Accuracy of on-call resident interpretation of CT angiography for intracranial aneurysm in subarachnoid hemorrhage.

Objective: The purpose of this article is to evaluate the accuracy of preliminary on-call radiology resident interpretation of CT angiography (CTA) compared with digital subtraction angiography (DSA) in detecting cerebral aneurysms in subarachnoid hemorrhage (SAH).
Materials and Methods: A retrospective review compared resident interpretations of head CTA performed after hours for SAH to the results of DSA. The sensitivity and specificity of resident interpretations were classified on a per-patient and per-aneurysm basis. The accuracy of resident interpretations was also determined according to aneurysm location and number.
Results: Between January 2007 and December 2009, 83 patients with SAH underwent both CTA and DSA. DSA documented an aneurysm in 53 of 83 patients. Per patient, residents identified at least one aneurysm in 46 of 53 patients (87%). Per aneurysm, resident sensitivity and specificity for detecting aneurysms of any size were 62% and 91%, respectively, which improved for aneurysms 3 mm or larger to 73% and 97%, respectively. The posterior communicating and intracranial internal carotid arteries were resident "blind spots," with aneurysms 3 mm or larger detected with sensitivities of 33% and 50%, respectively. In contrast, anterior communicating artery aneurysms were correctly identified 95% of the time. In only 35% of cases with multiple aneurysms did residents correctly identify more than one aneurysm.
Conclusion: The sensitivity of on-call resident interpretation of CTA for aneurysms in SAH is lower than expected, with a potential for delay in diagnosis and management in a small number of patients. Focused training to carefully review apparent blind spots and the frequency of multiple aneurysms may reduce inaccuracies.