Excessive outlier removal may result in cut points that are not suitable for immunogenicity assessments

Abstract

Cut point determination is an important aspect of immunogenicity assay development. The cut point can be influenced by a myriad of factors. Key among those is the analytical variability of the assay itself and biological variation due to test samples. Since a smaller cut point value may result in improved sensitivity, the existing procedures often employ statistical techniques such as outlier removal to produce a conservative cut point. Although such practices are intended to yield acceptable assay sensitivity, they may fail to fully account for biological variability in the data, thus generating higher than expected number of false positive results. In this paper, we introduce the concept of minimum cut point. It is defined as the cut point that is determined in the absence of biological variability. Under the log-normal assumption of the data used for cut point analysis, closed-form formulas are derived for the minimum cut point. This minimum cut point can be used to benchmark whether a cut point derived from a procedure can compromise assay specificity by being too low.