Identification of epigenetic genes for predicting prognosis and immunotherapy response of ovarian cancer

Background Epigenetic factors play a critical role in tumour development and progression. The aim of this study was to construct and validate a robust epigenetic gene set-based signature for predicting prognosis of ovarian cancer. Methods By using LASSO Cox regression model, we screened out the most useful prognostic epigenetic factors and a prognostic signature was developed based on them. Survival receiver operating characteristic was used to test the prognostic accuracy of signature in training and validation sets. The associations between the risk scores and immune cell infiltration, tumour purity, immune checkpoint inhibitor genes expression were also assessed in ovarian cancer . Results A total of 26 epigenetic factors were identified to develop the prognostic signature. In the training set, the prognosis of high-risk patients was strikingly poorer than that of low-risk patients (hazard ratio: 2.11, 95% confidence interval: 1.65-2.72, P < 0.001). Similar results were further observed in the internal validation set (hazard ratio: 1.69, 95% confidence interval: 1.07-2.63, P = 0.020) and external validation set (hazard ratio:1.95, 95% confidence interval: 1.41-2.69; P < 0.001). Survival receiver operating characteristic at 5 year showed the epigenetic signature (area under the curve = 0.700) performed better than other clinical features in predicting prognosis. Distinct difference in immune activation related pathways, immune cells infiltration, tumour purity reflected by immune and stromal score and immune checkpoint inhibitor genes gene expression was observed between high- and low-risk samples. Conclusions This study constructed an epigenetic signature that was capable of predicting postoperative outcomes and may also serve as potential biomarker for immunotherapy responses for ovarian cancer. The prognostic value of epigenetic factors was comprehensively evaluated in this study. An epigenetic signature that was capable of predicting postoperative outcomes and may also serve as potential biomarker for immunotherapy responses for ovarian cancer was successfully developed.