Cerebrospinal fluid amyloid and neurodegeneration biomarkers in the diagnosis and prognosis of amyotrophic lateral sclerosis: a multicohort observational study.

Abstract

Neurofilament light chain (NfL) is an established biomarker for amyotrophic lateral sclerosis (ALS). However, it remains unclear whether biomarkers reflecting amyloid-β metabolism (Aβ), tau (phosphorylated tau, pTau), and astrogliosis (glial fibrillary acidic protein, GFAP) can complement NfL in clinical applications. It is also unclear how Alzheimer's disease (AD) co-pathology affects disease progression in ALS. We studied 264 patients with ALS and 164 disease-control participants, analysing baseline cerebrospinal fluid (CSF) NfL, Aβ peptides, pTau181, pTau217, and GFAP. A two-step diagnostic strategy was examined in individuals with intermediate NfL values, using machine-learning models. Prognostic associations were evaluated using linear mixed-effects and Cox models adjusted for established predictors; survival was additionally examined in pooled analyses including an external ALS cohort (n = 153). Patients with ALS exhibited higher CSF NfL, Aβ38, Aβ40, Aβ42, pTau181, and pTau217 than controls, even within Aβ-negative subgroups. NfL discriminated ALS from controls with an AUC of 0.85 (0.82-0.88), which increased to 0.90 (0.88-0.92) in machine-learning models incorporating additional biomarkers. The machine-learning models maintained a high positive predictive value (0.90) for classifying cases with intermediate NfL values. Significant Aβ × NfL interactions were observed for longitudinal functional decline, and pooled survival analyses showed stronger associations between NfL and survival in the Aβ-positive ALS subgroup (14%) than in the Aβ-negative ALS subgroup (HR 4.02 [2.17-7.45] vs 1.62 [1.36-1.92]). Multiple biomarker assessments can reduce misclassification of individuals with intermediate NfL values. CSF Aβ positivity modified the association between baseline NfL and disease progression. Personalised diagnosis and prognosis based on fine-grained co-pathology evaluation may enhance clinical management of ALS. This study was primarily supported by JST SPRING (JPMJSP2125). A complete list of funding organisations is provided in Acknowledgements.