Comparison of Parametric and Cox Models in Analysis of Factors Influencing the survival of Women with Breast Cancer in Yazd city

Breast cancer is one of the most common cancers in women. One of the new methods of therapy are the use of tumor markers (ER, PR, Ki67, Her2). This study aimed to investigate the prognostic factors that affecting the survival of women with breast cancer in Yazd using parametric and Cox model as well as comparing the effectiveness of these models.
The 538 of the patients referred to Ramezanzade Radiotherapy Center who had breast cancer and recorded survival status as a census from the April 2005 until March 2012. The criterion to compare the performance of models was Akaike information criterion. Data was then analyzed in R software and 0.05 was considered as the significance level.
Findings: according to the Kaplan Meier method, Average of 1,2,3,5 and 8-years survival rate, and estimated, respectively, 0.976, 0.941, 0.898, 0.823 and 0.737. mean age of women with breast cancer was 48.03 ± 11.66 years and mean of survival time was 97.4±64.23. Ki67 variables for patients under 40 years of age (HR = 1.374, CI: 95% = [1.023, 3.536], and Her2 for patients aged 40 years and over (HR = 2.567, CI: 95% = [1.1767.5.598]) is significant on survival time.
Ki67 and Her2 variables had a positive effect on death risk.According to Akaike
information criterion we found that Cox model best fits for determining Prognostic factors influencing the diagnosis of retinopathy.