Immune infiltrate characterization and differential gene expression by RNA-sequencing analysis in patients with oncogene-addicted NSCLC with early progression under targeted therapy.

8535 Background: Despite remarkable advances with tyrosine kinase inhibitors (TKI), patients (pts) with oncogene addicted advanced non-small cell lung cancer (aNSCLC) eventually face resistance. Moreover, some pts exhibit early TKI resistance, often linked to a reduced response subsequent TKIs. These pts might have unique biology, and identifying post-resistance treatments for them is an unmet need. Here, we explore the transcriptomic profile of pts with aNSCLC and early TKI progression, as compared to long responders. Methods: We included pts with aNSCLC harboring EGFR mutations or oncogenic fusions ( ALK, ROS1, RET, NTRK), treated with TKI at Gustave Roussy and enrolled in the prospective MATCH R trial (NCT02517892). Clinical data on treatment outcomes and survival were collected. Pts with PFS 18 months as long responders (LR). At least one tissue biopsy was performed at PD, on which bulk RNA sequencing was carried out. We explored the transcriptomic differences between pts with EP and LR. Differential gene expression (DESeq2), pathway enrichment studies and identification with MSigDb were performed and compared between groups. Immune cell type contents were evaluated by CIBERSORT deconvolution method. Results: From 2015 to 2022, 37 pts were included. 68% were females, median age was 55 (IQR43-83). 18 carried EGFR mutations (EP N=5, 28%; LR N=13, 72%), 19 had fusions (EP N=10, 52%; LR N=9, 48%). All pts had progressed to targeted TKI (7 first generation EGFR-TKI, 11 osimertinib, 8 crizotinib, 7 second generation ALK-TKI, 2 lorlatinib, 2 pralsetinib, 1 repotrectinib). Immune suppressive macrophages were the most represented immune cells in EGFR+ and fusion+ PD samples (33.5% and 33.2% of immune cells). The composition of the immune infiltrate was not significantly different between EP and LR in EGFR+ and fusion+ pts. DESeq2 showed that EGFR+ EP had significantly lower EPCAM(log2FC: 1.60, padj=0.01), higher SOS1 (log2FC: -1.18, padj=0.01), higher JAK2(log2FC: -1.56, padj=0.02) expression, as compared to EGFR+ LR. In the fusion+ group, EP had higher FGFR3/4 (log2FC: -3.5, padj=0.02) and lower MTAP expression (log2FC: 1.75, padj=0.03). Hallmark analysis in the EGFR+ showed higher activation of TNFα, IFNγ response, and KRAS pathways in EP (NES: -1.72, p=0.03; NES=-1.78,p=0.02; NES:-2.07, p=0.01) versus LR. No differential activation was observed in the fusion+ group. Conclusions: Immune suppressive macrophages are the most represented immune cells at PD to TKI, irrespective of the treatment duration. Pts with EP seem to have a different biology, with transcriptomic features of stemness and immune suppression in the EGFR+ pts, and lower MTAP expression in the fusion+ group. This should be further explored for the development of novel therapeutics.