BackgroundOvarian cancer has the highest death rate of all fatal gynecological cancers. Increasing evidence has depicted the correlation between serous ovarian carcinoma prognosis and immune signature. Therefore, the aim of this study is to develop a robust prognostic immune-related gene pairs (IRGPs) signature for estimating overall survival (OS) of HGSOC.MethodsGene expression profiling and clinical information of serous ovarian carcinoma patients were derived from three public data sets, divided into training and validation cohorts. Immune genes significantly associated with prognosis were selected.ResultsAmong 1,534 immune genes, a 20 IRGPs signature was built which was significantly associated with OS in the training cohort (P=1.44×10−14; hazard ratio [HR] =3.05 [2.26, 4.10]). In the validation datasets, the IRGPs signature significantly divided patients into high- vs low- risk groups considering their prognosis (P=4.30×10−3; HR =1.48 [1.13, 1.95]) and was prognostic in multivariate analysis. Functional analysis showed that several biological processes, including EMT and TGF-β related pathways, enriched in the high-risk group. Macrophages M2 was significantly higher in the high-risk group compared with the low-risk group.ConclusionWe successfully constructed a robust IRGPs signature with prognostic values for serous ovarian carcinoma, providing new insights into post-operational treatment strategies.

Supplemental Digital Content is available in the text

Ovarian cancer is the leading cause of cancer-related death in gynecologic malignancies and consists of different histologic types. Histopathologic examination and accurate subtype diagnosis has become increasingly important in guiding patient management and, as such, is the most important currently available ovarian carcinoma "biomarker." In this study, we examined the expression of PAX2 and PAX8 by immunohistochemistry in 58 cases of ovarian serous tumors and 68 cases of ovarian mucinous tumors. The results demonstrated that PAX2 and PAX8 were detected in 100% (30/30) and 77% (23/30) of serous cystadenomas, 100% (16/16) and 94% (15/16) of borderline serous cystadenomas, and 100% (12/12) and 83% (10/12) of serous carcinomas, respectively. However, PAX2 and PAX8 were detected in only 0% (0/29) and 0% (0/29) of mucinous cystadenoma, 4% (1/24) and 4% (1/24) of borderline mucinous cystadenoma, and 0% (0/15) and 7% (1/15) of mucinous carcinoma, respectively. Further, there is a linear correlation between PAX2 and PAX8 (R=0.745; P<0.0001). Overall, our data indicated that PAX2 correlated with PAX8, and these 2 proteins differentially expressed in ovarian serous tumors and ovarian mucinous tumors. Thus, PAX2 and PAX8 are useful biomarker in the differential diagnosis of ovarian serous and mucinous tumors.