Performance of plasma pTau181 and pTau217 measured with fully automated LUMIPULSE G prototype immunoassays

Abstract

AbstractBackgroundThere is a need for scalable, robust and precise Alzheimer’s Disease (AD) plasma phosphorylated Tau (pTau) biomarker tests to support clinical trials and clinical practice. This study evaluated the performance of novel plasma pTau181 and pTau217 assays developed on the LUMIPULSE G1200 automated immunoassay platform and compared them to Quanterix homebrew assays.MethodHomebrew Quanterix Simoa assays for plasma pTau181 and pTau217 were transferred to the LUMIPULSE G instrument using recombinantly expressed versions of respectively α‐pTau181 mAb ADx252 and α‐pTau217 mAb RD‐084 (capture) both combined with an ALP‐conjugated Fab fragment digested from α‐N‐terminal Tau recombinant ADx204/RD‐073 (detector). Analytical performance evaluated specificity, LLOQ and precision. Performance in patient samples was evaluated using a prospective memory cohort of 42 biomarker confirmed AD patients in the prodromal or dementia stage as well as 35 healthy spouse controls. LUMIPULSE G results were compared with results from the Quanterix homebrew assays designed by ADx using the same antibody configuration.ResultRobust LUMIPULSE G1200 measurements (mean intra‐run CV 3%) distinguished AD patients from spouse controls with an area under the curve (AUC) of 0.81 and 0.88 for pTau181 and pTau217, respectively. The Quanterix homebrew assays demonstrated higher AUC of 0.98 for pTau181 (pDeLong = 0.002) and of 0.95 for pTau217 (Figure 1A). Addition of age, sex and ApoE4 status to the LUMIPULSE G‐based models increased the performance of both pTau181 and pTau217 to predict AD diagnosis – relatively to univariate models ‐ with AUCs of respectively 0.91 and 0.95 (Figure 1B, both pDeLong < 0.05), approximating Simoa‐based pTau performance. PTau values measured on both platforms correlated significantly (p<0.0001) for pTau181 (r = 0.66) and pTau217 (r = 0.77).ConclusionThe LUMIPULSE G platform is a scalable, automated, real‐time‐testing platform that is globally available. These data demonstrate that two assays on this platform, pTau181 and pTau217, are capable of distinguishing AD from controls in plasma. The sensitivity and specificity are slightly less on LUMIPULSE G when compared to an assay using the same antibodies developed on the Simoa instrument. Thus, additional experiments are required on the LUMIPULSE G to optimize analytical performance of these assays using an IVD‐grade automated system.