Clinical Significance of the Plasma Biomarker Panels in Amyloid-Negative and Tau PET-Positive Amnestic Patients: Comparisons with Alzheimer's Disease and Unimpaired Cognitive Controls

Abstract

The purpose of this study was to investigate whether plasma biomarkers can help to diagnose, differentiate from Alzheimer disease (AD), and stage cognitive performance in patients with positron emission tomography (PET)-confirmed primary age-related tauopathy, termed tau-first cognitive proteinopathy (TCP) in this study. In this multi-center study, we enrolled 285 subjects with young-onset AD (YOAD; n = 55), late-onset AD (LOAD; n = 96), TCP (n = 44), and cognitively unimpaired controls (CTL; n = 90) and analyzed plasma Aβ42/Aβ40, pTau181, neurofilament light (NFL), and total-tau using single-molecule assays. Amyloid and tau centiloids reflected pathological burden, and hippocampal volume reflected structural integrity. Receiver operating characteristic curves and areas under the curves (AUCs) were used to determine the diagnostic accuracy of plasma biomarkers compared to hippocampal volume and amyloid and tau centiloids. The Mini-Mental State Examination score (MMSE) served as the major cognitive outcome. Logistic stepwise regression was used to assess the overall diagnostic accuracy, combining fluid and structural biomarkers and a stepwise linear regression model for the significant variables for MMSE. For TCP, tau centiloid reached the highest AUC for diagnosis (0.79), while pTau181 could differentiate TCP from YOAD (accuracy 0.775) and LOAD (accuracy 0.806). NFL reflected the clinical dementia rating in TCP, while pTau181 (rho = 0.3487, p = 0.03) and Aβ42/Aβ40 (rho = -0.36, p = 0.02) were significantly correlated with tau centiloid. Hippocampal volume (unstandardized β = 4.99, p = 0.01) outperformed all of the fluid biomarkers in predicting MMSE scores in the TCP group. Our results support the superiority of tau PET to diagnose TCP, pTau181 to differentiate TCP from YOAD or LOAD, and NFL for functional staging.

Keywords: florzolotau; pTau181; primary age-related tauopathy; tau first cognitive tauopathy.

Figure 1

(A) Topographic display of Florzolotau (F18) in the three disease groups compared with the CTL; significance set as a vertex threshold of 3 with parametric Gaussian-based simulations and cluster-wise corrections. (B) Plasma biomarker comparisons in the four groups. (C) Z transformation showing disease-related panels; data presented as mean and standard error bars; LOAD: late-onset Alzheimer disease; YOAD: young-onset Alzheimer disease; TCP: tau-first cognitive proteinopathy; CTL: cognitively unimpaired controls. NFL: neurofilament light. ns: not significant. ** p < 0.01, *** p < 0.001, **** p < 0.0001.

Figure 2

The diagnostic accuracy of the predefined biomarkers and receiver operating characteristic (ROC) curves for binary classifications (disease groups versus CTL). The tested biomarkers were tau centiloid, amyloid centiloid, hippocampal volume (average of left and right hippocampus), pTau181, total tau, NFL, and Aβ42/40 ratio. Areas under the ROC curves (AUCs) of significant biomarkers are reported. LOAD: late-onset Alzheimer disease; YOAD: young-onset Alzheimer disease; TCP: tau-first cognitive proteinopathy; CTL: cognitively unimpaired controls. NFL: neurofilament light. Std. Error: standard error.

Figure 3

Interval plot (mean and 95% confidence interval) of clinical dementia rating (CDR) score and plasma biomarker levels. AD: Alzheimer disease; LOAD: late-onset Alzheimer disease; YOAD: young-onset Alzheimer disease; TCP: tau-first cognitive proteinopathy; NFL: neurofilament light. * p < 0.05, *** p < 0.001.