Evaluation of survival rate of patients with breast cancer using mixture cure model

Introduction & Aim: Survival analysis is a set of methods used for analyzing time to occurrence of a specific event. There are some occasions in which a proportion of the population of interest will never experience the event of interest. Two major approaches in cure rate models are Mixture and Non-mixture cure rate models and they developed based on different assumptions. Aim of the present research is to apply mixture cure model to analyze survival of patients with breast cancer.Methods: We studied 438 female patients with breast cancer who were visited and treated at cancer research center in Shahid Beheshti University of Medical Sciences, Tehran, Iran. These patients were diagnosed from 1992-2012 and followed up until October 2014. Mixture(in consideration of Bernoulli distribution for latent variable) cure rate model under assumption of Weibull distribution for survival time were fitted.Results: A total of 75(17.12%) deaths due to breast cancer were observed (non-censored observations). The mean of age at the time of diagnosis was 48.37 years. Kaplan-Meier survival curves were plotted for chickening the presence of cure. By assuming the Bernoulli distribution, Lymph node involvement, Type of surgery, stage ɪɪ and ɪɪɪ, Lymph node involvement, Historical grade, and Estrogen receptor positive were significant. Conclusion: Based on mixture cured model a significant proportion of the population was cure and regardless of the results is wrong conclusion.