How reliable is CT angiography in the etiologic workup of intracranial hemorrhage? A single surgeon's experience

Abstract Objectives CT angiography (CTA) is gradually replacing digital subtraction angiography (DSA) in the etiologic workup of intracranial hemorrhage, though debate persists regarding its diagnostic accuracy. To better define the role of CTA in a pragmatic, real-world setting, we reviewed the experience of a single dual vascular-endovascular neurosurgeon. Patients and methods Nontraumatic intracranial hemorrhage cases managed by the senior author over a 15-month period were retrieved from a prospectively maintained database. Cases where a hypertensive etiology was presumed were excluded. Demographics, intracranial hemorrhage pattern, CTA and DSA findings were recorded. Results The study cohort consists of 59 cases where both CTA and DSA were obtained, including 32 women and 27 men with mean age 50 years (18–83). Intracranial hemorrhage pattern was: aneurysmal subarachnoid hemorrhage (SAH) in 37, perimesencephalic SAH (PMSAH) in 8, intraparenchymal in 11, intraventricular in 2, subdural in 1. The overall yield of vascular imaging was 62.7 % (37/59): 29 saccular aneurysms, 4 dissecting aneurysms, 4 microarteriovenous malformations (microAVMs). The specificity and positive predictive value of CTA were 100 %. Its sensitivity and negative predictive value were only 89.2 % and 84.6 %, respectively. CTA missed 4 lesions: 2 dissecting aneurysms, 1 microAVM, 1 small saccular aneurysm. Of 8 patients with PMSAH, 3 (37.5 %) had a vascular lesion: 1 vertebrobasilar dissection, 1 cerebellar microAVM, 1 basilar tip aneurysm. Of those, 2 were missed by CTA. Conclusions DSA may identify a lesion in up to 15 % of intracranial hemorrhage cases with negative CTA. Excessive reliance on CTA can be potentially hazardous, especially in the setting of PMSAH.