Imaging biomarkers can be used to screen participants for Alzheimer’s disease clinical trials. To test the predictive values in clinical progression of neuropathology change (amyloid‐PET) or brain metabolism as neurodegeneration biomarker ([18F]FDG‐PET), we evaluated data from N = 268 healthy controls and N = 519 mild cognitive impairment subjects. Despite being a significant risk factor, amyloid positivity was not associated with clinical progression in the majority (≥60%) of subjects. Notably, a negative [18F]FDG‐PET scan at baseline strongly predicted clinical stability with high negative predictive values (>0.80) for both groups. We suggest [18F]FDG‐PET brain metabolism or other neurodegeneration measures should be coupled to amyloid‐PET to exclude clinically stable individuals from clinical trials.