A Novel Prognostic Model for Clear Cell Renal Cell Carcinoma with Differentially Expressed Immune-related lncRNA Pairs Based on the Cancer Genome Atlas

Abstract Objective: To construct a novel prognostic model of immune-related lncRNA (irlncRNA) pairs in clear cell renal cell carcinoma (ccRCC). Methods: RNA-seq and clinical data were retrieved from The Cancer Genome Atlas (TCGA). Differentially expressed irlncRNAs (DEirlncRNAs) were obtained by co-expression strategy with immune genes. A 0-1 matrix was constructed according to DEirlncRNAs relevant expression levels. Univariate cox regression was used to select potential target pairs. Lasso regression with cross validation and multivariate cox regression were carried out to extract the final biomarker pairs for risk score calculation. Through calculating the optimal cutoff of AUCs, patients were divided into high and low risk group. Model validation was conducted by independent prognostic analysis, survival analysis, tumor-infiltrating and chemosensitivity analysis. Results: A total of 42 DEirlncRNAs were identified and 12 target pairs were included to construct the final model. The risk score were both significantly different according to univariate (p<0.001, HR=1.391, 95%CI [1.313–1.475]) and multivariate cox regression (p<0.001, HR=1.3104, 95%CI [1.227-1.399]). The AUC reached 0.765 at 1-year, 0.724 at 3-year and 0.785 at 5-year. Patients in the high-risk group had significantly poor survival, higher level of CD8+T infiltration, lower drug sensitivity of sunitinib and temsirolimus but higher sensitivity of lapatinib and pazopanib.Conclusion: The novel prognostic model constructed by paring irlncRNAs showed an effective clinical prediction in ccRCC patients.