Modeling estrogen receptor pathways in breast cancer using an Artificial Neural Networks based inference approach

Estrogen receptor (ER) status is an important consideration in the prognosis and management of breast cancer patients, dictating treatment and patient management. While the prognosis of ER positive patients is generally poorer because of treatments such as Tamoxifen this situation has been reversed. Some detail is known of the ER pathway, however this has been based on reductionist studies of small numbers of markers. Here we present an Artificial Neural Network (ANN) using a feed forward back-propagation algorithm applied to a three layer multi-layer perceptron based approach that facilitates a wider more holistic approach to the identification of genes associated with ER status and the modeling of their interactions with one another in the context of a pathway.