Skip to main content
NCBI home page
Interventional Neuroradiology logoLink to Interventional Neuroradiology
. 2009 Jan 2;14(Suppl 2):53–63. doi: 10.1177/15910199080140S211

Meta-analysis on Diagnostic Accuracy of MR Angiography in the Follow-Up of Residual Intracranial Aneurysms Treated with Guglielmi Detachable Coils

Hsu-Huei Weng 1,2,3, Shaner-Yeun Jao 1, Chun-Yuh Yang 4, Yuan-Hsiung Tsai 1,a
PMCID: PMC3328063  PMID: 20557802

Summary

Patients with intracranial aneurysm after coiling with Guglielmi detachable coils (GDC) require imaging follow-up. The accuracy of noninvasive magnetic resonance angiography (MRA) techniques including time-of-flight (TOF) and contrast-enhancement (CE) have been compared with the gold standard digital subtraction angiography (DSA). We systematically reviewed the diagnostic accuracy of these imaging methods in follow-up study of patients with residual intracranial aneurysms after GDC treatment. The authors used MEDLINE, bibliographies, review articles, textbooks, and expert opinion to retrieve English-and non-English-language articles published from 1966 to December 2007. Sixteen suitable MRA original articles (14 TOF-MRA and six CE-MRA) with comparison to DSA have met the inclusion criteria. TOF-MRA had a pooled sensitivity of 90% (95% CI, 79% to 95%), a specificity of 95% (95% CI, 88% to 98%), and a diagnostic odds ratio (DOR) of 168.4 (95% CI, 60.3 to 470.3). CE-MRA had an overall sensitivity of 92% (95% CI, 79% to 97%), a specificity of 96% (95% CI, 91% to 98%), and a DOR of 280.4 (95% CI, 64.8 to 1212.6). The areas under two summary ROC curves of TOF-MRA and CE-MRA were 0.97 (95% CI, 0.96 to 0.99) and 0.98 (95% CI, 0.96 to 0.99), respectively. Compared with DS angiography, both TOF-MRA and CE-MRA can accurately depict the residual aneurysm. The diagnostic accuracy of TOF-MRA and CE-MRA tests offer comparable and equal results and may obviate the invasive DS angiography

Key words: aneurysm, coil, MR angiography, meta-analysis.

Introduction

Spontaneous subarachnoid hemorrhage (SAH) is a condition wherein an acute bleeding has occurred into the subarachnoid space surrounding the brain. Aneurysms are the cause of SAH in 85% of cases1. Spontaneous SAH accounts for only 5% of strokes1. Its' overall incidence is approximately 9 per 100,000 person-years 2.

Diagnosis of intracranial aneurysm requires several imaging tools such as digital subtraction angiography (DSA), computed tomographic angiography (CTA) and magnetic resonance angiography (MRA) 3. For the past decade, with the advancement of multislice CTA and MRA, particularly time-of-flight (TOF) MRA and contrast-enhanced (CE) MRA, the pre-treatment diagnosis of intracranial aneurysm have shown marked improvement4,5 with an accuracy of 90%6.

One of the most widely accepted treatment option for an intracranial aneurysm is the use of Guglielmi detachable coils (GDC)7. An imaging follow-up is often necessary to assess any residual aneurysm after coiling. Traditionally, the catheter-based DSA, however invasive, has been consistently considered by many as the gold standard for follow-up examination. It is also known that MRA is superior to CTA due to the absence of marked beam hardening and streak artifacts from endovascular coiling mass 8. Thus, noninvasive alternative imaging techniques such as TOF-MRA and CE-MRA8,9 have been tried to replace digital subtraction angiography in depicting post-coiling residual aneurysm. Their accuracy is still controversial.

The authors believe that the choice between TOF-MRA and CE-MRA for the demonstration of residual aneurysms should rely on the comparison of their diagnostic performance based on summary estimates of available reports. The purpose of this study is to systematically review the diagnostic accuracy of TOF-MRA and CE-MRA imaging techniques in comparison to reference DSA in the follow-up study of patients with residual intracranial aneurysms after coil treatment.

Materials and methods

Data Sources and search strategies

We conducted a literature search to identify articles published between January 1966 and December 2007 on studies that used MRA as diagnostic tests for residual aneurysm.

The search strategies in the PubMed and Cochrane Network databases were used with the following keywords:

1. ("intracranial aneurysm" [MeSH Terms] OR intracranial aneurysm [Text Word]) or ("intracranial aneurysm" [MeSH Terms] or intracranial aneurysm [Text Word]) and ("magnetic resonance angiography" [MeSH Terms] or magnetic resonance angiography [Text Word]) and coil [Text Word] and ("sensitivity and specificity" [TIAB] NOT Medline [SB]) or "sensitivity and specificity" [MeSH Terms] or sensitivity [Text Word])

2. ("intracranial aneurysm" [MeSH Terms] or intracranial aneurysm [Text Word]) or ("intracranial aneurysm" [MeSH Terms] or intracranial aneurysm [Text Word]) and ("magnetic resonance angiography" [MeSH Terms] or magnetic resonance angiography [Text Word]) and (Guglielmi detachable coil [Text Word]) and ("sensitivity and specificity" [TIAB] not Medline [SB]) or "sensitivity and specificity" [MeSH Terms] or sensitivity [Text Word])

3. Search ("sensitivity and specificity" [All Fields] or "sensitivity and specificity/standards" [All Fields]) or "specificity" [All Fields]) or "screening" [All Fields]) or "false positive" [All Fields]) or "false negative" [All Fields]) or "accuracy" [All Fields]) or ("predictive value" [All Fields] or "predictive value of tests" [All Fields]) or "predictive value of tests/standards" [All Fields]) or "predictive values" [All Fields]) or "predictive values of tests" [All Fields]) or ("reference value" [All Fields] or "reference values"" [All Fields]) or "reference values/standards" [All Fields]) or ("roc" [All Fields] or "roc analyses" [All Fields]) or "roc analysis" [All Fields]) or "roc and" [All Fields]) or "roc area" [All Fields]) or "roc auc" [All Fields]) or "roc characteristics" [All Fields]) or "roc curve" [All Fields]) or "roc curve method" [All Fields]) or "roccurves" [All Fields]) or "roc estimated" [All Fields]) or "roc evaluation" [All Fields]) or "likelihood ratio" [All Fields]) and ("intracranial aneurysm" [MeSH Terms] or intracranial aneurysm [Text Word]) and (Guglielmi detachable coil [Text Word]) and magnetic resonance

4. Search Sensitivity or specificity or intracranial aneurysm

5. Search Magnetic Resonance

6. Search 4 and 5

7. Search 6 and "Guglielmi detachable coil"

8. Search 7 or 2 or 1.

We exhaustively search using MEDLINE, bibliographies, review articles and textbooks to retrieve English-and non-English-language published articles. The title, abstract and full-text of all retrieved articles were evaluated to determine the eligibility for inclusion. The reference lists of the eligible articles were verified to recognize other relevant articles. Only data that contain the numbers of true-positive (TP), false positive (FP), false-negative (FN), and true-negative (TN) were eligible for inclusion. Diagnostic studies were included for evaluation followed by imaging examination with DSA and MRA (TOF or CE) in patients with intracranial aneurysm after coiling.

Study Selection

Sixteen suitable MRA primary articles (14 TOF-MRA and 6 CE-MRA) have met the inclusion criteria. We reviewed each article to extract the relevant study characteristics, quality, diagnostic accuracy and results.

Data extraction

The numbers of TP, FP, FN, and TN results in the detection of residual intracranial aneurysm were extracted. The authors defined residual aneurysm as the presence of residual neck and residual flow in the aneurysm. We chose the highest estimate to represent the best diagnostic performance of each article when there was more than one pair of sensitivity and specificity estimates due to their evaluation by more than one reviewer.

In addition, the following items were extracted:

(1) year of publication;

(2) place of journal publication (US. vs. Non-US.);

(3) study design (prospective, retrospective and unknown);

(4) description of patient population;

(5) MRI technique: magnetic field, type of sequences, and use of intravenous contrast medium (gadolinium);

(6) patient selection (consecutive or nonconsecutive);

(7) description of reference test (angiography) and the diagnostic test (MRI);

(8) interpretation of results (blinded or not blinded);

Statistical Analysis

For each study, sensitivities, specificities, and diagnostic odds ratios-with relevant 95% confidence intervals (CIs)-were recalculated from the true-positive, true-negative, false-positive, and false-negative results in the contingency tables. Because some studies reported a sensitivity or specificity of 100%, 0.5 was added to each cell of the 2-by-2 tables (number of true positives, false positives, true negatives, and false negatives) for all of the studies 10. The diagnostic odds ratio (DOR) is a simple statistic to express the discriminative power of a test and is calculated by the ratio of sensitivity/(1-sensitivity) over (1-specificity)/specificity. We applied the bivariate analysis of sensitivity and specificity using mixed model and calculated the diagnostic test data assuming binomial errors distribution11. Between-study variability was assessed assuming correlated normally distributed random effects for log-transformed sensitivity (logit-sensitivity) and log-transformed specificity (logit-specificity) with the degree of correlation between studies predictive of an implicit threshold effect11. In this method, the sensitivity and false positivity are transformed into its logits, defined as the natural log of the positivity rate/(1-positivity rate). A linear regression was then performed using the difference between the logits of the true positives and false positives as the dependent variable and the sum of the logits of the true positive and false positives as the independent variable. The y-intercept is a measure of the diagnostic odds ratio (DOR), and the slope is a measure of how the odds ratio is dependent on the threshold 12,13. We derived pooled sensitivity and specificity as functions of the estimated model parameters with associated 95% CIs. The comparison of significant difference between the pooled sensitivity, specificity and DOR were calculated by Reitsma's mixed model method 11. Summary receiver-operating characteristic (SROC) curves and their 95% CIs were constructed from the sensitivities and false positivities (1-specificities) of the studies, accounting for random thresholds across studies 10,12. The accuracy of the tests by the area under the SROC curve was calculated using the method of Walter 13.

Our model assessed the independent effect of each covariate on diagnostic discrimination as assessed by the DOR, the ratio of the positive and negative likelihood ratios. The DOR is a summary measure of the ability of the diagnostic test to discriminate between diseased and nondiseased individuals and reflects the trade-off between sensitivity and specificity. It is closely related to the ROC curve and the area under the curve (AUC). A DOR of 1 is equivalent to an AUC of 50%. The bigger the DOR, the bigger the AUC, and as the AUC grows to 100%, the DOR grows to infinity. It is an accepted and commonly used single measure of diagnostic accuracy in the comparison of diagnostic tests 14,15.

We conducted a visual examination of the funnel plots of sensitivity and specificity of studies to investigate the potential for publication bias. The overall heterogeneity of the results between the studies was assessed graphically by forest plots and statistically using the Q and I2 values. The Q value is Cochran's heterogeneity statistic under the degrees of freedom (df). The I2 value could quantify the proportion of total variation in the study estimates due to heterogeneity. A value of 0% indicates no observed heterogeneity, and values greater than 50% may be considered substantial heterogeneity. Formal investigation of heterogeneity is performed by multiple univariable bivariate meta-regression models.

Figure 1.

Figure 1

Open in a new tab

Forrest plots for (A) TOF-MRA and (B) CE-MRA illustrate the sensitivity results, with 95% CIs and summary results of the meta-analysis. The 95% CIs are indicated by the horizontal lines. The vertical broken line represents pooled sensitivity values, and the distortion of the diamond represents 95% CIs of the pooled results.

Figure 2.

Figure 2

Open in a new tab

Forrest plots for (A) TOF-MRA and (B) CE-MRA illustrate the specificity results, with 95% CIs and summary results of the meta-analysis. The 95% CIs are indicated by the horizontal lines. The vertical broken line represents pooled specificity values, and the distortion of the diamond represents 95% CIs of the pooled results.

Figure 3.

Figure 3

Open in a new tab

Forrest plots for (A) TOF-MRA and (B) CE-MRA illustrate the diagnostic odd ratio (DOR) results, with 95% CIs and summary results of the meta-analysis. The 95% CIs are indicated by the horizontal lines. The vertical broken line represents pooled DOR values, and the distortion of the diamond represents 95% CIs of the pooled results.

Figure 4.

Figure 4

Open in a new tab

Graph of summary receiver operating characteristic (SROC) curves for (A) time-of-flight MR angiography (TOF-MRA) and (B) contrast-enhancement MR angiography (CE-MRA) compared with the reference standard, intraarterial digital subtraction angiography (DSA). Both curves are concentrated in the upper left corner of the ROC space, with a large area under the curve. This indicates that both examinations performed well. The CE-MRA curve is marginally better than the TOF-MRA curve, but the difference is not significant. The small cross signs mean observed data. The diamond sign stands for the summary operating point.

The analyses and graphic illustration were performed by using STATA, version 9.2 (Stata-Corp., College Station, Tex., USA.) by means of the midas9, Metan, Metareg, Metainf, and Metabias programs. The Proc Mixed procedure in SAS version 9.1 for Windows (SAS Institute Inc., Cary, NC, USA.) was used to compare all bivariate models. We used p < .05 to indicate a significant difference.

Results

Overview of Studies

We identified 16 articles consisting of 14 and six studies for TOF-MRA and CE-MRA, respectively. There are four articles involving the comparison of diagnostic accuracy of both TOF-MRA and CE-MRA in the same paper. We extracted the information of MRA techniques, the year of publication, patient characteristics, design characteristics (retrospective or prospective), reference standard, and results from the 16 included articles. The characteristics are summarized in the Table 1.

Table 1.

Summary of studies evaluating TOF-MRA and CE-MRA for detecting residual cerebral aneurysm.

Study Year Journal Study Design Number
of
aneurysms		 Description
of patient
age & sex
distribution		 Magnetic
field
(Tesla)		 Blinded
analysis		 Consecu-
tive
TOF-MRA
Derdeyn20 1997 AJNR Retrospective 25 No 1.5T No  No 
Brunereau21 1999 JCAT Prospective 27 No  1T Yes No 
Kahara22 1999 AJNR Retrospective 20 Yes 1T Yes Yes
Anzalone16 2000 AJNR Retrospective 57 Yes 1.5T No  No 
Michardiere23 2001 J Neuroradiol Retrospective 25 Yes 1.5T No  No 
Weber24 2001 Eur Radiol Retrospective 56 Yes 1T No  No 
Nome25 2002 Acta Radiol Retrospective 79 No  1T No No 
Leclerc18 2002 AJNR Prospective 20 Yes 1.5T Yes No 
Cottier26 2003 Neuroradiology Prospective 73 Yes 1.5T Yes Yes
Yamada27 2004 AJNR Retrospective 51 No  1.5T Yes No 
Westerlaan28 2005 Neuroradiology Retrospective 31 Yes 1.5T No No 
Majoie29 2005 AJNR Prospective 21 Yes 3T Yes Yes
Farb17 2005 Neuroradiology Retrospective 31 Yes 1.5T Yes No 
Pierot9 2006 AJNR Prospective 40 Yes 1.5T Yes No 
CE-MRA
Anzalone 16 2000 AJNR Retrospective 57 Yes 1.5T No No 
Boulin30 2001 Radiology Prospective 80 Yes 1.5T Yes Yes
Leclerc 18 2002 AJNR Prospective 20 Yes 1.5T Yes No 
Farb17 2005 Neuroradiology Retrospective 36 Yes 1.5T Yes No 
Gauvrit31 2006 Stroke Retrospective 101 Yes 1.5T Yes Yes
Pierot9 2006 AJNR Prospective 42 Yes 1.5T Yes No 
Abbreviations: AJNR: American Journal of Neuroradiology; JCAT: Journal of Computer Assisted Tomography;
J Neuroradiol: Journal of Neuroradiology; Eur Radiol: European Radiology; Acta Radiol: Acta Radiologica; NA: Not available.
Open in a new tab

Pooling Sensitivities, Specificities, DORs and SROC curves and their diagnostic performance comparison

For the 14 TOF-MRA studies, the pooled sensitivity was 90% (95% CI, 79% to 95%), the pooled specificity, 95% (95% CI, 88% to 98%), and the pooled DOR, 168.4 (95% CI, 60.3 to 470.3). The 6 CE-MRA studies had an overall sensitivity of 92% (95% CI, 79% to 97%), a specificity of 96% (95% CI, 91% to 98%), and a DOR of 280.4 (95% CI, 64.8 to 1212.6). The areas under two summary ROC curves of TOF-MRA and CE-MRA were 0.97 (95% CI, 0.96 to 0.99) and 0.98 (95% CI, 0.96 to 0.99), respectively. No statistical difference was detected between TOF-MRA and CE-MRA among their summary sensitivities, specificities, DORs and areas under SROC curves. The results are shown in the Table 2.

Table 2.

Various pooled diagnostic metrics.

Imaging methods
Diagnostic metrics TOF-MRA (95% CI) CE-MRA (95% CI) p value
Sensitivity 90% (79% - 95%) 92% (79% - 97%) 0.545
Specificity 95% (88% - 98%) 96% (91% - 98%) 0.394
DOR 168.4 (60.3 - 470.3) 280.4 (64.8 - 1212.6) 0.240
Area under SROC curve 0.97 (0.96 - 0.99) 0.98 (0.96 - 0.99) 0.69
Abbreviations: TOF-MRA: time-of-flight MR angiography; CE-MRA: contrast-enhancement MR angiography;
95% CI: 95% confidence interval; DOR: diagnostic odds ratio; SROC: summary receiver operating characteristic.
Open in a new tab

Sources of Heterogeneity

There was high heterogeneity in the study-specific sensitivity for two MRA methods. Q, p, and I2 (95%CI) were, respectively, 42.62, <0.01, and 69.50 (95%CI, 52.73 to 96.27) for TOF-MRA. Q,p, and I2 (95%CI) for CE-MRA were, respectively, 19.85, <0.01, and 74.82 (95%CI, 54.20 to 95.43). Heterogeneity was present for between-study specificities for TOF-MRA, Q, p, and I2 (95%CI) were, respectively, 74.41, <0.01, and 82.53 (95% CI,74.23 to 90.83).There was no statistical evidence of heterogeneity for between-study specificities for CE-MRA, Q, p, and I2 (95% CI) were, respectively, 6.17, 0.29, and 18.99 (95% CI, 0.00 to 84.12).There was also high heterogeneity in the study-specific DOR for two methods. Q, p, and I2 (95% CI) were, respectively, 1195.95, <0.01, and 98.91 (95% CI, 98.69 to 99.13) for TOF-MRA; and 69.06, <0.01, and 92.76 (95% CI, 88.50 to 97.02) for CE-MRA.

Meta-regression analysis failed to show any evidence of heterogeneity related to the potential confounders on comparison of studies by means of dichotomization of the covariates into prospective versus retrospective (study design), patient selection, and US. versus non-US. (continent of study origin). The overall sensitivities and specificities were shown not to be different between the covariate terms.

Discussion

Despite being an invasive modality, digital subtraction angiography (DSA) is currently considered the most accurate diagnostic test for detection of intracranial aneurysm. On the other hand, MRA has been proposed as a non-invasive alternative to cerebral angiography. For patients with intracranial aneurysm after endovascular occlusion with GDCs, our analysis shows that the advantage of CE-MRA over TOF-MRA lies in the slight increase in pooled sensitivities, specificities, and DORs. However, no significant difference was noted in the pooled sensitivities, specificities, DORs and areas under ROC curves of TOF-MRA and CE-MRA for patients with cerebral aneurysm. Both TOF-MRA and CE-MRA were highly sensitive and specific, with a pooled sensitivity of 90% and 92% and an overall specificity of 95% and 96% respectively for the diagnosis of residual aneurysm. The DOR of CE-MRA is 280.4, higher than the ratio of TOF-MRA (168.4). The area under SROC curves for TOF-MRA and CE-MRA were also rather high (0.97 and 0.98, respectively).

However, a statistical pooling of the sensitivity and specificity of a test may not be reasonable12,14. Based on a neuroradiologist's point of view, the actual sensitivity and specificity represent a tradeoff or an exchange in the criteria for distinguishing an aneurysm from an artifact. By diminishing the criteria, a test will become more sensitive but less specific. Irwig and colleagues 12,14, also pointed out that a pooled estimate of sensitivities and specificities may give an inaccurate assessment of the diagnostic accuracy. The authors decided to use both statistical pooling and an SROC approach to evaluate the accuracy of both TOF and CE-MRA. Our analysis revealed a high sensitivity and specificity for detecting patients with intracranial aneurysm after coiling. The DOR application in diagnostic tests is constant, regardless of the diagnostic threshold. SROC analysis also suggested that TOF-MRA and CE-MRA seem to be highly precise with similar discriminatory diagnostic power. Therefore, DORs and SROC curves are highly advantageous in systematic reviews and meta-analyses.

Although some studies have shown that TOF-MRA is a good imaging sequence to follow up the condition of cerebral aneurysm in patients after coiling, others have suggested that subsequent evaluation using CE-MRA may be more reliable. Anzalone and colleagues stated that CE-MRA has advantages over TOF-MRA in evaluating residual patency in large and giant aneurysms and in discerning the distal branch arteries 16. Farb and colleagues also showed diagnostic superiority of CE-MRA over TOF-MRA17. CE-MRA has a higher sensitivity but slightly lower specificity than TOF- MRA18, while CE-MRA was not superior to TOF-MRA for discerning the residual neck or residual aneurysm, even though CE-MRA having several advantages over TOF-MRA, such as short echo time, fewer artifacts and better visualization of recanalization9. Our meta-analysis was able to prove the superior but equal discriminable diagnostic power of both TOF-MRA and CE-MRA. Given that TOF-MRA and CE-MRA being highly accurate for identifying residual cerebral aneurysm, it is reasonable to evaluate their clinical significance. From our analysis, the clinical efficacy of both TOF-MRA and CE-MRA is noticeable no matter if the test is positive or negative.

The major limitation of this study is interpretation of the data. There was significant heterogeneity in almost all diagnostic performance indexes. Thus the results and clinical applications should be interpreted with caution. Not all of the included studies provided real data on the patient number, so we had to retrieve the figures through sensitivities and specificities. We also recognize that our search was limited by the fact that we did not include gray literature such as conference proceedings. By including only published data, publication bias was evident, which tends to cause overestimation of diagnostic performance because of the greater likelihood of publication of positive rather than negative results 19. The design of the studies included in our meta-analysis does have a bias in favor of DSA because they used DSA as the "gold standard".

In summary, these results suggest that CE-MRA has a slightly better discriminatory power compared with TOF-MRA in diagnosing residual intracranial aneurysm. Both MRA techniques were highly sensitive and specific tests compared with DSA in the evaluation of residual aneurysm. Therefore, both TOF-and CE-MRA are comparable with DSA in terms of accuracy and can be considered a reliable substitute for DS angiography in discerning residual neck and residual aneurysm after endovascular GDC treatment.

Acknowledgments

Supported in part by Grants No. CMRPG660291 and CM-RPG660351 from the Chang Gung Memorial Hospital, Chiayi, Taiwan.

References

  • 1.van Gijn J, Kerr RS, Rinkel GJ. Subarachnoid haemorrhage. Lancet. 2007;369:306–318. doi: 10.1016/S0140-6736(07)60153-6. [DOI] [PubMed] [Google Scholar]
  • 2.de Rooij NK, Linn FH, et al. Incidence of subarachnoid haemorrhage: a systematic review with emphasis on region, age, gender and time trends. J Neurol Neurosurg Psychiatry. 2007;78:1365–1372. doi: 10.1136/jnnp.2007.117655. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.White PM, Teasdale EM, et al. Intracranial aneurysms: CT angiography and MR angiography for detection prospective blinded comparison in a large patient cohort. Radiology. 2001;219:739–749. doi: 10.1148/radiology.219.3.r01ma16739. [DOI] [PubMed] [Google Scholar]
  • 4.Westerlaan HE, Gravendeel J, et al. Multislice CT angiography in the selection of patients with ruptured intracranial aneurysms suitable for clipping or coiling. Neuroradiology. 2007;49:997–1007. doi: 10.1007/s00234-007-0293-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Westerlaan HE, van der Vliet AM, et al. Magnetic resonance angiography in the selection of patients suitable for neurosurgical intervention of ruptured intracranial aneurysms. Neuroradiology. 2004;46:867–875. doi: 10.1007/s00234-004-1260-9. [DOI] [PubMed] [Google Scholar]
  • 6.White PM, Wardlaw JM, Easton V. Can noninvasive imaging accurately depict intracranial aneurysms? A systematic review. Radiology. 2000;217:361–370. doi: 10.1148/radiology.217.2.r00nv06361. [DOI] [PubMed] [Google Scholar]
  • 7.Brisman JL, Song JK, Newell DW. Cerebral aneurysms. N Engl J Med. 2006;355:928–939. doi: 10.1056/NEJMra052760. [DOI] [PubMed] [Google Scholar]
  • 8.Wallace RC, Karis JP, et al. Noninvasive imaging of treated cerebral aneurysms, part I: MR angiographic follow-up of coiled aneurysms. Am J Neuroradiol. 2007;28:1001–1008. doi: 10.3174/ajnr.A0662. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Pierot L, Delcourt C, et al. Follow-up of intracranial aneurysms selectively treated with coils: Prospective evaluation of contrast-enhanced MR angiography. Am J Neuroradiol. 2006;27:744–749. [PMC free article] [PubMed] [Google Scholar]
  • 10.Moses LE, Shapiro D, Littenberg B. Combining independent studies of a diagnostic test into a summary ROC curve: data-analytic approaches and some additional considerations. Stat Med. 1993;12:1293–1316. doi: 10.1002/sim.4780121403. [DOI] [PubMed] [Google Scholar]
  • 11.Reitsma JB, Glas AS, et al. Bivariate analysis of sensitivity and specificity produces informative summary measures in diagnostic reviews. J Clin Epidemiol. 2005;58:982–990. doi: 10.1016/j.jclinepi.2005.02.022. [DOI] [PubMed] [Google Scholar]
  • 12.Irwig L, Tosteson AN, et al. Guidelines for meta-analyses evaluating diagnostic tests. Ann Intern Med. 1994;120:667–676. doi: 10.7326/0003-4819-120-8-199404150-00008. [DOI] [PubMed] [Google Scholar]
  • 13.Walter SD. Properties of the summary receiver operating characteristic (SROC) curve for diagnostic test data. Stat Med. 2002;21:1237–1256. doi: 10.1002/sim.1099. [DOI] [PubMed] [Google Scholar]
  • 14.Irwig L, Macaskill P, et al. Meta-analytic methods for diagnostic test accuracy. J Clin Epidemiol. 1995;48:119–130. doi: 10.1016/0895-4356(94)00099-c. discussion 131-112. [DOI] [PubMed] [Google Scholar]
  • 15.Lijmer JG, Bossuyt PM, Heisterkamp SH. Exploring sources of heterogeneity in systematic reviews of diagnostic tests. Stat Med. 2002;21:1525–1537. doi: 10.1002/sim.1185. [DOI] [PubMed] [Google Scholar]
  • 16.Anzalone N, Righi C, et al. Three-dimensional time-of-flight MR angiography in the evaluation of intracranial aneurysms treated with Guglielmi detachable coils. Am J Neuroradiol. 2000;21:746–752. [PMC free article] [PubMed] [Google Scholar]
  • 17.Farb RI, Nag S, et al. Surveillance of intracranial aneurysms treated with detachable coils: a comparison of MRA techniques. Neuroradiology. 2005;47:507–515. doi: 10.1007/s00234-005-1375-7. [DOI] [PubMed] [Google Scholar]
  • 18.Leclerc X, Navez JF, et al. Aneurysms of the anterior communicating artery treated with Guglielmi detachable coils: follow-up with contrast-enhanced MR angiography. Am J Neuroradiol. 2002;23:1121–1127. [PMC free article] [PubMed] [Google Scholar]
  • 19.Rothstein H, Sutton AJ, Borenstein M. Publication Bias in Meta-Analysis: Prevention, Assessment and Adjustments. Hoboken, NJ, USA.: John Wiley & Sons Inc; 2005. [Google Scholar]
  • 20.Derdeyn CP, Graves VB, et al. MR angiography of saccular aneurysms after treatment with Guglielmi detachable coils: preliminary experience. Am J Neuroradiol. 1997;18:279–286. [PMC free article] [PubMed] [Google Scholar]
  • 21.Brunereau L, Cottier JP, et al. Prospective evaluation of time-of-flight MR angiography in the follow-up of intracranial saccular aneurysms treated with Guglielmi detachable coils. J Comput Assist Tomogr. 1999;23:216–223. doi: 10.1097/00004728-199903000-00009. [DOI] [PubMed] [Google Scholar]
  • 22.Kahara VJ, Seppanen SK, et al. MR angiography with three-dimensional time-of-flight and targeted maximum-intensity-projection reconstructions in the follow-up of intracranial aneurysms embolized with Guglielmi detachable coils. Am J Neuroradiol. 1999;20:1470–1475. [PMC free article] [PubMed] [Google Scholar]
  • 23.Michardiere R, Bensalem D, et al. Comparison of MRA and angiography in the follow-up of intracranial aneurysms treated with GDC. J Neuroradiol. 2001;28:75–83. [PubMed] [Google Scholar]
  • 24.Weber W, Yousry TA, et al. Noninvasive follow-up of GDC-treated saccular aneurysms by MR angiography. Eur Radiol. 2001;11:1792–1797. doi: 10.1007/s003300000741. [DOI] [PubMed] [Google Scholar]
  • 25.Nome T, Bakke SJ, Nakstad PH. MR angiography in the follow-up of coiled cerebral aneurysms after treatment with Guglielmi detachable coils. Acta Radiol. 2002;43:10–14. doi: 10.1080/028418502127347574. [DOI] [PubMed] [Google Scholar]
  • 26.Cottier JP, Bleuzen-Couthon A, et al. Follow-up of intracranial aneurysms treated with detachable coils: comparison of plain radiographs, 3D time-of-flight MRA and digital subtraction angiography. Neuroradiology. 2003;45:818–824. doi: 10.1007/s00234-003-1109-7. [DOI] [PubMed] [Google Scholar]
  • 27.Yamada N, Hayashi K, et al. Time-of-flight MR angiography targeted to coiled intracranial aneurysms is more sensitive to residual flow than is digital subtraction angiography. Am J Neuroradiol. 2004;25:1154–1157. [PMC free article] [PubMed] [Google Scholar]
  • 28.Westerlaan HE, van der Vliet AM, et al. Time-of-flight magnetic resonance angiography in the follow-up of intracranial aneurysms treated with Guglielmi detachable coils. Neuroradiology. 2005;47:622–629. doi: 10.1007/s00234-005-1395-3. [DOI] [PubMed] [Google Scholar]
  • 29.Majoie CB, Sprengers ME, et al. MR angiography at 3T versus digital subtraction angiography in the follow-up of intracranial aneurysms treated with detachable coils. Am J Neuroradiol. 2005;26:1349–1356. [PMC free article] [PubMed] [Google Scholar]
  • 30.Boulin A, Pierot L. Follow-up of intracranial aneurysms treated with detachable coils: comparison of gadolinium-enhanced 3D time-of-flight MR angiography and digital subtraction angiography. Radiology. 2001;219:108–113. doi: 10.1148/radiology.219.1.r01mr06108. [DOI] [PubMed] [Google Scholar]
  • 31.Gauvrit JY, Leclerc X, et al. Intracranial aneurysms treated with Guglielmi detachable coils: imaging follow-up with contrast-enhanced MR angiography. Stroke. 2006;37:1033–1037. doi: 10.1161/01.STR.0000209236.06451.3b. [DOI] [PubMed] [Google Scholar]

Articles from Interventional Neuroradiology are provided here courtesy of SAGE Publications

RESOURCES