Abstract
Fibroblast growth factor (FGF) and its receptor (FGFR) play crucial roles in gastric cancer (GC). Long non-coding RNAs (lncRNAs) are defined as RNA molecules of around 200 nucleotides or more, which are not translated into proteins. As well-known regulatory factors, lncRNAs are considered as biomarkers for prognosis and treatment response in GC. It is of importance to identify FGF/FGFR-related lncRNAs in GC. Here, some FGF/FGFR-related lncRNAs were identified in GC based on the data from public databases, the Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO). Then a four-lncRNAs (FGF10-AS1, MIR2052HG, POU6F2-AS2, and DIRC1) risk score (RS) model was established for predicting GC's prognosis by using Cox analysis. According to the median value of RS, GC patients were divided into low and high RS group. Low RS group displayed high tumor mutation burden and infiltration of immune cells, as well as more sensitivity to immunotherapy or chemotherapy. High RS group showed high infiltration of stromal cells and more oncogenic signatures. In addition, a comprehensive analysis was carried out and found that high RS group may exhibit specific sensitivity to Panobinostat (histone deacetylases inhibitor) and Tivantinib (MET inhibitor). In summary, our study not only offers a novel personalized prognostication classification model according to FGF/FGFR-related lncRNAs, but also provides a new strategy for subclass-specific precision treatment in GC.