properties of nitrogen-containing type 316LN stainless steels, Corros. Sci. 49 (2) (2007) 481-496.

[38] D. Pierson, J. Edick, A. Tauscher, E. Pokorney, P. Bowen, J. Gelbaugh, J. Stinson, H. Getty, C.H. Lee, J. Drelich, A simplified in vivo approach for evaluating the bioabsorbable behavior of candidate stent materials, J. Biomed. Mater. Res. B Appl. Biomater. 100 (1) (2012) 58-67.

[39] H.-S. Han, S. Loffredo, I. Jun, J. Edwards, Y.-C. Kim, H.-K. Seok, F. Witte, D. Mantovani, S. Glyn-Jones, Current status and outlook on the clinical translation of biodegradable metals, Mater. Today 23 (2019) 57-71.

[40] J.E. Schaffer, Development and Characterization of Vascular Prosthetics for Controlled Bioabsorption, Purdue University, 2012.

[41] A.P. Avolio, M.F. O'rourke, K. Mang, P.T. Bason, B.S. Gow, A comparative study of pulsatile arterial hemodynamics in rabbits and Guinea pigs, Am. J. Physiol.-Legacy Content 230 (4) (1976) 868-875.

[42] M.P. Kwesiga, A.A. Gillette, F. Razaviamri, M.E. Plank, A.L. Canull, Z. Alesch, W. He, B.P. Lee, R.J. Guillory II, Biodegradable magnesium materials regulate ROSRNS balance in pro-inflammatory macrophage environment, Bioact. Mater. 23 (2023) 261-273.

[43] P. Hou, P. Han, C. Zhao, H. Wu, J. Ni, S. Zhang, J. Liu, Y. Zhang, H. Xu, P. Cheng, Accelerating corrosion of pure magnesium co-implanted with titanium in vivo, Sci. Rep. 7 (1) (2017) 41924.

[44] H. Lee, H.-j. Lee, H.-E. Kim, J. Kweon, B.-D. Lee, C. Lee, Oxidant production from corrosion of nano-and microparticulate zero-valent iron in the presence of oxygen: a comparative study, J. Hazard Mater. 265 (2014) 201-207.

[45] H. Ahsan, A. Ali, R. Ali, Oxygen free radicals and systemic autoimmunity, Clin. Exp. Immunol. 131 (3) (2003) 398-404.

[46] E. Scarcello, A. Herpain, M. Tomatis, F. Turci, P. Jacques, D. Lison, Hydroxyl radicals and oxidative stress: the dark side of Fe corrosion, Colloids Surf. B Biointerfaces 185 (2020) 110542.

[47] E. Scarcello, I. Lobysheva, C. Bouzin, P. Jacques, D. Lison, C. Dessy, Endothelial dysfunction induced by hydroxyl radicals-the hidden face of biodegradable Febased materials for coronary stents, Mater. Sci. Eng. C 112 (2020) 110938.

[48] J.W. Drelich, J. Goldman, Bioresorbable vascular metallic scaffolds: current status and research trends, Curr. Opin. Biomed. Eng. 24 (2022) 100411.

[49] J.-F. Zheng, Z.-W. Xi, Y. Li, J.-N. Li, H. Qiu, X.-Y. Hu, T. Luo, C. Wu, X. Wang, L.- F. Song, Long-term safety and absorption assessment of a novel bioresorbable nitrided iron scaffold in porcine coronary artery, Bioact. Mater. 17 (2022) 496-505.

[50] R.S. Schwartz, K.C. Huber, J.G. Murphy, W.D. Edwards, A.R. Camrud, R. E. Vlietstra, D.R. Holmes, Restenosis and the proportional neointimal response to coronary artery injury: results in a porcine model, J. Am. Coll. Cardiol. 19 (2) (1992) 267-274.

[51] T.A. Altes, H.J. Cloft, J.G. Short, A. DeGast, H.M. Do, G.A. Helm, D.F. Kallmes, Creation of saccular aneurysms in the rabbit: a model suitable for testing endovascular devices, Am. J. Roentgenol. 174 (2) (2000) 349-354.

[52] J. Cortese, E.A. Bayraktar, A.A. Oliver, C. Bilgin, Y.-H. Ding, Y. Ueki, S. Lortscher, W. Brinjikji, R. Kadirvel, D.F. Kallmes, Creating elastase aneurysms in rabbits: a video primer, J. Neurointerventional Surg. (2024).