Transcriptional Profile Of Immune Microenvironment And Their Prediction Role For The Prognosis Of Esophageal Squamous Cell Carcinoma

416 Background: Esophageal squamous cell carcinoma (ESCC) is one of the most common cancer in China. The genetic characterizations have already been described in many studies, but the immune microenvironment features were seldom reported. Our study was aimed to explore the relationship between the immune profile in stage IIIa ESCC and patients’ prognosis. Methods: 20 eligible stage IIIa ESCC patients with received surgery and radiotherapy (19 with and 1 without radiotherapy). RNA targeted sequencing was performed on 20 primary tumor specimens. Transcripts of 395 immune-related genes expressed in tumor were analyzed. The univariate and multivariate Cox proportional hazards analyses were examined the relations of 395 genes’ expression and the prognosis. Results: Patients were divided into two groups based on the median expression values of 395 genes. The univariate analyses showed 20 genes were significantly associated with overall survival (OS). Unsupervised hierarchical clustering analysis using 20 gene expression data revealed two distinct clusters (cluster 1 and 2). The cluster1 patients had high level expression of AXL, ADGRE5, CD40, CXCR6, MPO, etc., which were associated with tumor proliferation and migration. However, the cluster2 patients expressed higher NOS2, IL15, IKZF2, TNFRSF18 which were related with T lymphocyte activated. Therefore, the cluster2 patients had significantly longer OS than cluster 1. Moreover, the combined expression of 10 genes (CD40, ADGRE5, CSF1R, CCR1, IGSF6, MPO, MRC1, SRGN, CD63, LRP1) which had high correlation among 20 genes were significantly related to OS and DFS (disease free survival) in univariate analysis. The multivariate analyses demonstrated that 10 genes signature expression was the independent high risk factor for OS, and the low signature value was associated with better prognosis. Conclusions: In IIIa ESCC, the 20 genes can profile the tumor immune microenvironment, and were independently associated with the patients’ OS. The 10 genes’ signature which represents the malignancy and immune activation of tumor was an independent predictive factor for OS, which can identify the patients with favorable or poor prognosis.