CT angiography versus 3D rotational angiography in patients with subarachnoid hemorrhage

CT angiography (CTA) is increasingly used as primary diagnostic tool to replace digital subtraction angiography (DSA) in patients with subarachnoid hemorrhage (SAH). However, 3D rotational angiography (3DRA) has substituted DSA as a reference standard. In this prospective observational study, we compare CTA with 3DRA of all cerebral vessels in a large cohort of patients with SAH.

Of 179 consecutive patients with SAH admitted between March 2013 and July 2014, 139 underwent 64- to 256-detector row CTA followed by complete cerebral 3DRA within 24 h. In 86 patients (62 %), 3DRA was performed under general anesthesia. Two observers from outside hospitals reviewed CTA data.

In 118 of 139 patients (85 %), 3DRA diagnosed the cause of hemorrhage: 113 ruptured aneurysms, three arterial dissections, one micro-arteriovenous malformation (AVM), and one reversible vasoconstriction syndrome. On CTA, both observers missed all five non-aneurysmal causes of SAH. Sensitivity of CTA in depicting ruptured aneurysms was 0.88-0.91, and accuracy was 0.88-0.92. Of 113 ruptured aneurysms, 28 were ≤3 mm (25 %) and of 95 additional aneurysms, 71 were ≤3 mm (75 %). Sensitivity of depicting aneurysms ≤3 mm was 0.28-0.43. Of 95 additional aneurysms, the two raters missed 65 (68 %) and 58 (61 %). Sensitivity in detection was lower in aneurysms of the internal carotid artery than in other locations.

CTA had some limitations as primary diagnostic tool in patients with SAH. All non-aneurysmal causes for SAH and one in ten ruptured aneurysms were missed. Performance of CTA was poor in aneurysms ≤3 mm. The majority of additional aneurysms were not depicted on CTA.