The long term goal of the proposed project is to develop the software, image database and knowledge-base for an intelligent computer workstation that supports interpretation of mammograms for breast cancer detection. The workstation will reinforce perceptive and cognitive elements of mammography interpretation and provide a dynamic, interactive tool for teaching these skills. The system will be based on an existing prototype mammography expert system developed by investigators in the Department of Diagnostic Imaging at the Yale School of Medicine. Before undertaking full commercial development, we want to test the clinical appropriateness of the image recall strategies using an expanded image data base (phase I). We will also refine design specifications during this phase. The final system will draw from a large image database (greater than 1000 cases) and knowl- edge-base in order to assemble highly context sensitive, patient specific output that is expected to improve the accuracy of mammographic diagnosis. This project addresses the important public health problem of breast cancer, the most common nonpreventable cause of cancer death in women. Its training and decision support capabilities could potentially alleviate projected manpower shortages caused by current rapid increases in mammography volume, and also improve the quality of mammography interpretation.
| Swett, H A; Mutalik, P G; Neklesa, V P et al. (1998) Voice-activated retrieval of mammography reference images. J Digit Imaging 11:65-73 |
