1096P Identifying predictive biomarkers for long term response on immunotherapy in non-small cell lung cancer (NSCLC)

Abstract

Programmed Death 1 (PD-1) pathway targeting immune checkpoint inhibitors (ICIs) are standard of care for advanced NSCLC. Today, histological expression of PD-1 ligand (PD-L1) is the only approved predictive biomarker for response. Yet, ICI treatment can lead to durable or absence of response in patients with low/absent or high PD-L1 resp., indicating that PD-L1 expression is not an ideal biomarker. Further, it remains unclear whether we can define a subgroup that remains progression free after therapy withdrawal. This retrospective pilot study aspires to identify superior biomarkers for long term vs nonresponsive patients and to guide ICI therapy duration. Advanced NSCLC patients with complete and persistent metabolic response (‘responders’, n=5) versus absence of response (‘non-responders’ with >60% PD-L1, n=5) upon ICI therapy were included. Inflammation-related blood biomarkers were compared. From pre-treatment biopsies, RNA and DNA was extracted for respective transcriptional and genomic analysis. Differential gene expression analysis (DESeq2 package) and gene ranking-based gene set enrichment analysis were performed to identify gene oncology biological processes. We used the Mutect2T tool for DNA reads variant calling and tumor mutational burden (TMB) evaluation. While none of the investigated blood biomarkers showed a significant difference between both cohorts, RNA analysis showed 546 significantly expressed genes (p< 0,1). In brief, upregulated genes in the responder group were linked to immune effector response, phagocytosis and myeloid/mast cell activation. Significantly downregulated genes in the responder group were more abundant and mainly linked to epithelial development, cell-cell adhesion, cytoskeletal rearrangements, and endocrine metabolic processes. While no differences in the TMB profiles were observed, the non-responders harbored more mutations in the following genes: TP53, CDKN2A, MSH6, GNAQ and CSMD3. Evaluation of two extremes of the ICI response spectrum in NSCLC patients showed significant differences with biomarker potential in terms of patient stratification and duration of ICI therapy.