An Expert-Trained Consultation System for Mammography

This project will develop a computer-assisted consultation system for clinical radiologists in theclassification of benign and malignant (cancer) calcifications on mammograms. An automated screeningof digital mammograms can be further developed when this system is fully tested in various clinicalsettings. The success of this project will inspire revolutionary improvements in other cancer diagnosticprocedures through further research and development. We have demonstrated that the newly developedartificial visual neural networks and training methods (by presenting microcalcification patterns in thetraining session) can detect microcalcifications on mammograms. A similar method can be used for thedetection of lung nodules on chest radiographs. This feature extraction methods and artificial visualneural networks simulate radiologists' routine practices in reading mammograms. Radiologists' viewingpatterns and decision making processes will be modeled and converted to computer readable form.Preliminary studies have shown the promise of this approach. The Phase I study will address issuesrelated to the differentiation of malignant from benign microcalcifications based on radiographs takenfrom breast tissue specimens. Based on Phase I study, the plan of the Phase II study is to (i) conductthe research involving classification of microcalcifications on clinical mammograms, (ii) analyze variousfeatures of benign and malignant microcalcifications and seek potential correlation with the learningpatterns of the artificial visual neural network, and (iii) integrate the research outcome and implementa prototype consultation system for clinical use.