close

Вход

Забыли?

вход по аккаунту

?

Diagnosing cerebral aneurysms by computed tomographic angiography Meta-analysis.

код для вставкиСкачать
ORIGINAL ARTICLE
Diagnosing Cerebral Aneurysms by
Computed Tomographic Angiography:
Meta-Analysis
Jan Menke, MD,1 Jörg Larsen, MD,2 and Kai Kallenberg, MD3
Objective: Cerebral aneurysms can cause substantial morbidity and mortality, specifically if they rupture, leading to
nontraumatic subarachnoid hemorrhage (SAH). This meta-analysis summarizes evidence about the accuracy of
noninvasive computed tomographic (CT) angiography for diagnosing intracranial aneurysms in symptomatic patients.
Methods: Four databases including PubMed were searched without language restrictions from January 1995 to
February 2010. Two independent reviewers selected and extracted 45 studies that compared CT angiography with
digital subtraction angiography (DSA) and/or intraoperative findings in patients suspected of having cerebral
aneurysms. Data from eligible studies were used to reconstruct 2 2 contingency tables on a per-patient basis in at
least 5 diseased and 5 nondiseased patients, with additional data on a per-aneurysm basis when available.
Results: The 45 included studies generally were of high methodological quality. Among the 3,643 patients included,
about 86% had nontraumatic SAH, and 77% had cerebral aneurysms. Overall, CT angiography had a pooled
sensitivity of 97.2% (95% confidence interval, 95.8–98.2%) for detecting and specificity of 97.9% (95.7–99.0%) for
ruling out cerebral aneurysms on a per-patient basis. On a per-aneurysm basis, the pooled sensitivity was 95.0%
(93.2–96.4%), and the specificity 96.2% (92.9–98.0%). The diagnostic accuracy of CT angiography with 16- or 64-row
multidetector CT was significantly higher than that of single-detector CT, especially in detecting small aneurysms of
4mm in diameter.
Interpretation: CT angiography has a high accuracy in diagnosing cerebral aneurysms, specifically when using
modern multidetector CT. In the future, CT angiography may increasingly supplement or selectively replace DSA in
patients suspected of having a cerebral aneurysm.
ANN NEUROL 2011;69:646–654
I
ntracranial aneurysms are common lesions, with a prevalence of 1 to 5% in adults.1,2 In most cases, the cerebral
aneurysms are small and remain asymptomatic.1 However,
symptomatic cerebral aneurysms can cause substantial morbidity and mortality, specifically if they rupture.1–3 Ruptured cerebral aneurysms are the most common etiology of
nontraumatic subarachnoid hemorrhage (SAH).1–3 SAH is
usually well diagnosed by unenhanced cranial computed
tomography (CT).1–4 Rarely, the unenhanced CT scan is
falsely negative, but diagnostic lumbar puncture may then
indicate SAH by xanthochromia of the cerebrospinal fluid
(CSF) or by the evidence of siderophages in the CSF.1–3,5,6
In unruptured cerebral aneurysms, clinical symptoms predominately relate to the aneurysm’s mass effect. A classical
finding is third-nerve palsy in aneurysms of the posterior
communicating artery.1 Intra-arterial digital subtraction angiography (DSA) is the gold standard in the diagnostic
imaging of cerebral aneurysms.3 Occasionally, a follow-up
DSA is required to detect an aneurysm that was missed during the initial DSA, and occasionally neurosurgical findings
modify the initial diagnosis obtained from DSA. However,
DSA generally provides the correct diagnosis. Nonetheless,
DSA is invasive and more time-consuming than CT angiography. Most symptomatic patients receive a cranial unenhanced CT, which could be followed by CT angiography in
the case of nontraumatic SAH. However, this approach
requires that CT angiography has a high sensitivity and
specificity for diagnosing cerebral aneurysms.
Meta-analyses of the diagnostic accuracy of singledetector CT angiography in cerebral aneurysms have
View this article online at wileyonlinelibrary.com. DOI: 10.1002/ana.22270
Received Jun 21, 2010, and in revised form Sep 16, 2010. Accepted for publication Sep 17, 2010.
Address correspondence to Dr Menke, Diagnostic Radiology, University Hospital, 37075 Goettingen, Germany. E-mail: Menke-J@T-Online.de
From the 1Department of Diagnostic Radiology, University Hospital; 2Department of Clinical Radiology, Protestant Hospital; and
Neuroradiology, University Hospital, Goettingen, Germany.
Additional supporting information can be found in the online version of this article.
C 2011 American Neurological Association
646 V
3
Department of
Menke et al: CT in Cerebral Aneurysms
been published in 2000 and 2003.7–9 Since then, several
studies have been performed using multidetector CT,
which is generally known to be more accurate than single-detector CT. Therefore, a meta-analytic update of the
published evidence is useful.
The objective of this meta-analysis was to determine the sensitivity and specificity of CT angiography
for detecting or ruling out cerebral aneurysms in symptomatic patients compared to the reference standard of
DSA or neurosurgery, by summarizing diagnostic test accuracy studies published between 1995 and February
2010.
Subjects and Methods
This meta-analytic review was written with reference to the
PRISMA statement (Preferred Reporting Items for Systematic
reviews and Meta-Analyses) without prepublication of the
review protocol.10,11
Data Sources and Searches
The PubMed, Scopus, Biosis Previews, and Web of Science
electronic databases were searched without language restriction
for ‘‘cerebral aneurysm,’’ ‘‘CT angiography,’’ ‘‘sensitivity and
specificity,’’ and related terms from January 1995 to February
2010 (Supplementary Table 1). Additionally, reference lists of
retrieved articles were searched.
Study Selection
Two observers independently selected eligible publications with
disagreement resolved in consensus. The inclusion criteria were:
(1) the patients were clinically suspected of having a cerebral
aneurysm; (2) the diagnostic index test was contrast-enhanced
helical CT angiography; (3) the reference standard was DSA or
its combination with neurosurgical findings; (4) the studied
condition was having or not having a cerebral aneurysm; (5)
the study topic was the primary diagnosis of cerebral aneurysms, that is, not the diagnosis of residual or recurrent aneurysmal disease; (6) the study was not limited to specific aneurysm
types or locations; (7) 2 2 contingency tables could be reconstructed on a per-patient basis (optionally also on a per-aneurysm basis); (8) at least 5 patients with as well as 5 patients
without an aneurysm were included, so that the study provided
meaningful numbers for sensitivity and specificity; and (9) it
was likely that the patient enrollment was not arbitrary. The
latter criterion was considered fulfilled if a study was prospective, if the time period was reported from which the patients
were enrolled, or if consecutive patient enrollment was stated.
A study was excluded if any one of the inclusion criteria was
not met.
Data Extraction
Data from included studies were independently extracted by 2
observers using electronic forms, with disagreement solved in
consensus. Extracted study characteristics comprised details
April 2011
about study design, patients, index test, and reference standard.
The 2 2 count data were extracted on a per-patient basis
and, if reported, also on a per-aneurysm basis. The number of
true negatives on a per-aneurysm basis was obtained from the
number of true negatives on a per-patient basis, which has been
the approach in most of the included studies. The per-patient
analysis addresses the question of whether at least 1 aneurysm is
detected in patients who have 1 or more aneurysms, or whether
a patient is correctly identified as having no aneurysm. The
per-aneurysm analysis addresses the question of whether all of
the aneurysms are detected.
Assessment of Study Quality and Risk of Bias
On the study level, the methodological study quality and sources of bias were assessed by the 14 quality items of the QUADAS tool.12 The QUADAS item 4 was scored positive if the
delay between index test and reference test was 3 days in all
patients. For scoring QUADAS item 6, the combination of
DSA and/or neurosurgical findings was considered equivalent
to solely using DSA as the reference standard. For each study, a
quality score was calculated by assigning 1 point for each QUADAS item if fulfilled, 0.5 points if unclear, and 0 points if not
fulfilled. Low study quality was defined as a score <11 points,
and high study quality as a score between 11 and 14 points.
This quality assessment was used to determine a study’s strength
of evidence.
On the outcome level, the risk of publication bias was
assessed by a funnel plot and bivariate meta-regression of the
logarithm of the diagnostic odds ratio (LOR ¼ logit[sensitivity]
þ logit[specificity]) versus the effective sample size parameter
ESS(1/2).13–15 The regression slope of LOR versus ESS(1/2)
was tested for positivity with the significance level set to p <
0.10.13 This tests whether small publications report a higher
diagnostic accuracy than large publications. ESS(1/2) is the
inverse of the square root of ESS.13 Each study’s ESS was calculated by ESS ¼ (4 RP RN)/(RP þ RN), where RP ¼ positive reference tests and RN ¼ negative reference tests.13
Data Synthesis and Analyses
For all studies included, the individual sensitivities and specificities were recalculated from the 2 2 count data on a perpatient basis using the meta-analysis program Meta-DiSC.16
This program was also used to assess the between-study heterogeneity of sensitivity and specificity by I-squared statistics.16,17
Pooled summary estimates for sensitivity and specificity were
obtained from a bivariate random effects meta-analysis that
models the between-study heterogeneity with normally distributed random effects and accounts for a possible correlation
between the studies’ sensitivities and specificities.14,18 Similar
calculations were performed on a per-aneurysm basis.
To analyze potential sources of heterogeneity, a subgroup
analysis was performed on a per-patient basis for each of the
study characteristics as categorical covariate in a bivariate random effects meta-regression of sensitivity and specificity, except
if that covariate did not vary.14,15,19 For each numerical
647
ANNALS
of Neurology
covariate, the studies were ranked by that covariate and subdivided at the median to obtain 2 subgroups of approximately
similar size.
Parameters relating to the intravenous contrast medium
(CM) were divided into the injected iodine volume (¼ CM volume iodine content) and the iodine flow rate (¼ CM flow iodine content), because it is the iodine that causes the vascular
enhancement. Within the meta-regressions, the LOR of each
subgroup was estimated on the outcome level by LOR ¼ logit
(sensitivity) þ logit(specificity) to obtain a summary estimate of
the overall accuracy in addition to sensitivity and specificity.14,20
Three additional analyses were performed with binomial
fixed effects models for studies that had reported sufficient
data. These analyses assessed (1) the number of aneurysms per
diseased patient, (2) the intracranial location of aneurysms, and
(3) how the detectability of aneurysms depends on aneurysm
size and CT scanner class. For the latter analysis, the aneurysms
were stratified into 2 groups: (1) 4mm in diameter and (2)
>4mm in diameter.
Statistics (with p < 0.05) were computed with SAS version 9.2 (SAS Institute, Cary, NC), unless specified otherwise.
Results
Literature Search and Selection
The literature search identified 2,245 sources (Fig 1). By
reading titles and abstracts, 2,142 sources were excluded.
The remaining 103 sources were evaluated by their
full text. Of these, 58 sources were excluded for reasons
given in Figure 1. The remaining 45 studies were
included.21–65 Whereas 9 studies21–28,30 utilizing singledetector CT had been part of previous meta-analyses,7–9
36 studies29,31–65 utilizing single- or multi-detector CT
had not been meta-analyzed before.
Descriptive Statistics
The study characteristics of the 45 studies included are
given in Supplementary Table 2. Supplementary Table 3
summarizes the image types used to evaluate CT angiography. The 2 2 count data are presented in Supplementary Table 4. Optional count data on a per-aneurysm basis
in addition to count data on a per-patient basis were provided by 43 of the 45 studies. Twenty-six studies were prospective, 12 studies were retrospective, and in 7 studies
this was unclear. The 45 studies comprised 3,643 patients.
Twenty-four studies used DSA as the reference standard, whereas 21 studies used a combination of DSA and/or
intraoperative findings. Among 42 studies that reported sufficient details, DSA was part of the reference standard in 89%
of patients, whereas solely neurosurgical findings (without
DSA) were the reference standard in only 11% of patients.
Most studies (38 of 45) reported the utilization of
4-vessel DSA. This was performed by selective catherization of both carotid arteries and both vertebral arteries
648
FIGURE 1: Flow diagram of articles included in the metaanalysis. MCA 5 middle cerebral artery; PCoA 5 posterior
communicating artery; CTA 5 computed tomographic
angiography.
(or 1 vertebral artery, if retrograde opacification of the
contralateral vertebral artery was sufficient). The remaining 7 studies did not report such methodical details
about DSA. Ten of the newer studies selectively applied
rotational angiography in addition to DSA.
According to the reference standard, 853 patients
had no aneurysm, whereas 2,790 patients had one or
more. Among the 45 studies, the average prevalence of
nontraumatic SAH was 86%, and the average prevalence
of cerebral aneurysmal disease was 76.6%. Twenty-five
studies (2,144 patients) had limited their inclusion criteria to patients with nontraumatic SAH, and among these
pure SAH patients 23.3% (500 patients) had no cerebral
aneurysm, whereas 76.7% (1,644 patients) had cerebral
aneurysms according to the reference standard.
On average, the percentage of women was 59%
(range, 40–77%). Most studies reported summary statistics of the patient age. In these studies, the mean/median
patient age per study was on average 53 years (range,
45–62 years), all studies included elderly patients of 70
years as well as young patients of 40 years, and about
half of the studies also included children or adolescents
suspected of having cerebral aneurysms.
CT angiography was performed with single-detector
CT (16 studies), 4-detector CT (9 studies), 16-detector
Volume 69, No. 4
Menke et al: CT in Cerebral Aneurysms
TABLE : Summary Estimates of the Meta-Analysis (per patient and per aneurysm)
Scan
Studies, No.
Sensitivity, % (95% CI)
Specificity, % (95% CI)
LOR (95% CI)
All studies
45
97.2 (95.8–98.2)
97.9 (95.7–99.0)
7.38 (6.45–8.31)
1-row CT
16
93.3 (89.9–95.6)
95.9 (91.7–98.1)
5.79 (4.83–6.76)
4-row CT
9
97.3 (94.7–98.6)
94.6 (87.5–97.8)
6.45 (5.18–7.71)
16-row CT
12
98.4 (96.9–99.2)
99.7 (97.3–100)
9.85 (7.54–12.2)
64-row CT
8
99.2 (97.5–99.8)
99.2 (93.0–99.9)
9.64 (7.07–12.2)
All studies
43
95.0 (93.2–96.4)
96.2 (92.9–98.0)
6.19 (5.35–7.04)
1-row CT
15
91.8 (87.5–94.7)
94.0 (87.4–97.3)
5.17 (4.08–6.25)
4-row CT
8
92.8 (87.3–96.0)
90.8 (77.2–96.6)
4.85 (3.41–6.28)
16-row CT
12
96.9 (94.5–98.2)
98.0 (93.9–99.3)
7.29 (5.87–8.72)
64-row CT
8
97.8 (95.1–99.0)
98.7 (92.3–99.8)
8.12 (5.96–10.3)
Per patient
Per aneurysm
CI ¼ confidence interval; LOR ¼ logarithm of the diagnostic odds ratio; CT ¼ computed tomography.
CT (12 studies), or 64-detector CT (8 studies). On average, 100ml (range, 50–140) contrast medium with an iodine content of 320mg/ml (range 240–400) was intravenously injected at a flow rate of 3.7ml/s (range, 2.0–5.0)
during CT angiography. With respect to the injected iodine volume and iodine flow rate, these were on average
30g iodine (range, 15–42), with an average injection flow
rate of 1.2g iodine per second (range, 0.6–1.7). In 26
studies a single reader interpreted the CT angiography,
whereas in the other 19 studies 2 or more readers did this.
Study Quality, Publication Bias, and
Heterogeneity
The methodological quality of the 45 studies was generally high (Supplementary Fig 1 and Supplementary Table
5). The quality score was 13.0 to 13.5 points in 17 studies, 12.0 to 12.5 points in 19 studies, 11 to 11.5 points
in 7 studies, and 10.5 points in 2 studies. On the outcome level, the study quality score was not significantly
related to sensitivity (p ¼ 0.85) or specificity (p ¼ 0.45).
The funnel plot and regression test showed no significant
publication bias (p ¼ 0.71). On a per-patient and peraneurysm basis, the between-study heterogeneity of sensitivity and specificity was moderate, with I-squared ranging from 33% to 76%.
Meta-Analysis of Sensitivity and Specificity
The individual sensitivities and specificities of the 45 studies are summarized by forest plots with stratification for
the CT scanner class (Supplementary Fig 2). For diagnosing cerebral aneurysms by CT angiography, the meta-analysis of all 45 studies resulted in a pooled sensitivity of
April 2011
97.2% (95% confidence interval [CI], 95.8–98.2%) and
specificity of 97.9% (95.7–99.0%) on a per-patient basis
(Table). On a per-aneurysm basis, the pooled sensitivity of
all 45 studies was 95.0% (93.2–96.4%), and the pooled
specificity was 96.2% (92.9–98.0%).
Subgroup Analyses
The diagnostic accuracy (LOR) of CT angiography with
16-detector or 64-detector CT was significantly higher
than that of single-detector or 4-detector CT (see Table
and Supplementary Table 6). This trend is also evident
in Figure 2, which shows the meta-analytic results stratified for CT scanner class. Using CT images with submillimeter resolution (effective slice thickness <1mm)
was significantly more accurate than using thicker slices
for image interpretation (see Supplementary Table 6).
With high intravenous iodine flow rate, the sensitivity
was significantly higher (p ¼ 0.010), but the corresponding differences in the overall accuracy (LOR) did not
reach statistical significance (p ¼ 0.132). Studies with limitation to SAH populations showed significantly higher
specificity than studies without such limitation (p ¼
0.016), but the corresponding differences in the overall
accuracy (LOR) were equally not statistically significant
(p ¼ 0.143). The other study characteristics showed
some nonsignificant trends.
Images Types for Evaluation of CT Angiography
The combinations of image types used to evaluate CT angiography were manifold (see Supplementary Table 3), so
that meaningful subgroups could not be constructed for a
649
ANNALS
of Neurology
(mean sensitivity, 99%). The mean sensitivity for detecting
smaller aneurysms of 4mm diameter (n ¼ 830, 39% of
total) has gradually increased from about 75% in single-detector CT through 84% in 4-detector CT to about 94% in
16- and 64-detector CT. For all aneurysm sizes, the sensitivity of 16- or 64-detector CT was significantly higher than
that of single-detector CT (p < 0.05).
Discussion
FIGURE 2: Summary plot of sensitivity and specificity,
stratified for computed tomography (CT) scanner class. The
sensitivities and specificities of the 45 studies are
represented by colored symbols with stratification for
different CT scanner class. The corresponding colored
bowed curves are summary receiver operating characteristic
(ROC) curves. The central dot on each ROC curve represents
the bivariate summary estimate of sensitivity and specificity
of the random effects meta-analysis. Numerical summary
results are given in the Table. Pink squares 5 1 row; green
triangles 5 4 rows; blue dots 5 16 rows; red dots 5 64
rows.
Findings of the Meta-Analysis
According to this meta-analysis of 45 studies, CT angiography has a high overall sensitivity of about 97.2% (95%
CI, 95.8–98.2%) and a high specificity of about 97.9%
(95.7–99.0%) for diagnosing cerebral aneurysms on a
per-patient basis. On a per-aneurysm basis, the diagnostic
accuracy was only slightly lower.
In the subgroup analysis, CT angiography with 16and 64-detector CT was significantly more accurate than
single-detector or 4-detector CT, especially for detecting
small aneurysms of 4mm in diameter. Using thin
image slices with submillimeter resolution was more
accurate than using thicker slices for image interpretation. The latter and the former findings are related,
because submillimeter resolution is only provided by 16-
quantitative analysis. Nearly all studies (42 of 45) had
used the original axial source images or had used multiplanar reformations as a substitute. Maximum intensity projections were used in 33 of the 45 studies. Nearly all studies (43 of 45) used virtual 3-dimensional reconstructions.
When considering when the study began, until 2002 most
studies applied shaded surface display, and since 2003
most studies applied volume rendering for this purpose.
Additional Analyses
According to the reference standard, 84.3% of the diseased patients had 1 aneurysm, 11.2% had 2, and 4.5%
had 3 or more aneurysms. The intracranial locations of
the 1,848 aneurysms from 30 studies are summarized in
Supplementary Table 7. On average, approximately 70%
of aneurysms were located in the anterior circulation and
30% in the posterior circulation.
Figure 3 shows how the detectability of cerebral
aneurysms depends both on their size and on the applied
CT scanner class. This analysis was based on 2,141 aneurysms from 30 studies. Aneurysms of >4mm in diameter (n
¼ 1,311, 61% of total) were already well diagnosed by single-detector CT (mean sensitivity, 96%); however, 16-detector and 64-detector CT technology is even more accurate
650
FIGURE 3: Detectability of aneurysms, depending on
aneurysm size and computed tomography (CT) scanner
class. The sensitivity (vertical axis) for detecting cerebral
aneurysms depends on their size and on the applied CT
scanner class (horizontal axis). The aneurysms were
stratified into 2 groups: (1) 4mm in diameter (triangles),
and (2) >4mm in diameter (circles). Each symbol (triangle or
circle) represents the mean sensitivity of the corresponding
CT scanner class. The surrounding bands represent the
corresponding 95% confidence intervals.
Volume 69, No. 4
Menke et al: CT in Cerebral Aneurysms
and 64-detector CT, but not by single- or 4-detector CT
scanners.
For diagnosing cerebral aneurysms by CT angiography, the accuracy of 64-detector CT was not significantly
different from that of 16-detector CT. This is explained by
the fact that both scanner classes offer similar submillimeter
resolution of about 0.5 to 0.7mm per detector row. Clearly,
with otherwise similar scan parameters 64-detector CT is
up to 4-fold faster than 16-detector CT due to having more
detectors. However, the scan length in cerebral examinations
is relatively short (about 12–15cm) and may thus comfortably be scanned by 16-detector CT with submillimeter resolution and 0.5 seconds rotation time in <10 seconds.53
With bolus triggering, this is sufficiently fast for optimal
utilization of the iodinated contrast bolus. Scanning any
faster would have no particular advantage, unless temporally
resolved 4-dimensional CT angiography is intended.
Two further significant cofactors were identified. First,
higher flow rates of the intravenously injected iodine were
associated with higher sensitivity. This may be explained by
the associated higher iodine concentration in the blood,
which improves the contrast-to-noise ratio between the
arteries and the surrounding tissue. Second, the specificity
was higher in studies with a limitation to SAH populations
than in studies that also included patients suspected of having an unruptured aneurysm, but the reasons for this finding
are unclear. However, the overall accuracy (LOR) did not
differ significantly among the subgroups of both mentioned
cofactors, so that these findings may require further research.
The diagnostic accuracy of prospective studies was
not significantly different from studies with retrospective
or unknown design. This may indicate that, at least in this
case, well-performed retrospective studies can provide
results similar to those of prospective studies. The sensitivity and specificity of the index test was slightly lower in
studies that comprised more women or younger patients,
but these trends were insignificant. Studies with 1 experienced observer showed similar accuracies to those of studies with 2 or more observers. This is reassuring for clinical
settings where commonly only 1 experienced observer
may be available, especially in out-of-hours emergency settings. Contrast timing by real-time bolus tracking showed
the highest diagnostic accuracy and may be favored,
because it is easily applied. However, the differences versus
using test bolus timing or versus a fixed delay were not
statistically significant. Regarding the reference standard,
using solely DSA was not significantly different from
using DSA and/or intraoperative findings, so that both
approaches may be considered equivalently acceptable.
Most studies had used the original axial CT images
for interpretation, as recommended by guidelines.3 A substiApril 2011
tute is viewing thin axial slices of a multiplanar reformation.
Additionally, multiplanar reformation allows viewing at the
staple of original axial CT images from any direction,
including coronal and sagittal views. Many studies applied
maximum intensity projection, which depicts predominately
the contrast-filled arteries and their outpouchings, possibly
after virtual bone removal.64 Although maximum intensity
projection provides 2-dimensional images, it is a 3-dimensional image postprocessing technique.66 In accordance with
the guidelines, nearly all studies utilized virtual 3-dimensional images for aneurysm depiction.3,67 Until 2002, most
studies applied shaded surface display that, however, may
have an artificial appearance. Since about 2003, most studies applied volume rendering, which requires more powerful
computers but appears more natural.
Limitations
This meta-analysis has some limitations. Ten newer studies
used rotational angiography in addition to DSA, but this
novel technique was not available in older studies. The 45
included studies were limited to populations with high disease prevalence. Extrapolation of the meta-analytic results to
screening populations with their inherently much lower disease prevalence might be biased. Seventeen studies were
excluded because the required 2 2 count tables could not
be reconstructed. Perhaps the literature search did not find
all eligible studies, but if such omission is random then this
is less likely to have caused a systematic bias. Although the
funnel plot and regression test did not indicate significant
publication bias, there still could be some unnoticed publication bias. The 45 studies of this meta-analysis showed
high overall study quality according to the QUADAS assessment, which accounts for several sources of bias.12 The
reductions in study quality were predominately due to
unclear or missing information. Future studies of diagnostic
accuracy could avoid this by following the recommendations for reporting and quality assessment.12,68
Comparison to Other Meta-Analyses
Nine studies about single-detector CT angiography21–28,30
had been part of former meta-analyses.7–9 The present
meta-analysis adds 36 studies29,31–65 and includes CT angiography with modern multidetector CT, which justifies the
update. The inclusion of these additional studies allowed
the performance of subgroup analyses for covariates that
had not been meta-analyzed before. The present meta-analysis applies a bivariate random effects model that is now
recommended for obtaining summary estimates of sensitivity and specificity.18
The previous meta-analyses had been performed using
different fixed and random effects models.7–9 In patients
suspected of having a cerebral aneurysm, a previous meta651
ANNALS
of Neurology
analysis published in 2000 found a pooled sensitivity of
92% and specificity of 94% for single-detector CT angiography.7 A further meta-analysis published in 2003 reported
a similar sensitivity of 93% but a lower specificity of 88%
with single-detector CT when using unweighted meta-analytic calculations.8 Another meta-analysis published in 2003
focused on the relation between aneurysm size and diagnostic accuracy using single-detector CT angiography and
reported sensitivities ranging from 53% for 2mm aneurysms to 95% for 7mm aneurysms.9 Similarly, the first
meta-analysis published in 2000 found a low sensitivity of
61% for aneurysms 3mm in diameter and a high sensitivity of 96% for detecting aneurysms >3mm.7 The findings
of the 3 former meta-analyses are comparable to the singledetector findings of the present meta-analysis. In the literature, there is no previous meta-analysis of multidetector CT
angiography in cerebral aneurysms, so the current metaanalysis offers novel quantitative summary information.
Other Diagnostic Modalities
Noninvasive magnetic resonance angiography (MRA) also
provides high diagnostic accuracy in the detection of cerebral aneurysms. However, most patients with SAH are clinically not stable enough to be subjected to MRA, so that fast
CT angiography may be preferred in the acute setting.3
Intra-arterial DSA is currently the diagnostic gold
standard in cerebral aneurysmal disease and has a high
image resolution of 0.1 to 0.2mm, but provides only 2dimensional projection images. Currently it is being
studied whether the addition of rotational angiography
with 3-dimensional image reconstruction may provide a
better gold standard.1,55,69–74
DSA offers the option of directly proceeding with
endovascular coil embolization if an aneurysm is found.
However, in an emergency setting with aneurysmal SAH
in an unstable patient, CT angiography may be more readily available than DSA. In such a setting, the standard
unenhanced CT could be directly followed by CT angiography to assist the clinical decision between open neurosurgical clipping and neuroendovascular management.
Submillimeter CT angiography with 16- or 64-detector CT has a lower image resolution than DSA but
permits the 3-dimensional visualization of aneurysms and
assesses surrounding intracranial structures that are not
visible on DSA. Therefore, modern multidetector CT angiography with its 3-dimensional imaging capabilities
and traditional 2-dimensional DSA with its very high
diagnostic accuracy are complementary.
In the studies using 16- and 64-detector CT angiography, the findings were true positive or true negative in most
cases, but there was also a small false negative rate of unde652
tected aneurysm. However, this may also happen with DSA
as first-line method because a repeat DSA after about 1
week discloses a previously unrecognized cerebral aneurysm
in 1% to 2% of cases.3 Whether this additional small yield
is worth the costs and potential morbidity of the second
DSA is currently a source of controversy.3 In patients with
nontraumatic SAH, suspected cerebral aneurysm, and a negative first-line DSA, some institutions favor a second-line
DSA to possibly detect a missed aneurysm, whereas others
favor watchful waiting, but this is more a matter of patient
management than a matter of diagnostic accuracy. Such
controversy in patient management is not resolved by using
CT angiography. However, there could be consensus that
DSA may be replaced by high-quality CT angiography for
the initial diagnosis, specifically when DSA cannot be performed in a timely fashion.3 If such first-line CT angiography shows no cerebral aneurysm, then this might be treated
similarly to a negative first-line DSA according to the individual institution’s management standards, that is, some
institutions might prefer observation, whereas other institutions might proceed with a second-line DSA, possibly combined with rotational angiography.
Future Research
There is ongoing research into whether flat-panel
volumetric CT with its DSA-like image resolution could
further improve the imaging of small cerebral aneurysms.75–78 In clinically applied multidetector CT, the
trend is toward 128 and more detector rows.79–81 Future
research may show whether such advanced CT scanners
further improve the differential diagnosis in patients
suspected of having a cerebral aneurysm.
Conclusions
CT angiography has a high accuracy in diagnosing cerebral aneurysms, specifically when using modern multidetector CT. In the future, CT angiography may increasingly supplement or selectively replace DSA in patients
suspected of having a cerebral aneurysm.3
Potential Conflicts of Interest
Nothing to report.
References
1.
Brisman JL, Song JK, Newell DW. Cerebral aneurysms. N Engl J
Med 2006;355:928–939.
2.
van Gijn J, Kerr RS, Rinkel GJ. Subarachnoid haemorrhage. Lancet
2007;369:306–318.
3.
Bederson JB, Connolly ES Jr, Batjer HH, et al. Guidelines for the
management of aneurysmal subarachnoid hemorrhage. Stroke
2009;40:994–1025.
Volume 69, No. 4
Menke et al: CT in Cerebral Aneurysms
4.
Provenzale JM, Hacein-Bey L. CT evaluation of subarachnoid hemorrhage: a practical review for the radiologist interpreting emergency room studies. Emerg Radiol 2009;16:441–451.
26.
Velthuis BK, Rinkel GJ, Ramos LM, et al. Subarachnoid hemorrhage: aneurysm detection and preoperative evaluation with CT
angiography. Radiology 1998;208:423–430.
5.
Veuger AJ, Kortbeek LH, Booij AC. Siderophages in differentiation of
blood in cerebrospinal fluid. Clin Neurol Neurosurg 1977;80:46–56.
27.
6.
Ito U, Inaba Y. Cerebrospinal fluid cytology after subarachnoid
hemorrhage. J Neurosurg 1979;51:352–354.
Young N, Dorsch NW, Kingston RJ. Pitfalls in the use of spiral CT
for identification of intracranial aneurysms. Neuroradiology 1999;
41:93–99.
28.
Anderson GB, Steinke DE, Petruk KC, et al. Computed tomographic angiography versus digital subtraction angiography for
the diagnosis and early treatment of ruptured intracranial aneurysms. Neurosurgery 1999;45:1315–1320.
29.
Montes JM, Rubio MJ, Merino M, et al. Aneurismas cerebrales:
estudio mediante TC-helicoidal (angio-TC) con reconstrucciones
tridimensionales. Radiologia 1999;41:247–252.
30.
Korogi Y, Takahashi M, Katada K, et al. Intracranial aneurysms:
detection with three-dimensional CT angiography with volume
rendering—comparison with conventional angiographic and surgical findings. Radiology 1999;211:497–506.
7.
White PM, Wardlaw JM, Easton V. Can noninvasive imaging accurately depict intracranial aneurysms? A systematic review. Radiology 2000;217:361–370.
8.
Chappell ET, Moure FC, Good MC. Comparison of computed tomographic angiography with digital subtraction angiography in
the diagnosis of cerebral aneurysms: a meta-analysis. Neurosurgery 2003;52:624–631.
9.
van Gelder JM. Computed tomographic angiography for detecting cerebral aneurysms: implications of aneurysm size distribution
for the sensitivity, specificity, and likelihood ratios. Neurosurgery
2003;53:597–605.
31.
10.
Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred reporting
items for systematic reviews and meta-analyses: the PRISMA statement. Ann Intern Med 2009;151:264–269.
Milosevic Z. Acute subarachnoid haemorrhage: detection of
aneurysms of intracranial arteries by computed tomographic angiography. Radiol Oncol 1999;33:275–282.
32.
Cenni P, Acati R, Simonetti L, et al. L’angio-TC nel riscontro e nell’analisi delle caratteristiche delle lesioni aneurismatiche intracraniche dei pazienti con ESA. Riv Neuroradiol 2000;13:711–720.
33.
Pedersen HK, Bakke SJ, Hald JK, et al. CTA in patients with acute
subarachnoid haemorrhage. A comparative study with selective,
digital angiography and blinded, independent review. Acta Radiol
2001;42:43–49.
34.
Lukosevicius S, Vaitkevicius G. Galvos smegenu kompiuterines
tomografijos angiografija ivykus subarachnoidinei hemoragijai.
Medicina (Kaunas) 2002;38:140–146.
35.
White PM, Wardlaw JM, Lindsay KW, et al. The non-invasive
detection of intracranial aneurysms: are neuroradiologists any better than other observers? Eur Radiol 2003;13:389–396.
36.
Wintermark M, Uske A, Chalaron M, et al. Multislice computerized
tomography angiography in the evaluation of intracranial aneurysms: a comparison with intraarterial digital subtraction angiography. J Neurosurg 2003;98:828–836.
37.
Teksam M, McKinney A, Casey S, et al. Multi-section CT angiography for detection of cerebral aneurysms. Am J Neuroradiol 2004;
25:1485–1492.
38.
Karamessini MT, Kagadis GC, Petsas T, et al. CT angiography
with three-dimensional techniques for the early diagnosis of intracranial aneurysms. Comparison with intra-arterial DSA and the surgical findings. Eur J Radiol 2004;49:212–223.
39.
Dammert S, Krings T, Moller-Hartmann W, et al. Detection of intracranial aneurysms with multislice CT: comparison with conventional angiography. Neuroradiology 2004;46:427–434.
40.
Hoh BL, Cheung AC, Rabinov JD, et al. Results of a prospective
protocol of computed tomographic angiography in place of catheter angiography as the only diagnostic and pretreatment planning study for cerebral aneurysms by a combined neurovascular
team. Neurosurgery 2004;54:1329–1340.
11.
12.
13.
14.
Liberati A, Altman DG, Tetzlaff J, et al. The PRISMA statement for
reporting systematic reviews and meta-analyses of studies that
evaluate health care interventions: explanation and elaboration.
Ann Intern Med 2009;151:W65–W94.
Whiting P, Rutjes AW, Reitsma JB,et al. The development of
QUADAS: a tool for the quality assessment of studies of diagnostic accuracy included in systematic reviews. BMC Med Res Methodol 2003;3:25.
Deeks JJ, Macaskill P, Irwig L. The performance of tests of publication
bias and other sample size effects in systematic reviews of diagnostic
test accuracy was assessed. J Clin Epidemiol 2005;58:882–893.
Menke J. Bivariate random-effects meta-analysis of sensitivity and specificity with SAS PROC GLIMMIX. Methods Inf Med 2010;49:54–64.
15.
Menke J. Diagnostic accuracy of contrast-enhanced MR angiography in severe carotid stenosis: meta-analysis with metaregression
of different techniques. Eur Radiol 2009;19:2204–2216.
16.
Zamora J, Abraira V, Muriel A, et al. Meta-DiSc: a software for meta-analysis of test accuracy data. BMC Med Res Methodol 2006;6:31.
17.
Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring
inconsistency in meta-analyses. BMJ 2003;327:557–560.
18.
Leeflang MM, Deeks JJ, Gatsonis C, Bossuyt PM. Systematic reviews
of diagnostic test accuracy. Ann Intern Med 2008;149:889–897.
19.
Thompson SG, Higgins JP. How should meta-regression analyses
be undertaken and interpreted? Stat Med 2002;21:1559–1573.
20.
Glas AS, Lijmer JG, Prins MH, et al. The diagnostic odds ratio: a single
indicator of test performance. J Clin Epidemiol 2003;56:1129–1135.
21.
Vieco PT, Shuman WP, Alsofrom GF, Gross CE. Detection of circle
of Willis aneurysms in patients with acute subarachnoid hemorrhage: a comparison of CT angiography and digital subtraction
angiography. Am J Roentgenol 1995;165:425–430.
22.
Liang EY, Chan M, Hsiang JH, et al. Detection and assessment of
intracranial aneurysms: value of CT angiography with shaded-surface display. Am J Roentgenol 1995;165:1497–1502.
41.
Jayaraman MV, Mayo-Smith WW, Tung GA, et al. Detection of
intracranial aneurysms: multi-detector row CT angiography compared with DSA. Radiology 2004;230:510–518.
23.
Lenhart M, Bretschneider T, Gmeinwieser J, et al. Cerebral CT angiography in the diagnosis of acute subarachnoid hemorrhage.
Acta Radiol 1997;38:791–796.
42.
Tipper G, U-King-Im JM, Price SJ, et al. Detection and evaluation
of intracranial aneurysms with 16-row multislice CT angiography.
Clin Radiol 2005;60:565–572.
24.
Anderson GB, Findlay JM, Steinke DE, Ashforth R. Experience
with computed tomographic angiography for the detection of intracranial aneurysms in the setting of acute subarachnoid hemorrhage. Neurosurgery 1997;41:522–527.
43.
Pavcec Z, Zokalj I, Saghir H, et al. Evaluation of circle of Willis
aneurysms with spiral computed tomographic angiography. Coll
Antropol 2006;30:867–870.
44.
25.
Rohnert W, Hanig V, Hietschold V, Abolmaali N. Aneurysmanachweis
bei Subarachnoidalblutung—CT-Angiographie versus digitale Subtraktionsangiographie. Aktuelle Radiol 1998;8:63–70.
Kadri S, Brunel H, Bourbotte G, et al. L’angio-scanner cerebral
multibarrette peut-il supplanter l’angiographie conventionnelle
dans le diagnostic etiologique des hemorragies sous-arachnoidiennes? J Neuroradiol 2006;33:45–50.
April 2011
653
ANNALS
of Neurology
45.
Halaji A, Budiselic B, Kovacic-Dujmovic S, et al. Angiografija
kompjutoriziranom tomografijom u dijagnostici unutarlubanjskih
aneurizmi. Medicina (Rijeka) 2006;42:103–107.
62.
Li Q, Lv F, Li Y, et al. Evaluation of 64-section CT angiography for
detection and treatment planning of intracranial aneurysms by
using DSA and surgical findings. Radiology 2009;252:808–815.
46.
Agid R, Lee SK, Willinsky RA, et al. Acute subarachnoid hemorrhage: using 64-slice multidetector CT angiography to ‘‘triage’’
patients’ treatment. Neuroradiology 2006;48:787–794.
63.
Chen W, Wang J, Xing W, et al. Accuracy of 16-row multislice
computerized tomography angiography for assessment of intracranial aneurysms. Surg Neurol 2009;71:32–42.
47.
Yoon DY, Lim KJ, Choi CS, et al. Detection and characterization
of intracranial aneurysms with 16-channel multidetector row CT
angiography: a prospective comparison of volume-rendered
images and digital subtraction angiography. Am J Neuroradiol
2007;28:60–67.
64.
Wang YC, Liu YC, Hsieh TC, et al. Aneurysmal subarachnoid hemorrhage diagnosis with computed tomographic angiography and
OsiriX. Acta Neurochir (Wien) 2010;152:263–269.
65.
Zhang LJ, Wu SY, Niu JB, et al. Dual-energy CT angiography in
the evaluation of intracranial aneurysms: image quality, radiation
dose, and comparison with 3D rotational digital subtraction angiography. Am J Roentgenol 2010;194:23–30.
66.
Sun Y, Parker DL. Performance analysis of maximum intensity projection algorithm for display of MRA images. IEEE Trans Med
Imaging 1999;18:1154–1169.
67.
Dhawan AP. Medical image analysis. 1st ed. Hoboken, NJ: WileyInterscience, 2003:280–295.
68.
Bossuyt PM, Reitsma JB, Bruns DE, et al. The STARD statement
for reporting studies of diagnostic accuracy: explanation and elaboration. Ann Intern Med 2003;138:W1–W12.
69.
Buhk JH, Kallenberg K, Mohr A, et al. Evaluation of angiographic
computed tomography in the follow-up after endovascular treatment of cerebral aneurysms—a comparative study with DSA and
TOF-MRA. Eur Radiol 2009;19:430–436.
70.
Irie K, Murayama Y, Saguchi T, et al. Dynact soft-tissue visualization using an angiographic C-arm system: initial clinical experience
in the operating room. Neurosurgery 2008;62(suppl 1):266–272.
71.
van Rooij WJ, Sprengers ME, de Gast AN, et al. 3D rotational angiography: the new gold standard in the detection of additional
intracranial aneurysms. AJNR Am J Neuroradiol 2008;29:976–979.
72.
McKinney AM, Truwit CL, Palmer CS, Teksam M. Intracranial
aneurysms: is the diagnostic accuracy rate of multidetector CT angiography equivalent to that of three-dimensional rotational conventional angiography? Radiology 2008;246:982.
73.
Hochmuth A, Spetzger U, Schumacher M. Comparison of threedimensional rotational angiography with digital subtraction angiography in the assessment of ruptured cerebral aneurysms. AJNR
Am J Neuroradiol 2002;23:1199–1205.
74.
Brinjikji W, Cloft H, Lanzino G, Kallmes DF. Comparison of 2D digital
subtraction angiography and 3D rotational angiography in the evaluation of dome-to-neck ratio. AJNR Am J Neuroradiol 2009;30:831–834.
75.
Ross W, Cody DD, Hazle JD. Design and performance characteristics of a digital flat-panel computed tomography system. Med
Phys 2006;33:1888–1901.
76.
Guo YW, Ke YQ, Zhang SZ, et al. Combined application of virtual
imaging techniques and three-dimensional computed tomographic angiography in diagnosing intracranial aneurysms. Chin
Med J 2008;121:2521–2524.
Gupta R, Grasruck M, Suess C, et al. Ultra-high resolution flatpanel volume CT: fundamental principles, design architecture,
and system characterization. Eur Radiol 2006;16:1191–1205.
77.
58.
Uysal E, Oztora F, Ozel A, et al. Detection and evaluation of intracranial aneurysms with 16-row multislice CT angiography: comparison
with conventional angiography. Emerg Radiol 2008;15:311–316.
Zhang T, Schulze D, Xu X, Kim J. Tetrahedron beam computed
tomography (TBCT): a new design of volumetric CT system. Phys
Med Biol 2009;54:3365–3378.
78.
59.
Haghighatkhah HR, Sabouri S, Borzouyeh F, et al. Diagnostic accuracy of multi-slice computed tomographic (MSCT) angiography in
the detection of cerebral aneurysms. Iran J Radiol 2008;5:209–214.
Mitha AP, Reichardt B, Grasruck M, et al. Dynamic imaging of a
model of intracranial saccular aneurysms using ultra-high-resolution flat-panel volumetric computed tomography. Laboratory
investigation. J Neurosurg 2009;111:947–957.
79.
60.
Li Q, Lv F, Li Y, et al. Subtraction CT angiography for evaluation
of intracranial aneurysms: comparison with conventional CT angiography. Eur Radiol 2009;19:2261–2267.
Page M, Nandurkar D, Crossett MP, et al. Comparison of 4 cm Z-axis and
16 cm Z-axis multidetector CT perfusion. Eur Radiol 2010;20:1508–1514.
80.
Siebert E, Bohner G, Dewey M, et al. 320-slice CT neuroimaging:
initial clinical experience and image quality evaluation. Br J Radiol
2009;82:561–570.
81.
Murayama K, Katada K, Nakane M, et al. Whole-brain perfusion
CT performed with a prototype 256-detector row CT system: initial experience. Radiology 2009;250:202–211.
48.
49.
Lubicz B, Levivier M, Francois O, et al. Sixty-four-row multisection
CT angiography for detection and evaluation of ruptured intracranial aneurysms: interobserver and intertechnique reproducibility.
Am J Neuroradiol 2007;28:1949–1955.
Pozzi-Mucelli F, Bruni S, Doddi M, et al. Detection of intracranial
aneurysms with 64 channel multidetector row computed tomography: comparison with digital subtraction angiography. Eur J
Radiol 2007;64:15–26.
50.
Nijjar S, Patel B, McGinn G, West M. Computed tomographic angiography as the primary diagnostic study in spontaneous subarachnoid hemorrhage. J Neuroimaging 2007;17:295–299.
51.
Taschner CA, Thines L, Lernout M, et al. Treatment decision in
ruptured intracranial aneurysms: comparison between multi-detector row CT angiography and digital subtraction angiography. J
Neuroradiol 2007;34:243–249.
52.
El Khaldi M, Pernter P, Ferro F, et al. Detection of cerebral aneurysms in nontraumatic subarachnoid haemorrhage: role of multislice CT angiography in 130 consecutive patients. Radiol Med
2007;112:123–137.
53.
Papke K, Kuhl CK, Fruth M, et al. Intracranial aneurysms: role of
multidetector CT angiography in diagnosis and endovascular therapy planning. Radiology 2007;244:532–540.
54.
Romijn M, Gratama van Andel HA, van Walderveen MA, et al.
Diagnostic accuracy of CT angiography with matched mask bone
elimination for detection of intracranial aneurysms: comparison
with digital subtraction angiography and 3D rotational angiography. Am J Neuroradiol 2008;29:134–139.
55.
56.
57.
61.
654
McKinney AM, Palmer CS, Truwit CL, et al. Detection of aneurysms by 64-section multidetector CT angiography in patients
acutely suspected of having an intracranial aneurysm and comparison with digital subtraction and 3D rotational angiography. Am J
Neuroradiol 2008;29:594–602.
Kokkinis C, Vlychou M, Zavras GM, et al. The role of 3D-computed tomography angiography (3D-CTA) in investigation of
spontaneous subarachnoid haemorrhage: comparison with digital
subtraction angiography (DSA) and surgical findings. Br J Neurosurg 2008;22:71–78.
Gerardin E, Daumas-Duport B, Tollard E, et al. Usefulness of multislice computerized tomography angiography in preoperative diagnosis of ruptured cerebral aneurysms. J Neuroradiol 2009;36:
278–284.
Volume 69, No. 4
Документ
Категория
Без категории
Просмотров
0
Размер файла
308 Кб
Теги
angiography, meta, tomography, analysis, computer, diagnosis, aneurysms, cerebral
1/--страниц
Пожаловаться на содержимое документа