Plasma and CSF Neurofilament Light Chain in Amyotrophic Lateral Sclerosis: A Cross-Sectional and Longitudinal Study

Background: Neurofilament light chain (NfL) is a validated biofluid marker of neuroaxonal damage with great potential for monitoring patients with neurodegenerative diseases. We aimed to further validate the clinical utility of plasma (p) vs. CSF (c) NfL for distinguishing patients with Amyotrophic Lateral Sclerosis (ALS) from ALS mimics. We also assessed the association of biomarker values with clinical variables and survival and established the longitudinal changes of pNfL during the disease course. Methods: We studied 231 prospectively enrolled patients with suspected ALS who underwent a standardized protocol including neurological examination, electromyography, brain MRI, and lumbar puncture. Patients who received an alternative clinical diagnosis were considered ALS mimics. We classified the patients based on the disease progression rate (DPR) into fast (DPR > 1), intermediate (DPR 0.5-1), and slow progressors (DPR < 0.5). All patients were screened for the most frequent ALS-associated genes. Plasma and CSF samples were retrospectively analyzed; NfL concentrations were measured with the SIMOA platform using a commercial kit. Results: ALS patients (n = 171) showed significantly higher pNfL (p < 0.0001) and cNfL (p < 0.0001) values compared to ALS mimics (n = 60). Both cNfL and pNfL demonstrated a good diagnostic value in discriminating the two groups, although cNfL performed slightly better (cNfL: AUC 0.924 ± 0.022, sensitivity 86.8%, specificity 92.4; pNfL: AUC 0.873 ± 0.036, sensitivity 84.7%, specificity 83.3%). Fast progressors showed higher cNfL and pNfL as compared to intermediate (p = 0.026 and p = 0.001) and slow progressors (both p < 0.001). Accordingly, ALS patients with higher baseline cNfL and pNfL levels had a shorter survival (highest tertile of cNfL vs. lowest tertile, HR 4.58, p = 0.005; highest tertile of pNfL vs. lowest tertile, HR 2.59, p = 0.015). Moreover, there were positive associations between cNfL and pNfL levels and the number of body regions displaying UMN signs (rho = 0.325, p < 0.0001; rho = 0.308, p = 0.001). Finally, longitudinal analyses in 57 patients showed stable levels of pNfL during the disease course. Conclusion: Both cNfL and pNfL have excellent diagnostic and prognostic performance for symptomatic patients with ALS. The stable longitudinal trajectory of pNfL supports its use as a marker of drug effect in clinical trials.