Mammographic feature analysis of clustered microcalcifications for classification of breast cancer and benign breast diseases

The authors are developing a computer-aided-diagnosis approach of classifying breast cancer and benign breast disease based on clustered microcalcifications in mammograms. The classification (malignant versus benign) is made by an artificial neural network (ANN) using computer-extracted features of microcalcifications and of clusters as input. The final diagnostic recommendation is made by a radiologist who takes the computer-estimated probability of malignancy into consideration