Analytical Validation and Performance of a Blood-Based P-tau217 Diagnostic Test for Alzheimer Disease

Abstract

Background: Blood-based biomarkers, especially P-tau217, have been gaining interest as diagnostic tools to measure Alzheimer disease (AD) pathology.

Methods: We developed a plasma P-tau217 chemiluminescent immunoassay using 4G10E2 and IBA493 as antibodies, a synthetic tau peptide as calibrator, and the Quanterix SP-X imager. Analytical validation performed in a College of American Pathologists-accredited CLIA laboratory involved multiple kit lots, operators, timepoints, and imagers. Florbetapir positron emission tomography was used to quantify amyloid for clinical validation.

Results: Precision across 80 runs was ≤20% CV using 23 patient-derived samples ranging from 0.09 U/mL to 3.35 U/mL. No significant lot-to-lot differences were observed. There was no interference from purified tau (2N4R) or lipemia, but hemolysis greater than 2 + was not acceptable. Functional analytical sensitivity (lower limit of quantitation) was 0.08 U/mL. Linearity studies support the use of a standard 1:2 plasma dilution. Samples demonstrated stability at 7 freeze/thaw cycles, with room temperature and refrigerated stability established for up to 72 hours. The final analytical measurement range was 0.08 to 2.81 U/mL. The calibration curve maintained ≤20% CV for raw signal intensity and 80% to 120% relative error for back-fitted concentration using a log-log power regression. Initial clinical assessment using plasma samples from 1091 individuals screened in TRAILBLAZER-ALZ 2 demonstrated an area under the curve of 91.6% (95% CI 0.90-0.94) with brain amyloid as the comparator. Positive and negative predictive value was >90% and >85%, respectively.

Conclusions: Through analytical validation, this assay demonstrated robust performance across multiple lots, operators, and instruments and could be used as a tool for diagnosing AD.