Plasma biomarkers of amyloid, tau & neuroinflammation in Alzheimer's disease and corticobasal syndrome

Abstract

Background: Blood-based biomarkers (BBBMs) could significantly facilitate the diagnosis of Alzheimer's disease (AD) and non-AD dementia by providing less invasive alternatives to cerebrospinal fluid (CSF) and positron emission tomography (PET) imaging.

Objective: This study investigated how well the BBBMs-amyloid-β (Aβ) 1-42/1-40 ratio, phosphorylated tau181 (pTau181), apolipoprotein E4 (ApoE4), glial fibrillary acidic protein (GFAP), and neurofilament light chain (NfL)-reflect thorough clinical work-up validated by PET and CSF biomarkers in participants with AD (n = 27), Aβ-negative CBS (n = 26), and agematched healthy controls (HC) (n = 17).

Methods: Factor and correlation explored biomarker associations. Bayesian regression, backward selection regression, and ROC curve analysis were applied to identify optimal biomarker combinations and diagnostic cut-offs.

Results: In AD cases, pTau181 and ApoE4 levels were elevated, and the Aβ1-42/1-40 ratio was reduced. ROC analysis showed high accuracy for pTau181, ApoE4 and Aβ1-42/1-40 in discriminating AD from HC, with a combination significantly improving performance. However, limited fold change, and high variability reduced the diagnostic applicability of Aβ1-42/1-40 ratio. Elevated NfL levels were the most reliable biomarker for CBS-Aβ(-) cases. GFAP showed limited discriminatory power due to overlapping levels, suggesting that it may not serve as a disease-specific biomarker but may be indicative of general neurodegeneration.

Conclusions: This study highlights the diagnostic utility of pTau181, ApoE4 and the Aβ1-42/1-40 ratio for AD and NfL in the CBS-Aβ(-) cases and emphasizes the added value of combined biomarker models for group differentiation. Prospective studies will help validate these findings and refine clinical thresholds.

Keywords: Apolipoprotein E (ApoE4); Beta-amyloid 1-40 (Aβ1-40); Beta-amyloid 1-42 (Aβ1-42); Glial fibrillary acidic protein (GFAP); Neurofilament light chain (NfL); Non-Alzheimer's disease dementia; beta; Phosphorylated tau (pTau).

Fig. 1

Recruitment and Group Classification. Of 138 participants, n = 31 withdrew their consent, n = 21 were excluded from this analysis due to missing data, n = 4 due to a diagnosis of small vessel disease (SVD), n = 4 due to indeterminate disease. In n = 2 cases amyloid PET was available but not CSF, in n = 5 cases CSF was available but not amyloid PET. One amyloid PET-negative case had a CSF amyloid ratio below 0.055, presented clinically with CBS and was classified as a CBS-Aβ( +) case. N = 78 participants underwent magnet resonance imaging. 76% were right-handed, 9% left-handed, 5% indifferent and 10% missing data

Fig. 2

Blood-based Biomarkers—Raw Data and Group Differences. Blood-based biomarkers showing raw data and group differences for pTau181, Aβ1-42/1-40 ratio, ApoE4, NFL, and GFAP. Statistical significance: *p < 0.05, **p < 0.01, ***p < 0.001 (FDR-adjusted)

Fig. 3

CSF Markers—Raw Data and Group Differences. CSF markers illustrating raw data and group differences for Aβ1-42/1-40 ratio, pTau181, tTau, and sTREM. Statistical significance: *p < 0.05, **p < 0.01, ***p < 0.001 (FDR-adjusted)

Fig. 4

Correlation of Blood-based Biomarkers and CSF Markers. Correlation analysis of blood-based biomarkers with CSF markers. Statistical significance: *p < 0.05, **p < 0.01, ***p < 0.001 (FDR-adjusted)

Fig. 5

Blood biomarker-based Prediction versus CSF/PET-based Diagnosis. Comparison of blood biomarker-based predictions versus CSF/PET-based diagnoses. AUC values indicate the predictive power of scaled biomarkers and combined models