The Importance of Wall Apposition in Flow Diverters

The Importance of Wall Apposition in Flow Diverters Abstract BACKGROUND It is assumed that high pore densities in flow diverters (FDs) are beneficial for intracranial aneurysm (IA) healing. However, various animal studies are not conclusive on the issue, suggesting that other factors are in play. One important factor might be wall apposition. OBJECTIVE To (1) determine the relationship between FD pore density and aneurysm occlusion, and (2) determine the relationship between FD wall apposition and aneurysm occlusion. METHODS Saccular aneurysms were microsurgically created in the aorta of 36 Wistar rats. Twelve rats received a low pore density FD (10 pores/mm2), 12 rats received a high pore density FD (23 pores/mm2), and the remaining 12 rats served as a control group. Six animals from each group were sacrificed 1 and 3 mo after surgery. We determined aneurysm occlusion, the number of struts not in contact with the aorta wall, and the average distance from malapposed struts to aorta wall through histology. RESULTS No significant differences were found in aneurysm occlusion between the low pore density and high pore density groups (P > .05) after 1 and 3 mo of follow-up. The average number of malapposed struts was lower for the occluded aneurysm group (4.4 ± 1.9) compared to the nonoccluded aneurysm group (7.7 ± 2.6, P < .01). The average distance between malapposed struts and parent artery wall was lower for the occluded aneurysm group (33.9 μm ± 11.5 μm) than for the nonoccluded aneurysm group (48.7 μm ± 18.8 μm, P < .05). CONCLUSION Wall apposition is more important than pore density for aneurysm occlusion. Animal model, Endovascular, Flow diverter, Intracranial aneurysm, Rats, Vascular biology, Wall apposition ABBREVIATIONS ABBREVIATIONS FD flow diverter IA intracranial aneurysm An intracranial aneurysm (IA) is a bulging area in the wall of an intracranial artery that may be abnormally weak and might rupture.1 IAs occur in about 2% of the population2,3 and account for approximately 85% of the nontraumatic subarachnoid hemorrhages,4 which can have devastating consequences such as permanent neurological damage and even death.5 In case an unruptured IA is diagnosed and conservative treatment is not an option, a flow diverter (FD) can be placed in the parent artery covering the ostium of the aneurysm. Ideally, this leads to intra-aneurysmal thrombosis due to reduced blood flow in the aneurysm sac, and parent vessel reconstruction due to endothelial cell coverage of the implant.6,7 Porosity (the nonsolid fraction of the implant surface) and pore density (number of pores per mm2) at the ostium are generally considered to determine the efficacy of FDs.6,7 Good clinical results with minimal side branch occlusion are observed with an implant porosity in the 60% to 80% range.8 The exact effect of pore density, however, is less clear. Computational9 and experimental10 studies demonstrate that higher implant pore densities lead to lower intrasaccular blood flow, which should promote rapid IA occlusion. Animal studies are less conclusive as some show a (slight) beneficial effect,11,12 while another shows no effect13 of pore density at all. The discrepancy between these studies suggests that other factors are in play that cannot be fully controlled. One of these factors might be implant wall apposition. We therefore set 2 goals in the current study: (1) whether there is a relationship between FD pore density and aneurysm occlusion, (2) whether there is a relationship between FD wall apposition and aneurysm occlusion. We used a rat model in which the aneurysm is prone to expansion and rupture to implant FDs with varying pore density. METHODS Aneurysms were microsurgically created in 36 Wistar rats according to previously published procedures.14,15 One group (n = 12) received a low pore density FD (10 pores/mm2), 1 group (n = 12) received a high pore density FD (23 pores/mm2), and the remaining 12 rats served as a control group. Six animals from each of the 3 groups were sacrificed at 2 different time points: 1 and 3 mo after surgery. Macroscopic and microscopic evaluation followed. Animals All experiments were approved by institutional animal care and use committee. Six-to-eight weeks old, Wistar WU rats (n = 72, Charles River, Den Bosch, The Netherlands) were used in this study. To limit the introduction of unknown factors in our experiment, we used only male animals, based on the available literature.14,15 Potential gender-related differences need to be addressed in future studies. Animals were housed in a room with an average temperature of 22°C (±1°C) and a 50/50 light/dark cycle. Animals had access to a pellet diet and sterilized water ad libitum and received humane care in agreement with institutional guidelines. A thoracic aorta segment between the left subclavian artery and the upper intelcostal artery was harvested from 36 donor rats. These segments were then incubated overnight (10 h) in a 0.1% sodium dodecyl sulfate in milli-Q water solution at a temperature of 37°C. The incubation step results in a decellularized vessel wall, with an intact extracellular matrix. After the incubation step, the aortic segment was stored in PBS on ice until surgery later that day. A microsurgical side wall aneurysm was then created in 36 recipient rats, by attaching the decellularized aortic segment to the abdominal aorta between the renal arteries and aortic bifurcation. Successful completion of the surgery was marked by a pulsating aneurysm without seeping or oozing of blood. Immediately after anastomosis creation, surgery was completed in 12 of the 36 animals and no FD was implanted. These animals served as controls. The remaining 24 animals received either a low pore density FD (n = 12) or a high pore density FD (n = 12), which was implanted in the aorta as previously described.15 Randomization was achieved by allocating each subsequent animal to a predetermined cycle of experimental groups: control—low pore density FD—high pore density FD. FDs and Dual Anti-Platelet Therapy The low pore density implant consisted of 20 cobalt chrome wires and 4 platinum wires, with a porosity of approximately 80%. The high pore density implant consisted of 30 cobalt chrome wires and 6 platinum wires with a porosity of approximately 70%. The 2 implants used in this experiment were both Surpass FDs and were produced by Stryker Neurovascular (Fremont, California). All rats received daily dual anti-platelet therapy through oral gavage: 5 mg/kg clopidogrel16 and 1 mg/kg/d of acetylsalicylic acid.17 Dual anti-platelet therapy started at least 3 d before surgery and was continued for the rest of the experiment. Group allocation was unknown to animal laboratory personnel during the entire experiment. Tissue Harvesting and Processing One month (n = 18) and 3 mo (n = 18) after surgery, the rats underwent a second laparotomy. Through visual inspection, the position of the aneurysm and the FD was determined, and clamps were placed above and below to isolate this aortic section. The complete aortic segment, including FD and aneurysm, was subsequently excised, flushed with saline, and fixed overnight in 4% formalin. Tissues obtained from control rats were embedded in paraffin and sectioned with a microtome. All tissues containing an FD embedded in methacrylate and sectioned with a saw microtome (SP1600, Leica Microsystems, Eindhoven, Netherlands). All tissues were stained with hematoxylin and eosin and digitalized using the Vision-Tek slide scanner (Sakura Finetek, Alphen aan den Rijn, Netherlands). Outcome Measures and Statistical Analyses Histological specimens were obtained as axial sections (ie, perpendicular to the long axis of the aorta) from the aorta and the aneurysm. Measurements were performed at the level of the largest diameter of the aneurysm. All slides were coded to mask experimental group allocation and to minimize bias. We measured the following parameters from our histological sections: the percentage of (organized) thrombus within the aneurysm (aneurysm occlusion), the number of struts not in contact with the aorta wall (implant apposition), and the average distance from malapposed struts to aorta wall (Figure 1). All measurements were performed using FIJI (ImageJ version 1.47), an open-source platform for biological-image analysis.18 FIGURE 1. View largeDownload slide Schematic depiction of typical cross-section of an aorta with the implanted FD and the microsurgically created aneurysm. In this example, the aneurysm is fully occluded with thrombus (pink: organized thrombus, red: fresh thrombus). Perfect implant apposition can be seen on the left side of the cross section where all FD struts are in contact with the parent artery wall (insert A). Imperfect implant apposition can be seen on the right side of the cross section where 3 FD struts are not touching the vessel wall (insert B). The red lines depict the measurements to calculate the distance of a malapposed strut perpendicular to the parent artery wall (insert B). FIGURE 1. View largeDownload slide Schematic depiction of typical cross-section of an aorta with the implanted FD and the microsurgically created aneurysm. In this example, the aneurysm is fully occluded with thrombus (pink: organized thrombus, red: fresh thrombus). Perfect implant apposition can be seen on the left side of the cross section where all FD struts are in contact with the parent artery wall (insert A). Imperfect implant apposition can be seen on the right side of the cross section where 3 FD struts are not touching the vessel wall (insert B). The red lines depict the measurements to calculate the distance of a malapposed strut perpendicular to the parent artery wall (insert B). All statistical analyses were performed using GraphPad Prism (Version 5.03, GraphPad Software, Inc, LaJolla, California). Three groups were compared using a 1-way ANOVA with a post hoc Bonferroni multiple comparison test to pinpoint differences between groups when variances were equal. When variances were not equal, 3 groups were compared using the Kruskal–Wallis test with a post hoc Dunn's multiple comparison test to pinpoint differences between groups. When 2 groups were compared, a Student's t-test was used for analysis. If variances were not equal for both groups, an unpaired t-test with Welch's correction was performed. Level of statistical significance is indicated as <0.05 (*), <0.01 (**), or <0.0001 (***). RESULTS General Information, Mortality, and Complications Two of the 36 recipient rats died during follow-up (mortality: 5.6%). One rat died due to an intestinal perforation the night after surgery, probably due to intestinal manipulation during laparotomy. Another rat died 50 d after surgery, but the cause of death remained unknown, even after necropsy (Table). TABLE. Overview of encountered deaths and complications in each group. Animals included for histological analysis are also specified   Number of animals    1-mo follow-up  3-mo follow-up  Control group  Deaths: 0  Deaths: 1 intestinal perforation the night after surgery    Complications: 0  Complications: 0    Included for histological analysis: 6  Included for histological analysis: 5  Low pore density FD group  Deaths: 0  Deaths: 0    Complications: 0  Complications: 0    Included for histological analysis: 6  Included for histological analysis: 6  High pore density FD group  Deaths: 0  Deaths: 1 unknown cause of death 50 d after surgery    Complications: 1 FD migration, rat was excluded for measurements  Complications: 0    Included for histological analysis: 5  Included for histological analysis: 6 deceased rat was not excluded    Number of animals    1-mo follow-up  3-mo follow-up  Control group  Deaths: 0  Deaths: 1 intestinal perforation the night after surgery    Complications: 0  Complications: 0    Included for histological analysis: 6  Included for histological analysis: 5  Low pore density FD group  Deaths: 0  Deaths: 0    Complications: 0  Complications: 0    Included for histological analysis: 6  Included for histological analysis: 6  High pore density FD group  Deaths: 0  Deaths: 1 unknown cause of death 50 d after surgery    Complications: 1 FD migration, rat was excluded for measurements  Complications: 0    Included for histological analysis: 5  Included for histological analysis: 6 deceased rat was not excluded  View Large TABLE. Overview of encountered deaths and complications in each group. Animals included for histological analysis are also specified   Number of animals    1-mo follow-up  3-mo follow-up  Control group  Deaths: 0  Deaths: 1 intestinal perforation the night after surgery    Complications: 0  Complications: 0    Included for histological analysis: 6  Included for histological analysis: 5  Low pore density FD group  Deaths: 0  Deaths: 0    Complications: 0  Complications: 0    Included for histological analysis: 6  Included for histological analysis: 6  High pore density FD group  Deaths: 0  Deaths: 1 unknown cause of death 50 d after surgery    Complications: 1 FD migration, rat was excluded for measurements  Complications: 0    Included for histological analysis: 5  Included for histological analysis: 6 deceased rat was not excluded    Number of animals    1-mo follow-up  3-mo follow-up  Control group  Deaths: 0  Deaths: 1 intestinal perforation the night after surgery    Complications: 0  Complications: 0    Included for histological analysis: 6  Included for histological analysis: 5  Low pore density FD group  Deaths: 0  Deaths: 0    Complications: 0  Complications: 0    Included for histological analysis: 6  Included for histological analysis: 6  High pore density FD group  Deaths: 0  Deaths: 1 unknown cause of death 50 d after surgery    Complications: 1 FD migration, rat was excluded for measurements  Complications: 0    Included for histological analysis: 5  Included for histological analysis: 6 deceased rat was not excluded  View Large All remaining rats displayed normal behavior and did not seem to be hindered by the operation or the FD. An average body weight increase of 58.1 g (SD ± 17.3 g), and 137.9 g (SD ± 30.1 g) was seen after 1 and 3 mo of follow-up, respectively. Body weight between the groups showed no statistically significant differences at each time point (1-way ANOVA, Figure, Supplemental Digital Content 1). A complication was observed in 1 animal. Upon sacrifice at 1 mo, signs of hemorrhage were clearly present around the aneurysm. After histological sectioning, we concluded that the FD had migrated after implantation, which led to incomplete coverage of the aneurysm neck (Table). This animal was excluded from further analysis. Outcome Measures Stratified per Implant Aneurysm Occlusion After 1 mo of follow-up, complete aneurysm occlusion was seen in 0 out of 6 animals in the control group, 2 out of 6 animals in the low pore density group, and 1 out of 5 animals in the high pore density group. The average aneurysm occlusion was 58.2% (SD ± 13.5%), 96.6% (SD ± 5.6%), and 93.8% (SD ± 8.3%) for the control group, the low pore density implant group and the high pore density implant group, respectively (Figure 2). After 3 mo of follow-up, total aneurysm occlusion was seen in 0 out of 5 animals in the control group, 4 out of 6 animals in the low pore density group, and 4 out of 6 animals in the high pore density group. The average aneurysm occlusion was 40.1% (SD ± 23.1%), 96.2% (SD ± 6.9%), and 99.4% (SD ± 1.0%) for the control group, the low pore density implant group, and the high pore density implant group, respectively (Figure 2). For both time points, the low pore density and high pore density implant groups statistically differed from the control group (1 mo: Bonferroni's multiple comparison test P < .0001, 3 mo: Dunn's multiple comparison test < 0.05), but not from each other (1 mo: Bonferroni's multiple comparison test P > .05, 3 mo: Dunn's multiple comparison test >0.05, Figure 2). FIGURE 2. View largeDownload slide Outcome measures stratified per implant group. Aneurysm occlusion was similar in both implant groups, but significantly lower in the control group (*P < .05; ***P < .0001). FIGURE 2. View largeDownload slide Outcome measures stratified per implant group. Aneurysm occlusion was similar in both implant groups, but significantly lower in the control group (*P < .05; ***P < .0001). Outcome Measures Stratified per Aneurysm Occlusion Status Aneurysm Occlusion When analyzing the occlusion status, independent of follow-up time, 11 out of 23 animals had a nonoccluded aneurysm, in all cases observed as neck remnants, with an average occlusion of 93.5% (SD ± 6.9%; Figure 3). Of these 11 animals, 6 received a low pore density FD and 5 received a high pore density FD. FIGURE 3. View largeDownload slide Comparison of occluded vs nonoccluded aneurysms regarding aneurysm occlusion (left), the number of malapposed FD struts (middle) and the average distance between malapposed FD struts and the parent artery wall (right) independent of follow-up time (*P < .05 and **P < .01). FIGURE 3. View largeDownload slide Comparison of occluded vs nonoccluded aneurysms regarding aneurysm occlusion (left), the number of malapposed FD struts (middle) and the average distance between malapposed FD struts and the parent artery wall (right) independent of follow-up time (*P < .05 and **P < .01). Implant Apposion After 1 mo of follow-up, the average number of malapposed struts was 4.7 (SD ± 2.5) and 7.6 (SD ± 2.4) for the occluded aneurysm group and the nonoccluded aneurysm group, respectively (Figure, Supplemental Digital Content 2). After 3 mo of follow-up, the average number of malapposed struts was 4.3 (SD ± 1.8) and 7.8 (SD ± 3.3) for the occluded aneurysm group and the nonoccluded aneurysm group, respectively (P < .05, Figure, Supplemental Digital Content 2). The average number of malapposed struts, independent of follow-up time, was lower for the occluded aneurysm group (4.4 ± 1.9) than for the nonoccluded aneurysm group (7.7 ± 2.6; Figure 3, P < .01). Distance Malapposed Struts to Parent Artery Wall After 1 mo of follow-up, the average distance between malapposed struts and parent artery wall was 45.6 μm (SD ± 10.2 μm) and 50.0 μm (SD ± 17.2 μm) for the occluded aneurysm group and the nonoccluded aneurysm group, respectively (Figure, Supplemental Digital Content 2). After 3 mo of follow-up, the average distance between malapposed struts and parent artery wall was 29.5 μm (SD ± 8.9 μm) and 46.1 μm (SD ± 24.20 μm) for the occluded aneurysm group and the nonoccluded aneurysm group, respectively (Figure, Supplemental Digital Content 2). The average distance between malapposed struts and parent artery wall, independent of follow-up time, was lower for the occluded aneurysm group (33.9 μm ± 11.5 μm) than for the nonoccluded aneurysm group (48.7 μm ± 18.8 μm; Figure 3, P < .05). DISCUSSION In this study, we investigated the effect of FD pore density and FD wall apposition on aneurysm occlusion using a rat aneurysm model. Although treatment with FDs clearly led to massive intra-aneurysmal thrombosis compared to control animals, no distinct difference in aneurysm healing was observed for the different FD designs. Incomplete occlusion, observed as neck remnants, was more likely when FD apposition was worse: FDs in nonoccluded aneurysms displayed more malapposed struts and the average distance between malapposed strut and parent artery wall was higher compared to those in occluded aneurysms. Our findings support the results of Rouchaud et al19 who found that good wall apposition was strongly associated with complete aneurysm occlusion after FD therapy in a series of 41 rabbits. Our findings are in contrast with computational,9 experimental,10 and animal studies11,12 in which a beneficial effect of high pore density implants is demonstrated. Although these studies seem to contradict each other, it is possible that a higher FD pore density indeed results in faster aneurysm occlusion, but only if implant apposition is perfect. A hypothesis supported by Mut et al,9 who reported a case of incomplete aneurysm occlusion due to poor FD apposition resulting in residual blood flow to the aneurysm. Also, Miyachi et al20 reported “a small leak from the FD” at the aneurysm neck of nonoccluding IAs. Implant malapposition does not only affect aneurysm healing, but is also associated with serious adverse events such as thromboembolic complications21-23 and FD migration.23,24 Both adverse events can have devastating clinical consequences. Thromboembolisms can lead to ischemic stroke resulting in permanent neurological damage. Implant migration can lead to ostium exposure and subsequent aneurysm rupture.25-27 In one of the rats in our study, a high pore density FD migrated in the vessel, resulting in an incompletely sealed aneurysm and a subsequent rupture (Figure, Supplemental Digital Content 3). The animal survived as the hemorrhage was contained by peri-aneurysmal fibrous tissue. Upon sacrifice (1 mo), signs of the hemorrhage were clearly visible macroscopically and a torn wall and organized thrombus was seen microscopically. This case illustrates the risk associated with FD migration, a phenomenon which has been described previously in clinical studies.19,28,29 Meta-analyses show that about 20% to 25% of the IAs are not occluded 6 mo after FD treatment, that the risk of thromboembolic complications is around 5% to 6% and the risk for aneurysm rupture after FD is around 3%.30-33 It is unknown to what extent FD malapposition is responsible for these complications. One reason might be the use of 2-dimensional digital subtraction angiography, which is not the optimal method for assessing FD deployment and apposition.19,34,35 Newer imaging techniques, such as high-resolution contrast-enhanced cone beam computed tomography34,35 or optical coherence tomography,36,37 are better capable of visualizing the FD in the parent vessel and could therefore give a better overview of how the aforementioned complications and FD apposition are linked. Limitations This study has several limitations. First, it is expected that the microsurgery itself will elicit a tissue response, which partially overlaps with aneurysm healing due to the FD treatment. It is currently unknown whether aneurysm healing would be different if this overlap was not present. Further refinement of the animal model could be achieved by creating the aneurysm first and placing the FD at a later stage. Addition of angiography to the study protocol is essential, as aneurysm patency should be guaranteed before FD implantation. Furthermore, the aneurysm in our animal model is created in the abdominal space. Besides a difference in vessel structure, such as the presence of vasa vasorum in the aorta and absence in intracranial arteries,38,39 the peri-aneurysmal environment also differs. It is currently unknown if these differences can alter the healing process in the aneurysm after FD placement. Further use of the current model and comparison to results of IA models might elucidate the importance of these discrepancies. Another limitation is the lack of aneurysm expansion and aneurysm rupture in our study. This is in sharp contrast to the results reported by Marbacher et al.40 We expect that the dual-antiplatelet regime could have caused this discrepancy. In a similar rat model, in which a decellularized guinea pig aorta segment was implanted in the abdominal aorta, administering a P2Y12 antagonist led to significantly less growth of the decellularized segment.41 The authors suggested that the lack of thrombus renewal, which limits the intrathrombus build-up of polymorphonuclear cells and the proteases they contain, lead to a healing-type response including smooth muscle cell invasion and limited aneurysm wall damage. This raises the interesting question if dual-antiplatelet therapy is an important aspect of successful FD treatment. The absence of tortuosity in the abdominal aorta compared to the intracranial arteries is another drawback of our current animal model. However, this also makes our results more intriguing as it is expected that FD apposition should be optimal in a straight artery like the aorta. We observed that small variations in apposition alter aneurysm healing response in our animal model, apposition may be more challenging to achieve in the tortuous and sometimes even dysplastic or atherosclerotic intracranial arteries where these devices are used. More research is needed on how vessel tortuosity, dysplasticity, and atherosclerosis affects implant apposition and subsequent aneurysm healing. Finally, the study would benefit from the use of more animals. It would make the use of multivariate analyses methods possible, which could elucidate whether interaction was present between variables such as implant malapposition and average distance between malapposed struts and parent artery wall. Furthermore, use of more animals per time point or using more animals to introduce additional time points would help us understand if all aneurysms would eventually occlude or if neck remnants persisted. A shift towards complete occlusion during longer follow-up is now suggested by our data, as 3 out of 11 and 8 out of 12 aneurysms are occluded after 1 and 3 mo of follow-up, respectively. However, this finding is not statistically significant. CONCLUSION Based on our results, we conclude that implant apposition is more important than pore density for aneurysm occlusion. It seems that in this biological model malapposition can “overrule” the beneficial effects that high pore density implants can have on IA healing. Disclosures This study was funded by Stryker Neurovascular. Dr de Vries works as a consultant for Stryker Neurovascular. Dr Gounis has been a consultant on a fee-per-hour basis for Codman Neurovascular, InNeuroCo, Medtronic Neurovascular, and Stryker Neurovascular; holds stock in InNeuroCo; and has received research support from the Research support from the National Institutes of Health (NIH), Anaconda, Codman Neurovascular, Gentuity, InNeuroCo, Microvention, Medtronic Neurovascular, MIVI Neurosciences, Neuravi, Philips Healthcare, InNeuroCo, Rapid Medical, R92M, Stryker Neurovascular, The Stroke Project, and the Wyss Institute. REFERENCES 1. Frösen J, Tulamo R, Paetae A, Laaksamo E, Korja M, Laakso A. Saccular intracranial aneurysm: pathology and mechanisms. J Vasc Surg . 2012; 123( 6): 719- 727. 2. Vernooij MW, Ikram MA, Tanghe HL et al.   Incidental findings on brain MRI in the general population. N Engl J Med . 2007; 357( 18): 1821- 1828. Google Scholar CrossRef Search ADS PubMed  3. Peymani A, Adams HHH, Cremers LGM et al.   Genetic determinants of unruptured intracranial aneurysms in the general population. Stroke . 2015; 46( 10): 2961- 2964. Google Scholar CrossRef Search ADS PubMed  4. van Gijn J, Kerr RS, Rinkel GJE. Subarachnoid haemorrhage. Lancet . 2007; 369( 9558): 306- 318. Google Scholar CrossRef Search ADS PubMed  5. Brown RD, Broderick JP. Unruptured intracranial aneurysms: epidemiology, natural history, management options, and familial screening. Lancet Neurol . 2014; 13( 4): 393- 404. Google Scholar CrossRef Search ADS PubMed  6. Lieber BB, Sadasivan C. Endoluminal scaffolds for vascular reconstruction and exclusion of aneurysms from the cerebral circulation. Stroke . 2010; 41( 10 Suppl): S21- S25. Google Scholar CrossRef Search ADS PubMed  7. Wakhloo AK, Gounis MJ. Revolution in aneurysm treatment: Flow diversion to cure aneurysms: A paradigm shift. Neurosurgery . 2014; 61( Suppl 1): 111- 120. Google Scholar CrossRef Search ADS PubMed  8. Augsburger L, Farhat M, Reymond P et al.   Effect of flow diverter porosity on intraaneurysmal blood flow. Klin Neuroradiol . 2009; 19( 3): 204- 214. Google Scholar CrossRef Search ADS PubMed  9. Mut F, Raschi M, Scrivano E et al.   Association between hemodynamic conditions and occlusion times after flow diversion in cerebral aneurysms. J NeuroIntervent Surg . 2015; 7( 4): 286- 290. Google Scholar CrossRef Search ADS   10. Dholakia R, Sadasivan C, Fiorella DJ, Woo HH, Lieber BB. Hemodynamics of flow diverters. J Biomech Eng . 2017; 139( 2): 021002. Google Scholar CrossRef Search ADS   11. Sadasivan C, Cesar L, Seong J et al.   An original flow diversion device for the treatment of intracranial aneurysms. Evaluation of a rabbit elastase-induced model. Stroke . 2009; 40( 3): 952- 958. Google Scholar CrossRef Search ADS PubMed  12. Sadasivan C, Cesar L, Seong J, Wakhloo AK, Lieber BB. Treatment of rabbit elastase-induced aneurysm models by flow diverters: development of quantifiable indexes of device performance using digital subtraction angiography. IEEE Trans Med Imaging . 2009; 28( 7): 1117- 1125. Google Scholar CrossRef Search ADS PubMed  13. Hodis S, Ding YH, Dai D et al.   Relationship between aneurysm occlusion and flow diverting device oversizing in a rabbit model. J Neurointerv Surg . 2016; 8( 1): 94- 98. Google Scholar CrossRef Search ADS PubMed  14. Marbacher S, Marjamaa J, Abdelhameed E, Hernesniemi J, Niemelä M, Frösen J. The Helsinki rat microsurgical sidewall aneurysm model. J Vis Exp . 2014;( 92): e51071. 15. Aquarius R, Smits D, Gounis M, Leenders W, de Vries J. Flow diverter implantation in a rat model of sidewall aneurysm: a feasibility study. J Neurointerv Surg . 2018; 10: 88- 92. Available at: http://jnis.bmj.com/content/early/2017/02/08/neurintsurg-2016- 012878.long. Accessed November 6, 2017. Google Scholar CrossRef Search ADS PubMed  16. Niitsu Y, Sugidachi A, Ogawa T et al.   Repeat oral dosing of prasugrel, a novel P2Y12 receptor inhibitor, results in cumulative and potent antiplatelet and antithrombotic activity in several animal species. Eur J Pharmacol . 2008; 579( 1-3): 276- 282. Google Scholar CrossRef Search ADS PubMed  17. Guerrero A, Gonzáles-Correa JA, Arrebola MM, Muñoz-Marín J, Sánchez De La Cuesta F, De La Cruz JP. Antioxidant effects of a single dose of acetylsalicylic acid and salicylic acid in rat brain slices subjected to oxygen-glucose deprivation in relation with its antiplatelet effect. Neurosci Lett . 2004; 358( 3): 153- 156. Google Scholar CrossRef Search ADS PubMed  18. Schindelin J, Arganda-Carreras I, Frise E et al.   Fiji: an open-source platform for biological-image analysis. Nat Methods . 2012; 9( 7): 676- 682. Google Scholar CrossRef Search ADS PubMed  19. Rouchaud A, Brinjikji W, Lanzino G, Cloft HJ, Kadirvel R, Kallmes DF. Delayed hemorrhagic complications after flow diversion for intracranial aneurysms: a literature overview. Neuroradiology . 2016; 58( 2): 171- 177. Google Scholar CrossRef Search ADS PubMed  20. Miyachi S, Ohnishi H, Hiramatsu R, Izumi T, Matsubara N, Kuroiwa T. Innovations in endovascular treatment strategies for large carotid cavernous aneurysms – the safety and efficacy of a flow diverter. J Stroke Cerebrovasc Dis . 2017; 26( 5): 1071- 1080. Google Scholar CrossRef Search ADS PubMed  21. Byrne JV, Beltechi R, Yarnold JA, Birks J, Kamran M. Early experience in the treatment of intra-cranial aneurysms by endovascular flow diversion: a multicentre prospective study. PLoS One . 2010; 5( 9): e12492. Google Scholar CrossRef Search ADS PubMed  22. Klisch J, Turk A, Turner R, Woo HH, Fiorella D. Very late thrombosis of flow-diverting constructs after the treatment of large fusiform posterior circulation aneurysms. AJNR Am J Neuroradiol . 2011; 32( 4): 627- 632. Google Scholar CrossRef Search ADS PubMed  23. Cirillo L, Leonardi M, Dall’olio M et al.   Complications in the treatment of intracranial aneurysm with Silk stents: an analysis of 30 consecutive patients. Interv Neuroradiol . 2012; 18( 4): 413- 425. Google Scholar CrossRef Search ADS PubMed  24. Kühn AL, Wakhloo AK, Gounis MJ et al.   Use of self-expanding stents for better intracranial flow diverter wall apposition. Interv Neuroradiol . 2017; 23( 2): 129- 136. Google Scholar CrossRef Search ADS PubMed  25. Lubicz B, Collignon L, Raphaeli G et al.   Flow-diverter stent for the endovascular treatment of intracranial aneurysms. A prospective study in 29 patients with 34 aneurysms. Stroke . 2010; 41( 10): 2247- 2253. Google Scholar CrossRef Search ADS PubMed  26. Chalouhi N, Tjoumakaris SI, Gonzalez LF et al.   Spontaneous delayed migration/shortening of the pipeline embolization device: report of 5 cases. ANJR Am J Neuroradiol . 2013; 34( 12): 2326- 2330. Google Scholar CrossRef Search ADS PubMed  27. Mattingly T, Van Adel B, Dyer E et al.   Failure of aneurysm occlusion by flow diverter: a role for surgical bypass and parent artery occlusion. J Neurointerv Surg . 2015; 7( 4): e13. Google Scholar CrossRef Search ADS PubMed  28. Kulcsar Z, Houdart E, Bonafé A et al.   Intra-aneurysmal thrombosis as a possible cause of delayed aneurysm rupture after flow-diversion treatment. AJNR Am J Neuroradiol . 2011; 32( 1): 20- 25. Google Scholar PubMed  29. Fox B, Humphries WE, Doss VT, Hoit D, Elijovich L, Arthur AS. Rupture of giant vertebrobasilar aneurysm following flow diversion: mechanical stretch as a potential mechanism for early aneurysm rupture. J Neurointerv Surg . 2015; 7( 11): e37. Google Scholar CrossRef Search ADS PubMed  30. Arrese I, Sarabia R, Pintado R, Delgado-Rodriguez M. Flow-diverter devices for intracranial aneurysms: systematic review and meta-analysis. Neurosurgery . 2013; 73( 11): 193- 200. Google Scholar CrossRef Search ADS PubMed  31. Brinjikji W, Murad MH, Lanzino G, Cloft HJ, Kallmes DF. Endovascular treatment of intracranial aneurysms with flow diverters: a meta-analysis. Stroke . 2013; 44( 2): 442- 447. Google Scholar CrossRef Search ADS PubMed  32. Lv X, Yang H, Liu P, Li Y. Flow-diverter devices in the treatment of intracranial aneurysms: a meta-analysis and systematic review. Neuroradiol J . 2016; 29( 1): 66- 71. Google Scholar CrossRef Search ADS PubMed  33. Ye G, Zhang M, Deng L, Chen X, Wang Y. Meta-analysis of the efficiency and prognosis of intracranial aneurysm treated with flow diverter devices. J Mol Neurosci . 2016; 59( 1): 158- 167. Google Scholar CrossRef Search ADS PubMed  34. Kizilkilic O, Kocer N, Metaxas GE, Babic D, Homan R, Islak C. Utility of VasoCT in the treatment of intracranial aneurysm with flow-diverter stents. J Neurointerv Surg . 2012; 117( 1): 45- 49. 35. Flood TF, van der Bom IMJ, Strittmatter L et al.   Quantitative analysis of high-resolution, contrast-enhanced, cone-beam CT for the detection of intracranial in-stent hyperplasia. J Neurointerv Surg . 2015; 7( 2): 118- 125. Google Scholar CrossRef Search ADS PubMed  36. van der Marel K, Gounis MJ, Weaver JP et al.   Grading of regional apposition after flow-diverter treatment (GRAFT): a comparative evaluation of VasoCT and intravascular OCT. J Neurointerv Surg . 2016; 8( 8): 847- 852. Google Scholar CrossRef Search ADS PubMed  37. Marosfoi M, Clarencon F, Langan ET et al.   Acute thrombus formation on phosphorilcholine surface modified flow diverters. J Neurointerv Surg . 2017. Available at: http://jnis.bmj.com/content/early/2017/07/07/neurintsurg-2017-013175.long. Accessed November 6, 2017. 38. Aydin F. Do human intracranial arteries lack vasa vasorum? A comparative immunohistolochemical study of intracranial and systemic arteries. Acta Neuropathol . 1998; 96( 1): 22- 28. Google Scholar CrossRef Search ADS PubMed  39. Clower BR, Sullivan DM, Smith RR. Intracranial vessels lack vasa vasorum. J Neurosurg . 1984; 61( 1): 44- 48. Google Scholar CrossRef Search ADS PubMed  40. Marbacher S, Marjamaa J, Bradacova K et al.   Loss of mural cells leads to wall degeneration, aneurysm growth, and eventual rupture in a rat aneurysm model. Stroke . 2014; 45( 1): 248- 254. Google Scholar CrossRef Search ADS PubMed  41. Dai J, Louedec L, Philippe M, Michel JB, Houard X. Effect of blocking platelet activation with AZD6140 on development of abdominal aortic aneurysm in a rat aneurysmal model. J Vasc Surg . 2009; 49( 3): 719- 727. Google Scholar CrossRef Search ADS PubMed  Acknowledgements We would like to thank Dr Marjamaa and Dr Niemelä from the Neurosurgical Research Group, Helsinki, Finland for their help with the animal model. Supplemental digital content is available for this article at www.neurosurgery-online.com. Supplemental Digital Content 1. Figure. Animal body weight increase during follow-up. Animal body weight increase after 1 month (left) and 3 months (right) of follow-up. Supplemental Digital Content 2. Figure. Implant malapposition. Number of malapposed flow diverter struts (top row) and the average distance between malapposed flow diverter struts and the parent artery wall (bottom row) stratified per occlusion status at both follow-up moments. Supplemental Digital Content 3. Figure. Flow diverter migration. Macroscopic image (A) shows signs of hemorrhage outside the aneurysm pouch. Microscopic image of aortic cross-section (B) clearly shows the absence of the flow diverter. Detail (insert) shows a continuous aneurysm wall (arrows) until the wall is discontinued and shaped irregularly (asterisk), suggesting a rupture. COMMENT The authors describe their experience with flow diverters of different pore densities in the treatment of aneurysms on an animal model (in the aorta of 36 Wistar rats). The treatment arms were divided equally between aneurysms treated with low pore density flow diverters (10 pores/mm) and aneurysms treated with high pore density flow diverters (23 pores/mm) based on the assumption that high pore flow diverters are beneficial for intracranial aneurysm healing. Half of the animals in each group was sacrificed 1 month after treatment and the other half was sacrificed 3 months after treatment. There was no difference in aneurysm occlusion rate between low pore density and high pore density groups at 1 or 3 months after treatment. The number of stent struts not in contact with the parent artery wall and the average strut distance to the parent artery wall was higher in nonoccluded aneurysms at time of sacrifice. The authors concluded that implant apposition is more important than pore density for aneurysm occlusion in their animal model. These findings are significant because implant malapposition affects aneurysm healing and can be associated with thromboembolic complications as well as migration of flow diverter. In one of the rats in the study, a flow diverter migrated in the vessel resulting in an incompletely sealed aneurysm and subsequent rupture, which is a possible cause for delayed hemorrhage of elective aneurysms previously treated with flow diverters. Future clinical studies evaluating neuro-imaging modalities that visualize wall apposition after deployment of flow diverting stents and the relationship with outcomes (aneurysm obliteration, hemorrhagic complications) are necessary. Rafael A. Ortiz David J. Langer New York, New York Copyright © 2018 by the Congress of Neurological Surgeons http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Neurosurgery Oxford University Press

Loading next page...
 
/lp/ou_press/the-importance-of-wall-apposition-in-flow-diverters-l3TvhR0KUn
Publisher
Oxford University Press
Copyright
Copyright © 2018 by the Congress of Neurological Surgeons
ISSN
0148-396X
eISSN
1524-4040
D.O.I.
10.1093/neuros/nyy092
Publisher site
See Article on Publisher Site

Abstract

Abstract BACKGROUND It is assumed that high pore densities in flow diverters (FDs) are beneficial for intracranial aneurysm (IA) healing. However, various animal studies are not conclusive on the issue, suggesting that other factors are in play. One important factor might be wall apposition. OBJECTIVE To (1) determine the relationship between FD pore density and aneurysm occlusion, and (2) determine the relationship between FD wall apposition and aneurysm occlusion. METHODS Saccular aneurysms were microsurgically created in the aorta of 36 Wistar rats. Twelve rats received a low pore density FD (10 pores/mm2), 12 rats received a high pore density FD (23 pores/mm2), and the remaining 12 rats served as a control group. Six animals from each group were sacrificed 1 and 3 mo after surgery. We determined aneurysm occlusion, the number of struts not in contact with the aorta wall, and the average distance from malapposed struts to aorta wall through histology. RESULTS No significant differences were found in aneurysm occlusion between the low pore density and high pore density groups (P > .05) after 1 and 3 mo of follow-up. The average number of malapposed struts was lower for the occluded aneurysm group (4.4 ± 1.9) compared to the nonoccluded aneurysm group (7.7 ± 2.6, P < .01). The average distance between malapposed struts and parent artery wall was lower for the occluded aneurysm group (33.9 μm ± 11.5 μm) than for the nonoccluded aneurysm group (48.7 μm ± 18.8 μm, P < .05). CONCLUSION Wall apposition is more important than pore density for aneurysm occlusion. Animal model, Endovascular, Flow diverter, Intracranial aneurysm, Rats, Vascular biology, Wall apposition ABBREVIATIONS ABBREVIATIONS FD flow diverter IA intracranial aneurysm An intracranial aneurysm (IA) is a bulging area in the wall of an intracranial artery that may be abnormally weak and might rupture.1 IAs occur in about 2% of the population2,3 and account for approximately 85% of the nontraumatic subarachnoid hemorrhages,4 which can have devastating consequences such as permanent neurological damage and even death.5 In case an unruptured IA is diagnosed and conservative treatment is not an option, a flow diverter (FD) can be placed in the parent artery covering the ostium of the aneurysm. Ideally, this leads to intra-aneurysmal thrombosis due to reduced blood flow in the aneurysm sac, and parent vessel reconstruction due to endothelial cell coverage of the implant.6,7 Porosity (the nonsolid fraction of the implant surface) and pore density (number of pores per mm2) at the ostium are generally considered to determine the efficacy of FDs.6,7 Good clinical results with minimal side branch occlusion are observed with an implant porosity in the 60% to 80% range.8 The exact effect of pore density, however, is less clear. Computational9 and experimental10 studies demonstrate that higher implant pore densities lead to lower intrasaccular blood flow, which should promote rapid IA occlusion. Animal studies are less conclusive as some show a (slight) beneficial effect,11,12 while another shows no effect13 of pore density at all. The discrepancy between these studies suggests that other factors are in play that cannot be fully controlled. One of these factors might be implant wall apposition. We therefore set 2 goals in the current study: (1) whether there is a relationship between FD pore density and aneurysm occlusion, (2) whether there is a relationship between FD wall apposition and aneurysm occlusion. We used a rat model in which the aneurysm is prone to expansion and rupture to implant FDs with varying pore density. METHODS Aneurysms were microsurgically created in 36 Wistar rats according to previously published procedures.14,15 One group (n = 12) received a low pore density FD (10 pores/mm2), 1 group (n = 12) received a high pore density FD (23 pores/mm2), and the remaining 12 rats served as a control group. Six animals from each of the 3 groups were sacrificed at 2 different time points: 1 and 3 mo after surgery. Macroscopic and microscopic evaluation followed. Animals All experiments were approved by institutional animal care and use committee. Six-to-eight weeks old, Wistar WU rats (n = 72, Charles River, Den Bosch, The Netherlands) were used in this study. To limit the introduction of unknown factors in our experiment, we used only male animals, based on the available literature.14,15 Potential gender-related differences need to be addressed in future studies. Animals were housed in a room with an average temperature of 22°C (±1°C) and a 50/50 light/dark cycle. Animals had access to a pellet diet and sterilized water ad libitum and received humane care in agreement with institutional guidelines. A thoracic aorta segment between the left subclavian artery and the upper intelcostal artery was harvested from 36 donor rats. These segments were then incubated overnight (10 h) in a 0.1% sodium dodecyl sulfate in milli-Q water solution at a temperature of 37°C. The incubation step results in a decellularized vessel wall, with an intact extracellular matrix. After the incubation step, the aortic segment was stored in PBS on ice until surgery later that day. A microsurgical side wall aneurysm was then created in 36 recipient rats, by attaching the decellularized aortic segment to the abdominal aorta between the renal arteries and aortic bifurcation. Successful completion of the surgery was marked by a pulsating aneurysm without seeping or oozing of blood. Immediately after anastomosis creation, surgery was completed in 12 of the 36 animals and no FD was implanted. These animals served as controls. The remaining 24 animals received either a low pore density FD (n = 12) or a high pore density FD (n = 12), which was implanted in the aorta as previously described.15 Randomization was achieved by allocating each subsequent animal to a predetermined cycle of experimental groups: control—low pore density FD—high pore density FD. FDs and Dual Anti-Platelet Therapy The low pore density implant consisted of 20 cobalt chrome wires and 4 platinum wires, with a porosity of approximately 80%. The high pore density implant consisted of 30 cobalt chrome wires and 6 platinum wires with a porosity of approximately 70%. The 2 implants used in this experiment were both Surpass FDs and were produced by Stryker Neurovascular (Fremont, California). All rats received daily dual anti-platelet therapy through oral gavage: 5 mg/kg clopidogrel16 and 1 mg/kg/d of acetylsalicylic acid.17 Dual anti-platelet therapy started at least 3 d before surgery and was continued for the rest of the experiment. Group allocation was unknown to animal laboratory personnel during the entire experiment. Tissue Harvesting and Processing One month (n = 18) and 3 mo (n = 18) after surgery, the rats underwent a second laparotomy. Through visual inspection, the position of the aneurysm and the FD was determined, and clamps were placed above and below to isolate this aortic section. The complete aortic segment, including FD and aneurysm, was subsequently excised, flushed with saline, and fixed overnight in 4% formalin. Tissues obtained from control rats were embedded in paraffin and sectioned with a microtome. All tissues containing an FD embedded in methacrylate and sectioned with a saw microtome (SP1600, Leica Microsystems, Eindhoven, Netherlands). All tissues were stained with hematoxylin and eosin and digitalized using the Vision-Tek slide scanner (Sakura Finetek, Alphen aan den Rijn, Netherlands). Outcome Measures and Statistical Analyses Histological specimens were obtained as axial sections (ie, perpendicular to the long axis of the aorta) from the aorta and the aneurysm. Measurements were performed at the level of the largest diameter of the aneurysm. All slides were coded to mask experimental group allocation and to minimize bias. We measured the following parameters from our histological sections: the percentage of (organized) thrombus within the aneurysm (aneurysm occlusion), the number of struts not in contact with the aorta wall (implant apposition), and the average distance from malapposed struts to aorta wall (Figure 1). All measurements were performed using FIJI (ImageJ version 1.47), an open-source platform for biological-image analysis.18 FIGURE 1. View largeDownload slide Schematic depiction of typical cross-section of an aorta with the implanted FD and the microsurgically created aneurysm. In this example, the aneurysm is fully occluded with thrombus (pink: organized thrombus, red: fresh thrombus). Perfect implant apposition can be seen on the left side of the cross section where all FD struts are in contact with the parent artery wall (insert A). Imperfect implant apposition can be seen on the right side of the cross section where 3 FD struts are not touching the vessel wall (insert B). The red lines depict the measurements to calculate the distance of a malapposed strut perpendicular to the parent artery wall (insert B). FIGURE 1. View largeDownload slide Schematic depiction of typical cross-section of an aorta with the implanted FD and the microsurgically created aneurysm. In this example, the aneurysm is fully occluded with thrombus (pink: organized thrombus, red: fresh thrombus). Perfect implant apposition can be seen on the left side of the cross section where all FD struts are in contact with the parent artery wall (insert A). Imperfect implant apposition can be seen on the right side of the cross section where 3 FD struts are not touching the vessel wall (insert B). The red lines depict the measurements to calculate the distance of a malapposed strut perpendicular to the parent artery wall (insert B). All statistical analyses were performed using GraphPad Prism (Version 5.03, GraphPad Software, Inc, LaJolla, California). Three groups were compared using a 1-way ANOVA with a post hoc Bonferroni multiple comparison test to pinpoint differences between groups when variances were equal. When variances were not equal, 3 groups were compared using the Kruskal–Wallis test with a post hoc Dunn's multiple comparison test to pinpoint differences between groups. When 2 groups were compared, a Student's t-test was used for analysis. If variances were not equal for both groups, an unpaired t-test with Welch's correction was performed. Level of statistical significance is indicated as <0.05 (*), <0.01 (**), or <0.0001 (***). RESULTS General Information, Mortality, and Complications Two of the 36 recipient rats died during follow-up (mortality: 5.6%). One rat died due to an intestinal perforation the night after surgery, probably due to intestinal manipulation during laparotomy. Another rat died 50 d after surgery, but the cause of death remained unknown, even after necropsy (Table). TABLE. Overview of encountered deaths and complications in each group. Animals included for histological analysis are also specified   Number of animals    1-mo follow-up  3-mo follow-up  Control group  Deaths: 0  Deaths: 1 intestinal perforation the night after surgery    Complications: 0  Complications: 0    Included for histological analysis: 6  Included for histological analysis: 5  Low pore density FD group  Deaths: 0  Deaths: 0    Complications: 0  Complications: 0    Included for histological analysis: 6  Included for histological analysis: 6  High pore density FD group  Deaths: 0  Deaths: 1 unknown cause of death 50 d after surgery    Complications: 1 FD migration, rat was excluded for measurements  Complications: 0    Included for histological analysis: 5  Included for histological analysis: 6 deceased rat was not excluded    Number of animals    1-mo follow-up  3-mo follow-up  Control group  Deaths: 0  Deaths: 1 intestinal perforation the night after surgery    Complications: 0  Complications: 0    Included for histological analysis: 6  Included for histological analysis: 5  Low pore density FD group  Deaths: 0  Deaths: 0    Complications: 0  Complications: 0    Included for histological analysis: 6  Included for histological analysis: 6  High pore density FD group  Deaths: 0  Deaths: 1 unknown cause of death 50 d after surgery    Complications: 1 FD migration, rat was excluded for measurements  Complications: 0    Included for histological analysis: 5  Included for histological analysis: 6 deceased rat was not excluded  View Large TABLE. Overview of encountered deaths and complications in each group. Animals included for histological analysis are also specified   Number of animals    1-mo follow-up  3-mo follow-up  Control group  Deaths: 0  Deaths: 1 intestinal perforation the night after surgery    Complications: 0  Complications: 0    Included for histological analysis: 6  Included for histological analysis: 5  Low pore density FD group  Deaths: 0  Deaths: 0    Complications: 0  Complications: 0    Included for histological analysis: 6  Included for histological analysis: 6  High pore density FD group  Deaths: 0  Deaths: 1 unknown cause of death 50 d after surgery    Complications: 1 FD migration, rat was excluded for measurements  Complications: 0    Included for histological analysis: 5  Included for histological analysis: 6 deceased rat was not excluded    Number of animals    1-mo follow-up  3-mo follow-up  Control group  Deaths: 0  Deaths: 1 intestinal perforation the night after surgery    Complications: 0  Complications: 0    Included for histological analysis: 6  Included for histological analysis: 5  Low pore density FD group  Deaths: 0  Deaths: 0    Complications: 0  Complications: 0    Included for histological analysis: 6  Included for histological analysis: 6  High pore density FD group  Deaths: 0  Deaths: 1 unknown cause of death 50 d after surgery    Complications: 1 FD migration, rat was excluded for measurements  Complications: 0    Included for histological analysis: 5  Included for histological analysis: 6 deceased rat was not excluded  View Large All remaining rats displayed normal behavior and did not seem to be hindered by the operation or the FD. An average body weight increase of 58.1 g (SD ± 17.3 g), and 137.9 g (SD ± 30.1 g) was seen after 1 and 3 mo of follow-up, respectively. Body weight between the groups showed no statistically significant differences at each time point (1-way ANOVA, Figure, Supplemental Digital Content 1). A complication was observed in 1 animal. Upon sacrifice at 1 mo, signs of hemorrhage were clearly present around the aneurysm. After histological sectioning, we concluded that the FD had migrated after implantation, which led to incomplete coverage of the aneurysm neck (Table). This animal was excluded from further analysis. Outcome Measures Stratified per Implant Aneurysm Occlusion After 1 mo of follow-up, complete aneurysm occlusion was seen in 0 out of 6 animals in the control group, 2 out of 6 animals in the low pore density group, and 1 out of 5 animals in the high pore density group. The average aneurysm occlusion was 58.2% (SD ± 13.5%), 96.6% (SD ± 5.6%), and 93.8% (SD ± 8.3%) for the control group, the low pore density implant group and the high pore density implant group, respectively (Figure 2). After 3 mo of follow-up, total aneurysm occlusion was seen in 0 out of 5 animals in the control group, 4 out of 6 animals in the low pore density group, and 4 out of 6 animals in the high pore density group. The average aneurysm occlusion was 40.1% (SD ± 23.1%), 96.2% (SD ± 6.9%), and 99.4% (SD ± 1.0%) for the control group, the low pore density implant group, and the high pore density implant group, respectively (Figure 2). For both time points, the low pore density and high pore density implant groups statistically differed from the control group (1 mo: Bonferroni's multiple comparison test P < .0001, 3 mo: Dunn's multiple comparison test < 0.05), but not from each other (1 mo: Bonferroni's multiple comparison test P > .05, 3 mo: Dunn's multiple comparison test >0.05, Figure 2). FIGURE 2. View largeDownload slide Outcome measures stratified per implant group. Aneurysm occlusion was similar in both implant groups, but significantly lower in the control group (*P < .05; ***P < .0001). FIGURE 2. View largeDownload slide Outcome measures stratified per implant group. Aneurysm occlusion was similar in both implant groups, but significantly lower in the control group (*P < .05; ***P < .0001). Outcome Measures Stratified per Aneurysm Occlusion Status Aneurysm Occlusion When analyzing the occlusion status, independent of follow-up time, 11 out of 23 animals had a nonoccluded aneurysm, in all cases observed as neck remnants, with an average occlusion of 93.5% (SD ± 6.9%; Figure 3). Of these 11 animals, 6 received a low pore density FD and 5 received a high pore density FD. FIGURE 3. View largeDownload slide Comparison of occluded vs nonoccluded aneurysms regarding aneurysm occlusion (left), the number of malapposed FD struts (middle) and the average distance between malapposed FD struts and the parent artery wall (right) independent of follow-up time (*P < .05 and **P < .01). FIGURE 3. View largeDownload slide Comparison of occluded vs nonoccluded aneurysms regarding aneurysm occlusion (left), the number of malapposed FD struts (middle) and the average distance between malapposed FD struts and the parent artery wall (right) independent of follow-up time (*P < .05 and **P < .01). Implant Apposion After 1 mo of follow-up, the average number of malapposed struts was 4.7 (SD ± 2.5) and 7.6 (SD ± 2.4) for the occluded aneurysm group and the nonoccluded aneurysm group, respectively (Figure, Supplemental Digital Content 2). After 3 mo of follow-up, the average number of malapposed struts was 4.3 (SD ± 1.8) and 7.8 (SD ± 3.3) for the occluded aneurysm group and the nonoccluded aneurysm group, respectively (P < .05, Figure, Supplemental Digital Content 2). The average number of malapposed struts, independent of follow-up time, was lower for the occluded aneurysm group (4.4 ± 1.9) than for the nonoccluded aneurysm group (7.7 ± 2.6; Figure 3, P < .01). Distance Malapposed Struts to Parent Artery Wall After 1 mo of follow-up, the average distance between malapposed struts and parent artery wall was 45.6 μm (SD ± 10.2 μm) and 50.0 μm (SD ± 17.2 μm) for the occluded aneurysm group and the nonoccluded aneurysm group, respectively (Figure, Supplemental Digital Content 2). After 3 mo of follow-up, the average distance between malapposed struts and parent artery wall was 29.5 μm (SD ± 8.9 μm) and 46.1 μm (SD ± 24.20 μm) for the occluded aneurysm group and the nonoccluded aneurysm group, respectively (Figure, Supplemental Digital Content 2). The average distance between malapposed struts and parent artery wall, independent of follow-up time, was lower for the occluded aneurysm group (33.9 μm ± 11.5 μm) than for the nonoccluded aneurysm group (48.7 μm ± 18.8 μm; Figure 3, P < .05). DISCUSSION In this study, we investigated the effect of FD pore density and FD wall apposition on aneurysm occlusion using a rat aneurysm model. Although treatment with FDs clearly led to massive intra-aneurysmal thrombosis compared to control animals, no distinct difference in aneurysm healing was observed for the different FD designs. Incomplete occlusion, observed as neck remnants, was more likely when FD apposition was worse: FDs in nonoccluded aneurysms displayed more malapposed struts and the average distance between malapposed strut and parent artery wall was higher compared to those in occluded aneurysms. Our findings support the results of Rouchaud et al19 who found that good wall apposition was strongly associated with complete aneurysm occlusion after FD therapy in a series of 41 rabbits. Our findings are in contrast with computational,9 experimental,10 and animal studies11,12 in which a beneficial effect of high pore density implants is demonstrated. Although these studies seem to contradict each other, it is possible that a higher FD pore density indeed results in faster aneurysm occlusion, but only if implant apposition is perfect. A hypothesis supported by Mut et al,9 who reported a case of incomplete aneurysm occlusion due to poor FD apposition resulting in residual blood flow to the aneurysm. Also, Miyachi et al20 reported “a small leak from the FD” at the aneurysm neck of nonoccluding IAs. Implant malapposition does not only affect aneurysm healing, but is also associated with serious adverse events such as thromboembolic complications21-23 and FD migration.23,24 Both adverse events can have devastating clinical consequences. Thromboembolisms can lead to ischemic stroke resulting in permanent neurological damage. Implant migration can lead to ostium exposure and subsequent aneurysm rupture.25-27 In one of the rats in our study, a high pore density FD migrated in the vessel, resulting in an incompletely sealed aneurysm and a subsequent rupture (Figure, Supplemental Digital Content 3). The animal survived as the hemorrhage was contained by peri-aneurysmal fibrous tissue. Upon sacrifice (1 mo), signs of the hemorrhage were clearly visible macroscopically and a torn wall and organized thrombus was seen microscopically. This case illustrates the risk associated with FD migration, a phenomenon which has been described previously in clinical studies.19,28,29 Meta-analyses show that about 20% to 25% of the IAs are not occluded 6 mo after FD treatment, that the risk of thromboembolic complications is around 5% to 6% and the risk for aneurysm rupture after FD is around 3%.30-33 It is unknown to what extent FD malapposition is responsible for these complications. One reason might be the use of 2-dimensional digital subtraction angiography, which is not the optimal method for assessing FD deployment and apposition.19,34,35 Newer imaging techniques, such as high-resolution contrast-enhanced cone beam computed tomography34,35 or optical coherence tomography,36,37 are better capable of visualizing the FD in the parent vessel and could therefore give a better overview of how the aforementioned complications and FD apposition are linked. Limitations This study has several limitations. First, it is expected that the microsurgery itself will elicit a tissue response, which partially overlaps with aneurysm healing due to the FD treatment. It is currently unknown whether aneurysm healing would be different if this overlap was not present. Further refinement of the animal model could be achieved by creating the aneurysm first and placing the FD at a later stage. Addition of angiography to the study protocol is essential, as aneurysm patency should be guaranteed before FD implantation. Furthermore, the aneurysm in our animal model is created in the abdominal space. Besides a difference in vessel structure, such as the presence of vasa vasorum in the aorta and absence in intracranial arteries,38,39 the peri-aneurysmal environment also differs. It is currently unknown if these differences can alter the healing process in the aneurysm after FD placement. Further use of the current model and comparison to results of IA models might elucidate the importance of these discrepancies. Another limitation is the lack of aneurysm expansion and aneurysm rupture in our study. This is in sharp contrast to the results reported by Marbacher et al.40 We expect that the dual-antiplatelet regime could have caused this discrepancy. In a similar rat model, in which a decellularized guinea pig aorta segment was implanted in the abdominal aorta, administering a P2Y12 antagonist led to significantly less growth of the decellularized segment.41 The authors suggested that the lack of thrombus renewal, which limits the intrathrombus build-up of polymorphonuclear cells and the proteases they contain, lead to a healing-type response including smooth muscle cell invasion and limited aneurysm wall damage. This raises the interesting question if dual-antiplatelet therapy is an important aspect of successful FD treatment. The absence of tortuosity in the abdominal aorta compared to the intracranial arteries is another drawback of our current animal model. However, this also makes our results more intriguing as it is expected that FD apposition should be optimal in a straight artery like the aorta. We observed that small variations in apposition alter aneurysm healing response in our animal model, apposition may be more challenging to achieve in the tortuous and sometimes even dysplastic or atherosclerotic intracranial arteries where these devices are used. More research is needed on how vessel tortuosity, dysplasticity, and atherosclerosis affects implant apposition and subsequent aneurysm healing. Finally, the study would benefit from the use of more animals. It would make the use of multivariate analyses methods possible, which could elucidate whether interaction was present between variables such as implant malapposition and average distance between malapposed struts and parent artery wall. Furthermore, use of more animals per time point or using more animals to introduce additional time points would help us understand if all aneurysms would eventually occlude or if neck remnants persisted. A shift towards complete occlusion during longer follow-up is now suggested by our data, as 3 out of 11 and 8 out of 12 aneurysms are occluded after 1 and 3 mo of follow-up, respectively. However, this finding is not statistically significant. CONCLUSION Based on our results, we conclude that implant apposition is more important than pore density for aneurysm occlusion. It seems that in this biological model malapposition can “overrule” the beneficial effects that high pore density implants can have on IA healing. Disclosures This study was funded by Stryker Neurovascular. Dr de Vries works as a consultant for Stryker Neurovascular. Dr Gounis has been a consultant on a fee-per-hour basis for Codman Neurovascular, InNeuroCo, Medtronic Neurovascular, and Stryker Neurovascular; holds stock in InNeuroCo; and has received research support from the Research support from the National Institutes of Health (NIH), Anaconda, Codman Neurovascular, Gentuity, InNeuroCo, Microvention, Medtronic Neurovascular, MIVI Neurosciences, Neuravi, Philips Healthcare, InNeuroCo, Rapid Medical, R92M, Stryker Neurovascular, The Stroke Project, and the Wyss Institute. REFERENCES 1. Frösen J, Tulamo R, Paetae A, Laaksamo E, Korja M, Laakso A. Saccular intracranial aneurysm: pathology and mechanisms. J Vasc Surg . 2012; 123( 6): 719- 727. 2. Vernooij MW, Ikram MA, Tanghe HL et al.   Incidental findings on brain MRI in the general population. N Engl J Med . 2007; 357( 18): 1821- 1828. Google Scholar CrossRef Search ADS PubMed  3. Peymani A, Adams HHH, Cremers LGM et al.   Genetic determinants of unruptured intracranial aneurysms in the general population. Stroke . 2015; 46( 10): 2961- 2964. Google Scholar CrossRef Search ADS PubMed  4. van Gijn J, Kerr RS, Rinkel GJE. Subarachnoid haemorrhage. Lancet . 2007; 369( 9558): 306- 318. Google Scholar CrossRef Search ADS PubMed  5. Brown RD, Broderick JP. Unruptured intracranial aneurysms: epidemiology, natural history, management options, and familial screening. Lancet Neurol . 2014; 13( 4): 393- 404. Google Scholar CrossRef Search ADS PubMed  6. Lieber BB, Sadasivan C. Endoluminal scaffolds for vascular reconstruction and exclusion of aneurysms from the cerebral circulation. Stroke . 2010; 41( 10 Suppl): S21- S25. Google Scholar CrossRef Search ADS PubMed  7. Wakhloo AK, Gounis MJ. Revolution in aneurysm treatment: Flow diversion to cure aneurysms: A paradigm shift. Neurosurgery . 2014; 61( Suppl 1): 111- 120. Google Scholar CrossRef Search ADS PubMed  8. Augsburger L, Farhat M, Reymond P et al.   Effect of flow diverter porosity on intraaneurysmal blood flow. Klin Neuroradiol . 2009; 19( 3): 204- 214. Google Scholar CrossRef Search ADS PubMed  9. Mut F, Raschi M, Scrivano E et al.   Association between hemodynamic conditions and occlusion times after flow diversion in cerebral aneurysms. J NeuroIntervent Surg . 2015; 7( 4): 286- 290. Google Scholar CrossRef Search ADS   10. Dholakia R, Sadasivan C, Fiorella DJ, Woo HH, Lieber BB. Hemodynamics of flow diverters. J Biomech Eng . 2017; 139( 2): 021002. Google Scholar CrossRef Search ADS   11. Sadasivan C, Cesar L, Seong J et al.   An original flow diversion device for the treatment of intracranial aneurysms. Evaluation of a rabbit elastase-induced model. Stroke . 2009; 40( 3): 952- 958. Google Scholar CrossRef Search ADS PubMed  12. Sadasivan C, Cesar L, Seong J, Wakhloo AK, Lieber BB. Treatment of rabbit elastase-induced aneurysm models by flow diverters: development of quantifiable indexes of device performance using digital subtraction angiography. IEEE Trans Med Imaging . 2009; 28( 7): 1117- 1125. Google Scholar CrossRef Search ADS PubMed  13. Hodis S, Ding YH, Dai D et al.   Relationship between aneurysm occlusion and flow diverting device oversizing in a rabbit model. J Neurointerv Surg . 2016; 8( 1): 94- 98. Google Scholar CrossRef Search ADS PubMed  14. Marbacher S, Marjamaa J, Abdelhameed E, Hernesniemi J, Niemelä M, Frösen J. The Helsinki rat microsurgical sidewall aneurysm model. J Vis Exp . 2014;( 92): e51071. 15. Aquarius R, Smits D, Gounis M, Leenders W, de Vries J. Flow diverter implantation in a rat model of sidewall aneurysm: a feasibility study. J Neurointerv Surg . 2018; 10: 88- 92. Available at: http://jnis.bmj.com/content/early/2017/02/08/neurintsurg-2016- 012878.long. Accessed November 6, 2017. Google Scholar CrossRef Search ADS PubMed  16. Niitsu Y, Sugidachi A, Ogawa T et al.   Repeat oral dosing of prasugrel, a novel P2Y12 receptor inhibitor, results in cumulative and potent antiplatelet and antithrombotic activity in several animal species. Eur J Pharmacol . 2008; 579( 1-3): 276- 282. Google Scholar CrossRef Search ADS PubMed  17. Guerrero A, Gonzáles-Correa JA, Arrebola MM, Muñoz-Marín J, Sánchez De La Cuesta F, De La Cruz JP. Antioxidant effects of a single dose of acetylsalicylic acid and salicylic acid in rat brain slices subjected to oxygen-glucose deprivation in relation with its antiplatelet effect. Neurosci Lett . 2004; 358( 3): 153- 156. Google Scholar CrossRef Search ADS PubMed  18. Schindelin J, Arganda-Carreras I, Frise E et al.   Fiji: an open-source platform for biological-image analysis. Nat Methods . 2012; 9( 7): 676- 682. Google Scholar CrossRef Search ADS PubMed  19. Rouchaud A, Brinjikji W, Lanzino G, Cloft HJ, Kadirvel R, Kallmes DF. Delayed hemorrhagic complications after flow diversion for intracranial aneurysms: a literature overview. Neuroradiology . 2016; 58( 2): 171- 177. Google Scholar CrossRef Search ADS PubMed  20. Miyachi S, Ohnishi H, Hiramatsu R, Izumi T, Matsubara N, Kuroiwa T. Innovations in endovascular treatment strategies for large carotid cavernous aneurysms – the safety and efficacy of a flow diverter. J Stroke Cerebrovasc Dis . 2017; 26( 5): 1071- 1080. Google Scholar CrossRef Search ADS PubMed  21. Byrne JV, Beltechi R, Yarnold JA, Birks J, Kamran M. Early experience in the treatment of intra-cranial aneurysms by endovascular flow diversion: a multicentre prospective study. PLoS One . 2010; 5( 9): e12492. Google Scholar CrossRef Search ADS PubMed  22. Klisch J, Turk A, Turner R, Woo HH, Fiorella D. Very late thrombosis of flow-diverting constructs after the treatment of large fusiform posterior circulation aneurysms. AJNR Am J Neuroradiol . 2011; 32( 4): 627- 632. Google Scholar CrossRef Search ADS PubMed  23. Cirillo L, Leonardi M, Dall’olio M et al.   Complications in the treatment of intracranial aneurysm with Silk stents: an analysis of 30 consecutive patients. Interv Neuroradiol . 2012; 18( 4): 413- 425. Google Scholar CrossRef Search ADS PubMed  24. Kühn AL, Wakhloo AK, Gounis MJ et al.   Use of self-expanding stents for better intracranial flow diverter wall apposition. Interv Neuroradiol . 2017; 23( 2): 129- 136. Google Scholar CrossRef Search ADS PubMed  25. Lubicz B, Collignon L, Raphaeli G et al.   Flow-diverter stent for the endovascular treatment of intracranial aneurysms. A prospective study in 29 patients with 34 aneurysms. Stroke . 2010; 41( 10): 2247- 2253. Google Scholar CrossRef Search ADS PubMed  26. Chalouhi N, Tjoumakaris SI, Gonzalez LF et al.   Spontaneous delayed migration/shortening of the pipeline embolization device: report of 5 cases. ANJR Am J Neuroradiol . 2013; 34( 12): 2326- 2330. Google Scholar CrossRef Search ADS PubMed  27. Mattingly T, Van Adel B, Dyer E et al.   Failure of aneurysm occlusion by flow diverter: a role for surgical bypass and parent artery occlusion. J Neurointerv Surg . 2015; 7( 4): e13. Google Scholar CrossRef Search ADS PubMed  28. Kulcsar Z, Houdart E, Bonafé A et al.   Intra-aneurysmal thrombosis as a possible cause of delayed aneurysm rupture after flow-diversion treatment. AJNR Am J Neuroradiol . 2011; 32( 1): 20- 25. Google Scholar PubMed  29. Fox B, Humphries WE, Doss VT, Hoit D, Elijovich L, Arthur AS. Rupture of giant vertebrobasilar aneurysm following flow diversion: mechanical stretch as a potential mechanism for early aneurysm rupture. J Neurointerv Surg . 2015; 7( 11): e37. Google Scholar CrossRef Search ADS PubMed  30. Arrese I, Sarabia R, Pintado R, Delgado-Rodriguez M. Flow-diverter devices for intracranial aneurysms: systematic review and meta-analysis. Neurosurgery . 2013; 73( 11): 193- 200. Google Scholar CrossRef Search ADS PubMed  31. Brinjikji W, Murad MH, Lanzino G, Cloft HJ, Kallmes DF. Endovascular treatment of intracranial aneurysms with flow diverters: a meta-analysis. Stroke . 2013; 44( 2): 442- 447. Google Scholar CrossRef Search ADS PubMed  32. Lv X, Yang H, Liu P, Li Y. Flow-diverter devices in the treatment of intracranial aneurysms: a meta-analysis and systematic review. Neuroradiol J . 2016; 29( 1): 66- 71. Google Scholar CrossRef Search ADS PubMed  33. Ye G, Zhang M, Deng L, Chen X, Wang Y. Meta-analysis of the efficiency and prognosis of intracranial aneurysm treated with flow diverter devices. J Mol Neurosci . 2016; 59( 1): 158- 167. Google Scholar CrossRef Search ADS PubMed  34. Kizilkilic O, Kocer N, Metaxas GE, Babic D, Homan R, Islak C. Utility of VasoCT in the treatment of intracranial aneurysm with flow-diverter stents. J Neurointerv Surg . 2012; 117( 1): 45- 49. 35. Flood TF, van der Bom IMJ, Strittmatter L et al.   Quantitative analysis of high-resolution, contrast-enhanced, cone-beam CT for the detection of intracranial in-stent hyperplasia. J Neurointerv Surg . 2015; 7( 2): 118- 125. Google Scholar CrossRef Search ADS PubMed  36. van der Marel K, Gounis MJ, Weaver JP et al.   Grading of regional apposition after flow-diverter treatment (GRAFT): a comparative evaluation of VasoCT and intravascular OCT. J Neurointerv Surg . 2016; 8( 8): 847- 852. Google Scholar CrossRef Search ADS PubMed  37. Marosfoi M, Clarencon F, Langan ET et al.   Acute thrombus formation on phosphorilcholine surface modified flow diverters. J Neurointerv Surg . 2017. Available at: http://jnis.bmj.com/content/early/2017/07/07/neurintsurg-2017-013175.long. Accessed November 6, 2017. 38. Aydin F. Do human intracranial arteries lack vasa vasorum? A comparative immunohistolochemical study of intracranial and systemic arteries. Acta Neuropathol . 1998; 96( 1): 22- 28. Google Scholar CrossRef Search ADS PubMed  39. Clower BR, Sullivan DM, Smith RR. Intracranial vessels lack vasa vasorum. J Neurosurg . 1984; 61( 1): 44- 48. Google Scholar CrossRef Search ADS PubMed  40. Marbacher S, Marjamaa J, Bradacova K et al.   Loss of mural cells leads to wall degeneration, aneurysm growth, and eventual rupture in a rat aneurysm model. Stroke . 2014; 45( 1): 248- 254. Google Scholar CrossRef Search ADS PubMed  41. Dai J, Louedec L, Philippe M, Michel JB, Houard X. Effect of blocking platelet activation with AZD6140 on development of abdominal aortic aneurysm in a rat aneurysmal model. J Vasc Surg . 2009; 49( 3): 719- 727. Google Scholar CrossRef Search ADS PubMed  Acknowledgements We would like to thank Dr Marjamaa and Dr Niemelä from the Neurosurgical Research Group, Helsinki, Finland for their help with the animal model. Supplemental digital content is available for this article at www.neurosurgery-online.com. Supplemental Digital Content 1. Figure. Animal body weight increase during follow-up. Animal body weight increase after 1 month (left) and 3 months (right) of follow-up. Supplemental Digital Content 2. Figure. Implant malapposition. Number of malapposed flow diverter struts (top row) and the average distance between malapposed flow diverter struts and the parent artery wall (bottom row) stratified per occlusion status at both follow-up moments. Supplemental Digital Content 3. Figure. Flow diverter migration. Macroscopic image (A) shows signs of hemorrhage outside the aneurysm pouch. Microscopic image of aortic cross-section (B) clearly shows the absence of the flow diverter. Detail (insert) shows a continuous aneurysm wall (arrows) until the wall is discontinued and shaped irregularly (asterisk), suggesting a rupture. COMMENT The authors describe their experience with flow diverters of different pore densities in the treatment of aneurysms on an animal model (in the aorta of 36 Wistar rats). The treatment arms were divided equally between aneurysms treated with low pore density flow diverters (10 pores/mm) and aneurysms treated with high pore density flow diverters (23 pores/mm) based on the assumption that high pore flow diverters are beneficial for intracranial aneurysm healing. Half of the animals in each group was sacrificed 1 month after treatment and the other half was sacrificed 3 months after treatment. There was no difference in aneurysm occlusion rate between low pore density and high pore density groups at 1 or 3 months after treatment. The number of stent struts not in contact with the parent artery wall and the average strut distance to the parent artery wall was higher in nonoccluded aneurysms at time of sacrifice. The authors concluded that implant apposition is more important than pore density for aneurysm occlusion in their animal model. These findings are significant because implant malapposition affects aneurysm healing and can be associated with thromboembolic complications as well as migration of flow diverter. In one of the rats in the study, a flow diverter migrated in the vessel resulting in an incompletely sealed aneurysm and subsequent rupture, which is a possible cause for delayed hemorrhage of elective aneurysms previously treated with flow diverters. Future clinical studies evaluating neuro-imaging modalities that visualize wall apposition after deployment of flow diverting stents and the relationship with outcomes (aneurysm obliteration, hemorrhagic complications) are necessary. Rafael A. Ortiz David J. Langer New York, New York Copyright © 2018 by the Congress of Neurological Surgeons

Journal

NeurosurgeryOxford University Press

Published: Apr 4, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off