Diagnosing cerebral aneurysms by computed tomographic angiography: meta-analysis.

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 x 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 ≤ 4 mm 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.