Automated grading of cerebral vasospasm to standardize computed tomography angiography examinations after subarachnoid hemorrhage

Abstract

Background: Computed tomography angiography (CTA) is frequently used with computed tomography perfusion imaging (CTP) to evaluate whether endovascular vasospasm treatment is indicated for subarachnoid hemorrhage patients with delayed cerebral ischemia. However, objective parameters for CTA evaluation are lacking. In this study, we used an automated, investigator-independent, digital method to detect vasospasm, and we evaluated whether the method could predict the need for subsequent endovascular vasospasm treatment.Methods: We retrospectively reviewed the charts and analyzed imaging data for 40 consecutive patients with subarachnoid hemorrhages. The cerebrovascular trees were digitally reconstructed from CTA data, and vessel volume and the length of the arteries of the circle of Willis and their peripheral branches were determined. Receiver operating characteristic curve analysis based on a comparison with digital subtraction angiographies was used to determine volumetric thresholds that indicated severe vasospasm for each vessel segment.

Results: The automated threshold-based volumetric evaluation of CTA data was able to detect severe vasospasm with high sensitivity and negative predictive value for predicting cerebral hypoperfusion on CTP, although the specificity and positive predictive value were low. Combining the automated detection of vasospasm on CTA and cerebral hypoperfusion on CTP was superior to CTP or CTA alone in predicting endovascular vasospasm treatment within 24 h after the examination.

Conclusions: This digital volumetric analysis of the cerebrovascular tree allowed the objective, investigator-independent detection and quantification of vasospasms. This method could be used to standardize diagnostics and the selection of subarachnoid hemorrhage patients with delayed cerebral ischemia for endovascular diagnostics and possible interventions.