Prognostic and predictive value of a mRNA signature in high-grade serous ovarian carcinoma with an integrated computation analysis

Abstract

Abstract Ovarian cancer (OC) is the leading cause of death among women with gynecologic malignancies. High-grade serous ovarian carcinoma (HGSOC) is the deadliest subtype of OC, accounting for about 90 percent of all ovarian cancer subtypes. Recent studies have shown that HGSOC patients have mutations in proto-oncogenes within the genome. Genome-wide detection and diagnosis are helpful for the diagnosis and treatment of HGSOC. To explore the genomic and transcriptional characteristics of subtypes of HGSOC, achieve accurate typing of tumor types, and obtain genomic characteristics that can reflect the subtypes of HGSOC, Using NMF clustering, SAM, PAM and survival time analysis, copy number variation data and gene expression data of 698 HGSOC samples were analyzed and differential expression genes of different disease subtypes were enriched and analyzed. Functions of genes related to different disease subtypes were enriched. All patients with HGSOC could be stratified into three categories according to genetic variation information and gene expression value. There was significant difference in the survival time curves of patients in different subtypes. And we identified twenty-one genes as the ones with the strongest power to differentiate the samples, including FTH1, COL1A2, COL3A1, GFBP7, ACTB, SPARC, PTTG1IP, TIMP1 and HLA-DPA1. Furthermore, we found that JAK/STAT (Janus kinase and signal transducers and activators of transcription) signaling pathway changes obviously in different subtypes. By investigating the genetic features and gene expression features, subtypes of patients with HGSOC could be accurately judged, which is useful for selecting therapeutic methods.