Objectives: Oxygen extraction fraction (OEF) and cerebrovascular reserve (CVR) are both proven predictors of stroke risk in symptomatic patients with carotid occlusion. Accordingly, hemispheric comparisons of CVR and OEF are significantly correlated. However, there was also substantial disagreement: hemispheres identified as compromised by CVR were normal by OEF. Our aim was to determine whether regional comparisons could resolve the CVR-OEF discordance. We also studied the relationship between white matter (WM) infarction and hemodynamic compromise.
Methods: Quantitative CVR and OEF were measured in 12 symptomatic patients with internal carotid artery occlusion. CVR and OEF comparisons were made in the anterior watershed (AWS), middle cerebral artery (MCA) and WM territories using various thresholds for hemodynamic compromise. Associations with WM infarction were also recorded.
Results: Comparison of CVR and OEF for the AWS and MCA showed high sensitivity (100%) with specificities of 83 and 40%, respectively. There was also agreement (k=Cohen's Kappa) for the AWS (k=0.83) and MCA (k=0.39) territories. CVR-OEF discordance was reduced with regional analysis. Hemodynamic compromise was more often found in patients with WM infarction.
Discussion: Regional comparison of CVR and OEF reduced the discordance compared with hemispheric analysis, especially for the AWS territory. Despite the persistence of some regions with compromised CVR and normal OEF, CVR is able to identify all regions with elevated OEF making it a useful screening technology. Future studies are needed to understand whether those remaining regions with compromised CVR are also at increased stroke risk despite normal OEF.