Oral squamous cell carcinoma (OSCC) is one of the most common malignancies worldwide. Despite relative high 5-year survival rate, delayed diagnosis still hampers its prognosis improvement. Long noncoding RNA (lncRNA) has recently been demonstrated to involve in development of various cancers. This study aimed to identify a lncRNA-based signature by which we could accurately predict OSCC prognosis. We first downloaded the RNA sequencing datasets for OSCC from The Cancer Genome Atlas (TCGA) and divided samples into training set and validation set. Associations between lncRNAs' expression and OSCC overall survival (OS) were evaluated in training set. Further screening through Random Survival Forest (RSF) method identified a prognostic signature composed of three lncRNAs, including AC013268.5, RP11.65 L3.4, and RP11.15A1.7, and by which OSCC samples in training set could be divided into high-risk and low-risk group with significantly different OS (p < 0.001, hazard ratio [HR] = 1.873). Besides, reliability of the prognostic signature was confirmed in the validation set (p < 0.01, HR = 2.14). Receiver operating characteristic curve analysis showed the superiority of the signature in OSCC prognosis prediction than other clinical characteristics. Gene set enrichment analysis indicated that high-risk score was closely associated with metabolism and spliceosome related pathways. Our findings suggest that lncRNA should be an important biomarker for prognosis of OSCC patients.
Keywords: OSCC; TCGA; lncRNA; overall survival.