Impact of reference region on longitudinal florbetapir PET SUVR changes from the API ADAD Colombia Trial

Abstract

AbstractBackgroundWe previously examined the impact of using a cerebellar, pons, or cerebral white matter (WM) reference region (RR) on the ability of each to distinguish cortical measures of fibrillar amyloid‐(Aβ) deposition with baseline data in unimpaired presenilin (PSEN1) E280A mutation carriers and non‐carriers (NC) from the Alzheimer’s Prevention Initiative Autosomal Dominant Alzheimer’s Disease (API ADAD) Colombia Trial (NCT01998841). In this study, we extend our findings by comparing these RRs on their ability to distinguish longitudinal changes in cortical measures of serial Aβ‐PET scans from the completed clinical trial.MethodWe estimated template‐based (SPM12) cortical mean change using baseline, 2, and 5‐year florbetapir PET standard‐uptake value ratios (SUVRs) with three different RRs (whole cerebellum, pons, and WM) and compared carriers on placebo (placebo‐carriers, n = 82) with NC all on placebo (NC, n = 83). We calculated Cohen’s D with [95% CI] and standard deviation of NC as effect‐sizes for 2 and 5‐year cortical changes. Additionally, we characterized cerebellar Aβ burden in carriers on crenezumab (treatment‐carriers, n = 81), placebo‐carriers, and NC, using our previously reported method targeting the whole cerebellum region‐of‐interest (ROI) with pons as RR to estimate cerebellar‐to‐pons‐SUVRs and compare using a mixed‐model repeated‐measures ANOVA adjusted for age but not multiple comparisons.ResultAs expected, placebo‐carriers had significantly greater SUVR change compared with NC for all three RRs. However, the pons RR was best at detecting 2‐year changes (d = 1.75[1.39‐2.11]), and WM and pons were best at detecting 5‐year changes in cortical‐Aβ levels between placebo‐carriers and NC (pons d = 2.01[1.64‐2.39], WM d = 2.51[2.10‐2.92]). Additionally, treatment‐carriers had higher cerebellar‐to‐pons‐SUVRs compared with placebo‐carriers at baseline and 2‐years, but not 5‐years (p = .034, p = .027,& p = .06,respectively), both carrier groups had higher cerebellar‐to‐pons‐SUVRs compared to NC at each interval(p<.001). Cerebellar‐to‐pons‐SUVRs continued to increase within carrier groups at each interval(5‐year>2‐year>baseline,p<.05) and remained constant within the NC‐group(p>.05).ConclusionUse of pons for 2‐year and WM or pons RRs for 5‐year changes may improve the power to track longitudinal increases and evaluate Aβ‐modifying treatments in 2 and 5‐year studies, in this ADAD population. A cerebellar RR may be confounded by early Aβ deposition in the cerebellum which can artificially reduce average cortical Aβ‐PET SUVRs even in preclinical stages of ADAD.