Multi-omics identification of SPON1-related risk model for predicting prognosis and drug response in ovarian cancer.

Ovarian cancer (OC) is a highly lethal gynecological malignancy owing to late-stage diagnosis, high recurrence and metastasis rate. Spondin-1 (SPON1), a secreted extracellular matrix protein, has been demonstrated to be over-expressed in multiple tumors. However, the carcinogenic mechanism of SPON1 in OC remains unclear. Differential expression analysis was employed to identify SPON1-related genes from public databases. Immunohistochemistry (IHC) was utilized to detect SPON1 protein expression. Kaplan-Meier analysis was applied to evaluate the survival differences. Least absolute shrinkage and selection operator and Cox analysis were used to construct risk model. The C-index, receiver operating characteristic curve, calibration curve and decision curve analysis were performed to evaluate the predictive efficacy. Enrichment analysis revealed biological processes. Additionally, the relationships between risk score and tumor microenvironment (TME), somatic mutations and drug sensitivity were assessed. SPON1 was significantly upregulated in OC samples, and high SPON1 expression was associated with worse survival rate. Then a six-gene risk model was conducted in training set. The prognostic accuracy of model was validated using multiple external cohorts. Functional analysis indicated that multiple cancer progression and immune-related pathways were markedly associated with the risk score. The infiltration levels of most immune cells were decreased in the high-risk subgroup. High-risk patients displayed poor response to common drugs for chemotherapy and decreased tumor mutation burden in OC. In summary, we developed and validated a SPON1-based risk model for predicting the prognosis of OC patients and explored the differences in the TME and response to chemotherapy.