Abstract
OBJECTIVE: The diagnostic performance of the Amyloid Probability Score 2 (APS2) - the algorithmic result of the PrecivityAD2(™) blood test - was originally trained and validated in two cohorts of cognitively impaired (CI) individuals. Using an independent cohort to evaluate blood test reliability, we conducted an external diagnostic accuracy assessment of the validated APS2 cut point as it is currently applied in clinical practice. METHODS: Plasma biomarker ratios Aβ42/40 and p-tau217/np-tau217 (expressed as %p-tau217) were quantified and incorporated into the APS2 algorithm in samples obtained from 192 Alzheimer's Disease Neuroimaging Initiative participants with CI (70% mild cognitive impairment / 30% dementia). APS2 diagnostic performance was determined using amyloid positron emission tomography (PET) as the reference standard. Plasma biomarkers were quantified in a CLIA-certified, CAP-accredited laboratory (C2N Diagnostics, St. Louis, MO) using liquid chromatography-tandem mass spectrometry. RESULTS: APS2 values were significantly higher in the 56% of CI participants with a positive amyloid PET scan. Concordance with amyloid PET was high (AUC-ROC 0.95 (95%CI): 0.93-0.98); 54% of participants had a positive APS2. The previously validated APS2 cut point yielded an overall accuracy of 91% (95%CI: 86-94%), sensitivity 90% (95%CI: 83-94%) and specificity 92% (95%CI: 84-96%). INTERPRETATION: The PrecivityAD2 blood test's APS2 identified brain amyloid pathology with accuracy, sensitivity, and specificity ≥ 90% in this intended use population. This external validation reaffirms the diagnostic robustness of this blood biomarker test and supports its use as a confirmatory test, consistent with published expert recommendations, for assessment of presence or absence of brain amyloid pathology in symptomatic patients.