P1‐309: Correlation between baseline measures of amyloid‐positron emission tomography (PET) imaging, cerebral spinal fluid (CSF) biomarkers, and clinical scales in predementia Alzheimer's disease

CSF amyloid and tau biomarkers are strong predictors of AD progression from mild cognitive impairment (MCI). Emerging clinical evidence suggests that amyloid-PET imaging has similar utility. The correlation between CSF and PET amyloid biomarkers is expected to be high in AD. We performed an exploratory analysis limited to baseline data from predementia AD patients (identified by CSF biomarker measurements) enrolled in CN156018 who completed florbetapir (florbetapir F 18 18F-AV-45) amyloid-PET scans to examine: 1) concordance between pathologic CSF (low CSF Aß42, high total-tau t-tau) and florbetapir-PET visual analysis, 2) correlation between t-tau/Aß42 ratio and quantitative florbetapir PET standard uptake volume ratios (SUVRs), and 3) correlation between clinical rating scales and amyloid CSF and PET biomarkers. CN156018 is an ongoing, randomized, double-blind, placebo-controlled, Phase 2 study designed to assess the safety and tolerability of BMS-708163, a gamma secretase inhibitor, in patients with predementia AD. Approximately 100 patients will complete both CSF measurements and florbetapir PET imaging. Study entry criteria include: male and female outpatients 45-90 years of age, CSF Aß42 ≤ 200 pg/mL or t-tau/Aß42 ratio ≥ 0.39, MMSE score 24-30 (inclusive), subjective memory complaint, objective memory loss measured by education adjusted scores on Wechsler Memory Scale Logical Memory II or the Free Cued Selective Reminding Test (FCSRT), and Clinical Dementia Rating scale global score of 0.5 with a memory box score of = 0.5. Eighteen of the first 20 patients enrolled in CN156018 demonstrated concordance between pathologic CSF (low CSF Aß42 and high t-tau) and amyloid positive florbetapir-PET visual analysis at baseline. Quantitative analyses of amyloid burden using florbetapir-PET imaging at baseline demonstrated statistically significant correlations (Pearson correlation = 0.76 to 0.80, P < 0.05; N = 24) between SUVRs and CSF t-tau/Aß42 ratio at baseline. Among clinical scales, FCSRT demonstrated the highest correlation with both amyloid CSF and florbetapir-PET SUVR at baseline. Consistent with the diagnostic framework proposed by Dubois et al, the relationship between pathologic CSF (low CSF Aß42, high t-tau) and amyloid-PET uptake values observed in initial participants in CN156018 demonstrate that multiple supportive amyloid biomarkers can be used to potentially identify patients with predementia AD.