Cancer pathways analysis and correlation with survival in patients with advanced stage non-small cell lung cancer treated with PD-1 inhibitor.

e21007 Background: Immune checkpoint inhibitors (ICIs) targeting programmed cell death protein-1 (PD-1) have demonstrated activity in the first-line setting of advanced stage non-small cell lung cancer (NSCLC); however, a proportion of patients progresses to immunotherapy, hence the definition of prognostic and predictive factors is of interest. Potentially, the expression pattern of gene sets can serve as clinical indicator. We hereby present data from a pilot study of a larger project of gene expression analysis in patients with advanced NSCLC treated with ICIs, to explore the baseline expression level of single and grouped-in-pathways genes and to draw correlations with the overall survival (OS). Methods: We performed a retrospective transcriptome analysis of diagnostic cancer tissue from 14 consecutive patients with advanced stage NSCLC with PD-L1 expression > 50% treated with anti-PD-1 pembrolizumab as first-line. Two hundred ng of RNA was profiled by Nanostring technology using the human nCounter® PanCancer IO 360™ Panel, that includes 770 genes from 14 different immune cell types, common checkpoint inhibitors, Cancer-Testis antigens, and genes covering both the adaptive and innate immune response. Data analysis was performed by nCounter Advanced Analysis (version 2.0.134) plug in for nSolver Software and R package. Results: The median age of patients was 71 years (range 59-88), most were males (78%) with an ECOG performance status ranging from 0 to 1 (86%). All patients were smokers or former smokers. Seven out of 14 patients (50%) had a partial/complete response as best response, whereas the remaining population reported disease progression, 3 of whom died before the first CT-scan evaluation. The median OS was 12 months (range 1–40). The 770-tumor/immune gene expression profiling reported 11 genes (6 overexpressed and 5 down-modulated in patients with a survival under the median), which were significantly associated with poor outcome ( p-value < 0.01). Notably, when we adjusted the expression data to the site of biopsy (tumor tissue vs. neoplastic lymph node), we identified a 27-gene signature ( ATG5, BMI1, CASP1, CCL11, CCL8, CD164, CD86, CD97, EGR1, FADD, FOS, HLA-DQA1, HLA-DQB1, IFITM1, IRF3, LY96, MCAM, NFKBIA, RRAD, SMAD2, STAT6, TLR3, TNFAIP3, TNFRSF1A, TOLLIP, VEGFC, and YTHDF2) able to separate patients with a survival below or above the median. The in-silico prediction of this signature showed that the antigen processing pathway (HLA-DQA1, HLA-DQB1), already described as an immune checkpoint resistance mechanism, was significantly down-modulated in long survivors. Conclusions: By a multiplexed analysis of gene expression through messenger RNA digital detection, we identified a gene set that, if confirmed, might have a prognostic value in patients with advanced stage NSCLC with PD-L1 expression >50% treated with pembrolizumab as first-line.