Artificial Neural Networks Used in the Survival Analysis of Breast Cancer Patients: A Node-Negative Study

摘要: Artificial neural networks have been shown to be effective as general non-linear models with applications medical diagnosis, prognosis and survival analysis. This chapter begins a review of artificial used regression in the analysis breast cancer patients. These techniques are much interest because they allow modelling time-dependent hazards presence complex non-additive effects between covariates. First, role is introduced within context statistical methods parametric for cancer. Second, these applied study comprising node-negative patients order evaluate evidence improved or combination prognostic indices clinical environment. In particular, an early form which cells not yet spread regional lymph nodes. There determining relevant factors that can allocate into groups correlating risk disease relapse mortality following surgery. index then inform choice therapy. The Cox model Neural Networks (ANN), Partial Logistic Network Automatic Relevance Determination (PLANN-ARD) identify interpret group allocation. A monthly retrospective cohort 5-year follow-up conducted pathologically selected from two datasets collected Manchester Christie Hospital, UK.