Computed tomography angiography or magnetic resonance angiography for detection of intracranial vascular malformations in patients with intracerebral haemorrhage

Colin B Josephson, Philip M White, Ashma Krishan, Rustam Al-Shahi Salman

Research output: Contribution to journalLiterature reviewpeer-review

Abstract / Description of output

BACKGROUND: Intracranial vascular malformations (brain or pial/dural arteriovenous malformations/fistulae, and aneurysms) are the leading cause of intracerebral haemorrhage (ICH) in young adults. Early identification of the intracranial vascular malformation may improve outcome if treatment can prevent ICH recurrence. Catheter intra-arterial digital subtraction angiography (IADSA) is considered the reference standard for the detection an intracranial vascular malformation as the cause of ICH. Computed tomography angiography (CTA) and magnetic resonance angiography (MRA) are less invasive than IADSA and may be as accurate for identifying some causes of ICH.

OBJECTIVES: To evaluate the diagnostic test accuracy of CTA and MRA versus IADSA for the detection of intracranial vascular malformations as a cause of ICH.

SEARCH METHODS: We searched MEDLINE (1948 to August 2013), EMBASE (1980 to August 2013), MEDION (August 2013), the Database of Abstracts of Reviews of Effects (DARE; August 2013), the Health Technology Assessment Database (HTA; August 2013), (August 2013), and WHO ICTRP (International Clinical Trials Register Portfolio; August 2013). We also performed a cited reference search for forward tracking of relevant articles on Google Scholar (, screened bibliographies, and contacted authors to identify additional studies.

SELECTION CRITERIA: We selected studies reporting data that could be used to construct contingency tables that compared CTA or MRA, or both, with IADSA in the same patients for the detection of intracranial vascular malformations following ICH.

DATA COLLECTION AND ANALYSIS: Two authors (CBJ and RA-SS) independently extracted data on study characteristics and measures of test accuracy. Two authors (CBJ and PMW) independently extracted data on test characteristics. We obtained data restricted to the subgroup undergoing IADSA in studies using multiple reference standards. We combined data using the bivariate model. We generated forest plots of the sensitivity and specificity of CTA and MRA and created a summary receiver operating characteristic plot.

MAIN RESULTS: Eleven studies (n = 927 participants) met our inclusion criteria. Eight studies compared CTA with IADSA (n = 526) and three studies compared MRA with IADSA (n = 401). Methodological quality varied considerably among studies, with partial verification bias in 7/11 (64%) and retrospective designs in 5/10 (50%). In studies of CTA, the pooled estimate of sensitivity was 0.95 (95% confidence interval (CI) 0.90 to 0.97) and specificity was 0.99 (95% CI 0.95 to 1.00). The results remained robust in a sensitivity analysis in which only studies evaluating adult patients (≥ 16 years of age) were included. In studies of MRA, the pooled estimate of sensitivity was 0.98 (95% CI 0.80 to 1.00) and specificity was 0.99 (95% CI 0.97 to 1.00). An indirect comparison of CTA and MRA using a bivariate model incorporating test type as one of the parameters failed to reveal a statistically significant difference in sensitivity or specificity between the two imaging modalities (P value = 0.6).

AUTHORS' CONCLUSIONS: CTA and MRA appear to have good sensitivity and specificity following ICH for the detection of intracranial vascular malformations, although several of the included studies had methodological shortcomings (retrospective designs and partial verification bias in particular) that may have increased apparent test accuracy.

Original languageEnglish
Pages (from-to)CD009372
JournalCochrane Database of Systematic Reviews
Publication statusPublished - 2014


Dive into the research topics of 'Computed tomography angiography or magnetic resonance angiography for detection of intracranial vascular malformations in patients with intracerebral haemorrhage'. Together they form a unique fingerprint.

Cite this