Abstract
Beginning in 1989, in conjunction with the Tokyo University of Agriculture and Technology and the National Cancer Center Hospital East, we began studying computer-aided image-reading skills for CR mammography images. We developed an experimental system incorporating a detection algorithm in 1995, installed it at the National Cancer Center Hospital East in 1996, and began its evaluation. Furthermore, using prototype mammography CAD systems that had been developed based on routine-type usage patterns and that reflected the results of the evaluations made with the experimental system, evaluations were performed at nine medical institutions in Japan. By factoring in the results of studies and evaluations made so far, we developed an algorithm for detecting candidate breast cancer shadows and incorporated it into a mammography image reading system dubbed “MammoViewer MV-SR657.” This article provides explanations of this detection algorithm technology.
