Abstract 1626: Technical variability in NGS immune gene expression and mutation profiling has a nominal effect on tumor classification

Abstract

Abstract Background: A custom NGS cancer immune gene expression assay was developed which measures the transcript level of >350 genes involved in T-cell receptor signaling (TCRS), tumor infiltrating lymphocyte (TILs) complement as well as other key targets expected to predict the likelihood of patient response to checkpoint inhibitors (CPI). In parallel to the gene expression assay, mutational profiling was carried out using the 409 gene Comprehensive Cancer Panel (ThermoFisher). As variability between runs is common when performing NGS assays a detailed comparison of specific technical variations were assessed for their ability to effect gene expression and mutation profiles of clinical FFPE samples. Methods: Studies were designed to characterize the analytical performance of the immune response NGS assay using RNA and DNA from a subset of 300 FFPE tissues representing NSCLC, melanoma, renal cell carcinoma and bladder cancer. As part of the study, we tested the impact of variability in RNA and DNA input quantity at the library preparation step, sample batch size which affects mapped reads/sample and depth of coverage, and linearity of expression and sensitivity of mutation profiling through serial dilutions of pico-molar (pM) input of normalized library. PCA and unsupervised clustering was performed on samples with checkpoint inhibition, TCRS and TILs genes as well as mutational profiling to reveal sample groups with three distinct immune signatures (low, indeterminate and high). Further correlation and over-representation analysis was performed to determine impact of technical characteristics on these three immune signatures. Results: Immune signatures including mutation profiles and gene expression levels were maintained throughout variable RNA/DNA input amounts at the library generation level as well as with diminution of pM levels of library pooled at the sequencing step. Increase in the number of mapped reads and sequencing depth through decreasing the number of batched samples per sequencing run also did not affect the gene expression and mutation profile signatures of the FFPE derived samples. Conclusion: The gene expression and mutation profiles responsible for classifying FFPE samples using NGS are not affected by variation normally introduced in the technical workflow commonly associated with these platforms. The analytical assessment of input at the nucleic acid, library, and sample size level has shown the plasticity available when using amplicon based NGS technologies for classifying the immune gene expression signature as well as mutational profiles of FFPE derived clinical tumor samples. This flexibility increases the strength and utility of NGS-base gene expression profiling and mutational analysis of tumor samples for both basic research and clinical applications. Citation Format: Sean Glenn, Jeffrey Conroy, Blake Burgher, Sarabjot Pabla, Maochun Qin, Jon Andreas, Vincent Giamo, Marc Ernstoff, Mary Nesline, Ji He, Mark Gardner, Carl Morrison. Technical variability in NGS immune gene expression and mutation profiling has a nominal effect on tumor classification [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 1626. doi:10.1158/1538-7445.AM2017-1626