OBJECTIVETo investigate the factors affecting the survival and to predict the survival time of glioma.

METHODSGraph method was applied to fit the type of probability distribution of patients' postoperative survival time. As one suitable model, Weibull regression model tested by residual analysis was used to identify prognostic factors for postoperative survival and to predict patients' post-operative survival time.

RESULTSIn multivariate analysis of Weibull regression model, following characteristics showed significant influence on postoperative survival time including: sex, age, epilepsy, type of glioma, density of glioma, type of surgery, times of surgery and mode of postoperative therapy. According to predictive value, patients were clustered into three groups. The survival probability of one year for the first group was less than 50% and that of five years for the second group about 50%, for the third group around 80%.

CONCLUSIONSex, age, epilepsy, type of glioma, density of glioma, type of surgery, times of surgery and mode of postoperative therapy were important risk factors and Weibull regression model might help to predict patients' postoperative survival time.