The second assessment of the commutability of candidate reference materials for the harmonization of blood neurofilament light chain measurements

Abstract Objectives Blood neurofilament light (NfL) is a sensitive biomarker of neuroaxonal injury, but substantial differences in absolute concentrations across analytical platforms underscore the need for assay standardization. This study aimed to evaluate the commutability of candidate reference materials (RMs) generated by spiking serum pools with cerebrospinal fluid (CSF) to support global assay harmonization. Methods The study included 45 individual clinical serum samples and 30 candidate RMs. The RMs comprised three serum pools, each combined with each of three CSF pools at three concentration levels, in addition to the three unspiked serum pools. Commercial NfL measurement procedures (MPs) from six manufacturers (Roche, Fujirebio, Meso Scale Discovery, Bio-Techne, UmanDiagnostics, and Quanterix) were evaluated. Commutability was assessed using a difference in bias approach. Results Spearman correlations between MPs were excellent (ρ=0.96–0.99), although regression slopes differed substantially, with differences up to a 46-fold. Across the 450 RM assessments, 62 % were classified as commutable, 37 % as indeterminate, and 0.66 % as non-commutable. Approximately two-thirds of the uncertainty in the difference in bias was attributable to MP imprecision rather than sample specific effects, suggesting that standardization of NfL measurements is feasible. Conclusions Spiking serum pools with CSF appears to be an effective strategy for consistently producing commutable RMs for NfL measurement. Although the relatively high proportion of indeterminate results indicates a need for improved MP precision or additional replicate measurements, the study supports continued development of SI-traceable reference materials to facilitate standardization of NfL measurements in clinical practice.