Development and Validation of a Nomogram Predicting Over all and Cancer-specific Survival in Elder Patients With Brain Ependymoma

Background: Epithelioma originates from the ependymal epithelial cells on the surface of the ventricles of the brain and is a rare tumor lacking a prognostic prediction model. Our study aims to establish a column chart to predict the over all survival of patients with ependymoma. Method: The data used is from the SEER (Surveying, Epidemiology, and End Results; https://seer. cancer. gov/) database，Extract data from the database of patients with meningioma, including age,race, sex, grade, stage, surgery, tumor size, chemotherapy, radiation etc.Using Excel software,determine the optimal cutoff point, use univariate and multivariate cox regression to screen for risk factors in the over all and cancer specific survival prognostic models, establish nomogram models using meaningful variables in multivariate analysis, and use the consistency index (C-index) and receiver operating characteristic curve (ROC) to evaluate the area below the line (AUC). Then evaluate the effectiveness of one prediction of the model.Result:The median survival time of Over all Survival (OS) is 118 months, and no median survival time of Cancer-specific Survival(CSS) was observed.After univariate and multivariate analysis, it was found that age, grade, stage, surgery, tumor size, and primary site were statistically significant risk factors for OS prognosis.CSS univariate analysis found that age, grade, stage, tuber size, and primary site were statistically significant, while multivariate analysis found that age, grade, tuber size, and primary site were statistically significant.Conclusion: A prognostic column chart has been established for patients with ependymoma, providing a basis for clinical doctors to evaluate the prognosis of patients with ependymoma.