Pitfalls of regional cerebral blood flow analysis in cerebrovascular disease.

In 20 patients with unilateral cerebrovascular disease, regional cerebral blood flow (rCBF) was measured by F-18-fluoromethane inhalation and positron emission tomography (PET). The purpose was to study various strategies of rCBF data analysis that are currently used in SPECT and PET. Methods of quantitative data analysis standardize rCBF values relative to an intraindividual reference, and thus create ratios as estimators of rCBF in the ischemic regions-of-interest (ROI). These ratios were compared as well as the absolute rCBF values in the ischemic ROI graphically and by Spearman's rank correlation coefficient (r). The results demonstrated that previously reported methods of data analysis failed to represent rCBF in the ischemic ROI; r values were 0.467, -0.406, -0.453, and 0.329, respectively (N.S.). This failure was due to a more widespread reduction of rCBF even in patients with minimal ischemic deficits. In conclusion, previously reported strategies of rCBF analysis based on intraindividual standardization should not be used in patients with cerebrovascular disease since they may produce misleading results.