Modelos paramétricos de sobrevivência aplicados a dados de câncer

Abstract

Cancer has become one of the diseases responsible for a large number of deaths worldwide. According to theWorld Health Organization, this pathology is one of the most common malignant diseases in the world. Thus, it would be of interest to evaluate the factors capable of influencing the time until the death of patients with this disease. Therefore, this study aims to present a review of the main parametric models of survival used to describe the life time of patients with this pathology as well as the factors associated with it. And, based on these models, fitted them to a set of data from patients with lung cancer. The adjusted parametric models were Weibull, Gamma, Generalized Gamma and Log-normal, with the best model being selected using the AIC, SBC and GD information criteria. After selecting the best model, in this study the Generalized Gamma, residual analysis was performed and the accuracy of the model was evaluated. It was found that the Generalized Gamma model was the one that best fitted the data set, being thus used to describe the time until the death of patients with lung cancer and the factors associated with it. And, it was concluded that the factors that contributed to the patients’ survival time were gender, age, staging (TNM), histology and smoking.