Establishment and validation of a prognostic model based on grouping of vasculogenic mimicry-related genes in ovarian cancer

Abstract Background Vasculogenic mimicry (VM) is a vascular-like microcirculatory delivery system directly formed by tumor cells. It plays significant roles in tumor invasion and resistance to antiangiogenic drugs. However, there is no model based on VM for ovarian cancer (OC) prognosis. Results The occurrence and progression of OC is associated with vasculogenic mimicry-related genes (VMGs) interactions in a cellular network. The prognostic VMGs can distinguish between two subgroups with significantly different prognoses. Differentially expressed genes (DEGs) between the subgroups suggest that certain signaling pathways and cellular functions are involved in the formation or promotion of VM. Based on filtered DEGs, a prognostic model was constructed and demonstrated outstanding performance in internal and external validations. The model results do not affect the distribution of clinical features and can be used to construct a nomogram. There are differences in immune cell composition and immune infiltration between the high-risk group and the low-risk group. Differences also exist in their resistance and sensitivity to certain drugs. Finally, subgroups analysis demonstrated differences in ovarian cancer VM structures which may be associated with different expressions of VMGs and VM-related prognostic indices (VMRPI). Conclusions VM may play a crucial role in the development and tumor immune microenvironment of serous ovarian cancer. VMRPI holds promise as a valuable prognostic biomarker. This is the first time that it is used to identify high risk OC patients with ovarian cancer and may indicate responsiveness to specific drug treatments.