Comparative analytical performance of multiple plasma amyloid‐beta assays and their relationship to amyloid PET

Abstract

Background The National Institute on Aging and the Alzheimer’s Association framework for classifying Alzheimer’s disease (AD) utilizes measures of pathology for amyloid, tau, and neurodegeneration (ATN), that can identify participants for clinical trials. Currently, amyloid pathology is determined by costly PET or invasive CSF measurements. When applied to participant selection, these measures are associated with high screen failure rates and contribute to the high costs and long duration of recruitment of AD clinical trials. Recent reports indicate plasma amyloid beta (Aβ) peptides correlate highly with PET and CSF amyloid measures. This study aims to characterize novel assays as a screening tool that can quantify plasma Aβ peptides in a robust and reproducible manner. Method The project team consists of representatives from pharmaceutical industry, nonprofit/patient advocacy, and academic institutions. Six plasma Aβ assays were selected to be part of the comparison: three liquid chromatography–mass spectrometry (LC‐MS/MS) assays and three immunoassays. Plasma samples with corresponding amyloid PET data were selected from ADNI n=130 (50% Aβ+): cognitively normal n=54 (37% Aβ+), mild cognitive impairment n=54 (46% Aβ+), and AD n=22 (91% Aβ+). Each participant’s sample was tested in a blinded fashion on all six assays with analytical controls. Statistical tests were performed to identify which assays can significantly improve the Area Under the Receiver Operating Characteristic (AUROC) curve for predicting amyloid PET status compared to age and APOE ε4 status alone. Result The project tested 130 samples across assays with comparison to amyloid PET. Initial results from assay providers show AUROCs to predict amyloid status at levels consistent with performance in earlier publications. A comparison of results for the various assays will be presented with respect to their ability to classify amyloid PET status and whether they add significantly to the prediction of amyloid positivity beyond age and genotype. Conclusion Measurement of Aβ in plasma contributes to addressing the amyloid component of the ATN framework and can be a pre‐screen or a substitute for PET or CSF screening. Further clinical validation of these assays is planned to assess amyloidosis detection performance in a cognitively normal population.