Cohn P. Derdeyn, MD #{149}Christopher J. Moran, MD #{149}DeWitte T. Cross III, MDEric W. Sherburn, MD #{149}Ralph G. Dacey, Jr, MD

Intracranial Aneurysm: AnatomicFactors That Predict the Usefulnessof Intraoperative Angiography’

Index terms: Aneurysm, intracranial, 17.73 #{149}Angiography, intraoperative, 17.73

1 From the Section of Neuroradiolog� Mallinckrodt Institute of Radiology (C.P.D., C.J.M., D.T.C.);and the Department of Neurology and Neurosurgery (Neurological Surgery) (E.W.S., R.G.D.), Wash-ington University School of Medicine, 510 5 Kingshighway Blvd, St Louis, MO 63110 ;. ReceivedApril 23, 1997; revision requested June 13; revision received June 24; accepted July 8. Address reprintrequests to C.P.D. C.P.D. supported by a grant from the RSNA/Siemens Medical Systems Researchand Education Fund.

C RSNA, 1997

335

PURPOSE: To correlate the size andlocation of intracranial aneurysmwith the need to reposition the aneu-rysm clip after intraoperative angiog-raphy.

MATERIALS AND METHODS: In199 consecutive patients with 234clipped intracranial aneurysms, 273intraoperative angiographic studieswere retrospectively reviewed. Aneu-rysm size and location, determinedwith preoperative angiographic andsurgical reports, were correlated withthe frequency of clip repositioningbecause of parent- or branch-vesselcompromise or unexpected residualaneurysm.

RESULTS: Findings from intraopera-live angiograms resulted in clip re-positioning in 46 of 273 (16.8%) stud-ies. Clip repositioning was statistically

significantly less frequent with aneu-rysms of the posterior communicat-ing (three of 52 [5.7%] studies) andanterior choroidal (none of 12 stud-ies) arteries. High rates of clip repo-sitioning were found in aneurysmsof the superior hypophyseal artery(seven of 18 [38.9%1 studies), supe-nor cerebellar artery (three of five[60.0%] studies), and bifurcation ofthe internal carotid artery (three ofnine [33.3%] aneurysms). In 98 con-ventional follow-up angiographicstudies, seven (7%) false-negativecases with unsuspected aneurysmneck remnant were found.

CONCLUSION: The rate of clip re-positioning in aneurysms of the pos-tenor communicating or anteriorchoroidal arteries was less than thatat other locations (P < .05). Intraop-erative angiography may not be nec-essary when aneurysms are at thesetwo locations.

I NTRAOPERATWE angiography afterclipping of intracranial aneurysms

is an established, useful procedure

(1-4). The advantages of intraopera-tive angiography, compared withthose of conventional postoperativestudies, are all related to timing. De-spite limitations inherent in intraop-

emative angiography, informationgained with this technique can lead tomodifications at initial surgery thatprevent complications or obviate theneed for a second operation. Modifi-cations may involve the adjustment of

an aneurysm clip because of parent-or branch-vessel compromise or fillingof residual aneurysm.

The yield (ie, relative frequency ofabnormal studies that lead to changesin the surgical approach) of intraop-erative angiography, however, is reba-tively low. In a recent report (4) of thelargest series of which we are aware(100 intraoperative angiographic stud-ies after aneurysm clipping), a 12%frequency of clip repositioning or

other changes in therapy due to an-giographic findings was reported.This rate of revision is similar to thatreported in other series (1-3). Theyield of intraopemative angiography

could be improved if the techniquewere performed only in patients inwhom an abnormal study was likelyto be obtained.

The purpose of this study was todetermine if such a group of patientscould be identified preoperativeby onthe basis of aneurysm size and loca-tion.

Radiology 1997; 205:335-339

MATERIALS AND METHODS

After the clipping of 234 aneurysms in

199 patients between January 1994 andDecember 1996, 273 consecutive intraop-erative angiograms were obtained. Twofactors account for the difference betweenthe number of aneurysms and the numberof angiographic studies. First, repeatedangiographic studies after clip reposition-ing were considered separately from theinitial angiographic results during thesame surgical procedure. Second, in pa-tients with multiple clipped aneurysms,each angiogmaphic examination of atreated aneurysm was considered a sepa-rate study.

This series of patients did not includethose who were treated for a dissectinganeurysm (n = 4) or those who underwentemergent intraoperative angiography inthe operating room before craniotomy(n = 4). Also not included in this serieswere patients whose aneurysms weretreated with surgical techniques (eg, trap-ping or proximal ligation) other than clip-ping of the aneurysm neck. One patientunderwent surgical exploration withoutclipping. During the 3-year study period,six patients underwent craniotomy andsurgical clipping without follow-up intra-operative angiography.

A 5-F femoral sheath was placed duringpreoperative diagnostic angiography orwhile the patient was in the operatingroom (usually after the administration ofanesthesia and before surgery). While notin use, the sheath was continuouslyflushed with hepaminized saline at arterialpressure. A radiolucent operating table(Skytron, Grand Rapids, Mich) and car-bon-fiber head holder (Mayfield radiolu-cent skull clamp; Ohio Medical, Cincin-nati) were used in all patients. The femoral

Table IRevision of Oip Placement according to Aneurysm Location

Frequency of Frequency ofNo. of No. of Revisions No. of No. of Revisions

Arterial Location Studies Revisions (%)* Aneurysms Revisions (%)*

All sites 273 46 16.8 (±4.5) 234 30 12.8 (±4.2)Middle cerebral 60 14 23.3 (±10.6) 49 8 16.3 (±10.3)Anterior communi-

caring 52 8 15.4 (±9.8) 45 5 11.1 (±9.2)Posterior commu-

nicating 52 3 5.7 (±6.3)t 50 3 6.0 (±6.6)Ophthalmic 23 3 13.0 (±13.7) 20 2 10.0 (± 13.1)Superior hypophy-

seal 18 7 38.9 (±22.8)t 3 27.3 (±26.5)tPericallosal 14 3 21.4 (±22.8) 11 2 18.2 (±22.8)Basilar tip 14 1 7.7 (±13.4) 13 1 7.7 (±14.5)Anterior choroidal 12 0 0.0 (±25.0) 12 0 0.0 (±27.3)Bifurcation of

internal carotidartery

Superior cerebellar

125

33

25.0 (±24.5)60.0 (±42.9)

93

32

33.3 (±30.7)t66.6 (±53.3)t

Posterior inferiorcerebellar 4 0 0.0 (±75.0) 4 0 0.0 (±75.0)

Posterior cerebral 3 1 33.3 (±53.3) 3 1 33.3 (±53.3)Anterior cerebral

(Al segment) 2 0 0.0 2 0 0.0 (±0.0)Vertebrobasilar

junction 1 0 0.0 1 0 0.0 (±0.0)Anterior inferior

cerebellar I 0 0.0 1 0 0.0 (±0.0)

* Numbers in parentheses are 95% confidence leveL

t Rate of revision was statically significantly different than the overall rate (P < .05).

336 Radiology November 1997

sheath was draped to allow access duringangiography. Care was taken to avoid

placement of radiopaque materials overthe patient’s head, neck, and chest. The oper-ating room table was positioned to allowspace for the bedside angiographic unit.

Selective catheter placement of the de-

sired vessel was performed in standardfashion with the 5-F arterial sheath imme-diately before angiography. Manual injec-tions were performed. Three views of eachlesion were routinely obtained, includinganterior, lateral, and oblique projections,and attempts were made to duplicate use-fiil preoperative views. In many patients, theneurosurgical head-holding device preventedpredse duplication ofstandard views.

A bedside digital subtraction unit (OECDiasonics, Salt Lake City Utah) that con-sisted of a C-arm fluoroscope, a digital im-age processor and storage unit, and avideo monitor was used in all cases. Thisunit allows performance of routine fluo-roscopy and real-time digital subtractionangiography. The recorded images couldbe reviewed at different speeds and frame-by-frame. Permanent hard-copy imageswere made for the radiologic file with aphotographic unit. The preoperative diag-

nostic studies were available in the operat-ing room for comparison in all cases. Allstudies were interpreted by the attendingneuroradiobogist (C.P.D., C.J.M., D.T.C.),and results were discussed with the neuro-surgeon (R.G.D.). The femorab sheath wasremoved in either the recovery room or theintensive care unit, which allowed routineobservation by the nursing staff of thepuncture site and lower extremity.

The medical records, including surgicaland radiographic reports, were reviewedin all patients. Information gathered in-cluded (a) information about the intraop-erative or postoperative complicationspossibly attributable to angiography orsurgery, (b) the recorded findings of theintraoperative studies, and (c) the sun-geon’s notes about intraoperative decisionsmade on the basis of the angiographic results.

Preoperative and intraoperative angiograms

and surgical reports were reviewed to deter-mine aneurysm size and location. Measureddiameters of aneurysms of the anterior circu-lation were corrected for magnification withthe method described by Zubillaga and co-workers (5). With this method, the actual di-

ameter ofthe internal carotid artery proximal

to the bifurcation is assumed to be 3.4 mm. Aproximal basilar artery diameter of 3.3mmwas used in cases of aneurysm of the poste-

nor circulation (5). Aneurysm size wasgraded as follows: small (diameter < 10

mm), large (diameter, 10-25 mm), and gi-

ant (diameter > 25 mm). Statistical anaby-

sis was performed with the x2 test, withstatistically significant differences inferredat the P < .05 level. Aneurysm location

was categorized by using established con-ventions. Internal carotid artery aneu-rysms were named for the nearest branchvessel of origin (eg, superior hypophyseal

artery aneurysms) even if the vessel was

not visible on angiograms.At our institution, postoperative angiog-

raphy is generally reserved for patients

who do not respond to medical treatmentfor vasospasm and who are candidates forendovascular treatment. These postopera-tive studies were reviewed, and findingswere correlated with those from the intra-operative study to assess accuracy withintraoperative angiograms.

RESULTS

Aneurysm clips were repositioned

on the basis of intraoperative angio-graphic findings from 46 of 273 (16.8% ±4.5 [mean ± 95% confidence level])studies of 234 aneurysms (Table 1).

The repositioning rate per aneurysmwas lower (30 of 234 [12.8% ± 4.2]

aneurysms) than the 16.8% rate per an-giographic study. The 46 intraoperativeangiograms with findings that bed torevision of clip placement were pen-

formed in 30 cases of aneurysm. Two

on more revisions were necessary in 10

of these 30 aneurysms. Findings from

an initial study that bed to clip reposi-

tioning were, therefore, associated with astatistically significantly higher nate ofsubsequent revision of clip placement(10 of 30 [33.3% ± 16.9] aneurysms).

The frequency of clip repositioningwas statistically significantly higher

than average for aneurysms in the

superior hypophyseal and superior

cemebelbar arteries. The frequency of

clip repositioning was statistically sig-

nificantly higher for aneurysms at thebifurcation of the internal carotid am-

temy compared with the frequency forall clipped aneurysms, but not whenthe number of studies performed wasconsidered. High rates of clip reposi-honing that did not reach statisticalsignificance were also seen for aneu-

mysms in the middle cerebral artery

and pericallosal artery. The rate of cliprepositioning was also statisticallysignificantly more frequent for large

aneurysms (Table 2).

The mate of clip repositioning wasstatistically significantly lower than

average for aneurysms in the poste-non communicating artery and in theanterior choroidal artery. If these twosimilar and adjacent locations alongthe posterior wall of the supraclinoidinternal carotid artery proximal to thebifurcation are considered together,

the revision rate was even smaller(three of 64 [4.7% ± 5.1%] angiograms

and three of 60 [4.8% ± 5.3%] aneu-rysms) and achieved statistical signifi-cance for all aneurysms, as well as fortotal studies.

The reasons for clip repositioningcould be grouped into two categories:

(a) parent- or branch-vessel compro-mise and (b) residual aneurysm. The

category of residual aneurysm encom-passed three situations: persistent fill-

ing of the sac, filling of a neck rem-nant, or filling of an uncipped adjacent

lobe or of a second aneurysm. The rela-tive frequency of these two findings

Table 2

Revision of Clip Placement according to Aneurysm Size

No. of No. of Frequency of No. of No. of Frequency ofSize Studies Revisions Revisions (%)* Aneurysms Revisions Revisions (%)*

All sizes 273 46 16.8 (±4.5) 234 30 12.8 (±4.2)Small 240 37 15.4 (±4.6) 215 24 11.2 (±4.2)Large 22 8 36.4 (±20.1)t 19 5 26.3 (±19.8)tGiant 6 1 16.7 (±29.8) 5 1 20.0 (±35.1)

* Numbers in parentheses are 95% confidence level.t Rate of revision was statistically significantly different than the overall rate (P < .05).

Table 3Reasons for Revision of Clip Placement according to Aneurysm Location

Vessel Residual Vessel ResidualNo. of Compromise Aneurysm No. of Compromise Aneurysm

Arterial Location Studies (%) (%) Aneurysms (%) (%)

Allsites 273 17 29 234 11 19Middlecerebral 60 11 3 49 6 2Anterior communi-

cating 52 0 8 45 0 5Posterior communi-

cating 52 0 3 50 0 3Ophthalmic 23 0 3 20 0 2Superior hypophyseal 18 0 7 11 0 3Pericallosal 14 3 0 11 2 0Superior cerebellar 5 0 3 5 0 2Basilartip 14 1 0 13 1 0

Bifurcation of theinternal carotid 12 2 1 9 2 1

Posterior cerebral 3 0 1 3 0 1

p

.4�

a.

Figure 1. Unexpected filling of residual aneurysm at the superior hypophyseal artery after

initial clip placement. (a) Oblique lateral intraoperative angiogram demonstrates residual fill-

ing of the aneurysm sac (arrow). On the basis of this information, the jaws of the aneurysm clipwe’re opened, and the clip was extended anteriorly. A subtraction artifact (arrowhead) of the

fenestrated aneurysm clip overlies the supraclinoid carotid artery. (b) Same view from the re-

peated intraoperative angiogram demonstrates no filling of the aneurysm and no compromise

of internal carotid artery flow. Arrowhead = subtraction artifact due to the presence of the clip.

b.

Volume 205 #{149}Number 2 Radiology #{149}337

for each aneurysm location is shown

in Table 3. For internal carotid artery

aneurysms above the dunal ring butproximal to the bifurcation, all reposi-tionings were undertaken because of

residual aneurysm. At the superior

hypophyseal location, three of 11 an-

eurysms continued to fill after initial

clipping (Fig 1). In two of these three

aneurysms, multiple surgical adjust-

ments were needed to exclude the an-

eurysm from the circulation. All ante-nor communicating artery aneurysmsthat needed clip revision were for me-

sidual aneurysm, as well. At the tn-

furcation of the middle cerebral artery,most revisions were necessary to im-prove parent-or branch-vessel flow

(abnormal findings from 11 of 14 in-

traopenative angiogmams led to reposi-

tioning) (Fig 2).

Conventional angiography was pen-

formed in 98 of the 234 cases of clippedaneurysms. Unexpected residual aneu-rysm was noted in seven (7% false-nega-tive rate) cases. One of the seven false-

negative studies is shown in Figure 3.

No unexpected parent- or branch-vessel

compromise was seen. No complicationswere observed after intraoperative angi-ography.

DISCUSSION

The goal of surgery in cases of an-

eurysm is to exclude the aneurysmfrom the circulatory system without

compromising flow through parent orbranch vessels. The frequency of in-

complete aneurysm obliteration is notknown. Feuerberg et al (6) found me-

sidual aneurysm in one of 27 (4%)

treated intracranial aneurysms. Clipped

aneurysms with small (1-2 mm inlength) neck remnants have been

shown to enlarge and bleed (6-8).Branch-vessel occlusion caused by an

aneurysm clip may also complicatesurgery. After the clipping of 78 con-

secutive aneurysms, MacDonald et al(9) reported nine unexpected major-

vessel occlusions on postoperative

angiognams obtained urgently or rou-

tinely. Six of these occlusions resulted

in disability due to infarct. In addi-

tion, residual aneurysm was seen insix patients. Three of these six patients

with residual aneurysm underwentrepeated surgery. The frequency ofthese undesired events was 19.2% (15of 78 aneurysms)and was similar to

the rate of clip repositioning observedin our large series and to that reported

by other investigators (1-4).

The data from our investigationsuggest that certain aneurysm boca-

tions and sizes are associated with

particular rates of clip replacement

after intraopenative angiognaphy. A

low rate of clip repositioning after in-

tnaopenative angiography was ob-

served with aneurysms at the rela-

lively easily exposed and well-visualizedsupraclinoid segment of the internal

carotid artery. A change in clip place-

ment was not necessary in any of the

12 aneurysms of the anterior choroi-

dal artery. Revision of clip placementwas necessary in only three of the 50aneurysms of the posterior communi-cating artery. These findings contrast

with those of Alexander et al (4). In

their study of 100 craniotomies, cliprepositioning was necessary in five of27 aneurysms of the posterior commu-

nicating artery. Not all patients in

their series, however, underwent in-tmaoperative angiography; among pa-tients who did not undergo intraop-

erative angiognaphy were six patients

r

b.

,.

I... -

a.

�,1

I4

C.

338 Radiology November 1997

with aneurysms of the posterior com-

municating artery.

High frequencies of unexpected re-

sidual aneurysm that necessitated clipmanipulation were observed at the

superior hypophyseal artery, at the

bifurcation of the internal carotid an-

tery, and at the superior cerebellar ar-

tery. Aneurysms arising from the su-perior hypophyseal artery can be

difficult to �risualize; these aneurysms

are close to the cavernous sinus and

the relatively fixed position of the in-

ternal carotid artery at the durab ring,

which makes exposure of the aneu-rysm neck challenging. All revisions of

clip placement in aneurysms of the

superior hypophyseal artery were per-

formed for residual filling of the aneu-

rysm sac (Fig 1). Adequate exposure ofaneurysms of the superior cerebellan

artery and of the bifurcation of the

internal carotid artery may also be dif-

ficult. Parent- or branch-vessel com-

promise was a frequent finding after

clipping of aneurysms of the middle

cerebral artery. Aneurysms at this lo-

cation often involve origins of branch

vessels, and complete obliteration of

the aneurysm may be difficult without

compromising parent- on branch-yes-

sel flow.

Large aneurysms were associated

with higher rates of clip repositioning.

This finding is not surprising because

visualization of parent and branch

vessels can be more difficult with

larger aneurysms. Giant aneurysms

were not associated with high rates of

revision, but only six of these lesionswere included in this series. Alexander

et al (4) found a statistically signifi-

cantly higher frequency of clip reposi-

tioning in the 11 giant aneurysms they

studied.

No complications attributable to

intraoperative angiography were en-

countered in this retrospective review.

Derdeyn et ab (3) noted one possible

embolic complication in 87 transfemo-

ral catheter placements at intraopera-

tive angiography. Alexander et al (4)

reported one possible embolic compli-

cation in 100 transfemoral studies. The

working conditions in the operatingroom are different from those in the

angiography suite, and the apparentimprovement in the complication rate

of Alexander et al may reflect the ben-

efit of experience.The resolution on intraoperative

angiograms is not as high as that on

conventional angiograms. Small yes-sels such as the anterior choroidab an-tery and perforating vessels often are

not visualized on intraoperative an-

giograms. In addition, the head holder

can limit the angles at which angio-

Figure 2. Unexpected branch-vessel occlu-

sion after clipping of a complex aneurysm of

the middle cerebral artery. (a) Anteroposte-

nor preoperative angiogram of the right in-

ternal carotid artery shows a large, lobulated

aneurysm (arrow) at the trifurcation. An in-

tracerebral hematoma was present with mass

effect. Narrowing of the Ml and Al segments

of the proximal middle cerebral and anterior

cerebral arteries was consistent with vaso-

spasm. (b) Oblique anteroposterior initialintraoperative angiogram obtained after clip-

ping demonstrates no filling of the aneurysm

or of the previously visualized branches of

the middle cerebral artery. Arrow = subtrac-

tion artifact due to the presence of the aneu-

rysm clip. (c) Angiogram acquired after the

clip was repositioned shows restoration of

flow to the branch. Of necessity, the residual

aneurysm (arrow) was left in place.

graphic projections are obtained. Con-

yentional angiography in a dedicatedangiography suite offers biplane capa-

bility, which doubles the number of

views obtained with each injection

and provides higher resolution and

greaten flexibility with regard to pro-

jections. Despite the limitations of in-traoperative angiography, however,

diagnostic information regarding par-

ent- and branch-vessel patency and

the status of the clipped aneurysm is

usually obtained. The radiologist per-

forming these studies must be thor-

oughly familiar with the preoperative

cerebral angiogmam to ensure that in-

traoperative images will optimally

provide clinically relevant infonma-

tion.

An unexpected residual aneurysm

neck was identified on seven of the 98

(7()/()) postoperative angiognams in this

series; note that no vessel occlusions

were found. In general, the false-nega-

tive results obtained in cases of small

neck remnants were often attributableto limited resolution and limited

availability of angiographic projec-

tions with the current intraopenative

angiogmaphic equipment (Fig 3). The

actual rate of false-negative findingsfrom intraopemative angiography in

cases of an unsuspected residual an-

eurysm cannot be determined from

our data because only those patientssuspected of having vasospasm Un-

denwent conventional angiographyafter surgery. Our 7% nate is similar

to that found in other studies of this

technique. Martin et al (1) noted threefalse-negative intraopemative studiesin their series with angiognaphic fob-

low-up in 62 patients. Two residualaneurysm sacs were overlooked, and

a small residual nidus of an anteniove-nous malformation was overlooked in

a third case. In the series of Barrow etal (2), one residual aneurysm sac was

noted in 17 postoperative studies afteraneurysm clipping. With subsequent

angiography, Derdeyn et al (3) ob-

served two unexpected, small residual

aneurysms (false-negative studies) on

25 studies of aneurysms.While intraopenative angiogmaphy

cleanly offers several advantages in

a. b. C.

t_.

I Jfr�.d. e.

Figure 3. False-negative intraoperative angiogram. (a) Preoperative angiogram shows a largeinternal carotid artery aneurysm at the ophthalmic level. (b) On the anteroposterior intraopera-

tive angiogram, no definite residual aneurysm filling is depicted, and the study was consid-

ered to he normal. (c) A similar view from the postoperative angiogram shows definite re-sidual lateral filling (arrow) of the aneurysm beneath the clips. (d) Findings on the lateralintraoperative angiogram are unremarkable and are similar those on (e) the postoperative lat-

eral angiogram.

Volume 205 #{149}Number 2 Radiology #{149}339

the surgical treatment of aneurysms,

this procedure also has some limita-

tions. First, intraoperative angiogra-

phy necessitates more operating-room

time (average of 45 minutes [3]) and

consequently increases cost. Second,intraoperative angiography is associ-

ated with a small risk of angiographiccomplications. Third, intraoperativeangiographic equipment does not pro-

�‘ide the resolution available with

equipment in modem angiography

suites. The magnitude of the rate of

false-negative intraopenative angio-

grams in cases of aneurysm neck rem-

nant and the subsequent effect of this

rate on outcome are unknown. Finally,

the information gained with intnaop-

enative angiognaphy may not result in

improved outcome. For example, de-

spite identification of an abnormality

such as branch-vessel occlusion andrapid surgical correction, an infarct

may still occur (3).

The data from this study demon-strate that different aneurysm diam-

eters and locations are associated withdifferent rates of clip repositioning

after intraoperative angiogmaphy. The

yield of intraoperative angiognaphy is

high in cases of large aneurysms and

of aneurysms of the superior hy-

pophyseal artery, superior cerebellan

artery, and bifurcation of the internal

carotid artery. The need for intraop-

enative angiography should be antici-pated in these situations. The rate ofclip repositioning in aneurysms of the

posterior communicating and anterior

choroidal arteries is low, and intraop-

enative angiognaphy may not be nec-

essary, in these cases. #{149}

� � ReferencesD � Martin NA, Bentson J, Vinucla F, et al. In-

traoperative digital subtraction angiographyand the surgical treatment of intracranialaneurysms and vascular malformations. J Ne’u-rosurg 1990; 73:526-533.

2. Barrow DL, Boyer KL, Joseph GJ. lntraop-

erative angiography in the management ofneurovascular disorders. Neurosurgery

�, 1992; 30:153-159../. 3. Derdeyn CP. Moran CJ, Cross DT, Grubb RL,

Dacey RG. lntraoperative digital subtractionangiography: a review of 112 consecutive cx-aminations. AJNR 1995; 16:307-308.

4. Alexander TD, Macdonald RL, Weir B, Kow-

alczuk A. Intraoperative angiography inaneurysm surgery: a prospective study of100 craniotomies. Neurosurgery 1996; 39:10-18.

5. Zubillaga AF, Guglielmi G, Vinuela F, Duck-wiler GR. Endovascular occlusion of intra-

cranial aneurysms with electrically detach-able coils: correlation of aneurysm neck size

and treatment results. AJNR 1994; 15:815-820.

6. Feuerberg I, Lindquist C, Lindqvist M,Steiner L. Natural history of postoperative

aneurysm rests. J Neurosurg 1987; 66:30-34.7. Drake CG, Vanderlinden RG. The late con-

sequences of incomplete surgical treatmentof cerebral aneurysms. J Neurosurg 1967;27:226-238.

8. Lin T, Fox AJ, Drake CG. Regrowth of an-eurysm sacs from residual neck followinganeurysm clipping. J Neurosurg 1989; 70:556-560.

9. MacDonald R, Wallace M, Kestle J. Role ofangiography following aneurysm surgery.Neurosurg 1993; 79:826-832.