Data mining and mathematical model development

The treatment and diagnosis of breast cancer require difficult decision making based on multidisciplinary fields, such as image, clinical, and pathological findings as well as new diagnostic technologies and molecular biomarkers. Thus, mathematical models that can predict a specific status and outcome of treatment by efficient usage of these big data are required. Successful models with high accuracy and high generalization ability can help promote personalized medicine and provide benefits of medical economy. Here, we introduce the use of advanced computational data mining technologies using artificial intelligence or machine learning and describe two models that we developed to predict pathological complete response of neoadjuvant therapy and lymph node metastasis in patients with primary breast cancer. The development and validation protocols are also discussed.