Plasma Alzheimer Biomarkers and Diagnostic Decision-Making in Memory Clinics.

Abstract

Blood-based biomarkers (BBM) provide minimally invasive, scalable, lower-cost tools for identifying neurodegenerative diseases, but prospective data on their clinical validity in memory clinic settings are limited. To evaluate how a tailored plasma BBM panel (phosphorylated tau 181 [pTau181], glial fibrillary acidic protein [GFAP], and neurofilament light chain [NfL]) during multidisciplinary diagnostic meetings is associated with syndrome diagnosis, suspected etiology, and clinician confidence. This prospective diagnostic study enrolled consecutive patients whose BBM were presented during weekly multidisciplinary meetings after standard workup (clinical assessment, neuropsychological testing, and brain magnetic resonance imaging) from September 2023 to October 2024 at 3 academic memory clinics in the Netherlands. When available, cerebrospinal fluid (CSF) and amyloid positron emission tomography (PET) results were subsequently shown. Findings were categorized as high, intermediate, or low probability for Alzheimer disease (AD), frontotemporal lobar degeneration (FTD), or dementia with Lewy bodies (DLB). Data were analyzed from December 2024 to July 2025. BBM (pTau181, GFAP, and NfL) measured weekly and analyzed jointly as a diagnostic panel. Outcomes of interest were changes in suspected syndrome diagnoses, primary etiology, and clinician confidence before vs after BBM disclosure. A total of 450 patients (mean [SD] age, 66 [10] years; 183 [41%] female; mean [SD] MMSE score, 25 [5]) were enrolled. Among 356 patients (79%) with AD as the primary suspected etiology, assessment of BBM classified 149 patients (42%) as high, 101 patients (28%) as intermediate, and 106 patients (30%) as low probability of AD. Median (IQR) diagnostic confidence in the total cohort increased from 80% (70%-90%) to 90% (70%-90%) after BBM disclosure (P < .001), increasing in 207 patients (46%), unchanged in 175 patients (39%), and decreasing in 68 patients (15%). Following BBM disclosure, syndrome diagnoses were revised in 6 patients (1%) and primary etiology was revealed in 23 patients (5%): 11 diagnoses (2%) shifted to AD, 3 diagnoses (1%) from AD to no neurodegeneration, 2 diagnoses (<1%) from AD to FTD, and 7 diagnoses (2%) became unclear. Among 450 patients, 234 (52%) had CSF and amyloid PET results. Using this as reference, BBM analyzed with an amyloid-positive vs amyloid-negative tool identified 76% amyloid positives and 85% amyloid negatives for high- and low-probability results, respectively; intermediate results occurred in 24% and 39%, respectively. In this prospective diagnostic study, a BBM panel was associated with altered etiologic diagnoses in a few patients and was associated with increased diagnostic confidence overall. These findings suggest that BBM may help refine the diagnostic process within specialized academic memory clinics.