Abstract A033: Concordance of IHC and a new blood-based expression assay for the detection of PD-L1 in patients diagnosed with NSCLC

• Published in: Molecular cancer therapeutics 2018-01, Vol.17 (1_Supplement), p.A033-A033
• Main Authors: Pritchett, Michael A., Niu, Jiaxin, Jackson, Leisa, Mellert, Hestia, Pestano, Gary A.
• Format: Article
• Language: English
• ISSN: 1535-7163; 1538-8514
• DOI: 10.1158/1535-7163.TARG-17-A033

Introduction: Therapeutic options for patients with non-small cell lung cancer (NSCLC) continue to expand with the approval of immunotherapies. PD-L1 testing can be clinically challenging due to lack of tissue and complexities associated with immunohistochemistry (IHC) including multiple antibodies, various scoring methods, and heterogeneous expression. Moreover, various thresholds have been established for diagnostic tests being used in the context of different checkpoint inhibitors in order to direct clinical practice. Therefore, there is an unmet need for diagnostic tests that measure biomarkers in circulation. We hypothesized that a test that delivered PD-L1 results from plasma read out as continuous variables may be of increased utility in the selection of therapeutic options. Methods: We focused this test development on mechanisms of blood-based testing for sensitive measurement of circulating RNA using ddPCR. Specifically, we optimized methods for the detection of PD-L1 transcripts recovered from platelet-enriched plasma. Specimens for feasibility and development included tumor derived cell lines, activated and resting immune cells, normal donor plasma (n=38), and NSCLC (n = 79) donor plasma. To assess the potential for concordance with tissue testing we collected a total of 43 tissue and blood samples. Tissue results had been previously performed for PD-L1 using IHC (PharmDx 22c3). The analysis excluded samples that lacked detectable CK19 indicating that RNA of epithelial origin was not present in circulation (n=25), and two cases with exceptionally high PD-L1 expression in plasma, leaving 16 samples for further statistical analysis and concordance evaluation. Results: Assessing PD-L1 in circulation is complicated by its expression in both immune and cancer cells. Analytic performance was initially evaluated with cancer cell lines and lymphocytes and monocytes expressing variable levels of PD-L1. Of the 79 NSCLC donor specimens initially evaluated with the RNA blood test, we observed a frequency of 62% positive samples (n=49) with highly variable levels of plasma PD-L1 (2 - 774 copies). We then evaluated a subset of a sample cohort with an IHC tissue test result (n=16). Although there was poor concordance with a 50% positive IHC cut-off, when we used a variable threshold based on a logistic regression score for the blood assay and the 1% cut-off, concordance of up to 80% was observed between the two assays. Conclusions: We have developed sensitive