Abstract 1566: Establishment of analytical performance features of two immunohistochemistry assays for PD-L1 expression in lung cancer

Abstract Therapies that target PD-1/PD-L1 and overcome checkpoint inhibition have shown great promise in the treatment of a variety of cancers, with PD-L1 expression demonstrated as a biomarker of therapy response in a variety of tumor types. There are multiple IHC assays that are used to measure PD-L1 expression in formalin-fixed, paraffin-embedded tissues (FFPE) with significant variability in primary antibody, assay platform, scoring criteria, and assay cut-offs. This situation creates potential confusion in applications of PD-L1 immunohistochemistry (IHC) assays for the appropriate intended use. In this study we evaluated two assays (Dako pharmDx PD-L1 28-8 assay and Ventana PD-L1 SP263) in 45 non-small cell lung cancer samples encompassing predominately squamous cell carcinomas (35) and adenocarcinomas (10). PD-L1 protein expression, measured by the percentage of tumor cell showing positivity for PD-L1 staining, has been compared between two assays. PD-L1 results were interpreted by using various assay cutoffs. Correlation of PD-L1 expression measured by positivity in percentage of tumor cell staining with each of the two assays achieved a significant concordance (r2 = 0.91). Similar trends in percentage of positive of tumor cells were demonstrated across a broad dynamic range of staining. Cutoffs that have been established in clinical studies and referenced in the respective assay package inserts (>1% positivity percentage of tumor cells for 28-8 assay and >25% positivity percentage of tumor cells for the SP263 assay) were used as the reference, for interpreting the test results as positive or negative. When these established cutoffs were applied, a low-to-medium concordance between two assays has been demonstrated (r2 = 0.58). When the same cutoff (1% or 25%) was applied, the correlation has been improved markedly (r2 = 0.73 or 0.95 respectively). In this sample set the 25% cutoff produced the more consistent data when comparing the two methods. Changes in the classification of PD-L1 status was also observed for both Dako pharmDx PD-L1 28-8 assay (29%, 13/45) and Ventana PD-L1 SP263 (22%, 10/45) assays when the established cutoff was replaced. The findings of this study are consistent with others such as those from Blueprint phase I study. This comparison study showed that the two assays (Dako pharmDx PD-L1 28-8 assay and Ventana PD-L1 SP263) were analytically comparable. Further investigation is warranted to explore both the analytical and clinical performance of the various PD-L1 assays across multiple assay cut-points. Citation Format: Lei Zhang, Claire Begot, Robert McGee, Steven M. Anderson. Establishment of analytical performance features of two immunohistochemistry assays for PD-L1 expression in lung cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 1566.