Genomic relevance of FGF14 and associated genes on the prognosis of pancreatic cancer

Background: Fibroblast (FGFs) and insulin (IGF) growth factor pathways are among 10 most recurrently altered genomic pathways in pancreatic ductal adenocarcinoma (PDAC). However, the prognostic and therapeutic relevance of FGF and IGF pathways in PDAC is largely unknown. Methods: We investigated the relationship between fibroblast and insulin pathway gene expression and clinicopathological features in three independent transcriptomic cohorts of 532 PDAC patients. Furthermore, we have examined the coexpressed genes specific to the prognostic marker identified from these cohorts. Statistical tests including Fisher-exact\Chi-square, Kaplan–Meier, Pearson Correlation and cox regression analyses were performed. Additionally, pathway analysis of gene-specific co-expressed genes was also performed. Results: The dysregulation of six genes including FGF9, FGF14, FGFR1, FGFR4, IGF2BP2 and IGF2BP3 were significantly associated with different clinical characteristics (including grade, stage, recurrence and nodes) in PDAC cohorts. 11 genes (including FGF9, FGF13, FGF14, FGF17, FGFR1, FGFRL1, FGFBP3, IGFBP3, IGF2BP2, IGF2BP3 and IGFBPL1) showed association with overall survival in different PDAC cohorts. Interestingly, overexpression of FGF14 was found associated with better overall survival (OS) in all three cohorts. Of note, multivariate analysis also revealed FGF14 as an independent prognostic marker for better OS in all three cohorts. Furthermore, FMN2 and PGR were among the top genes that correlated with FGF14 in all 3 cohorts. Of note, overexpression of FMN2 and PGR was found significantly associated with good overall survival in PDAC patients, suggesting FMN2 and PGR can also act as potential markers for the prediction of prognosis in PDAC patients. Conclusion: FGF14 may define a distinct subset of PDAC patients with better prognosis. Moreover, FGF14-based sub-classification of PDAC suggests that FMN2 and PGR can be employed as good prognostic markers in PDAC and this classification may lead to new therapeutic approaches.