Abstract
Ovarian cancer (OC) has the highest mortality rate among all female reproductive system malignant tumors worldwide. In this study, we aimed to investigate OC from several perspectives by using machine learning. Our results showed that the mRNA expression-stemness index (mRNAsi) is closely related to clinical characteristics of OC patients, as OC patients with venous or lymphatic invasion had higher mRNAsi score compared to patients with no invasion. Furhter grade 3/4 patient group had higher mRNAsi scores compared to the grade1/2 group. We also found that mRNAsi is closely related to immune infiltration in OC. We also built a competing endogenous RNA network, which contained 4 miRNAs, 5 lncRNAs, and 1 mRNA, by using Cytoscape based on the differentially expressed genes of the high- and low-mRNAsi groups. Through Lassio regression, we also established a model including 7 lncRNAs and 2miRNAs, which could effectively categorize OC patients into two groups based on the median risk score. We then developed a nomogram model which could effectively forecast the overall survival rate of OC for 1-, 3-, and 5-year period. The models assessed in this study showed potential for clinical application in treatment decisions for OC.