Forty-five studies (3,643 participants) were included. Studies were considered to be of high methodological quality: 43 scored at least 11 points on the QUADAS assessment.
Based on patient level data, CT angiography had a sensitivity of 97% (95% CI 96% to 98%) and a specificity of 98% (95% CI 96% to 99%) for the diagnosis of cerebral aneurysm. Results on a per-lesion basis were slightly lower. Heterogeneity was moderate (I2=33% to 76%). The accuracy of multidetector CT angiography was significantly higher than that of single-detector CT, especially in detecting small aneurysms (<4mm diameter). Summary sensitivity was 99% (95% CI 97% to 100%) and specificity was 99% (95% CI 93% to 100%) for 64-row CT (eight studies). Data for other types of scanner were reported.
CT images with effective slice thickness of less than 1mm was significantly more accurate than thicker slices, sensitivity was greater with high intravenous iodine flow rate (p=0.010) and specificity was higher in studies restricted to patients with subarachnoid haemorrhage (p=0.016). There was no association between study quality score or other features assessed and sensitivity or specificity.
There was no evidence of publication bias (p=0.71).