Background: The effect of age on amyloid burden in Alzheimer's Disease patients and in the discriminative power of amyloid imaging needs to be determined.
      Methods: 246 Healthy Control (HC), 342 Mild Cognitive Impairment (MCI) and 138 AD with raw [18F]Florbetapir- and [18F]FDG-PET data were downloaded from the Alzheimer's Disease Neuroimaging Initiative (ADNI) database. Florbetapir data were analysed using a Florbetapir-PET population template and atlas. Two age groups were created; mean = 70.9 ± 3.5 (young) and 80.9 ± 3.9 (old) years, matched for gender and disease severity (MMSE). A Receiver Operating Characteristic Analysis derived a composite neocortex ratio (cctxr) cut-off of 1.34. The AD patients with Amyloid negative scans had their Florbetapir and FDG scan read visually.
      Results: The Florbetapir cut-off displayed lower sensitivity and specificity in the old ADNI group in comparison to the young one (87/87 vs 73/74). The young HC displayed significant lower amyloid in the Posterior Cingulate (1.16 vs 1.24) and the Putamen (1.18 vs 1.25) in comparison to the older HC. The cctxr values of the old AD patients (oAD) were bimodal in mixture model analysis whereas the values of the younger AD (yAD) patients followed a normal distribution. The yAD patients, whether analysed dichotomously or continuously, displayed significantly higher values in all ROI examined in comparison to the oAD, except in Putamen and Occipital lobe. The neocortical FDG pattern of amyloid negative AD patients (n=27; 10 yAD vs 17 oAD) were all consistent with AD; whereas 10 of the same individuals had negative Florbetapir scans when read visually. In agreement with the PET template analysis the amyloid negative oAD had a greater number of Florbetapir scans read as negative (n=8) in comparison to the yAD (n=2).
      Conclusions: The Florbetapir cut-off, as well as the visual read of the Florbetapir scans, displayed a diminished discriminative ability in the old ADNI group in comparison to the younger patients. The oAD patients showed less prominent amyloid burden and different distribution of their values in comparison to the yAD group. This might have important implications for the clinical diagnostic use of Florbetapir imaging in the oldest patients.