Value of FDG-PET in predicting dementia within 3 years among patients clinically diagnosed with mild cognitive impairment

1249 Objectives To assess the predictive value of clinically-evaluated FDG PET for imminent risk of conversion to dementia among patients clinically diagnosed with mild cognitive impairment (MCI), who were enrolled in the Alzheimer9s Disease Neuroimaging Initiative (ADNI) study and for whom longitudinal follow-up data for at least 3 subsequent years was available. Methods Baseline FDG PET scans from MCI patients, prospectively recruited for ADNI, were examined. The first 87 subjects (mean age 76 ± 6 years, range 55-88) to complete 3 years and whose data were acquired in DICOM format (or readily convertible to DICOM format) by October 2010 were included. Activities in 240 standardized regions of interest (sROIs) were automatically generated for each scan. Six dementia-specific posterior cortical brain volume activities were linearly combined and internally normalized to both pontine and whole brain volume activity to form a dementia prognosis (DP) index. Results All MCI subjects had a Clinical Dementia Rating (CDR) of 0.5 at time of baseline PET. By 3 years, 32 of the 87 MCI patients (37%) demonstrated progression (MCI-P) to dementia (CDR ≥1), while the other 55 (63%) demonstrated no progression (MCI-N), remaining stable (CDR 0.5) or reverting to a state of normal cognition (CDR 0). Progression to dementia was powerfully predicted by the DP index (p = 0.0001) alone. An average value of posterior cortical volumes of at least 4% above regional whole brain activity provided a greater than 93% probability of remaining cognitively stable for at least 3 years. Conclusions A normalized hypometabolism index, derivable automatically from routine regional analysis of brain PET scans, may provide substantial additive value above the initial clinical evaluation for accurately predicting which MCI individuals are at imminent risk of developing dementia. Research Support Funded by the National Institutes of Health, National Institute of Aging