Association of <i>DDR1</i> with immune exclusion and outcomes in non-small cell lung cancer.

e20553 Background: Discoidin domain receptor 1 (DDR1) is a collagen receptor with tyrosine kinase activity. Although DDR1 has been reported to implicated in cancer progression and metastases, the underlying mechanism remains poorly characterized and its prognostic value is also contradictory in non-small cell lung cancer (NSCLC). Methods: Clinical and DDR1 mRNA expression data of 1925 patients with NSCLC were obtained from 14 publicly available datasets utilizing an online tool KM-plotter. All patients were split into 2 groups by median expression of DDR1. The Kaplan-Meier curve and log-rank test, and Cox regression analyses were used to examine the association between DDR1 expression and overall survival (OS). The table shows 48 immune-regulatory genes classified into 4 signatures associated with activated T cells, immune cytolytic activity, and antitumor immunity. Expression data of these genes for 994 patients with NSCLC were gathered from the TCGA Pan-Cancer Atlas. Abundance data of various immune cell types in tumor microenvironment for the TCGA patients were inferred using xCell, and then immune infiltration extent was calculated as immune score. Associations of DDR1 with immune genes expression and immune score were determined with Spearman rank correlation. Results: DDR1 overexpression was significantly associated with inferior OS in this large-scale cohort (median, 62 vs 78.5 months; hazard ratio [HR], 1.16; 95% CI, 1.02-1.31; P = .02). Of note, multivariable Cox model indicated DDR1 overexpression was an independent risk factor for OS (HR, 1.41; 95% CI, 1.01-1.98; P = .046), controlling for potential confounders (histology, stage, sex, and smoking history). Of the 48 immune-regulatory genes, 44 (92%) showed significant negative correlations with DDR1 expression (all P < .05), implying the role of DDR1 in decreasing immune infiltration into tumor microenvironment. Further, higher immune score was observed in tumors underexpressing DDR1 (correlation coefficient, -0.44; P < .001), more directly demonstrating the link between DDR1 and immune infiltration. Conclusions: To our knowledge, this is the largest study to evaluate the prognostic role of DDR1 expression in NSCLC. Our results suggest that DDR1 overexpression may predict poor outcomes and a possible mechanistic explanation regarding this role may be immune exclusion instigated by DDR1. Thus, targeting DDR1 expression to abrogate immune exclusion warrants further investigation.[Table: see text]