Colen et al.

Neuroradiolog y • Original Research

MDCT Angiography for
Intracranial Aneurysm
Detection

Effectiveness of MDCT Angiography

for the Detection of Intracranial
Aneurysms in Patients with
Nontraumatic Subarachnoid
Hemorrhage
Teran W. Colen1 OBJECTIVE. CT angiography (CTA) is a noninvasive imaging technique used to evaluate
Lilian C. Wang1 cerebral vascular structures. Digital subtraction angiography (DSA), although invasive, is the
Basavaraj V. Ghodke2 gold standard for diagnosing intracranial aneurysms. The purpose of this study was to evaluate
Wendy A. Cohen2 the effectiveness of CTA in the detection of intracranial aneurysms for patients with nontrau-
William Hollingworth3 matic subarachnoid hemorrhage (SAH) in a level 1 trauma center.
MATERIALS AND METHODS. We evaluated the diagnostic accuracy of MDCT in
Yoshimi Anzai1
336 consecutive patients undergoing evaluation for nontraumatic SAH with both CTA and 3D
American Journal of Roentgenology 2007.189:898-903.

Colen TW, Wang LC, Ghodke BV, Cohen WA, DSA within 48 hours. The diagnostic performance of CTA was assessed by radiology reports
Hollingworth W, Anzai Y using DSA as the gold standard. Analyses were performed per aneurysm and per patient, the
results were stratified by aneurysm size and location, and the MDCT data—16-MDCT data
versus 4-, 8-, and 16-MDCT combined data—were compared.
RESULTS. The overall sensitivity and specificity of CTA per aneurysm was 83% (CI,
0.78–0.87) and 93% (0.88–0.97), respectively. CTA failed to detect 49 of the 284 aneurysms.
Thirty-nine (80%) of these 49 missed aneurysms were ≤ 3 mm, nine (18%) were 4–6 mm, and
one (2%) was 7–10 mm. The sensitivity and specificity of CTA per patient was 95%
(0.91–0.97) and 97% (0.92–0.99), respectively. Of 211 patients, a primary aneurysm was
missed on CTA in 11 patients.
CONCLUSION. CTA showed excellent diagnostic performance for aneurysm detection.
The high negative predictive value (91.2%) for the per-patient analysis indicates that CTA has
merit as a screening tool. Most aneurysms missed were ≤ 3 mm and in patients in whom a pri-
mary aneurysm had already been detected.

t is estimated that 10–15 million The results of numerous studies evaluating

Keywords: aneurysm, cerebral vasculature, CT
angiography, digital subtraction angiography, MDCT
angiography, neurovascular injury, rotational angiography,
subarachnoid hemorrhage
I persons living in the United States
have or will have an intracranial
aneurysm [1]. A ruptured intracra-
CTA in the detection of intracranial aneurysms
have been published [5–17]. More recent stud-
ies have suggested that with MDCT, the diag-
nial aneurysm, leading to subarachnoid hemor- nostic accuracy of CTA approaches that of
DOI:10.2214/AJR.07.2491 rhage (SAH), carries a mortality rate of be- DSA [18–23]. MDCT uses multiple detector
tween 30% and 60% in the first 30 days [2, 3]. elements aligned in the z-axis (usually 4, 8, 16,
Received September 26, 2006; accepted after revision Moreover, many patients who survive experi- 32, or 64 rows) and uses continuous scanning
May 13, 2007. ence severe debilitation and morbidity. Digital while the patient moves through the gantry.
1Department
subtraction angiography (DSA) has tradition- The volumetric data set is used to produce
of Radiology, University of Washington
Medical Center, 1959 NE Pacific St., RR215, Box 357115, ally been considered the gold standard for an- maximum-intensity-projection, volume-ren-
Seattle, WA 98195-7115. Address correspondence to eurysm detection. However, DSA is an inva- dered, 2D, or 3D representations of the data
Y. Anzai. sive and costly procedure associated with a with readily available postprocessing software.
2Department
0.07% rate of permanent neurologic complica- Some authors have asserted that improvements
of Radiology, Harborview Medical Center,
Seattle, WA.
tions for patients with SAH [4]. At some insti- in resolution and the ability of multiplanar re-
tutions, CT angiography (CTA) has replaced formations allow MDCT to detect aneurysms
3Harborview Injury Prevention Research Center, DSA in the detection and pretreatment evalua- as small as 3 mm [20–22].
Seattle, WA. tion of intracranial aneurysms. CTA is consid- In many of the previously published stud-
AJR 2007; 189:898–903
ered a less invasive alternative diagnostic test ies, CTA was reviewed by two or more expert
that is readily available, less expensive, and reviewers or by neurointerventional radiolo-
0361–803X/07/1894–898
able to give more anatomic information relat- gists and neurosurgeons in consensus. This
© American Roentgen Ray Society ing to other intracranial structures than DSA. may have resulted in a positive bias toward

898 AJR:189, October 2007

 MDCT Angiography for Intracranial Aneurysm Detection

the diagnostic accuracy of CTA. To accu- graphics, presence of SAH on unenhanced head CT January 2004, 16-MDCT replaced 4- and 8-MDCT
rately assess how well CTA performs in a or lumbar puncture, and size and location of aneu- for CTA examinations. Overall, 191 of the CTA ex-
general clinical setting, we analyzed radiol- rysm or aneurysms. If no aneurysm was detected, the aminations were 16-MDCT, 137 were 4-MDCT, and
ogy reports of 336 consecutive patients who alternative diagnosis to explain the presence of SAH eight were 8-MDCT.
presented with nontraumatic SAH. or clinical presentation was recorded.
Our institution is a high-volume level 1 The location of the aneurysm was categorized as DSA Protocol
trauma center where approximately 22,000 middle, anterior, or posterior cerebral artery (MCA, DSA imaging was performed using 3D rota-
head CT and 800 head CTA examinations are ACA, or PCA, respectively); internal carotid artery tional angiography (Integris V3000, Philips Medi-
performed each year. The standard of care at (ICA); anterior communicating artery; posterior cal Systems). Images were acquired in the standard
our institution for patients undergoing evalu- communicating artery; basilar artery; or other (e.g., projections (anteroposterior [AP], lateral, and AP
ation for nontraumatic SAH is to undergo posterior inferior cerebellar artery, vertebral artery, and lateral obliques). Three-dimensional rotational
CTA first. Subsequently, most of these pa- superior cerebellar artery). Detected aneurysms angiography was routinely performed when an an-
tients undergo 3D rotational DSA to better as- were divided into four size categories: ≤ 3, 4–6, eurysm was found to better characterize the aneu-
sess the relationship of the aneurysm to the 7–10, and > 10 mm, as measured on DSA. The treat- rysm and its relationship to parent vessels to deter-
parent vessel and adjacent vessels and to find ment option for all 336 patients was recorded (i.e., mine the appropriate treatment option.
potentially CTA-occult aneurysms. surgical clipping, endovascular coiling, or observa- Three-dimensional rotational angiography uses
The purpose of this study was to evaluate tion). For those who underwent surgery, concor- a mode over an angle of 180° at a frame rate of 12.5
the effectiveness of CTA performed with dance between the surgical findings and the CTA and frames per second and a rotation speed of up to 30°
MDCT in patients undergoing evaluation for DSA findings was evaluated. An attempt was made per second. During the run, iodinated contrast agent
nontraumatic SAH. The diagnostic perfor- to identify the primary aneurysms by size, morphol- was injected (e.g., 300 mg/mL at a rate of 4 mL/s
mance of CTA was assessed in 336 consecu- ogy, and location with respect to the distribution of for 6 seconds) to provide continuous filling of the
American Journal of Roentgenology 2007.189:898-903.

tive patients for an 18-month period using ret- SAH or parenchymal hemorrhage. The clinical out- vasculature. Acquisition took place in a single 180°
rospective review of radiology reports, with comes of patients with aneurysms were categorized rotational angiography scan. Three-dimensional
DSA as the gold standard. We also compared as death, some neurologic deficit, or no neurologic volumes were reconstructed on a workstation (Inte-
the diagnostic accuracy of CTA performed deficit. Cases with discordant CTA and DSA find- gris 3D-RA, Philips Medical Systems).
with 4-, 8-, and 16-MDCT versus CTA per- ings were reviewed in more detail.
formed with 16-MDCT alone. CTA and DSA Reporting
CT Scanner Protocols In this study, 10 attending neuroradiologists
Materials and Methods The 336 patients were evaluated using 4-, 8-, or with various degrees of experience and expertise in
Research Plan and Subjects 16-MDCT. Each CTA examination included unen- cerebral aneurysms generated CTA and DSA re-
The electronic database search engine (Folio hanced and contrast-enhanced head imaging. The ports. In most cases (334/336), the CTA examina-
Views, version 4.2, for Windows [Microsoft], Open CTA images were sent to PACS and postprocessing tion preceded the DSA examination, and the CTA
Market) identified 406 consecutive patients who workstations (Vitrea 2, version 3.2, Vital Images). images were available on the PACS station before
were evaluated for SAH and underwent both CTA The protocol for the CTA portion of the examina- the DSA images.
of the head and intracranial DSA within 48 hours tion was as follows: 110 mL of iodixanol (Visi-
from July 2003 to January 2005. Folio Views iden- paque, Nycomed) for 4- and 8-MDCT or 80 mL of Data Analysis
tifies cases in the radiology database by using ex- iohexol (Omnipaque, Nycomed) for 16-MDCT fol- Using DSA as the gold standard, we calculated
amination codes or by searching for keywords in lowed by 30 mL of saline infused at 3.0 mL/s for 4- sensitivity, specificity, positive predictive value
dictated reports. We excluded 36 patients with a and 16-MDCT and at 4.0 mL/s for 8-MDCT. Slice (PPV), and negative predictive value (NPV) on per-
history of trauma or if there was uncertainty about thickness was 1.25 mm for 4- and 8-MDCT and aneurysm and per-patient bases, and the results
whether SAH had been caused by trauma (e.g., pa- 0.625 mm for 16-MDCT, and the table interval was were stratified by aneurysm size and location. CIs
tients with a history of “found down” or “fall”). An- 0.8 mm for 4- and 8-MDCT and 0.625 mm for 16- of 95% and chi-square test values were also calcu-
other 34 patients who were undergoing a follow-up MDCT. Table speed was 7.5 mm per rotation for 4- lated. The diagnostic accuracy of 16-MDCT was
examination for a known condition, such as a MDCT, 13.5 mm per rotation for 8-MDCT, and compared with that of the combined scanner data.
treated aneurysm or arteriovenous malformation 13.75 mm per rotation for 16-MDCT. A setting of
(AVM), were also excluded. This resulted in 336 140 kV was used, and tube current was 300 mA for Results
patients who were evaluated for nontraumatic SAH 4-MDCT, 380 mA for 8-MDCT, and 350–380 mA Patient Population
and underwent CTA and DSA within 48 hours. The for 16-MDCT. The display field of view was 16 cm. Of the 336 patients evaluated in this study, an-
diagnostic performance of CTA was assessed by The CTA source images were reviewed by 10 di- eurysms were detected in 211 patients. Of the
retrospective review of radiology reports using agnostic neuroradiologists at the PACS station. Six 211 patients with aneurysms, there were 133
DSA as the gold standard. standard views of 3D reformatted images were cre- males and 78 females who ranged in age from 13
ated by CT technologists before the radiologists’ re- to 92 years, with a median age of 55 years. In the
Data Extraction view in gray-scale. On dedicated workstations sepa- 125 remaining patients, either there was an alter-
This study was approved by our institutional re- rate from the PACS station, the neuroradiologist used native diagnosis for clinical presentation (e.g.,
view board for a retrospective chart review and data postprocessing software to create additional 3D and AVM, vasculitis, infarct, tumor, moyamoya dis-
extraction. For the 336 patients, the following infor- color reformation images as desired. This additional ease) or no clear cause was found. These 125 pa-
mation was recorded by review of CTA and DSA re- postprocessing is not standard and was performed tients served as the negative cases for the sensi-
ports and electronic patient records: basic demo- only occasionally. At our institution, beginning in tivity and specificity calculations.

AJR:189, October 2007 899

 Colen et al.

Fig. 1—Bar graph shows Aneurysm Distribution

sensitivity, specificity, 100
93.2 92.9
96.3 96.3 Overall, 284 aneurysms were detected on
positive predictive value
(PPV), and negative 82.7 81.7
DSA. The size distribution of the aneurysms
predictive value (NPV) by 80 was as follows: 71 (25%) were ≤ 3 mm, 94
aneurysm for combined 71.7
69.0 (33%) were 4–6 mm, 75 (26%) were 7–10 mm,
MDCT data (gray bars)
separated from 16-MDCT 60
and 44 (15%) were > 10 mm. The distribution
Percent of aneurysm locations was as follows: 64 MCA
alone (black bars). There
was no significant (23%), 57 ICA (20%), 48 anterior communicat-
difference between 40
ing artery (17%), 41 posterior communicating
data sets.
artery (14%), 27 ACA (10%), 24 basilar artery
20 (8%), 11 PCA (4%), and 12 other (4%).

0 Discordance by Aneurysm
Sensitivity Specificity PPV NPV The overall sensitivity and specificity of CTA
per aneurysm was 83% (0.78–0.87) and 93%
(0.88–0.97), respectively. The PPV was 96%
(0.93–0.98), and the NPV was 72% (0.64–0.78).
Fig. 2—69-year-old man with right middle cerebral Sensitivity was then calculated for each size cate-
artery (MCA) trifurcation aneurysm.
gory. For aneurysms ≤ 3 mm, sensitivity was 45%.
A, Axial CT angiography (CTA) image shows
thrombosed, partially calcified, 20-mm right MCA For aneurysms 4–6, 7–10, and > 10 mm, sensitiv-
trifurcation aneurysm (arrow). ity was 90%, 99%, and 100%, respectively
American Journal of Roentgenology 2007.189:898-903.

B, Selected digital subtraction angiography image from (p < 0.001). The sensitivity, specificity, PPV, and
right internal carotid artery injection shows large right MCA
aneurysm (black arrow) seen on CTA (A). A second 4-mm NPV of CTA were not significantly altered when
aneurysm that was not seen on CTA is seen at origin of right 16-MDCT data were separated from the com-
MCA–anterior cerebral artery bifurcation (white arrow). bined MDCT data for subgroup analyses (Fig. 1).
CTA failed to detect 49 (17%) of the 284 an-
eurysms identified on DSA. Thirty-nine (80%)
of these 49 aneurysms were ≤ 3 mm; nine (18%)
were in the 4- to 6-mm range, and one (2%) was
in the 7- to 10-mm range. None of the aneurysms
that was > 10 mm were missed on CTA (Fig. 2).
The locations of CTA-missed aneurysms are il-
lustrated in Figure 3 (listed above the total num-
ber of CTA-detected aneurysms in that location).
Of the 49 CTA misses, 16 (33%) were aneu-
rysms of the MCA; 12 (24%), ICA; six (12%),
posterior communicating artery; five (10%),
PCA; five (10%), ACA; four (8%), other; and
one (2%), anterior communicating artery. There
were a total of nine false-positives on CTA: two
A B each of the posterior communicating artery, basi-
lar artery, and ICA; and one each of the MCA,
ACA, and PCA (Fig. 4).

70

60
16
No. of Aneurysms

50 12 1

40 6

30
5 0
48 45 47
20
Fig. 3—Bar graph shows number of missed aneurysms on CT angiography (CTA)
35 (gray) compared with reference standard (digital subtraction angiography) listed
10 22 24
4
over total number of CTA-detected aneurysms in that location (white). Most missed
5 aneurysms were of middle cerebral artery (MCA), which was also most common
8 6 aneurysm location. Highest percentage of missed aneurysms were of posterior
0
MCA ICA Acom Pcom ACA Basilar Other PCA cerebral artery (PCA), with 45% of aneurysms in that location missed. ICA = internal
carotid artery, Acom = anterior communicating artery, Pcom = posterior
Location communicating artery. ACA = anterior cerebral artery.

900 AJR:189, October 2007

 MDCT Angiography for Intracranial Aneurysm Detection

Discordance by Patient
Among 336 patients, 211 patients were
found to have at least one aneurysm and 125 pa-
tients did not have an aneurysm. When calcu-
lated on a per-patient basis (i.e., whether a pa-
tient was found to have an aneurysm on CTA),
the sensitivity was 95% (0.91–0.97), specificity
was 97% (0.92–0.99), PPV was 98.0%
(0.95–0.99), and NPV was 91.2% (0.86–0.96).
The sensitivity, specificity, PPV, and NPV of
CTA were not significantly altered in subgroup
analyses when 16-MDCT data were separated
from the combined MDCT data.
There were 49 false-negatives and nine false-
positives on CTA, resulting in 58 discordant an-
eurysms in 46 patients. Thirty-one (67%) of the
A B 46 patients had multiple aneurysms. In 15 of the
46 patients, there was discordance between CTA
and DSA findings with regard to detection of a
primary aneurysm (Table 1). Of these 15 pa-
tients, CTA findings were false-negative in 11
American Journal of Roentgenology 2007.189:898-903.

patients and false-positive in four. Therefore, 11

(22%) of the 49 aneurysms missed on CTA were
primary aneurysms thought to be responsible for
SAH, and 38 (78%) of 49 aneurysms were sec-
ondary or multiple aneurysms. Thus, CTA
missed 11 (5.2%) of 211 primary aneurysms.
Of the 11 patients with false-negative CTA
examinations for primary aneurysms, five under-
went surgery and one underwent endovascular
coiling, confirming the presence an aneurysm.
C The remaining five patients were observed. The
clinical outcome of nine of these 11 patients re-
sulted in no neurologic deficit. The other two pa-
tients, one who had a 2-mm ICA aneurysm and
another who had a 3-mm posterior communicat-
ing artery aneurysm missed on CTA, died due to
massive brain swelling that occurred immedi-
ately postoperatively and to myocardial infarc-
tion thought to be due to a preexisting heart con-
dition, respectively (Fig. 5). Thus, two (0.5%)
of 336 patients evaluated had primary aneu-
rysms that were missed on CTA and negative
clinical outcomes.

Discussion
Several articles dealing with the diagnostic
accuracy of CTA compared with DSA as the
gold standard have been published in the liter-
ature. The accurate interpretation of their re-
sults requires an understanding of the study de-
sign and methods used. The studies that were
D
Fig. 4—64-year-old man with 5-mm basilar tip aneurysm on CT angiography (CTA). interpreted by one or two expert blinded re-
A, Axial unenhanced head CT image shows acute hemorrhage within basal cisterns. viewers using dedicated 2D and 3D reformat-
B, Axial CTA image shows prominent basilar artery tip (arrow). ted images often resulted in the highest sensi-
C, Three-dimensional reconstructed image from CTA shows prominent basilar artery tip (arrow), which was
tivity. Individual behavior may have been
described as 5-mm basilar tip aneurysm.
D, Selected digital subtraction angiography (DSA) image from left vertebral artery injection does not reveal aneurysm. altered because reviewers were aware that their
Follow-up DSA (not shown) performed did not reveal aneurysm. interpretation of the CTA examination was be-

AJR:189, October 2007 901

 Colen et al.

TABLE 1: Cases with Discordant CT Angiography (CTA) and Digital Our results are concordant with the published
Subtraction Angiography (DSA) Findings data showing increased sensitivity of CTA with
Discordant CTA and DSA Findings for Aneurysm Detection Value increasing aneurysm size, with the threshold for
No. of aneurysms missed on CTA / total no. of aneurysms detected on DSA 58 / 284 detection being > 3 mm. The sensitivity of CTA
for the detection of aneurysms ≤ 3 mm was 45%
No. of discordant cases / total no. of cases 46 / 336
and that for aneurysms > 3 mm was 95.3%.
No. of patients with multiple aneurysms 31
Eighty percent of the aneurysms missed on CTA
No. of patients with primary aneurysms 15 in our study were ≤ 3 mm. Only one
False-negative primary aneurysms 11 aneurysm > 6 mm was missed on CTA, and that
False-positive primary aneurysms 4 aneurysm was a cavernous ICA aneurysm in a
patient with six other aneurysms, including
multiple bilateral ICA aneurysms.
ing studied. This “Hawthorne effect” is fre- CTA. Dedicated viewing on a 3D workstation The diagnostic accuracy of CTA was much
quently seen in the well-controlled research might have improved the diagnostic accuracy higher when calculated on a per-patient ba-
setting. In our study, data were collected from of CTA. However, this was not routinely per- sis—approaching 95% sensitivity and 97%
multiple radiologists’ interpretations of routine formed in our institution (beyond the six stan- specificity—than on a per-aneurysm basis.
clinical cases, which reflects the “real” diag- dard technologist-generated reconstructions). More important, the NPV of CTA in this
nostic accuracy and effectiveness of CTA, in- The degree of scrutiny of the CTA might have study was 91.2%, which is critical for the use
stead of its efficacy. been compromised as a result of the reviewers’ of CTA in the context of a screening test.
Our study showed a sensitivity of 83% and knowledge that most patients at our institution Thirty-one (67%) of the 46 patients with an-
specificity of 93% per aneurysm, which was with SAH undergo DSA after CTA. This eurysms missed on CTA were patients with
American Journal of Roentgenology 2007.189:898-903.

slightly lower in sensitivity but higher in speci- knowledge could have affected missing small multiple aneurysms. One should not underes-
ficity than reported in more recent studies using incidental aneurysms in patients in whom a timate, however, the importance of detecting
MDCT. In a meta-analysis of 21 published primary aneurysm had already been detected multiple aneurysms. Secondary aneurysms
studies of more than 1,250 patients, Chappel et on CTA. An alternative explanation for the can be treated at the same time as the primary
al. [5] found CTA sensitivity to range from 75% lower sensitivity of CTA in our study may be aneurysm, either by a surgical or endovascu-
to 100%, with a cumulative sensitivity of 93%, due to advances in the accuracy of the gold lar approach. Given the morbidity and costs
and specificity to range from 50% to 100%, standard itself. Most of our patients underwent associated with cerebral aneurysm treatment,
with a cumulative specificity of 88%. 3D rotational DSA performed by dedicated accurate detection of all aneurysms before
The slightly lower sensitivity of our study neurointerventional radiologists. The use of making a treatment decision is essential. Sim-
(83%) compared with others reported in the lit- 3D DSA may lead to the detection of smaller ilarly, false-positive CTA results could have
erature might be explained by multiple factors; aneurysms than standard DSA [24, 25]. Our resulted in unnecessary surgical exploration
CTA was interpreted as a part of a clinical ex- study showed a higher specificity than most of associated with potentially substantial mor-
amination by one of our 10 neuroradiologists the published data, suggesting that false-posi- bidity and mortality if CTA were to com-
with variable degrees of experience and exper- tive cases in a “real” clinical setting are fewer pletely replace DSA for the diagnostic
tise in the diagnosis of cerebral aneurysms on than that by the expert reviewers. workup for patients with SAH.

Fig. 5—67-year-old woman with 3-mm posterior communicating artery aneurysm.

A, Axial unenhanced head CT image shows acute hemorrhage within basal cisterns.
B, Axial CT angiography image does not reveal aneurysm.
C, Selected digital subtraction angiography image from left internal carotid artery injection reveals 3-mm posterior
communicating artery aneurysm (arrow).

A B C

902 AJR:189, October 2007

 MDCT Angiography for Intracranial Aneurysm Detection

The subgroup analyses comparing 16- tervention. International Study of Unruptured In- perience with computed tomographic angiography
MDCT data with the combined data set tracranial Aneurysms Investigators. N Engl J for the detection of intracranial aneurysms in the
showed no significant difference in accuracy. Med 1998; 339:1725–1733 [Erratum in N Engl J setting of acute subarachnoid hemorrhage. Neuro-
Technologic improvement, particularly in 3D Med 1999; 340:744] surgery 1997; 41:522–527
volume-rendering or reconstruction tech- 2. Sacco RL, Wolf PA, Bharucha NE, et al. Subarach- 14. Zouaoui A, Sahel M, Marro B, et al. Three-dimen-
niques, may improve the accuracy of CTA in noid and intracerebral hemorrhage: natural history, sional computed tomographic angiography in de-
the detection of aneurysms < 3 mm. prognosis, and precursive factors in the Framing- tection of intracranial aneurysms in acute subarach-
Our study has several limitations. It is a ret- ham Study. Neurology 1984; 34:847–854 noid hemorrhage. Neurosurgery 1997; 41:125–130
rospective review using radiology reports 3. Jevela S, Porras M, Poussa K. Natural history of un- 15. Ogawa T, Okudera T, Noguchi K, et al. Intracranial
rather than prospective blinded reviewers. The ruptured intracranial aneurysms: probability and aneurysms: evaluation with three-dimensional CT
interpretation of CTA examinations was not risk factors for aneurysm rupture. Neurosurgery Fo- angiography. Am J Neuroradiol 1996; 17:447–454
uniform among our 10 neuroradiologists. cus 2000; 8:Preview 1 16. Kouskouras C, Charitanti A, Giavroglou C, et al. In-
Many of our neuroradiologists view a set of 4. Cloft HJ, Joseph GJ, Dion JE. Risk of intracra- tracranial aneurysms: evaluation using CTA and
source images with only a few default 3D re- nial angiography in patients with subarachnoid MRA—correlation with DSA and intraoperative
formations. Also, we enrolled only patients hemorrhage, cerebral aneurysm, and arterio- findings. Neuroradiology 2004; 46:842–850
who underwent both CTA and DSA for aneu- venous malformation: a meta-analysis. Stroke 17. Hoh BL, Cheung AC, Rabinov JD, et al. Results of
rysm detection, targeting a population with a 1999; 30:317–320 a prospective protocol of computed tomographic
high prevalence of aneurysm. For this reason, 5. Chappel ET, Moure FC, Good MC. Comparison of angiography in place of catheter angiography as the
our study likely underestimates the number of computed tomographic angiography with digital only diagnostic and pretreatment planning study for
true-negative CTA results by excluding patients subtraction angiography in the diagnosis of intra- intracranial aneurysms by a combined neurovascu-
who may have undergone evaluation with CTA cranial aneurysms: a meta-analysis. Neurosurgery lar team. Neurosurgery 2004; 54:1329–1340
American Journal of Roentgenology 2007.189:898-903.

alone. Despite these limitations, to our knowl- 2003; 52:624–631 18. Kato Y, Sano H, Katada K, et al. Application of
edge our study includes the largest number of 6. Goddard AJ, Tan G, Becker J. Computed tomogra- three-dimensional CT angiography (3D-CTA)
consecutive patients in the literature from a sin- phy angiography for the detection and characteriza- to intracranial aneurysms. Surg Neurol 1999;
gle institution where all patients underwent tion of intra-cranial aneurysms: current status. Clin 52:113–121
both CTA with MDCT and 3D rotational DSA Radiol 2005; 60:1221–1236 19. Kato Y, Nair S, Sano H, et al. Multi-slice 3D-CTA:
within 48 hours. Our results provide a fair rep- 7. Dammert S, Krings T, Moller-Hartmann W, et al. an improvement over single slice helical CTA for
resentation of how well CTA performs in a gen- Detection of intracranial aneurysms with multislice cerebral aneurysms. Acta Neurochir (Wien) 2002;
eral clinical setting and reflect the effectiveness CT: comparison with conventional angiography. 144:715–722
of CTA for the detection of aneurysms in pa- Neuroradiology 2004; 46:427–434 20. Forsting M. CTA of the ICA bifurcation and in-
tients with nontraumatic SAH. 8. Karamessini MT, Kagadis GC, Petsas T, et al. CT tracranial vessels. Eur Radiol 2005; 15[suppl
In conclusion, the results of our study angiography with three-dimensional techniques for 4]:D25–D27
showed that CTA was accurate for the detec- the early diagnosis of intracranial aneurysms: com- 21. Tipper G, U-King-Im JM, Price SJ, et al. Detection
tion of intracranial aneurysms in a routine clin- parison with intra-arterial DSA and the surgical and evaluation of intracranial aneurysms with 16-
ical setting at a high-volume level 1 trauma findings. Eur J Radiol 2004; 49:212–223 row multislice CT angiography. Clin Radiol 2005;
center. The majority of aneurysms that were 9. Kangasniemi M, Makela T, Koskinen S, et al. De- 60:565–572
missed on CTA were ≤ 3 mm (80%) and were tection of intracranial aneurysms with two-dimen- 22. Wintermark M, Uske A, Chalaron M, et al. Multi-
found in patients with multiple aneurysms sional and three-dimensional multislice helical slice computerized tomography angiography in the
(67%). The high NPV (91.2%) yielded from computed tomographic angiography. Neurosurgery evaluation of intracranial aneurysms: a comparison
our per-patient analysis suggests that CTA has 2004; 54:336–340 with intraarterial digital subtraction angiography. J
merit in the context of screening. The diagno- 10. Kato Y, Katada K, Hayakawa M, et al. Can 3D-CTA Neurosurg 2003; 98:828–836
sis of aneurysms ≤ 3 mm on CTA to assess pre- surpass DSA in diagnosis of cerebral aneurysm? 23. Villablanca JP, Jahan R, Hooshi P, et al. Detection
operative mapping of all aneurysms remains Acta Neurochir (Wien) 2001; 143:245–250 and characterization of very small intracranial an-
challenging. Future studies are needed to de- 11. Anderson GB, Steinke DE, Petruk KC, et al. Com- eurysms by using 2D and 3D helical CT angiogra-
termine whether additional 3D viewing or 64- puted tomographic angiography versus digital sub- phy. Am J Neuroradiol 2002; 23:1187–1198
MDCT improves the diagnostic accuracy of traction angiography for the diagnosis and early 24. Ahmetoglu A, Kosucu P, Dinc H, Sari A, Kuzeyli
CTA with MDCT and provides adequate char- treatment of ruptured intracranial aneurysms. Neu- K, Gümele HR. Multi-slice CT angiography in the
acterization of the morphology of aneurysms rosurgery 1999; 45:1315–1320 diagnosis and characterization of cerebral aneu-
for preoperative imaging workup. 12. van Gelder JM. Computed tomographic angiogra- rysms [in Turkish]. Tani Girisim Radyol 2003;
phy for detecting intracranial aneurysms: implica- 9:302–308
tions of aneurysm size distribution for the sensitiv- 25. Abe T, Hirohata M, Tanaka N, et al. Clinical benefits
References ity, specificity, and likelihood ratios. Neurosurgery of rotational 3D angiography in endovascular treat-
1. [No authors listed]. Unruptured intracranial an- 2003; 53:597–605 ment of ruptured intracranial aneurysms. Am J Neu-
eurysms: risk of rupture and risks of surgical in- 13. Anderson GB, Findlay JM, Steinke DE, et al. Ex- roradiol 2002; 23:686–688

AJR:189, October 2007 903