Plasma biomarkers for Alzheimer's disease: a field-test in a memory clinic

Abstract

Background: The key Alzheimer's disease (AD) biomarkers are traditionally measured with techniques/exams that are either expensive (amyloid-positron emission tomography (PET) and tau-PET), invasive (cerebrospinal fluid Aβ₄₂ and p-tau₁₈₁), or poorly specific (atrophy on MRI and hypometabolism on fluorodeoxyglucose-PET). Recently developed plasma biomarkers could significantly enhance the efficiency of the diagnostic pathway in memory clinics and improve patient care. This study aimed to: (1) confirm the correlations between plasma and traditional AD biomarkers, (2) assess the diagnostic accuracy of plasma biomarkers as compared with traditional biomarkers, and (3) estimate the proportion of traditional exams potentially saved thanks to the use of plasma biomarkers.

Methods: Participants were 200 patients with plasma biomarkers and at least one traditional biomarker collected within 12 months.

Results: Overall, plasma biomarkers significantly correlated with biomarkers assessed through traditional techniques: up to r=0.50 (p<0.001) among amyloid, r=0.43 (p=0.002) among tau, and r=-0.23 (p=0.001) among neurodegeneration biomarkers. Moreover, plasma biomarkers showed high accuracy in discriminating the biomarker status (normal or abnormal) determined by using traditional biomarkers: up to area under the curve (AUC)=0.87 for amyloid, AUC=0.82 for tau, and AUC=0.63 for neurodegeneration status. The use of plasma as a gateway to traditional biomarkers using cohort-specific thresholds (with 95% sensitivity and 95% specificity) could save up to 49% of amyloid, 38% of tau, and 16% of neurodegeneration biomarkers.

Conclusion: The implementation of plasma biomarkers could save a remarkable proportion of more expensive traditional exams, making the diagnostic workup more cost-effective and improving patient care.

Keywords: CSF; PET; alzheimer's disease; dementia; memory.

Figure 1

Performance of 95%-sensitivity and 95%-specificity plasma biomarkers cutoffs over traditional (A) amyloid (amyloid-PET and CSF Aβ₄₂), (B) tau (tau-PET and CSF p-tau₁₈₁), and (C) neurodegeneration (MRI MTA, FDG-PET) biomarkers. Amyloid-PET positivity: visual reading. CSF Aβ₄₂ positivity: <880.5 pg/mL. Tau-PET positivity: Braak stages IV-VI. CSF p-tau₁₈₁ positivity: >80.5 pg/mL. MTA positivity: age-based cut-off. FDG-PET positivity: SUVr <1.21. One plasma p-tau₁₈₁/Aβ₄₂ value (24.0) and two plasma NfL values (188.1 pg/mL and 260.1 pg/mL) were not displayed to improve data visualisation (but were included in the analyses). CSF, cerebrospinal fluid; CU, cognitively unimpaired; FDG, fluorodeoxyglucose; MCI, mild cognitive impairment; NfL, neurofilament light; PET, positron emission tomography; SUVr, standardised uptake value ratio.

Figure 2

Proportion of traditional exams that could be saved by using plasma biomarkers with 95%-sensitivity or 95%-specificity cut-offs. Values indicate the proportion of saved traditional exams and the ‘error rate’. The proportion of saved traditional exams was defined as the number of patients identified with the use of plasma biomarkers with 95%-sensitivity and 95%-specificity cut-offs (ie, true negatives+false negatives resulting from the 95%-sensitivity cut-off, and true positives+false positives resulting from the 95%-specificity cut-off) out of the total number of participants with plasma and traditional biomarkers. The ‘error rate’ is defined as the number of traditional exams incorrectly saved (ie, false negatives resulting from the 95%-sensitivity cut-off+false positive resulting from the 95%-specificity cut-off) out of the total number of the saved traditional exams. CSF, cerebrospinal fluid; FDG, fluorodeoxyglucose; NfL, neurofilament light; PET, positron emission tomography.