High Precision β‐Amyloid 1‐40, β‐Amyloid 1‐42, and Ptau181 Plasma Assay Performance in NCRAD Biomarker Assay Laboratory
Kristen A. Russ, Emily Smith, Ralitsa V. Kostadinova, Sujuan Gao, Steve Brown, Liana G. Apostolova, Shannon L Risacher, Andrew J. Saykin, Tatiana M. Foroud, Jeff L. Dage- Psychiatry and Mental health
- Cellular and Molecular Neuroscience
- Geriatrics and Gerontology
- Neurology (clinical)
- Developmental Neuroscience
- Health Policy
- Epidemiology
Abstract
Background
To achieve the robust and reliable implementation of assays in the National Centralized Repository for Alzheimer’s Disease and related Dementias (NCRAD), new assays and platforms must be assessed for performance characteristics. The goal of this study is to assess the [Fujirebio Lumipulse] G1200 performance in comparison to existing platforms
Method
Plasma samples utilized for assay precision measurements were collected from healthy control subjects and patients with Alzheimer’s disease (AD) by the Indiana Biobank. Five samples were chosen to span the assay range and were assayed on two separate days. Clinical utility was assessed with plasma samples (N = 179) collected by the Indiana Alzheimer’s Disease Research Center (IADRC). A subset (N = 92) of these samples were also assayed on the Neurology 4‐Plex E Advantage Kit on the [Quanterix] Simoa‐HDx platform. All samples were assayed on the [Fujirebio Lumipulse] G β‐Amyloid 1‐40, β‐Amyloid 1‐42, and Ptau181 plasma assays.
Result
[Fujirebio Lumipulse] G β‐Amyloid 1‐40, β‐Amyloid 1‐42, and Ptau181 plasma assays had an average inter‐assay coefficient of variation of 2.6%, 2%, and 2% respectively (Figure 1). In samples with both [Fujirebio Lumipulse] and Simoa‐HDx pTau181 data, linear regression showed an R^2 of 0.80 and slope of 1.67. Bland‐Altman analysis showed a consistent bias of 54%. Once results were standardized by Z‐score, slope was near one, y‐intercept near zero, demonstrating comparability of results. Area Under Receiver Operating Characteristic curve (AUC) for IADRC samples with PET data was 0.84 for Aβ 42/40 ratio alone and 0.87 including age, sex and APOE status. AUC for Ptau181 alone was 0.83 and 0.89 including age, sex and APOE status. Assays were able to detect differences in clinical diagnosis or amyloid status (Figure 2A and B).
Conclusion
The [Fujirebio Lumipulse] assays exhibited high precision. When standardized by Z‐score, the [Fujirebio Lumipulse] Ptau181 assay was comparable to the Simoa‐HDx Ptau181v2 assay allowing comparison of data between assay platforms. AUCs for amyloid PET were higher when considering age, sex and APOE status in the model. The assays were able to differentiate between clinical diagnoses and amyloid status suggesting suitability for implementation.