When the cerebral metabolic rate of glucose (CMRglc) is to be measured, arterial blood sampling is usually required for fluorine-18 fluorodeoxyglucose (FDG) and positron emission tomography (PET) studies. However, blood sampling is inconvenient because it requires several staff members and is invasive for patients.To assess cerebral glucose metabolism by a noninvasive and simplified method, the authors used the standardized uptake value (SUV), which requires no input function or blood sampling. The study participants included 18 healthy volunteers (4 men and 14 women; mean +/- SD age, 68.2+/-6.3 years), 18 patients with mild Alzheimer's disease (AD) (4 men and 14 women; mean +/- SD age, 68.8+/-7.3 years), and 18 patients with moderate AD (5 men and 13 women, mean +/- SD age, 69.5+/-8.5 years). Regional CMRglc and regional cerebral SUV were measured in the three groups using FDG PET, and the correlation between global CMRglc and global SUV was estimated.The correlation coefficients of global CMRglc and global SUV in the healthy volunteers, mild AD patients, and moderate AD patients were 0.82, 0.67, and 0.62, respectively. Compared with the healthy persons, the patients with mild AD showed significantly decreased CMRglc in the temporal, frontal, and parietal cortices, but they did not show significantly decreased SUV in any region. Patients with moderate AD had significantly decreased CMRglc in the temporal, frontal, occipital, parietal, and sensorimotor cortices and significantly decreased SUV in the temporal, frontal, occipital, and parietal cortices.The SUV would be useful as a semiquantitative index of cerebral glucose metabolism only in healthy persons or those with mild AD.