MammoNet (Mammography Decision Support) is a Bayesian-network model of breast disease to assist mammographers in formulating diagnoses and planning additional diagnostic procedures.

MammoNet is an ongoing project. The ongoing refinement of MammoNet includes variables and states to the model, acquiring conditional-probability data from large case series, and rigorous testing and evaluation. The developers are considering the addition of demographic features such as race and geographic locations, and patient-history features such as diet, body disposition, history of hormone therapy, and previous cancers. The granularity of the models variables could be increased by partitioning the breast cancer node into more than the current two states to represent the numerous types of cancer and benign conditions. Links between MammoNet and a database are being developed, to allow collection and analysis of a large set of clinical cases. The aim is to create a decision support tool to improve the diagnostic accuracy and cost-effectiveness of screening mammography. Such a decision aid must be reliable, integrated with a clinical database and reporting system, and able to generate explanations to the physicians who use it.

MammoNet is an ongoing project.

MammoNet is being developed by the MIDAS Consortium, Milwaukee, Wisconsin

Office of Clinical Informatics, 9200 West Wisconsin Avenue, Milwaukee, Wisconsin 53226 USA, Phone: (414) 257-5828, Fax: (414) 257-5186, Internet: http://www.mcw.edu/oci/ , or http://www.mcw.edu/midas/.