Ovarian cancer is the second most prevalent gynecologic malignancy, and ovarian serous cystadenocarcinoma (OSCA) is the most common and lethal subtype of ovarian cancer. Current screening methods have strong limits on early detection, and the majority of OSCA patients relapse. In this work, we developed and cross-validated a method for detecting gene expression biomarkers able to discriminate OSCA tissues from healthy ovarian tissues and other cancer types with high accuracy. A preliminary ranking-based approach was applied, resulting in a panel of 41 over-expressed genes in OSCA. The RNA quantity gene expression of the 41 selected genes was then cross-validated by using NanoString nCounter technology. Moreover, we showed that the RNA quantity of eight genes (ADGRG1, EPCAM, ESRP1, MAL2, MYH14, PRSS8, ST14 and WFDC2) discriminates each OSCA sample from each healthy sample in our data set with sensitivity of 100% and specificity of 100%. For the other three genes (MUC16, PAX8 and SOX17) in combination, their RNA quantity may distinguish OSCA from other 29 tumor types.
Keywords: NanoString technology; TCGA transcriptomes; biomarkers; digital RNA detection; ovarian serous cystadenocarcinoma (OSCA).