This cross-sectional study (n=924 ADNI participants; CU, MCI, and dementia) tested whether brain-derived (BD) p-tau217—measured in CSF and plasma via the Alamar NULISAseq panel—better reflects continuous amyloid PET burden (Centiloids) and classifies amyloid positivity across CL thresholds of 20, 40, 60, and 80, compared to conventional p-tau217.
Plasma BD p-tau217 showed the strongest continuous association with amyloid PET (R²=0.63, RMSE=24.5), outperforming conventional plasma p-tau217 (R²=0.56, RMSE=27.0) and CSF biomarkers (R²=0.37, RMSE=31.8). For binary classification, plasma BD p-tau217 AUCs ranged 90.0–93.2 across all thresholds, significantly higher than conventional plasma (88.0–91.4; DeLong p<0.001 at all thresholds), while CSF BD and conventional CSF p-tau217 showed no significant difference (AUCs ~80–83).
- Research-only assay platform (NULISAseq) not yet clinically scalable. - Cross-sectional design; no longitudinal BD p-tau217 data to assess tracking of Aβ progression. - Single cohort (ADNI) with highly selected participants, limiting generalizability to community populations.
Plasma BD p-tau217 offers modestly better amyloid staging than conventional plasma p-tau217—especially at low Centiloid thresholds relevant for preclinical trial enrichment—but individual-level error (±~25 CL) means it complements rather than replaces amyloid PET. CSF BD processing adds no meaningful benefit over standard CSF p-tau217.
Explore related topics
