Breast Cancer Survival Analysis Model

Abstract

(1) Background: Breast cancer (BC)—a leading cause of mortality in women globally—accounts for more than two million cases annually. BC was the most common cancer in Taiwan in 2015 and ranks among the top 10 malignancies in Taiwan. (2) Methods: We established a collection of BC survival and metastasis analyses using the Kaplan–Meier, logarithmic test, and Cox proportional hazard models to investigate the association among BC stages, different treatment modalities, and survival rate of patients with BC at various follow-up intervals. We also evaluated whether clinical prognostic factors had univariate and multivariate effects on the survival of patients with BC. Finally, we performed a metastasis analysis using the survival transition rate values of BC stages to develop a Markov chain and semi-Markov simulation model for BC and BC metastasis analysis, respectively. (3) Results: The Kaplan–Meier survival analysis revealed that the risk of BC treated with surgery was lower than that of those who did not receive surgery and the recommended treatment methods should be ranked by survival as follows: surgery, hormone therapy, chemotherapy, and radiation therapy (in descending order of risk). This is attributed to the predicted survival rate which ranges from 99.6% to 91.2%. Moreover, Cox’s treatment method considered the patient’s attributes and revealed a significant difference (p = 0.001). The Markov chain analyses determined the chance of metastasis at each stage, indicating that the lower the stage of BC, the greater the survival rate. (4) Conclusions: Patients’ treatment is influenced by different BC stages, and earlier detection presents better chances of survival and a greater probability of treatment success.

Keywords

• breast cancer
• Cox proportional hazard
• Markov chain
• semi-Markov chain
• survival analysis