Blood Derived Amyloid Biomarkers for Alzheimer’s Disease Prevention

Abstract

Background: Reliable, widely accessible and affordable biomarkers for predicting Alzheimer’s disease (AD) brain pathology status are a necessity to aid development of prevention strategies in cognitively healthy at-risk older adults, at the right timepoint. Measurements of the key neuropathological hallmark beta-amyloid (Aβ) by PET neuroimaging or cerebrospinal fluid measures reflect its accumulation in the brain, yet recent methodological advancements now enable blood-based measures reflecting cerebral amyloid burden. Objectives: The current study validated the capacity of plasma Aβ42/Aβ40 measured using six different assays to predict amyloid positivity in a subgroup of cognitively unimpaired (CU) participants in the ADNI study and assessed its ability to discriminate CU from AD cases. We also explored economic viability of using two different plasma amyloid assays for pre-screening in AD prevention trials and as routine clinical diagnostic tool, versus amyloid PET alone. Design: A cross-sectional analysis of plasma and brain amyloid data, including comparative cost analysis of the plasma biomarkers in relation to brain amyloid PET. Setting: Alzheimer’s Disease Neuroimaging Initiative (ADNI). Participants: ADNI participants consisting of 115 CU, mild cognitive impairment and AD cases who had plasma Aβ42/Aβ40 measured with six platforms. Measurements: Plasma Aβ42/Aβ40 was measured via six different platforms: three immunoassays (Roche, Quanterix and ADx Neurosciences) and three mass spectrometry (MS) based assays (WashU, Shimadzu and Gothenburg). Aβ-PET imaging was conducted within three months of plasma sampling using [18F]florbetapir. Results: There was a weak to moderate correlation of plasma Aβ42/Aβ40 ratio between platforms. The MS-based WashU test had the highest capacity to discriminate between CU and AD (area under the curve, AUC = 0.734, 95% CI: 0.613-0.854; P = 0.008). Within the CU group, the WashU plasma amyloid test had the best discriminative capacity to distinguish Aβ+ from Aβ– (AUC = 0.753, 95% CI: 0.601-0.905; P = 0.003) closely followed by the immunoassay from Roche (AUC = 0.737, 95% CI: 0.597-0.877; P = 0.006). The exploratory economic analyses showed that the use of Roche or WashU plasma amyloid assay as a pre-screening tool prior to Aβ-PET scans for clinical trial recruitment significantly reduced total screening cost (saving up to $5882 per recruited patient) expected in an AD prevention trial. Conclusions: With few available treatment strategies, dementia prevention is a global priority. CU individuals at risk for AD are the target population for dementia prevention but have been poorly studied. Our findings confirming diagnostic value of ultrasensitive immunoassays and high-performance immunoprecipitation coupled with MS for measurement of plasma Aβ42/Aβ40 to detect PET amyloid positivity in CU participants allude to potential clinical utility of this biomarker. Plasma Aβ42/Aβ40 could be optimal for pre-selecting at-risk candidates for more invasive and expensive investigations across AD prevention clinical trials and clinical care for a rapidly ageing population.