乳房X線写真上の腫瘤陰影の良・悪性鑑別ロジックに関する研究

抄録

One of the important subjects in constructing the computer-aided diagnosis system for detecting breast cancers is to get the knowledge of the diagnostic process when experts diagnose breast cancers by using mammograms. In this study, we investigated the classification logic for the diagnosis of breast masses on mammograms in cooperation with a mommographic specialist. From many features of the masses such as “sharp outline” and “stellate boundary”, we selected 13 features and weighted them by integer values from -1 to 2. The classification was achieved based on the total amount of the weighted values for all features. We tested our classification logic by use of 103 mammograms, and except the very difficult three cases the classification rate for defining the mass as malignant or benign was very high (sensitivity of 84% and specificity of 96%).