A prognostic model based on tumor microenvironment-related lncRNAs predicts therapy response in pancreatic cancer

Abstract

Pancreatic cancer is an aggressive malignant tumor with high mortality and a low survival rate. The immune and stromal cells that infiltrate in the tumor microenvironment (TME) significantly impact immunotherapy and drug responses. Therefore, we identify the TME-related lncRNAs to develop a prognostic model for predicting the therapy efficacy in pancreatic cancer patients. Firstly, we identified differentially expressed genes (DEGs) for weighted gene co-expression network analysis (WGCNA) to identify the TME-related module eigengenes. According to the module eigengenes, the TME-related prognostic lncRNAs were screened through the univariate Cox, least absolute shrinkage and selection operator (LASSO), and multivariate Cox analyses to construct a prognostic risk score (RS) model. Next, the predictive power of this model was evaluated by the time-dependent receiver operating characteristic (ROC) curve and Kaplan-Meier analyses. In addition, functional enrichment, immune cell infiltration, and somatic mutation analyses were performed. Finally, tumor immune dysfunction and exclusion (TIDE) score and drug sensitivity analyses were applied to predict therapy response. In this study, 11 TME-related prognostic lncRNAs were identified to develop the prognostic RS model. According to the RS, the low-risk patients had a better prognosis, lower rates of somatic mutation, lower TIDE scores, and higher sensitivity to gemcitabine and paclitaxel compared to high-risk patients. The findings above suggested that low-risk patients may benefit more from immunotherapy, and high-risk patients may benefit more from chemotherapy. Within this study, we established a prognostic RS model based on 11 TME-related lncRNAs, which may help improve clinical decision-making.

Keywords: Pancreatic cancer; Prognostic model; Therapy response; Tumor microenvironment.