Construction of a prognostic risk model for uveal melanoma based on immune-related long noncoding RNA.

Uveal melanoma (UM) is a common health challenge worldwide as a prevalent intraocular malignancy because of its high mortality rate. However, clinical workers do not have an accurate prognostic tool now. Immune function is closely related to tumor development. Interestingly, researchers have identified that long noncoding RNAs (lncRNAs) are tightly associated with biological processes at the cellular level, particularly their involvements in immune response and its regulation of the growth of tumor cells. Hence, lncRNAs may be involved in the progression of uveal melanoma. UM patients' RNA expression matrices were extracted from TCGA database. The targeted immune genes were filtered by weighted correlation network analysis and the immune-related lncRNAs with a high prognostic relevance were obtained by Cox regression analysis and least absolute shrinkage and selection operator regression analysis. Each sample was scored according to those lncRNA expression and divided into high-risk and low-risk group. We confirmed the sensitivity and independence of our risk model compared to the tumor mutation burden score. Finally, we demonstrated the clinical relevance of our model by examining its sensitivity to different drugs. The risk score based on our risk model was significantly independent of other clinical parameters in either univariate (hazard ratio = 109.852 [15.738-766.749], P value < .001) or multivariate (hazard ratio = 114.075 [15.207-855.735], P value < .001) analyses. The ROC curves of this model imply high predictive accuracy for 1-year, 3-year, and 5-year survival (1-year area under the curve [AUC] = 0.849, 3-years AUC = 0.848, and 5-years AUC = 0.761). Our study revealed that immune-related lncRNAs are significant in the clinical diagnosis, treatment and prognosis of UM patients. We successfully constructed a lncRNA-based prognostic risk model which may serve as a future reference for the diagnosis and prognosis of UM. Based on this model we also validated the sensitivity of some cancer drugs, which has implications for the future immunotherapy and drug development.