Association of Plasma Neurofilament Light Chain With Diagnostic Classification in a Neuromuscular Referral Cohort: A Cross-Sectional Study.

Abstract

Plasma neurofilament light chain (pNfL) is a blood-based biomarker of axonal injury elevated in several neuromuscular disorders. While previous studies have compared individual diseases with healthy controls, less is known about how pNfL performs across clinically relevant neuromuscular presentations encountered in subspecialty practice, particularly in distinguishing active from inactive neuropathies. We evaluated whether abnormal age-adjusted pNfL concentrations are associated with clinically relevant diagnostic categories and whether pNfL discriminates between active and inactive neuropathies. We conducted a cross-sectional study of patients evaluated by neuromuscular specialists at Mayo Clinic (Rochester, MN) between April 2022 and April 2024, in whom pNfL was ordered as part of routine clinical care. Diagnoses were assigned to prespecified categories by consensus review, including motor neuron disease (MND), active neuropathy, inactive neuropathy, and structural etiologies. Clinically active neuropathy was defined using prespecified, diagnosis-specific operational criteria incorporating treatment status, clinical relapse or progression, time from symptom onset, and pathologic evidence of inflammation. Abnormal pNfL was defined using previously established age-adjusted 97.5th percentile reference limits. Multinomial logistic regression evaluated associations between abnormal pNfL and diagnostic category, adjusting for neuropathy impairment score and symptom duration. Receiver operating characteristic (ROC) analyses assessed discrimination between prespecified groups. Among 306 patients (mean age 62.2 years, 41.5% female), 179 (58.5%) had MND, 28 (9.1%) active neuropathy, 75 (12.1%) inactive neuropathy, and 24 (7.8%) structural etiologies. Median pNfL concentrations differed across diagnostic groups (<i>p</i> < 0.0001) and were highest in MND and active neuropathy. After adjustment, abnormal pNfL was associated with a higher relative likelihood of classification as MND (relative risk ratio [RRR] 18.47, 95% CI 8.77-38.86) and active neuropathy (RRR 7.30, 95% CI 2.50-21.34) compared with inactive neuropathies. No association was observed with structural etiologies. Discrimination between active and inactive neuropathy was moderate (area under the curve [AUC] 0.79, 95% CI 0.70-0.88), and discrimination for MND vs all other diagnoses was good (AUC 0.85, 95% CI 0.80-0.90). In this cross-sectional cohort, abnormal age-adjusted pNfL concentrations were more frequently observed in diagnostic categories characterized by clinically active axonal injury than in inactive neuropathies and structural conditions. These findings suggest a potential role for pNfL as an adjunctive tool in neuromuscular diagnostic evaluation. Interpretation is limited by the cross-sectional design and convenience sampling. This study provides Class II evidence that abnormal age-adjusted pNfL distinguishes active neuropathies from inactive neuropathies with moderate diagnostic accuracy.