Abstract
OBJECTIVES: Cerebral vascular reactivity (CVR) is an independent predictor of poor prognosis for patients with symptomatic middle cerebral artery (MCA)/internal carotid artery (ICA) occlusion. This study assessed CVR impairment using resting-state blood oxygenation level-dependent (RS BOLD) MRI in this population and aimed to identify feasible pre-surgical markers. METHODS: The study included 45 patients with symptomatic unilateral MCA/ICA occlusion and 20 healthy controls. Of these, 16 underwent extracranial-intracranial (EC-IC) bypass with follow-up MRI. RS BOLD data were collected using a 3.0 T PET/MR system. Voxel-wise CVR maps were computed with a general linear model, and the mean CVR in the MCA perfusion territory was calculated. Interhemispheric CVR asymmetry index (CVR-AI) was assessed. Group differences were analyzed, and ROC analysis was used to evaluate CVR's diagnostic performance. Pearson correlation was used to assess relationships between baseline CVR and post-surgical changes. RESULTS: Affected-side CVR was significantly lower compared to the contralateral side and controls (P < 0.01). Individual analysis showed significant differences in the proportions of decreased, normal, and increased CVR between the affected and contralateral sides (P < 0.0001). CVR-AI showed high predictive efficiency for discriminating patients from controls (AUC = 0.762). Post-surgery, the affected-side CVR increased significantly, whereas CVR-AI decreased but remained higher than in controls (P < 0.05). Preoperative CVR-AI positively correlated with post-surgery affected-side CVR increase (R = 0.725) and CVR-AI decrease (R = 0.635). CONCLUSION: This study proposes CVR-AI as a novel imaging marker for pre-surgical evaluation in patients with unilateral intracranial arterial occlusion.