Diagnostic utility of neurofilament markers for MND is limited in restricted disease phenotype and for differentiation from compressive myeloradiculopathies.

Abstract

Misdiagnosis is frequent in early motor neuron disease (MND), typically compressive radiculopathy, or in patients with restricted MND phenotype. In this retrospective, single tertiary centre study, we measured levels of neurofilament light (NfL) and phosphorylated neurofilament heavy (p-NfH) chain in cerebrospinal fluid (CSF) and of p-NfH in serum with commercially available ELISA kits and assessed their respective diagnostic performance as a marker of MND. The entire study population (n = 164) comprised 71 MND patients, 30 patients with compressive myelo- or radiculopathy, and 63 disease controls (DC). Among MND patients, we specified subgroups with only lower motoneuron involvement (MND-LMN, n = 15) and with confounding nerve roots or spinal cord compression (MND-C, n = 18), representing clinical diagnostic pitfalls. MND-LMN displayed significantly lower CSF NfL (p = 0.003) and p-NFH (p = 0.017), but not serum p-NfH (p = 0.347) levels compared to other MND patients (n = 56). The discriminative ability (area under the curve-AUC) of both CSF Nfs towards all MND patients was comparable to each other but significantly higher than that of p-NfH in serum (ps < 0.001). AUC of both CSF Nfs between MND-LMN and DC and also between MND-C and myelo-/radiculopathies were reduced, as compared to AUC between other MND and DC or myelo-/radiculopathies, respectively. Our results suggest that both Nfs in CSF represent a reliable diagnostic marker in a general MND population, fulfilling Awaji criteria. As for diagnostic pitfalls, and also for p-NfH in serum, their discriminative ability and, therefore, clinical utility appears to be limited.