Comparison of conventional region of interest and statistical mapping method in brain single-photon emission computed tomography for prediction of hyperperfusion after carotid endarterectomy

OBJECTIVE: Hyperperfusion after carotid endarterectomy (CEA) has been proposed as the main mechanism of intracerebral hemorrhage, which is a disastrous complication. The aim of this study was to compare the predictive value of cerebral blood flow (CBF) abnormalities for hyperperfusion after CEA with the conventional region of interest (ROI) analysis and statistical brain mapping analysis. METHODS: For 46 patients with unilateral carotid stenosis of 70% or more, CBF and cerebral vasoreactivity were investigated with resting and acetazolamide-challenge single-photon emission computed tomography before CEA and 1 day after CEA. Three-dimensional sterectactic surface projection (3-D-SSP) analysis of CBF changes was performed by automatic quantification with a predefined template. RESULTS: Logistic regression analysis demonstrated CBF reduction (z-score) to be the only significant variable for postoperative hyperperfusion on 3-D-SSP with thalamic normalization but no significant variable with the ROI method. Receiver operating characteristic (ROC) analysis demonstrated significant improvement in the predictive value of CBF reduction (z-score) on 3-D-SSP (area under the ROC curve = 0.93) in comparison with the ROI method (area under the ROC curve = 0.78) (P = 0.049). According to the optimal cutoff values provided by ROC analysis, patients were categorized into two groups: Type I (CBF decrease < 20%, n = 23) and Type 11 (CBF decrease >= 20%, n = 23) on ROI analysis and Type A (z-score <= 2, In = 40) and Type B (z-score > 2, n = 6) on 3-D-SSP. There was a significant difference in incidence of hyperperfusion between Type A (1 of 40) and Type B (5 of 6) on 3-D-SSP (P = 0.00003) but not between Type I (I of 23) and Type 11 (5 of 23) on ROI analysis. Cerebral vasoreactivity did not show significant value in the prediction of hyperperfusion with either the ROI or the 3-D-SSP method. CONCLUSION: Objective assessment of CBF status, especially baseline CBF reduction (z-score), with automatic quantification by 3-D-SSP with normalization had a higher diagnostic value than conventional ROI analysis to identify patients at risk for hyperperfusion after CEA.