P2–285: Voxel–based analyses of pittsburgh Compound–B and FDG in AD

Region–of–interest (ROI) results for the PET amyloid–beta imaging agent Pittsburgh Compound–B (PIB) were previously reported and showed significantly greater PIB retention in Alzheimer's disease (AD) subjects (relative to controls) in posterior cingulate/precuneus, frontal, parietal, and temporal cortices (Wilcoxon, 1–sided, p<0.05). The present study was undertaken to evaluate voxel–based statistical methods for the assessment of PIB retention and to compare these results to FDG studies of glucose metabolism performed in the same subjects on the same day. PET studies were performed (ECAT HR+) in 10 mild AD subjects (age: 69±8 years; MMSE: 25±3) and 11 control subjects (age: 74±6 years; MMSE: 29±1). Parametric images of PIB retention were generated using the Logan graphical analysis with cerebellar (reference region) data as input. The final measure of PIB retention was the distribution volume ratio (DVR). FDG standardized uptake value (SUV) images were generated by summing the activity from 40–60 min post–injection and then normalizing this to the cerebellar SUV value. Data were compared using statistical parametric mapping (SPM) with the false discovery rate correction (FDR) for multiple comparisons. The voxel–based PIB results were consistent with previous ROI results, as AD subjects showed marked retention in primary cortical areas (relative to controls), with the most significant results in: frontal cortex p=1.19e–10; parietal/temporal cortex p=1.19e–10; and posterior cingulate/precuneus p=1.20e–10. No group differences were evident for white matter or cerebellum. AD subjects showed marginally significant decreases in FDG uptake in the same cortical areas (corrected p–values: parietal/temporal p=0.098; frontal p=0.098; posterior cingulate/precuneus p=0.098). The PIB SPM results were more robust than the FDG results in distinguishing AD and control subjects, as the former reached greater significance and detected differences with a larger spatial extent. These results indicate that voxel–based methods will be useful for future larger longitudinal studies of amyloid–beta deposition that could improve AD diagnosis and anti–amyloid therapy assessment.