Immunogenicity of tislelizumab: clinical summary of ADA incidence and feasibility of assessing the suitability of validation cut points using statistical methods.

Abstract

Anti-drug antibody (ADA) is a critical indicator of the immunogenicity of biotherapeutics. To investigate the immunogenicity of tislelizumab, a humanized anti-PD-1 (programmed cell death 1) antibody, 3815 patients were analyzed for ADA incidence. The incidence of ADA is reported for the first time as 19.2 %. For any immunogenicity study, appropriate cut points are essential to accurately distinguish true ADA-positive samples from varying backgrounds across populations. Apart from the standardized false positive rate (FPR) method, statistical methods may be more efficient and powerful in evaluating the appropriateness of cut points. However, a detailed protocol for this approach is not yet in consensus. Therefore, by utilizing extensive data from tislelizumab clinical trials, this study explored the feasibility of assessing the suitability of validation cut points (VCPs) using statistical methods. ADA response datasets from 21 clinical tumor populations were evaluated for distribution differences in comparison to the validation population using statistical methods. Fourteen datasets exhibited significant differences from the validation dataset in both medians and variances, with FPRs exceeding the 2-11 % range. Seven datasets only differed significantly in medians, with five having out-of-range FPRs and their positive rates narrowly affected by cut point alteration. The results showed that statistical methods can effectively assess the suitability of VCPs: in-study cut points are recommended for datasets with significantly different median and variance, while VCPs may be appropriate when only medians differ.