Cerebral vasospasm diagnosis by means of angiography and blood velocity measurements

Publisher Website
Google Scholar
Abstract
We investigated 76 patients with known subarachnoid haemorrhage (SAH) in order to compare the results of angiography and non-invasive Doppler recordings of cerebral artery blood velocity in the diagnosis of cerebral vasospasm. One radiologist and one neurovascular surgeon assessed angiographic spasm visually on a four-level scale. The radiologist's ratings were the term of reference for the study. When there was angiographic spasm of the middle cerebral artery (MCA), the MCA blood velocity was higher and the blood velocity in the distal extracranical internal carotid artery (ICA) was lower than when MCA spasm was scored as absent. Analysis by Kappa statistics, a measure for the agreement between two independent judges with correction for random coincidence, revealed moderate agreement between angiographic spasm and the absolute MCA blood velocity (Kappa=0.47). However, there was substantial agreement (Kappa=0.64) between angiographic spasm and the index calculated from dividing the blood velocity in the MCA (VMCA) by the blood velocity in the ipsilaterial ICA (VICA). The results indicate that this VMCA/VICA index gives more appropriate information on MCA spasm. Congenitally asymmetric circles of Willis with one wide dominant ACA showed normal blood velocities. In asymmetry induced by vasospasm, the diameter of the major ACA was normal or even reduced, and the blood velocity was significantly elevated. Hence, the agreement between blood velocity recordings and angiographic findings was substantial (Kappa=0.64) when considering together the findings from both sides of the anterior circle of Willis. The agreement between angiographic spasm of the posterior cerebral artery (PCA) and the PCA blood velocity was substantial (Kappa=0.68). The two observers' rating of angiographic spasm agreed substantially on the MCA (Kappa=0.67) and the ACA (Kappa=0.71). Moderate agreement (Kappa=0.59) was obtained for the PCA. These findings show that assessing cerebral vasospasm angiographically is difficult. The present study therefore indicates that with regard to diagnosing cerebral vasospasm, the reliability of blood velocity measurements match the reliability of judging angiograms visually.