Breast Cancer Diagnosis and Prognosis

Breast cancer accounts for the second most cancer diagnoses among women and the second most cancer deaths in the world. Until recently, evasive surgical biopsy was the only accurate and reliable diagnosis and prognosis tool. Of late, machine learning techniques have been applied to these tasks. We present a series of papers applying machine learning techniques to both problems of diagnosis and prognosis. One neural network diagnosis tool achieves 97.5% predicted accuracy and 100% observed accuracy. Prognosis poses a much more difficult problem due to the censored nature of the data. For most patients, we know only the disease-free survival time as opposed to the time of recurrence. One prognosis tool we present shows that good prognosis can be achieved without lymph node dissection.