Modeling and analysis of survival to patients of pancreatic cancer in krasnoyarsk territory

Abstract

Any serious medical research involves the use of information technologies of statistical analysis in processing the results of clinical observations, experimental data and measurements. In the paper the survival analysis patients by the medical database of pancreatic cancer in Krasnoyarsk Territory is represented. Survival analysis is based on the survival function. Survival function modeling was performed on the basis of three approaches: parametric, semi-parametric and non-parametric. The survival model was fitted on the base of the selection of distributions. The closest distribution is the Weibull distribution. Semiparametric modeling made it possible to construct a survival curve for a combination of prognostic factors based on the Cox proportional risk regression model. Two parameters were identified as prognostic for survival: distant metastases, number of chemotherapy courses taken. An assessment was made of the overall survival of patients with pancreatic cancer and an assessment of the survival of patients after surgical and chemotherapeutic treatment in various subgroups by age and gender using the non-parametric Kaplan-Meier method #CSOC1120.