Applying logistic model on breast cancer metastatic with missing data : penalized profile likelihood

  Breast cancer is one of the most common and worrisome diseases among women all over the world which becomes more difficult with localized relapses and metastases in this disease and the chance of survival is reduced. therefore the effect of various factors on the incidence of metastases can have an important role in the treatment and prevention of disease progression. In this study, 1959 women with breast cancer who referred to Omid Hospital in Mashhad from 2001 to 2011 were selected and evaluated. To analyze the data, after the imputation of missing data, the method of penalized profile likelihood was used and data analysis was done in the software R. The results of the analysis showed that family background, clinical stage, Her2 protein, and T (tumor size) variables had a significant effect on the incidence of metastasis in breast cancer patients. According to the results, the stage code and her2 variables, tumor size and family history in all conditions of analysis had a significant effect on the incidence of metastasis in patients with breast.