Blood-Brain Barrier Permeability in Aneurysmal Subarachnoid Hemorrhage: Correlation With Clinical Outcomes.

OBJECTIVE: The current understanding of the utility of blood-brain barrier permeability (BBBP) evaluation with extended-pass perfusion CT in the clinical setting is limited. We assessed whether BBBP parameters evaluated with perfusion CT correlate with unfavorable clinical outcomes in patients with aneurysmal subarachnoid hemorrhage (SAH).

MATERIALS AND METHODS: A retrospective cross-sectional analysis was performed of 22 patients who underwent perfusion CT on days 0-3 after SAH. Extended perfusion CT data were postprocessed into BBBP quantitative maps of kep (washout rate constant of contrast agent from the extravascular extracellular space [EES] to the intravascular space), PS (permeability surface area product), Ktrans (volume transfer constant from blood plasma to EES), and Ve (EES volume per unit tissue volume) using Olea Sphere software (version 2.0). Clinical outcomes data regarding the presence of permanent neurologic deficits and modified Rankin scores were collected at discharge. ROC curve analyses and unpaired t tests were performed.

RESULTS: The 22 patients were stratified on the basis of their clinical outcomes of permanent neurologic deficit and modified Rankin scores. In patients with unfavorable clinical outcomes (i.e., patients with permanent neurologic deficit and those with modified Rankin scores of 3-6), PS and Ve were significantly increased and kep and Ktrans were significantly decreased. A multiparameter ROC curve analysis combining the four parameters yielded an AUC of 0.80 for permanent neurologic deficit and an AUC of 0.89 for modified Rankin scores of 3-6.

CONCLUSION: SAH patients with unfavorable outcomes had significantly elevated PS and Ve and significantly decreased kep , indicating increased BBBP, compared with SAH patients with unfavorable outcomes. Evaluation of BBBP parameters may provide prognostication of unfavorable outcomes in patients with SAH and help guide management.