Spatial coefficient of variation of arterial spin labeling magnetic resonance imaging can predict decreased cerebrovascular reactivity measured by acetazolamide challenge single-photon emission tomography.

Abstract

The aim of this study was to investigate whether the spatial coefficient of variation of arterial spin labeling (ASL-CoV) acquired in clinical settings can be used to estimate decreased cerebrovascular reactivity (CVR) measured with single-photon emission computed tomography (SPECT) and acetazolamide challenge in patients with atherosclerotic stenosis of intra- or extracranial arteries. We evaluated the data of 27 atherosclerotic stenosis patients who underwent pseudocontinuous ASL and SPECT. After spatial normalization, regional values were measured using the distributed middle cerebral artery territorial atlas of each patient. We performed comparisons, correlations, and receiver operating characteristic (ROC) curve analyses between ASL-cerebral blood blow (CBF), ASL-CoV, SPECT-CBF and SPECT-CVR. Although the ASL-CBF values were positively correlated with SPECT-CBF values (r = 0.48, 95% confidence interval (CI) = 0.28-0.64), no significant difference in ASL-CBF values was detected between regions with and without decreased CVR. However, regions with decreased CVR had significantly greater ASL-CoV values than regions without decreased CVR. SPECT-CVR was negatively correlated with ASL-CoV (ρ = -0.29, 95% CI = -0.49 - -0.06). The area under the ROC curve of ASL-CoV in predicting decreased CVR (0.66, 95% CI = 0.51-0.81) was greater than that of ASL-CBF (0.51, 95% CI = 0.34-0.68). An ASL-CoV threshold value of 42% achieved a high specificity of 0.93 (sensitivity = 0.42, positive predictive value = 0.77, and negative predictive value = 0.75). ASL-CoV acquired by single postlabeling delay without an acetazolamide challenge may aid in the identification of patients with decreased CVR on SPECT.