Construction of a Prognostic Evaluation Model for Stomach Adenocarcinoma on the Basis of Immune-Related lncRNAs.

Abstract

Progression, prognosis, and therapeutic strategy of stomach adenocarcinoma (STAD) have a close connection with tumor microenvironment (TME). Thus, it is pivotal to delve into the TME and immune-related genes, which may bring possibilities for improving patient's prognosis. TCGA-STAD dataset was analyzed to acquire differentially expressed lncRNAs in tumor samples, which were overlapped with the immune-related lncRNA datasets in the ImmLnc database. Twenty-six lncRNAs related to STAD immunity and patient's prognosis were acquired by univariate Cox analysis. Following lncRNA expression patterns, STAD samples could be classified into two clusters with completely different immune patterns. We performed multivariate Cox regression analysis on lncRNAs to identify 7-feature lncRNAs and constructed a corresponding prognostic model. The model validity was verified by survival analysis and ROC curve in validation and training sets. To explore connection between model and TME and tumor drug resistance, this study analyzed differences in immune cell infiltration between samples from high- and low-risk groups and then revealed immune cells follicular helper with significant differences in tumor tissue infiltration. Analysis of resistance to chemotherapeutic drugs revealed that samples in the high-risk group had resistance to cisplatin, doxorubicin, bleomycin, and gemcitabine. Through univariate and multivariate Cox analyses, we manifested that risk score could be an independent prognostic factor. Combining risk score and clinical factors, a nomogram was constructed to accurately predict patient's prognosis. This model can effectively predict prognosis, TME, and drug resistance of STAD patients, which may provide a reference for tumor development evaluation and precise treatment for clinical STAD.