Predicting cognitive decline in older people by structural and molecular imaging

Availability of possible disease modifying treatments and the recognition of predementia stages of Alzheimer's disease (AD) have raised awareness for the prognostic and predictive role of biomarkers, particularly imaging markers.The positive predictive value of amyloid PET for the transition to prodromal AD or AD dementia in cognitively normal people is below 25%. Evidence for tau PET, FDG-PET and structural MRI is even more limited. In people with mild cognitive impairment (MCI), imaging markers yield positive predictive values above 60% with moderate advantages for amyloid PET over the other modalities and an added value for the combination of molecular with downstream neurodegeneration markers.In cognitively normal people, imaging is not recommended for individual prognosis due to lack of sufficient predictive accuracy. Such measures should be restricted to risk enrichment in clinical trials. In people with MCI, amyloid PET and, to a somewhat lesser extent, tau PET, FDG-PET, and MRI yield relevant predictive accuracy for clinical counseling as part of a comprehensive diagnostic program in tertiary care units. Future studies should focus on the systematic and patient-centered implementation of imaging markers in evidence-based care-pathways for people with prodromal AD.