A clinical prognostic model of oxidative stress-related genes linked to tumor immune cell infiltration and the prognosis of ovarian cancer patients.

ABSTRACT

Background: According to statistics, ovarian cancer (OV) is the most prevalent type of gynecologic malignancy and has the highest mortality rate of all gynecologic tumors. Although several studies have shown that oxidative stress (OS) contributes significantly to the onset and progression of cancer, the role of OS in OV needs to be investigated further. Thus, it is critical to comprehend the function of OS-related genes in OV. Methods: In this study, all data related to the transcriptome and clinical status of the patients were retrieved from "The Cancer Genome Atlas" (TCGA) and "Gene Expression Omnibus" (GEO) databases. Using the unsupervised cluster analysis technique, all patients with OV were classified into two different subtypes (categories) based on the OS gene. All hub genes were screened using the weighted gene co-expression network analysis (WGCNA). Since the hub genes and the differentially expressed genes (DEGs) in both categories were found to intersect, the univariate Cox regression analysis was implemented. A multivariate Cox analysis was also performed to construct a novel clinical prognosis model, which was validated using data from the GEO cohort. In addition, the relationship between risk score and immune cell infiltration level was evaluated using CIBERSORT. Finally, qRT-PCR was used to confirm the expression of the genes used to construct the model. Results: Two subtypes of OS were obtained. The findings indicated that OS-C1 had a better survival outcome than OS-C2. The results of WGCNA yielded 112 hub genes. For univariate COX regression analyses, 49 OS-related trait genes were obtained. Finally, a clinical prognostic model containing two genes was constructed. This model could differentiate between patients with OV having varying years of survival in the TCGA and GEO cohorts. The model risk score was verified as an independent prognostic indicator. According to the results of CIBERSORT, many tumor-infiltrating immune cells were found to be significantly related to the risk score. Furthermore, the results revealed that patients with low-risk OV in the CTLA4 treatment group had a high likelihood of benefiting from immunotherapy. qRT-PCR results also showed that the expression of MARVELD1 and VSIG4 was high in the OV samples. Conclusions: Analysis of the results suggested that the newly developed model, which contained two characteristic OS-related genes, could successfully predict the survival outcomes of all patients with OV. The findings of this study could offer valuable information and insights into the refinement of personalized therapy and immunotherapy for OV in the future.