Identification of a nomogram based on long non-coding RNA to improve prognosis prediction of esophageal squamous cell carcinoma

Purpose: Esophageal squamous cell carcinoma (ESCC) remains a common aggressive malignancy in the world. Several long non-coding RNAs (lncRNAs) are reported to predict the prognosis of ESCC. Therefore, an in-depth research is urgently needed to further investigate the prognostic value of lncRNAs in ESCC. Results: From the training set, we identified a eight-lncRNA signature (including AP000487, AC011997, LINC01592, LINC01497, LINC01711, FENDRR, AC087045, AC137770) which separated the patients into two groups with significantly different overall survival (hazard ratio, HR = 3.79, 95% confidence interval, 95% CI [2.56-5.62]; P < 0.001). The signature was applied to the validation set (HR = 2.73, 95%CI [1.65-4.53]; P < 0.001) and showed similar prognostic values. Stratified, univariate and multivariate Cox regression analysis indicated that the signature was an independent prognostic factor for patients with ESCC. A nomogram based on the lncRNAs signature, age, grade and stage was developed and showed good accuracy for predicting 1-, 3- and 5-year survival probability of ESCC patients. We found a strong correlation between the gene significance for the survival time and T stage. Eight modules were constructed, among which the key module most closely associated with clinical information was identified. Conclusions: Our eight-lincRNA signature and nomogram could be practical and reliable prognostic tools for esophageal squamous cell carcinoma. Methods: We downloaded the lncRNA expression profiles of ESCC patients from Gene Expression Omnibus (GEO) and The Cancer Genome Atlas (TCGA) datasets and separated to training and validation cohort. The univariate, least absolute shrinkage and selection operator (LASSO) and multivariate Cox regression analysis were used to identify a lncRNA-based signature. The predictive value of the signature was assessed using the Kaplan-Meier method, receiver operating characteristic (ROC) curves and area under curve (AUC). Weighted gene co-expression network analysis (WGCNA) was applied to predict the intrinsic relationship between gene expressions. In addition, we further explored the combination of clinical information and module construction.