BASELINE FLORBETAPIR AMYLOID PET STANDARD UPTAKE VALUE RATIO (SUVR) CAN PREDICT CLINICAL PROGRESSION IN PRODROMAL AD

Previous analyses suggest Aβ-related amyloid load is not correlated with cognitive change in Alzheimer's disease (Daniela et al 2014); primarily based on studies in patients with clinical dementia. This finding could be related to the initiation of a cascade of cognitive impairment caused by amyloid saturation, thus no longer responsive to change, i.e. much like reaching a plateau of a dose-response curve. We therefore hypothesized that there might be a correlation prior to amyloid plaque saturation. In addition, we hypothesized that correlation would not be seen at lower plaque loads, as cognition is not sufficiently impaired at this early stage of disease. Here we show an optimal range of baseline amyloid PET Standard Uptake Value Ratios (SUVR) can predict cognitive decline over 24 months in prodromal AD (pAD) subjects. Data from MCI subjects who were positive for amyloid (prodromal AD; pAD) in the Alzheimer's Disease Neuroimaging Initiative (ADNIMERGE, N=132; downloaded March, 2018; data collection and sharing funded by ADNI National Institutes of Health Grant U01 AG024904) were used to define a dataset with the following baseline characteristics: MMSE ≥24; global CDR=0.5; CDR-SB memory box ≥ 0.5; PET florbetapir composite cortex SUVr ≥ 1.13 (whole cerebellum reference region). Both CDR-SB and PET SUVr change from baseline (CFB) over 24 months were compared to baseline PET SUVr across multiple statistical models. Models estimated 3 ranges of PET SUVr values predicting linear increase in CFB CDR-SB and linear decrease in CFB SUVR across baseline SUVR range 1.2 to 1.6. Baseline SUVR below 1.2 predicts amyloid plaque load continues to build, with no change in CFB CDR-SB at 24 months as expected. In contrast, baseline SUVR above 1.6 shows no further increase in amyloid plaque load or no change in CDR-SB. The present study identifies an optimal baseline SUVR range of 1.2 to 1.6 that predicts linear decline on CFB CDR-SB over 24 months in pAD subjects. This suggests that baseline plaque load can predict rate of clinical progression, at least in pAD. These findings may have implications for study enrichment considerations.