An Integrated Nomogram Combining LncRNAs Classifier and Clinicopathologic Factors to Predict the Recurrence of Head and Neck Squamous Cell Carcinoma

Long non-coding RNAs (lncRNAs) which have little or no protein-coding capacity, due to their potential roles in the cancer disease, caught a particular interest. Our study aims to develop an lncRNAs-based classifier and a nomogram incorporating the lncRNAs classifier and clinicopathologic factors to help to improve the accuracy of recurrence prediction for head and neck squamous cell carcinoma (HNSCC) patients. The HNSCC lncRNAs profiling data and the corresponding clinicopathologic information were downloaded from TANRIC database and cBioPortal. Using univariable Cox regression and Least absolute shrinkage and selection operator (LASSO) analysis, we developed 15-lncRNAs-based classifier related to recurrence. On the basis of multivariable Cox regression analysis results, a nomogram integrating the genomic and clinicopathologic predictors was built. The predictive accuracy and discriminative ability of the inclusive nomogram were confirmed by calibration curve and a concordance index (C-index), and compared with TNM stage system by C-index, receiver operating characteristic (ROC) analysis. Decision curve analysis (DCA) was conducted to evaluate clinical value of our nomogram. Consequently, fifteen recurrence-free survival (RFS) -related lncRNAs were identified, and the classifier consisting of the established 15 lncRNAs could effectively divide patients into high-risk and low-risk subgroup. The prediction ability of the 15-lncRNAs-based classifier for predicting 3- year and 5-year RFS were 0.833 and 0.771. Independent factors derived from multivariable analysis to predict recurrence were number of positive LNs, margin status, mutation count and lncRNAs classifier, which were all embedded into the nomogram. The calibration curve for the recurrence probability showed that the predictions based on the nomogram were in good coincide with practical observations. The C-index of the nomogram was 0.76 (0.72-0.79), and the area under curve (AUC) of nomogram in predicting RFS was 0.809, which were significantly higher than traditional TNM stage and 15-lncRNAs-based classifier. Decision curve analysis further demonstrated that our nomogram had larger net benefit than TNM stage and 15-lncRNAs-based classifier. The results were confirmed externally. In summary, a visually inclusive nomogram for patients with HNSCC, comprising genomic and clinicopathologic variables, generates more accurate prediction of the recurrence probability when compared TNM stage alone, but more additional data remains needed before being used in clinical practice.