INTERVENTIONAL NEURORADIOLOGY

Aneurysmal wall enhancement and perianeurysmaledema after endovascular treatment of unruptured cerebralaneurysms

I-Chang Su & Robert A. Willinsky & Noel F. Fanning &

Ronit Agid

Received: 25 September 2013 /Accepted: 19 March 2014# Springer-Verlag Berlin Heidelberg 2014

AbstractIntroduction Perianeurysmal edema and aneurysm wall en-hancement are previously described phenomenon after coilembolization attributed to inflammatory reaction.We aimed todemonstrate the prevalence and natural course of these phe-nomena in unruptured aneurysms after endovascular treat-ment and to identify factors that contributed to theirdevelopment.Methods We performed a retrospective analysis of consecu-tively treated unruptured aneurysms between January 2000and December 2011. The presence and evolution of wallenhancement and perianeurysmal edema on MRI afterendovascular treatment were analyzed. Variable factors werecompared among aneurysms with and without edema.Results One hundred thirty-two unruptured aneurysms in 124patients underwent endovascular treatment. Eighty-five(64.4 %) aneurysms had wall enhancement, and 9 (6.8 %)aneurysms had perianeurysmal brain edema. Wall enhance-ment tends to persist for years with two patterns identified.Larger aneurysms and brain-embedded aneurysms were sig-nificantly associated with wall enhancement. In all edemacases, the aneurysms were embedded within the brain andhad wall enhancement. Progressive thickening of wall en-hancement was significantly associated with edema. Edema

can be symptomatic when in eloquent brain and stabilizes orresolves over the years.Conclusions Our study demonstrates the prevalence andsome appreciation of the natural history of aneurysmal wallenhancement and perianeurysmal brain edema followingendovascular treatment of unruptured aneurysms. Aneurys-mal wall enhancement is a common phenomenon whileperianeurysmal edema is rare. These phenomena are likelyrelated to the presence of inflammatory reaction near theaneurysmal wall. Both phenomena are usually asymptomaticand self-limited, and prophylactic treatment is notrecommended.

Keywords Aneurysm . Aneurysmal wall enhancement .

Endovascular coiling . Perianeursymal edema

Introduction

Endovascular embolization is an efficient and increasinglywidely used technique to treat intracranial aneurysms. Duringintraluminal procedures, there is usually no direct manipula-tion or mechanical irritation of the brain, and theoretically, thebrain parenchyma is less affected compared to the open sur-gical procedures. This is supported by the fact that mostunruptured saccular aneurysms are located within the sub-arachnoid space and do not directly involve the brain paren-chyma. Recently, however, it was shown that endovasculartreatment could have local or global intracranial effects, and itwas suggested that inflammation can occur followingendovascular coil embolization [1, 2]. For example, the useof modified bioactive coils was shown to be associated withthe development of hydrocephalus and aseptic meningitis [2].In addition, aneurysms coiled with high-packing density wereshown to be associated with aneurysmal wall enhancementafter coiling [1].

I.<C. Su : R. A. Willinsky : R. Agid (*)Division of Neuroradiology, Department of Medical Imaging,Toronto Western Hospital, 399 Bathurst St., 3 McLaughlin Wing,Room #425, Toronto, ON M5T 2S8, Canadae-mail: ronitagid@hotmail.com

I.<C. SuDivision of Neurosurgery, Department of Surgery, Taipei CathayGeneral Hospital, Taipei, Taiwan

N. F. FanningDepartment of Interventional Neuroradiology, Cork UniversityHospital, Cork, Ireland

NeuroradiologyDOI 10.1007/s00234-014-1355-x

De novo perianeurysmal edema after coil embolization isanother recently reported post-coiling phenomenon [3, 4]. It isusually restricted to the area adjacent to the coiled aneurysmand can lead to neurological symptoms when eloquent brain isaffected [3]. There is some evidence that perianeurysmaledema occurs in aneurysms with mural enhancement aftercoiling, but the relationship between enhancement and edemaremains unexplored [1]. Possible contributing factors to ede-ma development have been proposed, but most of them arebased on limited case reports [3, 5].

In order to understand the determinants of aneurysmal wallenhancement and perianeurysmal edema after endovasculartreatment, we evaluated the incidence and natural course ofthese phenomena in a cohort of unruptured aneurysms treatedconsecutively in our center. We also analyzed aneurysm- andtreatment-related factors that may contribute to the develop-ment of both phenomena.

Methods

Study design

The study included consecutive aneurysms treated byendovascular means between January 2000 and December2011 in our institution. We chose to include in this study onlyunruptured saccular aneurysms that underwent endovascularembolization. We excluded ruptured aneurysms since ruptureand hemorrhage were reported to cause aneurysmal wallenhancement and brain edema prior to treatment [6]. For thesame reason of possible wall enhancement and edema prior totreatment, we excluded giant or partially thrombosed aneu-rysms, fusiform/dissecting aneurysms (these also could havebright signal in the wall prior to contrast administration),aneurysms associated with pretreatment perianeurysmal ede-ma, and aneurysms previously treated in other institutes. Pa-tients were also excluded if they lacked follow-upMR recordsand if the aneurysms were treated by parent vessel occlusion.

Aneurysm-related characteristics

The following aneurysm-related characteristics were record-ed: (1) the aspect ratio of the aneurysm was calculated as theratio of the maximal dimension of aneurysmal dome to neckon three-dimensional angiograms, and the ratio was dichoto-mized as <2 or ≥2 [1]; (2) aneurysm volume was measured byusing the formula (4/3)Π(a/2)(b/2)(c/2), where a, b, and crepresented the width, length, and the height of the aneurysmon three-dimensional angiograms, respect ively.Multilobulated aneurysms were calculated as the sum of theindividual compartment [1]; (3) the type of aneurysm, fromhemodynamic point of view, was categorized into terminal(e.g., aneurysms at the carotid termination, basilar tip, and

middle cerebral artery (MCA) bifurcation) and nonterminal(e.g., aneurysms at the paraophthalmic artery and posteriorcommunicating artery) types [7]; and (4) Aneurysms werecategorized as embedded in brain tissue or not. To be regardedas embedded in brain, an aneurysm needed to have no CSFsignal, or only a thin CSF cleft, observed on MRI T2-weighted image (T2WI) between the aneurysmal dome andthe brain (Fig. 1a, b). Usually, embedded aneurysms alsoindented the adjacent brain parenchyma (Fig. 1c).

Procedure-related factors

In terms of coil types, aneurysms were categorized into thosewhich were treated only with bare platinum coils and thosetreated with both bare and modified platinum coils(HydroCoil or Matrix coil). Packing density was measuredby using the formula as follows: “packing density=total vol-ume of the coil/aneurysmal volume=Π(d/2)2(l)/aneurysmalvolume,” where d and l represented the diameter and thelength of each coil, respectively, at the fully expanded state[1], High-packing ratio was defined as a packing density>22 % [8].

Radiological follow-up

All patients were followed clinically in our multidisciplinaryclinic. Our imaging follow-up protocol for unruptured coiledaneurysms included baseline multisequence contrast-enhancedMR (CEMR) andMRA at 6 months after treatment,followed by the sameMR sequences every 1 to 2 years. Someof the patients (57 out of 132 (43.2 %)) received an additional,earlier CEMR and MRA. This was according to operators’preference and occurred mostly in cases with residual aneu-rysmal filling, observed in the control catheter angiogramsimmediately after coil embolization. Patients with MR abnor-malities or MRA suggesting of aneurysm recanalization un-dergo MR follow-up every 6 months or catheter angiogramfor confirmation. Before 2010, CEMRA at 1.5 Twas used [9].Since 2010, we alternately used CEMRA at 1.5 T or 3D TOFMRA at 3 T for follow-up of coiled aneurysms. Thus,gadolinium-enhanced sequences are not always available.

MRI examinations were evaluated for aneurysmal wallenhancement, de novo perianeurysmal edema, and aneurys-mal recanalization. All studies included pre and post contrastT1-weighted images (WI), and enhancement was called pos-itive only when there was no hyperintensity in the aneurysmalwall prior to contrast administration. Wall enhancement wasdefined as positive when a complete spherical enhancement ofthe coiled aneurysmal wall was noted on T1WI CEMR(Figs. 2d and 3a). Fluid-attenuated inversion recovery(FLAIR) images were evaluated for the presence ofperianeurysmal brain edema. The edema had to be vasogenic

Neuroradiology

(without diffusion restriction) and located within the braintissue surrounding the aneurysm (Fig. 2f, h).

Statistical analysis

Categorical and continuous variables were expressed as per-centage and mean±standard deviations, respectively.Aneurysm-specific, procedure-related, and postembolization-related findings were compared between groups using chi-square (or Fisher’s exact test if appropriate) for categoricalvariables and Mann–Whitney U test for continuous variables.Multivariate logistic regression models were used to evaluatethe odds ration (OR) and 95 % confidence interval (CI) ofeach factor in predicting aneurysmal wall enhancements. Sig-nificance was defined as p<0.05. Data were analyzed withcommercially available software Number Crunching Statisti-cal System (NCSS) 2004.

Results

In the study period, 724 aneurysms were treated byendovascular means in our institution; out of which, 201 wereunruptured. Out of these, 132 unruptured aneurysms wereincluded in our final analysis, and 69 aneurysms were exclud-ed. These included dissecting, giant, or partially thrombosed

aneurysms (n= 31, six of which had pretreatmentperianeurysmal edema); aneurysms retreated after priorendovascular or surgical treatment in another institution (n=11); and cases with no available posttreatment MRI (n=27).

The MR follow-up period in our study ranged from 6 to118 months with a mean follow-up of 35.1±27.9 months.Maximal aneurysm size ranged from 2.8 to 18 mm with amean maximum aneurysm size of 8.3±3.3 mm. Aneurysmswere treated with coiling alone in 103 cases (78.0 %), stent- orflow diverter-assisted coiling in 18 cases (13.6 %), and flowdiverter alone in 11 cases (8.3 %). High-packing density (ofabove 22 %) was achieved in 74 (56.1 %) aneurysms. Mod-ified coils were used in 21 (15.9 %) aneurysms. Duringfollow-up, 76 (57.6 %) aneurysms remained completely oc-cluded, and small residuals remained stable in 45 (34.1 %)aneurysms. Eleven (8.3 %) aneurysms had major recanaliza-tion; of which, 8 aneurysms underwent additionalendovascular or surgical treatments.

Aneurysmal wall enhancement

Posttreatment CEMRI/MRA were available for analysis in92.9, 71.8, and 47.0 % of aneurysms within 6 months, be-tween 6 and 24 months, and >24 months after treatment,respectively. Wall enhancement was observed in 85 (85/132,64.4 %) aneurysms. Table 1 summarizes the prevalence of

Fig. 1 Examples of aneurysmlocation in relation to theparenchyma of the brain.Multiplanar T2-weighted MRIimages demonstrated aneurysmswhich were embedded (a–c) andnot embedded (d) in theparenchyma of the brain. Theembedded aneurysms showabsence of CSF signal (solidarrow in (a)) or a very thin CSFcleft (dotted arrow in (b) and (c))between the aneurysm and thebrain. They also show indentationof the adjacent brain parenchymaby the aneurysm (asterisk in (c))

Neuroradiology

posttreatment aneurysmal wall enhancement according to an-eurysm location in our study cohort. Multivariate regressionanalysis (Table 2) demonstrated that larger aneurysms (OR=1.4, 95 % CI 1.1–1.6) and brain-embedded aneurysms (OR=4.6, 95%CI 1.6–13.3) were both significantly associated withhigher chance of wall enhancement, while long-term use ofanti-inflammatory agents (n=33, OR=0.9, 95 % CI 0.3–2.9)was not associated with less chance of wall enhancement.

We also found that the pattern of aneurysmal wall enhance-ment may change over time. Among the 85 aneurysms withposttreatment wall enhancement, interval CEMR follow-upwas available in 62 (72.9 %). In these, we observed twopatterns of wall enhancement evolution. Type I (Fig. 2) wascharacterized as “early progressive enhancement” pattern,with an initial progressive thickening of the wall enhance-ment, observed in at least two consecutive CEMR studiesperformed within 12 months of treatment, and subsequentdecrease of enhancement. Type I pattern of wall enhancementwas relatively rare and observed in 9 (9/62, 14.5 %) aneu-rysms. Type II (Fig. 3) was a “plateau” pattern of enhancementin which the thickness of enhancement persisted over monthsand years with a stable appearance or only a slight decrease inintensity with time. Type II pattern of wall enhancement wasmore common and observed in 53 (53/62, 85.5%) aneurysms.There were no statistically significant differences between thetwo types of wall enhancement in terms of aneurysm volume,

aspect ratio, aneurysmal types, packing density, or type ofcoils.

Perianeurysmal edema

Ten out of 132 (7.6 %) unruptured aneurysms were associatedwith perianeurysmal edema after endovascular treatment. Inone aneurysm, edema did not develop until 44 months aftercoiling, when major recanalization occurred. This was pre-sumably secondary to the massive regrowth and thrombusformation within the aneurysm and not related to theendovascular procedure itself. Therefore, true de novoperianeurysmal edema attributed to the endovascular proce-dure was observed in 9 (9/132, 6.8 %) aneurysms. Table 3summarizes the characteristics of the aneurysms which devel-oped perianeurysmal edema after embolization.

Cases 1 to 8 were treated with coils, and posttreatmentedema was either observed as early as 1 day after treatment(in two aneurysms) or between 1 and 6 months after treatment(in six aneurysms). In these cases, the edema involved non-eloquent regions of the brain. They were all asymptomatic anddid not require treatment. Maximal perianeurysmal edemawas encountered between 1 and 6 months after treatment.After maximal edema was reached, three edema cases (cases3, 6, 8) remained stable with a follow-up of 6 to 36months andfour edemas cases (cases 2, 4, 5, 7) resolved spontaneously at

Fig. 2 Type I “early progressive enhancement” pattern of aneurysmalwall enhancement. Pretreatment T2WI demonstrated a leftparaophthalmic aneurysm embedded within the brain parenchyma (a).MRA showed the aneurysm pre (b) and post (c) endovascular coiling.T1WI CEMR 1 day after coiling demonstrated early aneurysmal wallenhancement (d). T1WI CEMR6months after coiling showed thickening

of the wall enhancement (e). This was associated with prominentvasogenic edema around the aneurysm, as seen on FLAIR (f), withoutdiffusion restrictions on DWI. Follow-up MR, 1 year posttreatment,demonstrated decrease of the wall enhancement (g) and stable appearanceof perianeurysmal edema (h). The patient remained asymptomaticthroughout the follow-up period

Neuroradiology

5, 6, 11, and 61 months. In case 1, edema was identified2 months after initial coiling at the same time with earlyaneurysmal recanalization. It completely resolved 6 monthsafter the aneurysm was recoiled and never recurred up to5 years. Case 9 was unique to other cases in which edemadid not develop after initial coiling but rather after the aneu-rysm was retreated with a flow diverter (pipeline) 14 monthsafter coiling because of recanalization. Perianeurysm edemadeveloped 6 months after this second treatment. The edema inthis patient was in the mesencephalon and required steroidtreatment for new onset of diplopia. In the last MR follow-up

(12 months after pipeline treatment), the edema is partiallyresolved and the patient was free of symptoms.

All nine aneurysms that developed perianeurysmal edemawere embedded within the brain parenchyma. They represent-ed 15 % of the overall 60 embedded aneurysms included inthis study. All nine aneurysms that developed perianeurysmaledema also had aneurysmal wall enhancement (10.6 % of 85wall-enhancing aneurysms). We analyzed all the brain-embedded wall-enhancing aneurysms (n=53) in our study(Table 4) and found that only aneurysms with early progres-sion of wall enhancement (type I pattern) were significantly

Table 1 Prevalence of aneurys-mal wall enhancement afterendovascular treatment accordingto aneurysm location

Location Aneurysms with wall enhancement Total number

No. Percentage

Paraophthalmic 30 73.2 41

Posterior communicating artery 16 66.7 24

Basilar tip 14 63.6 22

Terminal ICA 12 92.3 13

Posterior circulation (exclude basilar tip) 5 55.6 9

Anterior communicating artery 4 26.7 15

Intracavernous 3 42.8 6

MCA bifurcation 1 50.0 2

Total 85 63.9 132

Fig. 3 Type II “plateau” patternof aneurysmal wall enhancement.This case demonstratedenhancement in the wall of abasilar tip aneurysm. T1WICEMRI 1 day post-coilingshowed circumferentialenhancement of the aneurysmalwall (a). Wall enhancement wasstable in thickness andappearance at 6 months (b) andslightly decreased at 2 years (c)and 5 years (d) post endovascularcoiling. There was no associatededema

Neuroradiology

associated with the development of perianeurysmal edema(57.1 vs 12.5 %, p=0.02). Terminal aneurysms also showeda borderline significant association with edema formation(31.6 vs 8.8 %, p=0.05). There were no significant associa-tions among other factors, such as aneurysm volume, type ofcoils, packing density, or long-term use of anti-inflammatoryagents.

Discussion

Aneurysmal wall enhancement and de novo perianeurysmaledema after endovascular treatment are sparsely described inthe literature, and their natural course is largely unknown [1,3]. In order to better understand the pathomechanism of thesephenomena, it is important to understand the histopathologicalchanges that occur in the aneurysmal wall after endovasculartreatment. In the early post-coiling stage (within days), acti-vated platelets are recruited to release inflammatory mediatorsand stimulate thrombus formation [4, 10]. Thereafter, a strongforeign body inflammatory reaction is elicited in order to

eliminate or isolate the coils and to heal the aneurysm [11].In the following 28 days, neutrophils and macrophages are thefirst-line inflammatory cells that participate in the thrombusformation within the aneurysm [12, 13]. At least 2 monthslater, multinucleated giant cells (so-called foreign body giantcells) form and surround the coil mass [11]. This foreign bodyinflammatory reaction persists over months and evolves in acentripetal fashion from the interface between the coil massand the aneurysmal wall toward the aneurysmal core [12]. Asa result, intra-aneurysm thrombus gradually converts into amature fibrocellular granulation tissue, and the aneurysmalwall itself is densely infiltrated with chronic inflammatorycells, fibrovascular tissues, and vaso vasora [12, 14]. Thesehistopathological hallmarks act immediately to prevent aneu-rysm rupture, but the entire healing process might takemonthsor even years to complete [15].

Early aneurysmal wall enhancement was previously report-ed to likely represent the normal healing process of early acuteinflammatory reaction [1]. Our series clearly demonstratesthat this phenomenon of wall enhancement is common aftercoiling of unruptured aneurysms (observed in 64.4 % of thestudy cohort). We found that higher aneurysmal volumes andaneurysms embedded within the brain parenchyma are factorssignificantly and independently associated with wall enhance-ment. These associations suggest that larger aneurysms inducea higher amount of normal inflammation, while aneurysmsspatially close to the brain tissue can presumably recruitvascular channels more easily from the pial surface. Bothfactors in turn lead to higher visibility of contrast enhancementon MR.

We also demonstrated that the enhancement pattern mightchange over time. Over time, 85.5 % of aneurysms with wallenhancement had a stable enhancement appearance (type II).This persistent long-lasting wall enhancement is probablyattributed to the enriched vaso vasorum and fibrovascular

Table 3 Characteristics of nine aneurysms with perianeurysmal edema after embolization

No. Aneurysm location Maximal dome size(mm)

Wall enhancementpattern

Time of 1stedema

Time of maximal edema(months)

Edemaevolution

MR follow-up(months)

1 Terminal ICA 16 Type II 2 months 2 Resolved 67

2 Terminal ICA 10 Type II 2 months 2 Resolved 74

3 Terminal ICA 11 Type I 5 months 5 Stable 36

4 Terminal ICA 11 Type II 6 months 6 Resolved 44

5 Paraophthalmic 14 Type II 4 months 6 Resolved 78

6 Paraophthalmic 15 Type I 1 day 6 Stable 12

7 MCA bifurcation 12 N.A. 1 month 1 Resolved 64

8 Posteriorcommunicating

8 Type I 1 day 6 Stable 6

9a Basilar tip 12 Type I 6 months 6 N.A. 6

N.A. not availablea Timing of edema development was counted from the treatment for the aneurysm recanalization, which occurred 14 months after initial coiling

Table 2 Multivariate logistic regression for predictors of aneurysmalwall enhancement after endovascular treatment

Variables OR 95 % CI p value

Maximal aneurysm dimension 1.4 1.1, 1.6 <0.01

Terminal aneurysms 1.4 0.5, 4.1 0.50

Embedded aneurysms 4.6 1.6, 13.3 <0.01

Aspect ratio ≧2 1.3 0.5, 3.3 0.60

Aneurysmal recanalization 1.9 0.3, 14.0 0.51

High-packing density (>22 %) 2.7 0.9, 7.6 0.07

Use of modified coils 1.3 0.3, 5.8 0.71

Use of anti-inflammatory agents 0.9 0.3, 2.9 0.88

Neuroradiology

tissues around the aneurysm wall in the chronic stage of thehealing process [12]. In 14.5 % of aneurysms with wallenhancement, the enhancement thickened significantly withinto the first 12 months (type I). This pattern most likely indi-cated a profound foreign body inflammatory reaction presum-ably mediated by leukocytes and giant cells. The high associ-ation of perianeurysmal edema with this specific type ofenhancement supports this hypothesis. Though we did notidentify any significant factors that were specifically associat-ed with type I pattern, it is reasonable to believe that increasedthickening of wall enhancement is a sign of more significantinflammation. The presence of an inflammatory reaction inthe aneurysmal wall could be potentially proven by aneurys-mal wall imaging using modern ferumoxytol-enhanced MRtechnique (i.e., to label macrophages within the aneurysmalwall) [16, 17].

Vasogenic edema in the brain is caused by the disruptionand increased permeability of the blood–brain barrier [18].Thus, only aneurysms that are spatially intimate with the brainwould be able to develop associated brain edema [3]. Theo-retically, larger aneurysms and aneurysms projecting towardthe brain are more prone to be embedded within the adjacentbrain parenchyma. According to this rational, certain aneu-rysms such as carotid termination aneurysms and superiorlyprojected paraophthalmic aneurysms carry higher risk of ede-ma because they invariably project into the brain. Other an-eurysms, such as superior hypophyseal aneurysms, posteriorcommunicating aneurysms, and middle cerebral bifurcationaneurysms, carry a minimal risk for edema because they arevery rarely embedded within the brain tissue. The highestoccurrence rate of perianeurysmal edema in carotid termina-tion aneurysms in our study population and the fact there wereno edema cases among non-embedded aneurysms support thishypothesis. Secondly, inflammatory mediators play a majorrole in blood–brain barrier disruption [18]. Therefore, aneu-rysms which release more inflammatory mediators aftercoiling should be more likely to impair the blood–brain barrier[4]. This explains our finding that perianeurysmal edemaoccurs merely in wall-enhancing aneurysms and also is

usually associated with type I enhancement evolution pattern.Gradual resolution of edema, either by steroid or throughnatural course, likely indicates that the blood–brain barrier isrepaired over time as the inflammation process subsides.

In addition, three specific mechanisms, including volumeeffect, water hammer effect, and the use of modified coils,have been proposed to induce edema [4, 19]. The formeraffects the brain before the inflammatory process reaches itspeak, while the latter two are typical examples of hemody-namic and chemical stimulants that lead to overt inflammatoryprocess, respectively. Sudden volume expansion induced bythe coil–thrombus complex can lead to mechanical irritationand mass effect on adjacent brain. This volume effect ispresumed to develop and subside within a short period oftime. Therefore, this effect could explain the edema thatdeveloped in two of our cases very shortly after coiling. Waterhammer effect, a kind of hemodynamic stimulant, occurswhen pulsatile blood flow strokes the coils and transmits thepressure to the aneurysm wall and surrounding brain tissue [4,20]. Since the same mechanism is also used to explain coilcompactions observed in aneurysmal recanalization,perianeurysmal edema has been proposed to be associatedwith higher recanalization rates [20]. This association couldexplain the evolution in our case 1 in whom additional embo-lization neutralized the effect and improved the edema, but theassociation between recanalization and perianeurysmal edemais not statistically significant in our overall analysis. Slightlyhigher association between terminal aneurysms and edemaformation partially supports the hypothesis that hemodynamicstress and water hammer effect with subsequentperianeurysmal inflammation might play a role inperianeurysmal edema formation [21]. HydroCoils andMatrixcoils are thought to induce a more pronounced inflammatoryreaction, enhanced aneurysm thrombosis, and improved long-term obliteration rate [1, 2]. We, however, did not observe ahigher association between the use of modified platinum coilsand perianeurysmal edema development. This observationwas supported by previous reports that observed similar de-gree of inflammatory reactions in aneurysms treated with bare

Table 4 Comparison of edemaand non-edema cases in wall-en-hancing aneurysms embeddedwithin the brain parenchyma (n=53)

Edema (+) (n=9) Edema (−) (n=44) p value

Terminal aneurysm 31.6 % 68.4 % 0.05

Enhancement evolution 0.02Type I 57.1 % 42.9 %

Type II 12.5 % 87.5 %

Maximal dome size (mm) 12.2±2.4 10.7±3.7 0.27

High-packing density (>22 %) 20.6 % 79.4 % 0.46

Use of modified coils type 18.2 % 81.8 % 1.00

Aspect ratio ≧2 18.9 % 81.1 % 0.71

Aneurysmal recanalization 28.6 % 71.4 % 0.59

Use of anti-inflammatory agents 21.4 % 78.6 % 0.68

Neuroradiology

or modified platinum coils [22]. Our case 9, in whom edemadid not occur after initial treatment with bare platinum coilsbut developed after the second treatment with a flow diverter,further demonstrates that overt inflammation process couldoccur secondarily to thrombus formation rather than the coilitself. Therefore, at least in terms of progressive wall enhance-ment and perianeurysmal edema, the use of modified coilalone did not explain the observation of overt inflammatoryreaction.

As reported in previous studies, our study shows thatneurological manifestations as a result of perianeurysmal ede-ma are entirely dependent on the location of affected brain [1,5]. When the involved brain is not eloquent, the edema has noclinical symptoms and appears to be self-limited. Based onour observations, one could now predict that aneurysms whichdevelop progressive thickening of wall enhancement are athigher risk to develop brain edema when they are anatomical-ly embedded within the brain parenchyma. However, sincemostly perianeurysmal edema is a self-limited subclinicalprocess, there is no need for prophylactic treatment for thesepatients unless symptoms appear. Similarly, our study showsthat the phenomenon of aneurysmal wall enhancement itselfdoes not warrant any treatment (e.g., steroids), as it is acommon finding following coil embolization (64.4 % in ourseries) without clinical symptoms. Thus, we do not suggesttreating either of these phenomena in a prophylactic manner.

Our study has several limitations. One major limitation isthe fact that most of our patients do not have a baselinecontrast-enhanced MR study prior to coiling for comparison.In order to overcome this limitation, we included onlyunruptured saccular aneurysms and excluded ruptured, giant,partially thrombosed, and fusiform/dissecting aneurysms toreduce to minimum the chance of pretreatment wall enhance-ment and edema. Recent papers show that unruptured saccularaneurysms do not have wall enhancement [6]. Also, due tochanges in our follow-up imaging policy, 27.1 % of aneu-rysms had only one CEMR after the treatment, and follow-upof wall enhancement was not done in a predefined timeinterval. Both factors limit our evaluation of enhancementevolution. Secondly, this study lacks histopathological studiesof our observed findings. All our hypotheses are based on thepathological reports from the literature. Finally, the number ofedema cases is too small for multivariate regression analysisand possible confounding effects of each variable cannot bedetermined.

Summary and conclusions

Our study demonstrates the prevalence and some appreciationof the natural history of aneurysmal wall enhancement andperianeurysmal brain edema following endovascular treat-ment of unruptured aneurysms. Aneurysmal wall

enhancement is a common phenomenon that in most casesremains stable over years. A small percentage of aneurysmsdevelop progressive thickening of wall enhancement. Thissubset of patients is at higher risk to develop perianeurysmalbrain edema when the aneurysm is anatomically embeddedwithin the brain parenchyma. Posttreatment perianeurysmalbrain edema is a rarer phenomenon. Edema stabilizes orsubsides within months to years and can cause neurologicalsymptoms only when located in eloquent brain. Both phe-nomena are likely related to the presence of inflammatoryreaction near the aneurysmal wall. Treatment is not warrantedfor these phenomena unless the patient develops symptoms.

Ethical standards and patient consent We declare that all human andanimal studies have been approved by the University Health NetworkEthics Committee and have therefore been performed in accordance withthe ethical standards laid down in the 1964 Declaration of Helsinki and itslater amendments. We declare that patient consent was waived in thisstudy.

Conflict of interest We declare that we have no conflict of interest.

References

1. Fanning NF, Willinsky RA, ter Brugge KG (2008) Wall enhance-ment, edema, and hydrocephalus after endovascular coil occlusion ofintradural cerebral aneurysms. J Neurosurg 108(6):1074–1086. doi:10.3171/JNS/2008/108/6/1074

2. Im SH, HanMH, Kwon BJ, Jung C, Kim JE, Han DH (2007) Asepticmeningitis after embolization of cerebral aneurysms using hydrogel-coated coils: report of three cases. AJNR Am J Neuroradiol 28(3):511–512

3. Horie N, Kitagawa N, Morikawa M, Tsutsumi K, Kaminogo M,Nagata I (2007) Progressive perianeurysmal edema induced afterendovascular coil embolization. Report of three cases and review ofthe literature. J Neurosurg 106(5):916–920. doi:10.3171/jns.2007.106.5.916

4. Tomokiyo M, Kazekawa K, Onizuka M, Aikawa H, Tsutsumi M,Ikoh M, Kodama T, Nii K, Matsubara S, Tanaka A (2007)Mechanisms of perianeurysmal edema following endovascular em-bolization of aneurysms. Interv Neuroradiol 13(Suppl 1):145–150

5. Craven I, Patel UJ, Gibson A, Coley SC (2009) Symptomaticperianeurysmal edema following bare platinum embolization of asmall unruptured cerebral aneurysm. AJNR Am J Neuroradiol30(10):1998–2000. doi:10.3174/ajnr.A1643

6. Matouk CC, Mandell DM, Gunel M, Bulsara KR, Malhotra A,Hebert R, Johnson MH, Mikulis DJ, Minja FJ (2013) Vessel wallmagnetic resonance imaging identifies the site of rupture in patientswith multiple intracranial aneurysms: proof of principle.N e u r o s u r g e r y 7 2 ( 3 ) : 4 9 2 – 49 6 . d o i : 1 0 . 1 2 2 7 /NEU .0b013e31827d1012, discussion 496

7. Steiger HJ, Poll A, Liepsch DW, Reulen HJ (1988) Haemodynamicstress in terminal aneurysms. Acta Neurochir (Wien) 93(1–2):18–23

8. Chalouhi N, Dumont AS, Hasan D, Tjoumakaris S, Gonzalez LF, StarkeRM,DalyaiR, ElMoursi S, RosenwasserR, Jabbour P (2012) Is packingdensity important in stent-assisted coiling?Neurosurgery 71(2):381–386.doi:10.1227/NEU.0b013e31825c36dd, discussion 386-387

9. Farb RI, Nag S, Scott JN,Willinsky RA,Marotta TR,MontaneraWJ,Tomlinson G, Terbrugge KG (2005) Surveillance of intracranial

Neuroradiology

aneurysms treated with detachable coils: a comparison of MRAtechniques. Neuroradiology 47(7):507–515. doi:10.1007/s00234-005-1375-7

10. Diegelmann RF, Evans MC (2004) Wound healing: an overview ofacute, fibrotic and delayed healing. Front Biosci 9:283–289

11. Szikora I, Seifert P, Hanzely Z, Kulcsar Z, Berentei Z, Marosfoi M,Czirjak S, Vajda J, Nyary I (2006) Histopathologic evaluation ofaneurysms treated with Guglielmi detachable coils or matrix detach-able microcoils. AJNR Am J Neuroradiol 27(2):283–288

12. Bavinzski G, Talazoglu V, Killer M, Richling B, Gruber A, Gross CE,Plenk H Jr (1999) Gross and microscopic histopathological findingsin aneurysms of the human brain treated with Guglielmi detachablecoils. J Neurosurg 91(2):284–293. doi:10.3171/jns.1999.91.2.0284

13. Killer M, Arthur AS, Barr JD, Richling B, Cruise GM (2010)Histomorphology of thrombus organization, neointima formation,and foreign body response in retrieved human aneurysms treatedwith hydrocoil devices. J Biomed Mater Res B Appl Biomater94(2):486–492. doi:10.1002/jbm.b.31660

14. Bocher-Schwarz HG, Ringel K, Bohl J, Filippi R, Kempski O,Perneczky A (2002) Histological findings in coil-packed experimen-tal aneurysms 3 months after embolization. Neurosurgery 50(2):379–384, discussion 384-375

15. Roth C, Struffert T, Grunwald IQ, Romeike BF, Krick C, PapanagiotouP, Krampe P, Reith W (2008) Long-term results with Matrix coils vs.GDC: an angiographic and histopathological comparison.Neuroradiology 50(8):693–699. doi:10.1007/s00234-008-0392-8

16. Hasan DM, Chalouhi N, Jabbour P, Magnotta VA, Kung DK, YoungWL (2013) Imaging aspirin effect on macrophages in the wall ofhuman cerebral aneurysms using ferumoxytol-enhanced MRI: pre-liminary results. J Neuroradiol 40(3):187–191. doi:10.1016/j.neurad.2012.09.002

17. Hasan DM, Mahaney KB, Magnotta VA, Kung DK, Lawton MT,Hashimoto T, Winn HR, Saloner D, Martin A, Gahramanov S, DosaE, Neuwelt E, YoungWL (2012)Macrophage imagingwithin humancerebral aneurysms wall using ferumoxytol-enhanced MRI: a pilotstudy. Arterioscler ThrombVasc Biol 32(4):1032–1038. doi:10.1161/ATVBAHA.111.239871

18. Stamatovic SM, Dimitrijevic OB, Keep RF, Andjelkovic AV (2006)Inflammation and brain edema: new insights into the role ofchemokines and their receptors. Acta Neurochir Suppl 96:444–450

19. Marden FA, Putman CM (2008) Perianeurysm edema with second-generation bioactive coils. Surg Neurol 69(6):627–632. doi:10.1016/j.surneu.2007.01.069, discussion 632

20. Slob MJ, van Rooij WJ, Sluzewski M (2005) Coil thickness andpacking of cerebral aneurysms: a comparative study of two types ofcoils. AJNR Am J Neuroradiol 26(4):901–903

21. Castro M, Putman C, Radaelli A, Frangi A, Cebral J (2009)Hemodynamics and rupture of terminal cerebral aneurysms. AcadRadiol 16(10):1201–1207. doi:10.1016/j.acra.2009.03.022

22. White JB, Cloft HJ, Kallmes DF (2008) But did you use HydroCoil?Perianeurysmal edema and hydrocephalus with bare platinum coils.AJNR Am J Neuroradiol 29(2):299–300. doi:10.3174/ajnr.A0877

Neuroradiology