Abstract 1993: A comprehensive benchmarking of paired whole exome and transcriptome biomarker analysis of response to cancer immunotherapy

Abstract

Abstract Recent advances in cancer immunotherapy have revolutionized cancer treatment. Notably, pembrolizumab - an anti-PD1 checkpoint inhibitor - has been approved as first-line treatment for metastatic non-small cell lung cancer. Clinical response to checkpoint inhibitors has been demonstrated in other cancer types; however, durable response is not observed in all patients. As a result, development of new biomarkers for response to checkpoint inhibitors has become an active area of research. Due to the complexity of the anti-tumor immune response, a single biomarker may not accurately predict patient response to immunotherapy. In parallel, dramatic decreases in the costs of next generation sequencing (NGS) have enabled broad whole exome sequencing (WES) and whole transcriptome sequencing (WTS)-based studies, enabling a new phase in both academic and clinical cancer research. Despite the proven utility of NGS-based biomarker analysis, the accuracy of combined WES/WTS results compared to other assays, specifically in clinical tumor tissues, has not been demonstrated. In this study, we perform comprehensive benchmarking of several immuno-oncology (IO) biomarkers on cell lines, fresh frozen tumor tissues, and formalin-fixed paraffin-embedded (FFPE) tumor tissues. WES and WTS libraries were generated using Illumina Nextera Flex for Enrichment, TruSight Oncology, and TruSeq Stranded Total RNA library prep methods, and sequenced on a NovaSeq 6000. Using an internally developed bioinformatics pipeline, the accuracy of several IO biomarker measurements was evaluated. Specifically, we investigated the performance of tumor-only (T) and/or paired tumor-normal (TN) WES/WTS in assessing tumor mutational burden (TMB), microsatellite instability (MSI), tumor purity, copy number variants (CNV), human leukocyte antigen (HLA) type, fusions, and tumor infiltrating lymphocytes (TILs) levels, among other features. Our benchmarking results demonstrated correlations of 0.99 and 0.98 for TMB as measured by TN and T, respectively (truth: FOCR reported TMB); 100.0% PPV and NPV for MSI status as measured by both TN and T (truth: MSI-PCR); 0.86 correlation for tumor purity estimates based on TN (0.64 based on T; truth: cell-line titration levels); 100% PPV for both TN and T CNV calling (truth: ddPCR); 92% accuracy for both TN and T HLA typing (truth: IHW HLA types); 92.9% sensitivity for fusion calling (truth: ddPCR); and R2 values of 0.87, 0.70, and 0.48 for CD19+, CD4+, and CD8+ TIL levels, respectively (truth: FACS). Taken together, this study demonstrates WES/WTS can not only be utilized for broad exome- and transcriptome-wide biomarker discovery, but can also be an accurate, comprehensive alternative to iteratively testing IO biomarkers. Future studies should further standardize assay and analysis approaches for NGS-based biomarker measurements to ensure consistent and accurate reporting. Citation Format: Mahdi Golkaram, Michael Salmans, Raakhee Vijayaraghavan, Shannon Kaplan, Robert Haigis, Joyee Yao, Kristina Kruglyak, Li Liu, Traci Pawlowski, Sven Bilke, Shile Zhang. A comprehensive benchmarking of paired whole exome and transcriptome biomarker analysis of response to cancer immunotherapy [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 1993.