Abstract
JAMA Neurol. 2026 Aug 10. doi: 10.1001/jamaneurol.2026.2650. Online ahead of print.
ABSTRACT
IMPORTANCE: Phosphorylated tau-217 (p-tau217) is now an established plasma biomarker for assessing amyloid-β pathology in individuals at risk of Alzheimer disease. However, its performance in identifying advanced neocortical neurofibrillary tangle burden remains suboptimal. Precise assessment of tau pathology is increasingly critical for the rational implementation of anti-amyloid therapies and developing anti-tau interventions. Improved biofluid biomarker-based tau staging could enhance patient stratification and optimize participant selection for clinical care and therapeutic trials.
OBJECTIVE: To develop and validate a multiprotein plasma panel to improve identification of neocortical tau pathology beyond p-tau217 alone.
DESIGN, SETTING, AND PARTICIPANTS: This multicenter cohort study included 2 independent observational cohorts, Swedish BioFINDER study and Translational Biomarkers in Aging and Dementia (TRIAD). Cross-sectional clinical data and blood samples were collected between 2017 and 2024. Participants included 560 individuals spanning the clinical spectrum from cognitively unimpaired to dementia. These data were analyzed from January 2025 to May 2026.
EXPOSURES: Plasma concentrations of 125 proteins measured using Nucleic Linked Immuno-Sandwich Assay central nervous system panel.
MAIN OUTCOME AND MEASURE: Advanced tau pathology defined as tau positron emission tomography (PET) uptake within Braak stage V and VI regions. Predictive performance of biomarker models was evaluated using area under the receiver operating characteristic curve (AUC).
RESULTS: The study included 560 amyloid-positive participants (BioFINDER: n = 431; mean [SD] age, 73.6 [7.0] years; 212 female [49.2%] and 219 male [51.8%]; TRIAD: n = 129; mean [SD] age, 70.4 [8.3] years; 76 female [58.9%] and 53 male [41.1%]). Using multivariable logistic regression approaches, a 7-protein panel was found in BioFINDER to identify tau PET uptake within Braak stage V and VI regions. When compared with p-tau217 (AUC, 0.86-0.88; 95% CI, 0.82-0.94), this multiprotein panel was associated with improved identification in both discovery and validation cohorts (AUC, 0.92-0.94; 95% CI, 0.89-0.98; DeLong P < .001). This reduced the proportion of individuals classified within the intermediate-risk range (between paired sensitivity and specificity thresholds) in the validation cohort by 14.7% to 21.0%.
CONCLUSIONS AND RELEVANCE: This multicohort study demonstrated how including additional plasma proteins significantly enhanced the performance of p-tau217 in predicting advanced tau pathology among amyloid-positive individuals. This suggests a multiprotein approach may offer a viable and scalable alternative to tau PET staging in clinical or research settings.
PMID:42573994 | DOI:10.1001/jamaneurol.2026.2650
UK DRI Authors
