NEUROFILAMENT LIGHT CHAIN AND PHOSPHOTAU/TAU RATIO AS CSF BIOMARKERS IN FRONTOTEMPORAL DEMENTIA

Abstract

Neurofilament light chain (NfL) in cerebrospinal fluid (CSF) is elevated and the phosphotau/tau ratio (p/t-tau) decreased in frontotemporal dementia (FTD), and are both potential biomarkers for disease severity and prognosis. Additionally, p/t-tau may discriminate FTD with TDP43 inclusions (FTLD-TDP) from FTD with tau pathology (FTLD-tau). We examined the diagnostic potential of these individual biomarkers in a large and clinically well-defined FTD cohort. CSF NfL and p/t-tau levels were compared between 45 cognitively healthy control subjects, 181 behavioral variant FTD (bvFTD), 16 FTD with motor neuron disease (FTD-MND), 36 semantic dementia (SD), 19 progressive non-fluent aphasia (PNFA), 42 corticobasal syndrome (CBS), and 64 progressive supranuclear palsy (PSP) patients. In 73 patients, a definite FTD was diagnosed (underlying pathology: TDP n=50, tau n=23). CSF NfL levels were higher in all patient groups than in controls (p<0.001, sensitivity 79%, specificity 89%). CSF NfL levels were equally elevated in bvFTD, SD, PNFA, CBS, and PSP; FTD-MND patients had higher NfL levels than all other patients. CSF p/t-tau was lower in all patients than controls (p<0.001, sensitivity 73%, specificity 93%), with the lowest values in FTD-MND patients. CSF NfL did not discriminate between underlying TDP and tau pathology (p=0.08). The p/t-tau ratio had a high specificity (81%) and moderate sensitivity (65%) to discriminate FTLD-TDP from FTLD-tau patients. In all patients combined, High NfL and low p/t-tau levels were associated with a high CDR-SB (rs=0.38, p=0.005 for NfL, rs=-0.29 for p/t-tau) and with a poor survival (estimated hazard ratio on tertiles 1.7 for NfL and 0.7 for p/t-tau). We confirmed that CSF NfL and the p/t-tau ratio are potential biomarkers for disease severity and prognosis in FTD and are therefore interesting surrogate endpoints in clinical trials. Both biomarkers discriminated FTD from controls, but not the individual subtypes, apart from FTD-MND. Additionally, the p/t-tau ratio was specific to discriminate TDP from tau pathology in vivo.