Radiology: CTA highly accurate in detecting cerebral aneurysms
Multidetector CT angiography (CTA) stands as a quick, non-invasive and highly accurate procedure for diagnosing cerebral aneurysms in patients with acute subarachnoid hemorrhage (SAH), concluded a systematic review of 50 clinical studies published in the January issue of Radiology.
Although CTA of the intracranial vessels represents a component of the routine imaging and treatment algorithm in many facilities for patients presenting with SAH (usually caused by the rupture of an intracranial aneurysm), substantial debate persists as to the detail and overall value yielded by CTA, with many investigators arguing that cerebral angiography is still necessary, explained Henriëtte E. Westerlaan, MD, of the department of radiology at University Medical Center Gronningen, in Gronningen, the Netherlands, and co-authors.
The authors sought to contribute to the debate with a systematic review and meta-analysis of the sensitivity and specificity of CTA in the diagnosis of cerebral aneurysms. The authors identified 50 studies, including 4,097 patients, through a comprehensive search of MEDLINE and EMBASE of articles that assessed the value of CTA in patients with proven SAH published between 1997 and 2009. Study quality was assessed (for reference and variate analysis) using the Quality Assessment of Diagnostic Accuracy Studies (QUADAS) tool, with a minimum score of 0 and a maximum of 14.
Overall sensitivity for all studies was moderately heterogeneous and ranged from 86 percent to 100 percent, with a combined value of 98 percent. Low heterogeneity of specificity was observed, with a pooled value of 100 percent. The average QUADAS score for all included studies was 11.
A total of 71 ruptured aneurysms were missed using CTA, while 15 false-positives were recorded. Of the 40 missed aneurysms with corresponding size data, 19 were smaller than 5 mm and four were 5 to 10 mm.
Most of the reviewed studies (30 of 50) used a four-detector row CT scanner, while 11 were performed with a 16- or 64-detector row scanner. The four-detector and the 16- and 64 detector studies were each found to have a pooled sensitivity of 98 percent and specificity of 100 percent.
Westerlaan and colleagues argued that their results showed "that CT angiography has a very high diagnostic value for the detection of ruptured intracranial aneurysms." Citing previous studies, they also noted that reader experience and perceptual accuracy was an important factor in the detection and depiction of intracranial aneurysms on CTA.
The authors also concluded that because aneurysms could be retrospectively viewed for many of the false-negative studies, those interpretations could "be categorized as perceptual in nature and could have been substantially bypassed by double reading," a more common practice in many European settings, such as mammography. Calculating that double reading could reduce the false-negative ratio to 1 percent, Westerlaan and colleagues recommended "a second re-evaluation by an experienced (neuro) radiologist for every negative CT angiography result, thus minimizing the risk of missing a ruptured aneurysm and its risk of rebleeding and associated high mortality."
The authors identified publication bias as a threat to the validity of their meta-analysis, considering that smaller studies and studies with negative results are less likely to be accepted for publication. Given their statistical analyses that were "strongly suggestive" of publication bias, they acknowledged that their high sensitivity and specificity figures might have been overestimates.
Nevertheless, given these high values for CT and the relatively high risk (up to 2.3 percent) of complications associated with selective cerebral angiography, Westerlaan and colleagues concluded that their results "prove that CT angiography is an accurate tool in the diagnosis of ruptured intracranial aneurysms and can be integrated as a primary examination tool into the imaging and treatment algorithm for patients with SAH at presentation. The chance of missing a ruptured aneurysm at CT angiography is no more than 2 percent."
Although CTA of the intracranial vessels represents a component of the routine imaging and treatment algorithm in many facilities for patients presenting with SAH (usually caused by the rupture of an intracranial aneurysm), substantial debate persists as to the detail and overall value yielded by CTA, with many investigators arguing that cerebral angiography is still necessary, explained Henriëtte E. Westerlaan, MD, of the department of radiology at University Medical Center Gronningen, in Gronningen, the Netherlands, and co-authors.
The authors sought to contribute to the debate with a systematic review and meta-analysis of the sensitivity and specificity of CTA in the diagnosis of cerebral aneurysms. The authors identified 50 studies, including 4,097 patients, through a comprehensive search of MEDLINE and EMBASE of articles that assessed the value of CTA in patients with proven SAH published between 1997 and 2009. Study quality was assessed (for reference and variate analysis) using the Quality Assessment of Diagnostic Accuracy Studies (QUADAS) tool, with a minimum score of 0 and a maximum of 14.
Overall sensitivity for all studies was moderately heterogeneous and ranged from 86 percent to 100 percent, with a combined value of 98 percent. Low heterogeneity of specificity was observed, with a pooled value of 100 percent. The average QUADAS score for all included studies was 11.
A total of 71 ruptured aneurysms were missed using CTA, while 15 false-positives were recorded. Of the 40 missed aneurysms with corresponding size data, 19 were smaller than 5 mm and four were 5 to 10 mm.
Most of the reviewed studies (30 of 50) used a four-detector row CT scanner, while 11 were performed with a 16- or 64-detector row scanner. The four-detector and the 16- and 64 detector studies were each found to have a pooled sensitivity of 98 percent and specificity of 100 percent.
Westerlaan and colleagues argued that their results showed "that CT angiography has a very high diagnostic value for the detection of ruptured intracranial aneurysms." Citing previous studies, they also noted that reader experience and perceptual accuracy was an important factor in the detection and depiction of intracranial aneurysms on CTA.
The authors also concluded that because aneurysms could be retrospectively viewed for many of the false-negative studies, those interpretations could "be categorized as perceptual in nature and could have been substantially bypassed by double reading," a more common practice in many European settings, such as mammography. Calculating that double reading could reduce the false-negative ratio to 1 percent, Westerlaan and colleagues recommended "a second re-evaluation by an experienced (neuro) radiologist for every negative CT angiography result, thus minimizing the risk of missing a ruptured aneurysm and its risk of rebleeding and associated high mortality."
The authors identified publication bias as a threat to the validity of their meta-analysis, considering that smaller studies and studies with negative results are less likely to be accepted for publication. Given their statistical analyses that were "strongly suggestive" of publication bias, they acknowledged that their high sensitivity and specificity figures might have been overestimates.
Nevertheless, given these high values for CT and the relatively high risk (up to 2.3 percent) of complications associated with selective cerebral angiography, Westerlaan and colleagues concluded that their results "prove that CT angiography is an accurate tool in the diagnosis of ruptured intracranial aneurysms and can be integrated as a primary examination tool into the imaging and treatment algorithm for patients with SAH at presentation. The chance of missing a ruptured aneurysm at CT angiography is no more than 2 percent."