Revolutionizing ESCC prognosis: the efficiency of tumor-infiltrating immune cells (TIIC) signature score

Abstract Background Patients suffer from esophageal squamous cell carcinoma (ESCC), which is the ninth highly aggressive malignancy. Tumor-infiltrating immune cells (TIIC) exert as major component of the tumor microenvironment (TME), showing possible prognostic value in ESCC. Methods Transcriptome data and scRNA-seq data of ESCC samples were extracted from the GEO and TCGA databases. Tissue Specific Index (TSI) was defined to identify potential TIIC-RNAs from the TME. Twenty machine learning algorithms were further applied to evaluate the prognostic efficacy of TIIC signature score. Gene colocalization analysis was performed. Differences in CNV on chromosomes and SNP sites of prognostic model genes were calculated. Results The most reliable model of TIIC signature score was developed based on three prognostic TIIC-RNAs. It showed a higher C-index than any other reported prognostic models. ESCC patients with high TIIC signature score showed poorer survival outcomes than low TIIC signature score. The activity of most immune cells decreased with the increase of TIIC score. TIIC signature score showed difference in the expression levels and methylation levels of DEGs. There was also significant different correlation with the degree of CNV amplification and CNV deletion of the immune checkpoint genes. Gene colocalization analysis showed two prognostic model genes (ATP6V0E1 and BIRC2). MR analysis found that rs148710154 and rs75146099 SNP sites of TIIC-RNA gene had a significant correlation between them gastro-oesophageal reflux and ESCC. Conclusion TIIC signature score was the first time developed which provided a novel strategy and guidance for the prognosis and immunotherapy of ESCC. It also gave the evidence in the important role of immune cells from the TME in the treatment of cancers.