The preliminary study of 18F-FDG brain PET in diagnosis of alzheimer's disease

Objective: To investigate the imaging characteristics and diagnostic criteria of 18F-FDG brain PET in diagnosis of Alzheimer's disease (AD). Methods: The sutdy included 12 normal subjects, 12 patients with AD and 11 patients with non-AD dementia. 40 min after intravenous administration of 18F-FDG, brain scan was performed using Siemens ECAT47 scanner. The transaxial, coronal and sagittal images were then reconstructed by computer. At the same time, semiquantitative analysis was also applied to help evaluation using the ratio of mean radioactivity of cerebral lobe to cerebellum (Rcl/cb). Results: In normal subjects PET scan showed clear images of cerebral cortex, basal ganglia, thalamus and cerebellum with symmetrical distribution of radioactivity. PET images from Alzheimer's disease patients were classified into 3 patterns: bilateral parietal hypometabolism in 5 cases, bilateral temporo-parietal hypometabolism in 4 cases and unilateral temporo-parietal hypometabolism in 3 cases. The Rcl/cb of AD patients in parietal and temporal lobe was significantly decreased than normal subjects (P<0.05). PET images from non-AD dementia were also classified into 3 patterns; multiple and asymmetrical patch foci with decreased radioactivity in 8 cases, bilateral temporo-parietal with diffuse cortical hypometabolism in 2 cases, and normal imaging in 1 case. The hypometabolic involvement was found in accord with severity of dementia. The Rcl/cb was also reflecting thedementia degree. Compared with MRI imaging , 12 patients with AD had cerebral hypometabolism but only 10 had hippocampus atrophy. 10 patients with non-AD dementia had local structural foci seen in MRI, including old hemorrhage, infarction and encephalomalacia, but these lesions were not found in AD. Conclusions: Based on excluding cerebral structural lesions which are better detected by MRI, bilateral or unilateral parietal or temporo-parietal hypometabolism found in FDG PET can be considered indicative of Alzheimer's disease. Semiquantitative analysis of the images yielded can help to evaluate the dementia degree