High Concordance Observed Between Local and Central INI1 Immunohistochemistry Results Obtained From Clinical Studies EZH-102 and EZH-202

Abstract Objectives A defining feature of epithelioid sarcomas, malignant rhabdoid tumors, renal medullary carcinomas, and poorly differentiated chordomas is the loss of INI1 expression in tumor cells. Dependence on EZH2 through transcriptional repression caused by aberrant H3K27me3 is one consequence of INI1 expression loss. Tazemetostat, a potent, selective, oral EZH2 inhibitor, has demonstrated tumor regressions in INI1-negative preclinical models and durable objective responses in adult and pediatric patients with INI1-negative tumors. Methods The phase 1 pediatric EZH-102 (NCT02601937) and phase 2 adult EZH-202 (NCT02601950) multicenter trials evaluate the effect of tazemetostat on patients with INI1-negative tumors. Trial eligibility required local pathological documentation demonstrating INI1 loss via an INI1 immunohistochemistry (IHC) test performed in a CAP/CLIA-certified laboratory or molecular evidence supporting biallelic loss of INI1. To confirm local evidence of tumor INI1-negative status, all enrolled patients were required to provide archival tumor for central INI1 testing. The Children’s Hospital Los Angeles performed central confirmation using their validated INI1 IHC protocol. Results A survey of active clinical trial sites demonstrated that INI1 IHC methods varied among local pathology laboratories, primarily through differences in the choice of INI1 antibody clone or autostainer selected. Central INI1 IHC testing of EZH-102 (n = 35) and EZH-202 (n = 97) cases from 40 clinical sites was completed using a single protocol. Of the 128 patients enrolled based on local INI1 IHC, concordance with central INI1 IHC results was 98% (126/128). Four patients were enrolled based on local genetic data supporting biallelic loss with all four demonstrating tumor INI1 loss when centrally tested. Conclusion The high concordance rate between local pathology labs and central testing demonstrates that differences in methods at local laboratories to identify INI1-negative tumors through the INI1 IHC tests are likely comparable between institutions enrolling patients to EZH-102 and EZH-202.