Eur. Radiol. (2001) 11: 123±130 Ó Springer-Verlag 2001

N. Young N. W. C. Dorsch R. J. Kingston G. Markson J. McMahon

Received: 13 April 1999 Revised: 22 March 2000 Accepted: 3 May 2000

)

N. Young ( ) ´ R. J. Kingston ´ G. Markson Department of Radiology, Westmead Hospital, P. O. Box 533, Wentworthville NSW 2145, Australia N. W. C. Dorsch ´ J. McMahon Department of Neurosurgery, Westmead Hospital, P. O. Box 533, Wentworthville NSW 2145, Australia

NE UR OR A DI O LOG Y

Intracranial aneurysms: evaluation in 200 patients with spiral CT angiography

Abstract The goal of this study was to assess the usefulness of spiral CT angiography (CTA) with three- dimensional reconstructions in defining intracranial aneurysms, particularly around the Circle of Willis. Two hundred consecutive patients with angiographic and/or surgical correlation were studied between 1993 and 1998, with CTA performed on a GE HiSpeed unit and Windows workstation. The following clinical situations were evaluated: conventional CT suspicion of an aneurysm; follow-up of treated aneurysm remnants or of untreated aneurysms; subarachnoid haemorrhage (SAH) and negative angiography; family or

Introduction In the past half decade there have been numerous publications describing the use of spiral, or helical, computed tomography angiography (CTA) in search of intracranial aneurysms, particularly of the Circle of Willis (COW) [1, 2, 3, 4]. This is an attempt to determine whether or not spiral CT can be an effective method of aneurysm evaluation, being less invasive than angiography. Three-dimensional rotation permits multi-planar viewing of the aneurysms; however, the number of patients reported have been relatively few. Although small aneurysms have been detected, the reliability of detection is suspect, and problems have also been encountered with skull-base bones and metal-clip artefacts. The largest number of patients reported in the literature is 40 [1]. Larger patient numbers are needed to be able to assess the efficacy of CTA as compared with the standard of angiography.

past aneurysm history; and for improved definition of aneurysm anatomy. Spiral CTA detected 140 of 144 aneurysms, and an overall sensitivity of 97 %, including 30 of 32 aneurysms 3 mm or less in size. In 38 patients with SAH and negative angiography, CTA found six of the seven aneurysms finally diagnosed. There was no significant artefact in 17 of 23 patients (74 %) with clips. The specificity of CTA was 86 % with 8 falsepositive cases. Spiral CTA is very useful in demonstrating intracranial aneurysms. Key words Spiral CT ´ Intracranial aneurysm

We report our experience in 200 consecutive patients with angiographic and/or surgical correlation.

Materials and methods Between 1993 and 1998 over 900 spiral CTAs of the intracranial arteries were performed at our institution. This report includes the first 200 patients with spiral CT identification and also with angiography and/or surgical comparisons, in studying aneurysms around the Circle of Willis. There were 132 females and 68 males, with ages ranging from 16±81 years (mean age 51 years). The first 85 patients were studied retrospectively, from 1993 until mid-1996, and the rest of the patients were entered into a prospective database. Patients were divided into six clinical groups, depending on their reason for study referral, which was primarily from the Department of Neurosurgery: 1. Group 1: Patients with preceding conventional CT or MRI performed for neurological symptoms, e. g. headaches, in whom

124

a Fig. 1 a This man had a conventional CT for headache, which was suspicious for an aneurysm (group-1 patient). Superior view of the Circle of Willis. There is a wide-neck saccular aneurysm arising from the posterior, left intracranial internal carotid artery (straight arrow). Superior to this is a smaller aneurysm arising from the termination of the internal carotid artery (arrowhead) b A lateral view of the left internal carotid angiogram shows a larger aneurysm (arrow) and the superior smaller aneurysm (arrowhead). No other aneurysms were found. Surgery confirmed these findings

there was suspicion of an aneurysm (e. g., Fig. 1). These patients did not present with subarachnoid haemorrhage (SAH). 2. Group 2: Patients with known untreated aneurysms referred for follow-up, with surgery refused or decided against. In these cases the unruptured aneurysms had been detected at the same time as a ruptured aneurysm in patients presenting with SAH. 3. Group 3: Patients previously treated for aneurysms and referred for evaluation of possible aneurysm sac remnants. All patients had correlating angiograms, with maximal interval between the studies being 1 month. 4. Group 4: Patients with known subarachnoid haemorrhage (SAH), in whom intra-arterial angiography had failed to identify any aneurysm. Patients proceeded directly to CTA if the angiography was negative. 5. Group 5: Patients with a family history of intracranial aneurysms, or with previously treated aneurysms, who were referred for screening for aneurysms. These patients had correlative angiograms following their CTA studies, whereas others, not reported here, did not. This group is distinct from group 1 in that these individuals were referred for CTA in order to detect a possible aneurysm, whereas in group 1 there had been no reason to suspect an aneurysm until it was suspected on a standard CT scan or MRI. 6. Group 6: Patients with usually large ( > 1 cm in diameter) or giant aneurysms, in whom angiography had not optimally demonstrated their anatomy, e. g. neck location, proximity of crucial vessels, prior to surgery. Surgical findings were then correlated retrospectively with the preceding CTA and angiography studies by the operating surgeon who graded the CTA information as being worse than, equal to, or better than the angiography information. Patients in groups 4 and 6 had their spiral CTA performed following angiography. In groups 1, 2, 3 and 5 the radiologist evaluating

b the CTA studies was blinded to the angiographic findings. In groups 4 and 6 the angiographer and attendant neurosurgeon cooperatively determined the adequacy of angiographic information. In group 4 the CT radiologist knew that no aneurysm had been elucidated by angiography. In group 6 the CT radiologist was blinded to the angiographic data. The interval between CTA and angiography in group 4 was within 24 h. In groups 1, 2, 5 and 6 it was within 2 weeks. All spiral CTAs were performed on a GE HiSpeed unit, with post-processing done using a GE Advantage windows workstation. Planning of the volume to be scanned was by use of an initial lateral scout image. One hundred twenty millilitres of non-ionic 300 contrast was given by peripheral electric pump injection at a rate of 4 ml/s. A scan delay of 15 s was normally used, but if there were any features of cardiovascular compromise this was increased to 20 s. No test timing injection was used, and we did not have access to any proprietary bolus timing software. Normally a volume depth of 3 cm was imaged at 1-mm collimation with a 1:1 pitch. This enabled visualisation of the Circle of Willis and adjacent arteries. If an aneurysm was suspected lower in the vertebrobasilar system, then a scan depth of 6 cm was used. This was not routinely employed because of the significant postprocessing time required to cut away the skull-base bones while preserving the vertebral and basilar arteries. Manual imaging manipulation on the workstation enabled removal of bones (threshold above 400 Hounsfield units) and of the cerebral tissues (threshold below 100 Hounsfield units). Metal surgical clips were also removed by threshold techniques. Surfacerendered displays (SSD) of the arteries were obtained in all studies. All axial acquisition images were examined. If SSD and axial source images did not provide the quality of information required, maximum intensity projection (MIP) images were also obtained. The computer processing of images required an average 15 min of workstation time, increasing to 20±30 min in more complex cases such as vertebrobasilar cases. Spiral CTAs were routinely reported from workstation viewing because the vessels could be rotated 360  in any plane. Each study was reviewed by one radiologist highly experienced in diagnostic cerebral angiography. Two radiologists shared the total study population work, and in each case agreed on the CTA findings. Cerebral angiography was performed mainly by a highly experienced radiologist, using selective catheter techniques and standard imaging planes on a Phillips V3000 digital subtraction unit. Non-ionic contrast was used in all cases. If standard views were suspicious but not optimal for aneurysm identification, then other views, e. g. submentovertical or reverse oblique, were obtained.

125

Table 1 Aneurysm numbers identified by angiography/surgery (clinical groups) Aneurysm size Group 1 Group 2 Group 4 Group 5 Group 6 No aneurysm £ 3 mm 3 mm to 1 cm > 1 cm

23 16 42 22

± 10 21 4

31 4 3 0

2 2 2 1

± 0 2 14

Angiography was reviewed at least twice, once by the performing angiographer, once by the referring neurosurgeon, when applicable, and finally by another, reviewing radiologist. As with CTAs, a consensus was finally agreed by all viewers. In patients who underwent operations the neurosurgeon studied the pre-operative CTA and angiogram studies, and thus was well able to correlate the imaging with his own surgical findings, including the artery of origin of the aneurysm, and its size and shape. He then established if the CTA/angiography findings were inferior to or equal to his surgical findings. For analysis aneurysms were divided into size groupings of £ 3 mm, 3 mm to 1 cm and > 1 cm. Although Juvela et al. [5] suggests that risk of SAH is not directly related to aneurysm size, others [6] argue that there is a differential rate of rupture between aneurysms above and below the 1-cm level. In our initial work [7] we were surprised at the ability of spiral CT in defining tiny aneurysms of 3 mm or less, so we have kept 3 mm as another marker level, as do Strayle-Batra et al. [2] in their work.

Results Table 1 details the efficacy of aneurysm identification by CTA, as compared with angiography and surgery, in the various clinical groups 1, 2, 4, 5 and 6. In group 1, which included 87 patients, there were two thrombosed aneurysms which we categorised into the ªno aneurysmº group. Nine patients had two aneurysms, 2 patients had three aneurysms and 1 patient had four aneurysms. Of the 16 patients in this group with £ 3 mm diameter aneurysms confirmed on angiography none have undergone surgery; they are currently on a followup programme. Twenty of the 42 patients with 3-mm to 1-cm aneurysms and 16 of 22 with > 1-cm aneurysms have had surgical treatment, the rest having regular follow-up. This group included five false-positive cases and three false negatives, two aneurysms of £ 3-mm diameter and one of 3-mm to 1-cm diameter. In clinical group 2 there were 28 patients. Seven had two aneurysms, accounting for the total of 35 aneurysms. None of these patients have undergone surgery for these aneurysms and all are being closely followed. In group 3, 23 of 25 patients had prior clipping surgery. Metal artefacts on CTA degraded interpretation ability in 6 of the 23 (26 %). Eleven patients had aneurysm remnants on angiography, with CTA able to identify these remnants in 9 patients. Fourteen patients had

no angiographic remnants and satisfactory CTA images were obtained in 10 patients. Each of these patients were evaluated on CTA with clips and skull-base bones included, and then with clip and bone thresholds removed from processed images (e. g. Figs. 2, 3). In the 19 patients with non-degrading artefacts, only one had multiple clips, with the others having single clips. In the six with significant artefacts, four had horizontal clips and 4 patients had multiple clips (three being horizontal). Clips that were horizontally placed (in the same plane as the acquisition images) and multiple clips caused more artefacts. In group 4 there were 38 patients with CT proven SAH. Each had angiography (some multiple) prior to CTA studies. Seven patients (18 %) had aneurysms found on CTA, and proceeded to surgery. Four aneurysms were very small, ² 3 mm in diameter. The aneurysm was found in each case to have caused the SAH, and all patients proceeded to surgery. Three aneurysms were 3 mm to 1 cm in size and all were treated by surgery. One patient with false-negative angiography and CTA had three further negative angiograms and another CTA study performed while critically ill in ICU with cerebral vasospasm. The shortest interval between one of these angiograms and the CTA study was 1 day. After the patient had recovered with medical therapy and papaverine angioplasty, the ruptured aneurysm on the left internal carotid artery was detected on a fourth angiogram study. It was presumed thrombosed during the previous negative studies. In the 6 patients of group 5, five aneurysms were found, including two in 1 patient. The 1 patient with a > 1-cm sized aneurysm proceeded to surgery. The others are being closely followed with periodic CTAs. There were two presumed false positives in this group, i. e. aneurysms that were perceived on CTA but not confirmed on digital subtraction angiography. In group 6 all studies permitted better anatomical delineation of aneurysm configuration and relationship with adjacent arteries than was seen on the preceding angiography, which was correlated with simultaneous viewing of CTA and angiographic results by the radiologist. These arose from the multiplane computer rotations available on workstation viewing, which proved very helpful in surgical planning. In all cases surgery was performed, with confirmation of the CTA findings. Table 2 details our false-positive CTA cases. Apart from the errors in cases 1, 2 and 3, all suspected aneurysms on CTA were 3 mm or less in diameter. Angiograms were performed within 2 weeks before CTA in cases 1±7. In case 8 the CTA was 1 day after angiography. Cases 1±4 occurred within 6 months of starting this service in 1993 and may, in part, be attributed to inexperience. Cases 5±8 were spread throughout our later experience (e. g. Fig. 4). Case 8 was a patient with proven SAH and negative angiography, and was the only

126

2a

2b

3a

3b

3c

3d

Fig. 2 a This man had previous anterior communicating artery (A Com A) aneurysm surgery (group-3 patient). Superior view from a follow-up spiral CT, with both bones of the skull base and the surgical clip (arrow) included. No aneurysm remnant is seen and no significant artefact formation is present. b The same spiral CT, with skull base bones and aneurysm clip at A Com A region (arrow) excluded Fig. 3 a This woman presented with subarachnoid haemorrhage. Oblique view of a right internal carotid angiogram shows a saccular anterior communicating artery (A Com A) aneurysm (arrow). There is a spasm of the right A1 segment. b Angulated submentico

vertical view from a left internal carotid run, showing a wide neck aneurysm arising from the left middle cerebral artery bifurcation. c This is a follow-up spiral CT after surgery to the ruptured A Com A aneurysm (group 3). Superior view with skull base bones and surgical clip (arrow) included. No significant artefact and no remnant aneurysm is seen at the site of surgery. The left middle cerebral artery bifurcation aneurysm is identified (arrowhead). d The same spiral CT study, with skull base bones and surgical clip removed. At the site of surgery (arrow) there is loss of definition of the arteries adjacent to the clip site. Arrowhead shows the left middle cerebral artery aneurysm, which was then treated by later follow-up surgery

127

Table 2 False-positive spiral CT angiography (CTA) cases Case no.

Findings

1 2 3

Small ªanatomicalº bulge of the basilar tip Pituitary stalk mistaken for an aneurysm Incorrect reconstructions of overlapping middle cerebral vessels Suspected small anterior communicating artery aneurysm Suspected small left A1 aneurysm Suspected small right posterior communicating artery aneurysm (Fig. 4) Suspected small basilar tip aneurysm Suspected small right internal carotid artery aneurysm after subarachnoid haemorrhage

4 5 6 7 8

Cases 1±7 had normal angiography; case 8 had negative surgical finding

false-positive case to proceed to surgery, when no aneurysm was found. The patient recovered well clinically and remains well on clinical follow-up after 2 years. The other patients also remain clinically well on follow-up, with periods ranging from 1±4 years. Fig. 4 a A female patient with a family history of aneurysm (group-5 patient). Superior view of the CTA shows suspicion of a small, saccular aneurysm in the region of the right posterior communicating artery (P Com A) origin (arrow). Arrowhead points to the right middle cerebral artery radicals. b,c Lateral and oblique views of a follow-up right internal carotid artery angiogram study. No aneurysm is seen, and in particular no P Com A aneurysm is seen. Vertebral angiography was also normal

Table 3 Efficiency of spiral CTA compared with surgery/angiography findings (excluding clinical group-3 patients) Aneurysm size

No. of aneurysms (found by CTA/ total number

Sensitivity Specificity (%) (%)

£ 3 mm 3 mm to 1 cm > 1 cm Overall

30/32 69/71 41/41 140/144

94 97 100 97

Table 3 details our overall results of CTA with respect to differing aneurysm sizes. Sensitivity ranged from 94 to 100 %, being lowest for aneurysms £ 3 mm in diameter and highest for the large aneurysms. There was a total of four missed aneurysms in our study population. The 3 patients in clinical group 1 had angiography within 2 weeks of their CTA studies. All three then proceeded to surgery. The single patient in clinical group 4 proceeded to surgery to treat the ruptured aneurysm found on the fourth angiogram. The lower specificity of 86 % reflected the 8 false-positive cases.

a

b

86

c

128

Table 4 Previous spiral CT reports (by other groups) Reference

No. of patients

No. of aneurysms

Sensitivity (%)

Strayle-Batra et al. [2] Liang et al. [3] Preda et al. [4] Anderson et al. [1]

16 23 26 40

20 17 22 43

85 88 100 86

The MIP reconstructions were used in further delineation of anatomy when the aneurysm was located on the internal carotid arteries adjacent to the skull-base bones, and when cutting away the bones for SSD images incurred significant degradation. This occurred in 3 patients in group 1, 2 patients in group 2 and 1 patient in group 6.

Discussion Traditionally, angiography has been the accepted standard for the detection and anatomical delineation of intracranial aneurysms; however, this technique has some established limitations. The false-negative rate for angiography ranges up to 10 % in situations of subarachnoid haemorrhage, with causes including small aneurysm size, aneurysm thrombosis, observer error, suboptimal projections and vessel spasm [8, 9, 10]. Additionally, there is a risk in undergoing angiography, an invasive procedure, with a reported overall neurological complication rate of 1.2 % by ACAS [11], although there is a generally accepted low ( ² 0.5 %) rate of permanent neurological deficits [12]. Hence, angiography has been limited to use in patients with a high suspicion for intracranial aneurysm, e. g. in clinically documented SAH, and restricted in other clinical settings, e. g. those with a strong family history of aneurysms; however, aneurysm-induced intracranial bleeding is a devastating condition in usually otherwise well people. Spiral CT angiography has potential as a modality which can complement angiography in the detection and anatomical evaluation of aneurysms. There has been a relatively recent surge of interest in utilizing spiral CTA for evaluating intracranial aneurysms [1, 2, 3, 4]. Table 4 details the published data of other groups working in this area. There have also been reports of CTA using conventional rather than spiral CT, with the largest report being from Hope et al. [13], with a sensitivity of detection of 90 % in a group of 80 patients with a total of 94 aneurysms. A comment by Brown et al. [14] suggests that this does not significantly sacrifice image quality as compared with spiral CTA. A reply from Marks and Rubin in the same publication remarked that spiral CT does have inherent advantages, including better longi-

tudinal spatial resolution, volume acquisition allowing overlapping reconstruction, and rapid coverage in the long axis allowing imaging at peak contrast vascular enhancement. In our group of 200 patients, spiral CTA detected 140 of 144 aneurysms, an overall sensitivity of 97 %. Not unexpectedly, this was best (100 %) with the larger ( > 1 cm diameter) aneurysms. There was still a very high (94±97 %) sensitivity of detection of the smaller aneurysms. This is higher than reported by others, with Anderson et al. [1] and Strayle-Batra et al. [2] describing sensitivities of 57 and 25 %, respectively, for aneurysms 3 mm or less in size. Because of this, Anderson et al. [1] warn against using CTA as a screening tool for asymptomatic patients with a family history of aneurysm. Of more concern in our results is a relatively low specificity of 87 % due to 8 false-positive cases, listed in chronological order in Table 2. The first 4 cases occurred early in our experience and are ascribed to a learning curve. The last 4 cases were spread over our second and third years of study. A possible explanation in some cases includes sub-optimal three-dimensional reconstruction, with ªcut-offº of branching vessels simulating a small aneurysm. Patients 1±7 had negative angiography and the eighth patient underwent exploratory craniotomy after presenting with SAH and negative angiography. The single false-positive case in Anderson's series [1] was a suspected 2-mm aneurysm. Because of the very severe consequences that can ensue from the presence of an aneurysm, we have formed a policy of having a high level of alertness for any aneurysm on our CTA studies, with the advantage of maintaining our high sensitivity rate and accepting the drawback of lower specificity. Juvela et al. [5] followed 142 patients with 181 unruptured aneurysms over a 30-year period. They found a cumulative haemorrhage rate of 25.6 % at 20 years and 32.4 % at 30 years. Importantly, aneurysm size did not accurately predict the likelihood of rupture, with 67 % of those that ruptured being 6 mm or smaller. They firmly concluded that patients with unruptured aneurysms should undergo therapy, if technically and clinically feasible, with greater benefit for the younger patient. Admittedly, most of the aneurysms followed in that study were in patients with previously ruptured aneurysms. In contrast, a recent multi-centre study [6] reported a very low rate of aneurysm rupture when size was less than 1 cm and when there was no preceding history of SAH. Our current policy is to follow up with progress CTAs any asymptomatic patient with a study suspicious of possible small aneurysms. We have not found the presence of subarachnoid blood to be an impediment to CTA. We agree with the opinion [1, 13, 14] that with good contrast enhancement the cerebral arteries can be well differentiated from any adjacent subarachnoid blood. In our clinical group 4,

129

CTA found six of seven aneurysms missed by initial angiography. We feel that this is due primarily to the multiplane rotations available on workstation viewing. The missed case was due to thrombosis or to intense vessel spasm around the aneurysm, precluding contrast filling of the aneurysm sac, and as noted above, several angiograms were negative in this patient during that period; hence, we highly recommend use of CTA in SAH when initial angiography is unhelpful. Of patients with previous aneurysm surgery with metal clips, we found no significant artefact formation in 74 %. We imaged each of these patients with metal clip both included and excluded from the reconstructions. We found that the more horizontally positioned clips and the presence of multiple clips were the main causes of significant artefact formation. We agree with van Loon et al. [16] who found satisfactory imaging by CTA in 11 of 13 patients with titanium clips and recommend spiral CTA for routine use in the evaluation of the post-operative patient. However, they also state that angiography should be used if CTA reveals an abnormality or when there is significant clip artefact. In our patients we use angiography in the former situation only if further surgery is contemplated. The multiplanar rotation capability on spiral CTA allowed better anatomical delineation of the more complex (and usually very large) aneurysms in our clinical group 6. Evaluation of the size of the aneurysm neck, its relationship to adjacent vessel origins and branches, and displacement and proximity of important nearby arteries have been invaluable in pre-operative planning. We normally employ a SSD technique and examination of the axial source images, but do not hesitate to utilize the MIP techniques if required. A deficiency of our CTA protocol is the routine volume data sets which we limit to a 3-cm depth to analyse the Circle of Willis and adjacent vessels. We do not routinely view the lower basilar artery and vertebral arteries, because of the significant time incurred in manually cutting away the bone close to these arteries. Other groups [2, 3, 14] have had similar experiences, with a major disadvantage of CTA being the time-consuming process in removing the bones of the skull base, with a possible risk of also cutting away any adjacent aneurysm. Up to 10 % of aneurysms can arise outside the Circle of Willis [16], so if required we do scans to a depth of 6 cm and carefully view the vertebrobasilar system with skull-base bones included, and then manually excluded. As yet we have not had a case of missed aneurysm in the vertebrobasilar system. Our other major technical difficulty lies in the timeconsuming task of post processing. This is a problem also encountered by other groups performing this work [2, 13, 15]. Extreme care and diligence are required to remove skull base bones without removing vessels and any possible aneurysm.

We primarily utilize SSD models because its depth and shading of images allows the best evaluation of aneurysms, particularly small aneurysms, near branching vessels, thus providing the most useful images for surgical planning. In this we totally agree with Kallmes et al. [18] who also assert that SSD images are generally superior to MIP images; however, SSD images do depend on the threshold chosen, so that if it is incorrect it can lead to loss of information, e. g. small vessels and aneurysms [13]. Another potential for CTA in aiding therapy management may lie in helping to determine the size and form of platinum coils to be used in cases of endovascular treatment of aneurysms. We usually adopt a scan delay of 15 s because this allows optimal opacification of the arteries of the Circle of Willis and adjacent vessels, while minimising any contamination of images by venous filling. During this study we did not have access to sophisticated time-delay contrast-timing software, e. g. G. E. Smart-Prep. We have not found this to be of great concern. Brink et al. [19] have suggested that 1-mm scan collimation results in more noise contamination than wider, e. g. 2 mm, collimation; however, we tend to agree with Kallmes et al. [18] who suggest that a narrower collimation improves diagnostic accuracy, more than offsetting any decrease in signal-to-noise ratio. On the other hand, we do not routinely use a smaller, e. g. 0.5 mm, collimation, because our initial experience did not show a perceptible improvement in image quality.

Conclusion Spiral CTA is no longer an experimental imaging modality in the assessment of intracranial aneurysms. We show it to have a very high sensitivity, even for very small aneurysms. It has roles in evaluating selected patients with other imaging and/or clinical suspicion of aneurysms, for follow-up of known aneurysms, for postsurgical follow-up, in search of suspected ruptured aneurysms when angiography has been negative, and for defining the anatomical relationships of aneurysms. We do have a lower than optimal specificity rate, but we accept this since it can increase sensitivity, and because of the devastating results that can accompany the presence of an undetected aneurysm. We do not suggest that CTA supplants good angiography, particularly in cases of SAH, but do suggest that spiral CTA has a strong and valuable role, complementing traditional angiography, in this difficult field of clinical medicine.

130

References 1. Anderson GB, Findlay JM, Steinke DE, Ashforth R (1997) Experience with computed tomographic angiography for the detection of intracranial aneurysms in the setting of acute subarachnoid hemorrhage. Neurosurgery 41:522±528 2. Strayle-Batra M, Skalej M, Wakhloo AK, Ernemann V, Klier R, Voigt K (1998) Three-dimensional spiral CT angiography in the detection of cerebral aneurysms. Acta Radiol 39:233±238 3. Liang EY, Chan M, Hsiang JHK, Walkden SB, Poon WS, Lam WWM, Metremeli C (1995) Detection and assessment of intracranial aneurysms: value of CT angiography with shaded surface display. AJR 165:1497±1502 4. Preda L, Gaetani P, Rodriguez y Baena R, Maggio EM di, La Fianza A, Dove R, Fulle I, Solcia M, Cecchini A, Infuso L, Campani R (1998) Spiral CT angiography and surgical correlations in the evaluation of intracranial aneurysms. Eur Radiol 8:739±745 5. Juvela S, Porras M, Heiskanen O (1993) Natural history of unruptured intracranial aneurysms: a long-term follow-up study. J Neurosurg 79:174±182 6. The International Study of Unruptured Intracranial Aneurysms Investigators (1998) Unruptured intracranial aneurysms: risk of rupture and risks of surgical intervention. N Engl J Med 339:1725±1733

7. Dorsch NWC, Young N, Kingston R, Compton J (1995) Early experience with spiral CT in the diagnosis of intracranial aneurysms. Neurosurgery 31:230±238 8. Tarr RW, Perl J II, Masaryk TJ (1994) Neuroradiological aspects of aneurysms. In: Ratcheson RA, Wirth FP (eds) Concepts in neurosurgery, vol 6. Ruptured cerebral aneurysms: perioperative management. Williams and Wilkins, Baltimore, pp 182±203 9. Weir B (1987) Aneurysms affecting the nervous system. Williams and Wilkins, Baltimore, pp 575±577 10. Smith RR, Zubkov YN, Tarassoli Y (1994) Cerebral aneurysms. Microvascular and endovascular management. Springer, Berlin Heidelberg New York, pp 29±30 11. Executive Committee for the Asymptomatic Carotid Atherosclerosis Study (1995) Endarterectomy for asymptomatic carotid artery stenosis. J Am Med Assoc 273:1421±1428 12. Pryor JC, Setton A, Nelson PK, Berenstin A (1996) Complications of diagnostic cerebral angiography and tips on avoidance. Neuroimaging Clin North Am 6:751±758 13. Hope JKA, Wilson JL, Thomson FJ (1995) Three-dimensional CT angiography in the detection and characterization of intracranial berry aneurysms. Am J Neuroradiol 17:439±445

14. Brown JH, Lutrin ES, Lev MH, Taveras JM (1997) CT angiography of the Circle of Willis: Is spiral technology always necessary? Am J Neuroradiol 18:1794±1797 15. Vieco PT, Shuman WP, Alsofrom GF, Gross CE (1995) Detection of Circle of Willis aneurysms in patients with acute subarachnoid haemorrhage: a comparison of CT angiography and digital subtraction angiography. AJR 165:425±430 16. Van Loon JJL, Yousry TA, Fink V, Seelos KC, Jurgen Reulen H, Steiger HJ (1997) Post-operative spiral computed tomography and magnetic resonance angiography after aneurysm clipping with titanium clips. Neurosurgery 41:851±857 17. Hacker RJ, Krall JM, Fox JL, Data I (1983) Intracranial aneurysms. Fox JL (ed) Springer, Berlin Heidelberg New York, pp 19±62 18. Kallmes DF, Evans AJ, Woodcock RJ, Dmary RA, Dix JE, McNulty BC, Holder CA, Dion JE (1996) Optimization of parameters for the detection of cerebral aneurysms: CT angiography of a model. Radiology 200:403±405 19. Brink JA, Lim JT, Wang G, Heiken JP, Deyoe LA, Vannier MW (1995) Technical optimization of spiral CT for depiction of renal artery stenosis: in vitro analysis. Radiology 194:157±163