Comparison of neuroimaging modalities for the prediction of conversion from mild cognitive impairment to Alzheimer's dementia

In this study we compared Pittsburgh compound-B (PIB) positron emission tomography (PET) amyloid imaging, fluorodeoxyglucose PET for metabolism, and magnetic resonance imaging (MRI) for structure to predict conversion from amnestic mild cognitive impairment (MCI) to Alzheimer's dementia using data from the Alzheimer's Disease Neuroimaging Initiative cohort. Numeric neuroimaging variables generated by the Alzheimer's Disease Neuroimaging Initiative-funded laboratories for each neuroimaging modality along with apolipoprotein-E genotype (n = 29) were analyzed. Performance of these biomarkers for predicting conversion from MCI to Alzheimer's dementia at 2 years was evaluated in 50 late amnestic MCI subjects, 20 of whom converted. Multivariate modeling found that among individual modalities, MRI had the highest predictive accuracy (67%) which increased by 9% to 76% when combined with PIB-PET, producing the highest accuracy among any biomarker combination. Individually, PIB-PET generated the best sensitivity, and fluorodeoxyglucose PET had the lowest. Among individual brain regions, the temporal cortex was found to be most predictive for MRI and PIB-PET.