Background and purpose: The presence of low-density areas on CT is used in clinical decision-making regarding treatment of angiographic vasospasm as well as in research as a surrogate marker for severity of angiographic vasospasm. We assess the interobserver variability in attributing hypodensities on CT to angiographic vasospasm-related delayed ischemic neurological deficit.
Methods: Three experienced reviewers, 2 neurosurgeons, and a neuroradiologist independently reviewed CT scans of 413 patients enrolled in the Clazosentan to Overcome Neurological iSChemia and Infarction OccUrring after Subarachnoid hemorrhage (CONSCIOUS-1) trial, who universally underwent catheter angiography to determine severity of angiographic vasospasm. Interobserver variability was calculated using the κ statistic and the χ(2) test was used to determine associations between dichotomized outcomes.
Results: There was considerable interobserver variability in attributing CT hypodensities to vasospasm-related delayed ischemic neurological deficit (κ=0.51-0.78; 95% CI, 0.35-0.90). Patients with hypodensities attributed to delayed ischemic neurological deficit were significantly more likely to have severe angiographic vasospasm (P=0.001), but a substantial proportion of these patients (19%) also had mild or no spasm. CT hypodensities had a sensitivity and specificity of 41% and 93%, respectively, in identifying patients with severe angiographic vasospasm, even with expert consensus that these represent angiographic vasospasm-related delayed ischemic neurological deficit.
Conclusions: We find considerable interobserver variability in attributing CT hypodensities to angiographic vasospasm and propose that they may not be a robust marker of severity of angiographic vasospasm, even with unanimous expert agreement that they are a result of vasospasm-related delayed ischemic neurological deficit.
Clinical trial registration: URL: www.clinicaltrials.gov. Unique identifier: NCT00111085.