Multi-Center Study on Alzheimer's Disease Using FDG PET: Group and Individual Analyses

Abstract

Using statistical parametric mapping (SPM), we evaluate the feasibility and accuracy of ¹⁸F -2-deoxy-2-fluoro-D-glucose (FDG) positron emission tomography (PET) for clinical diagnosis with characteristic hypo-metabolic regions, and examine the consistency of cerebral hypo-metabolism in Alzheimer's disease (AD) cross multicenters at both the group and the individual levels. Four groups of scan data including 39 AD patients and 52 healthy control subjects derived from three centers were analyzed and comparisons between patient subgroups or individual patient and relevant control population were performed using Two Sample T-test. In the group analysis, the hypo-metabolic regions of AD patients obtained from different PET centers were similar and consistent. The common hypo-metabolic cerebral areas were located bilaterally in the posterior cingulate and medial parietal cortex, temporo-parietal cortex, prefrontal cortex, and the middle and inferior temporal gyrus (uncorrected, p<0.001). In the analysis of each individual subject, the location of declined posterior cingulate and medial parietal cortex, temporo-parietal cortex and temporal lobe were found highly consistent with relevant characteristic regions obtained in the group analysis and were selected for diagnosic purposes. Complete typical hypo-metabolic pattern was observed in 67% and 54% of AD patients in two sets of 3D scans, respectively. Only 27% and 33.3% patients showed full typical pattern in two sets of 2D scans. The results indicated that FDG PET measures and SPM can 4 provide a valuable reference for clinical diagnosis of AD patients. The potential influence of acquisition mode on the clinical diagnosis of AD was suggested for further evaluation.