Recent advances in automated immunoassays have enabled sensitive detection of AD neuropathological markers Aβ42/40 and pTau217 in blood plasma, however, most exhibit <90% sensitivity and specificity with the application of a single cutoff.  Further characterization of increased diagnostic accuracy in pre- and symptomatic AD subjects is needed.

Multiple analytical approaches, based on plasma pTau217 and Aβ42/40 measurements, are evaluated using discriminatory performance for amyloid PET with the goal of developing a confirmatory test for Aβ pathology demonstrating >90% sensitivity and specificity. 

One hundred-ninety-seven participants from the AIBL cohort were selected representing a cross-sectional AD continuum population: cognitive unimpaired (CU) Aβ- (n=75), CU Aβ+ (n=48), mild cognitive impairment (MCI) Aβ+ (n=26), and AD Aβ+ (n=48) with amyloid PET positivity prevalence of 62.5%.  EDTA-plasma samples were analyzed with Lumipulse pTau217 and Aβ42/40 assays. Multivariable models and ratios combining biomarkers, and +/- adjustment for demographic variables, were tested vs. single measurements for ROC-AUC concordance with amyloid PET, applying single vs. dual cutoff approaches.

The improved performance of the pTau217/Aβ42 ratio relative to pTau217 alone, comparable to more complex multivariable models, meets a recommended performance of a confirmatory AD blood biomarker test of >90% sensitivity and specificity. Further characterization will be performed in a cognitive impaired intent to treat population.