Ovarian cancer is one of the most lethal gynecological malignancies. Factors contributing to its adverse outcomes include that its initial symptoms are not always obvious, making it difficult to detect early. Hence, at the time of diagnosis, most patients are already in at least an advanced stage [13]. As reported, despite benefiting from first-line therapy, 75% of patients with advanced ovarian cancer (stage III or IV) have tumor relapse at a median of 15 months from diagnosis. Additionally, patients with early-stage disease (stage I or II) have long-term survival rates (>10 years) of 80–95%. By contrast, patients with advanced disease (stage III or IV) tended to have a 10–30% long-term survival rate [14]. Consequently, it is urgent that new targeted therapies be developed to improve the treatment efficacy of ovarian cancer. Recent research has identified the tumor microenvironment as an important part of the tumorigenesis of ovarian cancer and a potential target for therapeutic intervention [15,16,17]. In clinical practice, patients presenting with the same clinical and pathological parameters can exhibit markedly different outcomes despite receiving identical treatments. Effective biomarkers that can distinguish high-risk groups in these patients are constantly emerging, providing effective help for more accurate diagnosis and treatment.

In our study, the expression of 604 ovarian cancer-related differential genes in two databases, TCGA and GTEx, and their relationship with OS were systematically investigated. After collecting clinical information of OV patients and using Cox univariate regression and lasso regression feature selection algorithm to analyze 604 differentially expressed characteristic genes in ovarian cancer, a total of 31 prognostic genes were screened (p value < 0.05), a new prognostic model was established for the first time and validated in an external cohort. The results confirmed the relationship between 17 genes and ovarian cancer, including KRT81, COL16A1, ADH1B, UGT2B17, PDK4, ALDH1L1, DNAJB1, CRISPLD2, ATP1A2, HSPB7, MCC, APOD, OPEP3, 1D01, REN, CXCL9, KLHDC8A. In the current study, univariate Cox regression analysis, LASSO regression analysis, and multivariate Cox regression analysis were used to screen and identify prognostic models, and external datasets were used for robustness validation, which made the model more reasonable and accurate. Therefore, in contrast to the paper-based approach of previous studies, our model obtained more reliable AUC values of 1-year, 3-year, and 5-year survival (AUC-1 year: 0.740 vs. 0.717, 3 years: 0.618 vs. 0.574, 5 years: 0.688 vs. 0.575).

A review of the TCGA data revealed that nine of the 17 genes were significantly associated with the overall prognosis of ovarian cancer. This finding suggests that the model has good prognostic predictive value. COL16A1 is found to be one of the important genes for predicting OV survival, which is highly expressed in cancer tissues compared with normal tissues, and its expression level of COL16A1 is significantly correlated with progression-free survival [18, 19]. ALDH1L1 expression in a variety of cancers is often associated with cancer-specific cellular metabolic pathways, such as one-carbon metabolism and folate metabolism, that are critical for cell proliferation and survival. Research on the specific role and potential therapeutic targets of ALDH1L1 in ovarian cancer is still in its preliminary stages, and further studies are needed to explore its exact role in ovarian cancer development. CRISPLD2 is part of a glucocorticoid-inducible gene and cytoskeletal network that mediates the morphogenesis of CRISP and LCCL domains, which have been proven to play important roles in immune responses, inhibition of inflammation, or cell motility [20,21,22,23,24]. However, the role of CRISPLD2 in OV has not been characterized. ATP1A2 gene encodes the α-2 subunit of Na+/K+-ATPase, while cardiogenic glycoside is a specific inhibitor of Na+/K+-ATPase [25]. The antitumor effect of cardiogenic glycoside has been reported for breast cancer, colon cancer, pancreatic cancer, lung cancer, prostate cancer, glioblastoma, and melanoma [26, 27]. Several studies have shown that Na+/K+-ATPase plays a role in the pathophysiology of ovarian cancer [28]. As reported by Huang et al., ATP1A2 overexpression is closely associated with poor prognosis in OSC patients with FIGO stage III, histological grade G3, p53 mutation, and age ≥ 60 years [29]. In ovarian cancer, IDO1 (indoleamine 2,3-dioxygenase 1) affects the immune system mainly through its role in the tryptophan metabolic pathway, and its activity is strongly correlated with the severity of the disease and prognosis. IDO1 induces suppression of immune cell function, particularly on effector T cells and natural killer cells, by catalysing the conversion of tryptophan to kynurenine (Kyn), while promoting the activation of regulatory T cells. Studies have shown that high expression of IDO1 in ovarian cancer is associated with poor disease prognosis. For example, increased expression of IDO1 in ovarian cancer induces PD-1 expression in T cells through activation of the aromatic hydrocarbon receptor (AhR), which reduces the prognostic benefit of tumour-infiltrating CD8 + T cells and further exacerbates immune escape [30]. In ovarian cancer, the role of REN and its potential clinical significance need to be clarified by further research. Current research on REN in ovarian cancer focuses on understanding its role in the tumour microenvironment and how it may offer potential therapeutic strategies by modulating the renin-angiotensin system. In ovarian cancer, the role of REN and its potential clinical significance need to be clarified by further research. Current research on REN in ovarian cancer focuses on understanding its role in the tumour microenvironment and how it may offer potential therapeutic strategies by modulating the renin-angiotensin system (PMID: 38535506). In ovarian cancer, expression of CXCL9 (C-X-C motif chemokine 9) has been associated with improved patient survival. Studies have shown that high expression of CXCL9 is associated with enhanced T-cell infiltration and is more highly expressed in ovarian cancer tissues than in normal tissues, suggesting that it may play an active role in anti-tumour immunity. For example, a study from the TCGA-OV database found that the median survival time of a patient group with high CXCL9 expression was longer than that of a group with low expression, implying that high CXCL9 expression is positively associated with survival in ovarian cancer patients [31]. In addition, CXCL9 has been found to have potential in promoting the effects of anti-PD-L1 therapies. It acts by increasing T-cell infiltration in the tumour microenvironment, which is critical in enhancing the effectiveness of immunotherapy. The function and potential clinical significance of HSPB7 and MCC in specific studies of ovarian cancer remain unclear. But their role in other cancers suggests their importance in tumor biology. Further studies are expected to reveal the specific function and its potential as a therapeutic target.

Our study found through functional analysis that the genomes associated with the high-risk group were involved in TNF-α signaling via NFKB/UV-responsive/supplement/IL2_STAT5 signaling/IL6_JAK_STAT3 signaling [32]. Enrichment analysis of immune subtype-related molecular pathways showed that the b subtype had an increased fraction of IL-6-JAK-STAT3 molecular pathway enrichment, which was closely associated with m2 macrophage infiltration [33]. JAK/STAT pathways play an essential role in regulating the tumor immune microenvironment, the activation of which was correlated with the drug resistance of OV patients [34]. Furthermore, a recent study focusing on single-cell sequencing of high-grade ovarian cancer illustrated that the JAK-STAT pathway in ovarian cancer cells was abnormally activated, and thus, small-molecule inhibitors of the JAK-STAT pathway were expected to be applied in the clinic [35]. Tumor-associated macrophages (TAM) play a crucial role in the development and progression of OV. The ratio of M1 and M2 macrophages is able to serve as a powerful prognostic indicator for OV patient [36]. In this study, we identified that M1 macrophages, M2 macrophages, and follicular helper T cells were associated with risk scores. Compared with the low-risk group, the proportion of M1 macrophages and follicular helper T cells infiltrating decreased, and the proportion of M2 macrophages infiltrating increased in the high-risk group. M1 macrophages could stimulate helper T cells to secrete cytokines such as IL-12, IL-23, and TNF-α, which are related to improving the long-term survival of OV patients [37]. The tumor-promoting effect of M2 macrophages has been widely reported. Monocytes or macrophages in the OV microenvironment can be polarized to M2 macrophages, which can further form an immune microenvironment to promote the progression of OV [38]. Therefore, promoting the transformation of M2-type macrophages into M1-type macrophages is the top priority of OV immunotherapy.

Given the fact that immune checkpoints have important guiding significance in individualized immunotherapy, we simultaneously analyzed key immune checkpoints in ovarian cancer patients [39,40,41]. CIBERSORT, as an effective algorithm, is used to compute fractional gene expression data for a subset of characteristic cells from batch samples obtained [42]. In this study, the infiltration of 21 immune cell types in the low-risk group and the high-risk group was evaluated using the above-mentioned software. In addition, we also analyzed the relationship between 21 immune cell types and 17 key immune checkpoint regulators in OV samples (including B7-H3, B7-H4, CD27, CD270, CD40, CD58, CD70, CD86, CTLA4, ICOS, IDO1, LAG3, PD-L1, PD-L2, TIGIT and TIM-3), as well as studied the expression of immune regulators in high-risk and low-risk populations. It was found that risk scores were closely related to key immune checkpoints such as CD274, IDO1, and TGFB1. There were significant differences in immune rejection between high and low risk groups. In general, the complexity of tumor-infiltrating immune cells broadly affects the immune status of the host and plays an important role in the immunotherapy response. Previous studies have found that immune checkpoint modulators can mediate the function of TICs [43, 44]. In our study, we found that Tfh cells and m1 macrophages were positively correlated with most regulators, while m2 macrophages were negatively correlated. The high expression of most regulatory proteins is related to the good outcome of OC. Therefore, we speculated that the presence of immunogenic TME in low-risk samples may be one of the reasons for better survival. This speculation was further demonstrated via the high expression of regulators in low-risk populations.

Optimal cytoreductive surgery combined with platinum-based chemotherapy with a carboplatin-paclitaxel regimen is currently the most prevalent standard of care for ovarian cancer [45]. However, with the development of chemotherapy-resistant and refractory diseases, the sensitivity of chemotherapy has decreased. Therefore, we further analyzed the sensitivity of high-risk and low-risk groups to chemotherapeutic drugs using the predictive model of the Genomics of Drug Sensitivity in Cancer (GDSC). It was concluded that several drugs had significant effects on the expression of both high-risk and low-risk groups, indicating that the high-risk and low-risk groups were sensitive to commonly used chemotherapeutic drugs (p < 0.05). The prediction model of chemotherapeutic drugs in the TCGA dataset showed that their sensitivity to chemotherapeutic drugs was in the order of Bexarotene >Imatinib >Dasatinib >Docetaxel >Pyrimethylamine >Bleomycin. Combination therapy of lapatinib and/or erlotinib combined with bexarotene is effective in overcoming lapatinib and/or erlotinib resistance in vivo, and can be further tested in preclinical and clinical trials of ovarian cancer and other types of cancer [46]. Imatinib is a tyrosine kinase inhibitor recommended for patients with advanced epithelial ovarian cancer who express PDGFRA [47]. Safra et al. performed intermittent weekly paclitaxel testing with imatinib in 14 patients with ovarian cancer and achieved objective responses in four of them [48]. Dasatinib is the most active compound approved by the FDA (80% inhibition of cell proliferation). The synergy between PARP inhibitor olaparib and approved multi-kinase inhibitor dasatinib has been confirmed by previous research, the interaction of which is particularly active in triple-negative breast cancer and ovarian cancer cell lines [49]. Dasatinib was studied in combination with standard cytotoxic chemotherapy (carboplatin and paclitaxel) in the latest phase I clinical trial, and the results suggest that this drug combination regimen can be administered safely, with some evidence of clinical efficacy [50]. Docetaxel has a similar drug mechanism to paclitaxel. A number of trials have demonstrated that the second-generation docetaxel can be used as a substitute for paclitaxel [51]. We revisited the GDSC database to do an in-depth analysis of the risk score and drug sensitivity. As in Figure S1, we show six drugs. AZ6102 is a potent TNKS1/2 inhibitor that has 100-fold selectivity against other PARP family enzymes [52]. Temozolomide and carmustine are alkylating agents, which are non-specific agents acting on the cell cycle, and DNA is identified as the main target of them [53]. Cediranib is an angiogenesis inhibitor that has been shown in previous ICON6 trials to improve survival outcomes in patients with recurrent epithelial ovarian cancer [54]. Platinum emerged as the mainstay of OV treatment in frontline therapy, and unfortunately, increased carboplatin/cisplatin exposure increases the risk of platinum resistance or hypersensitivity. Several studies have demonstrated the safety and efficacy of oxaliplatin in recurrent ovarian cancer, both alone and in combination regimens, demonstrating a good tolerability profile [55]. The sensitivity of the risk score to the standard treatment groups of platinum, paclitaxel, and olaparib is substantial.