Mining Breast Cancer Classification Rules from Mammograms

Abstract

Breast cancer is a leading cause of cancer death in women. Early diagnosis and treatment are crucial to reduce the mortality rate and increase patients’ lifespan. Mammography is effective in early detection. This study proposes a computer-aided diagnosis system based on the mini-Mammographic Image Analysis Society database for analyzing mammograms. After selecting the regions of interest, we computed three typical features: the shape, spatial, and spectral domain features. We then applied the structural equation model to obtain relations between the features and the breast tissue type, lesion class, and tumor severity after feature extraction by information gain. Finally, we used the decision tree and classification and regression tree to construct computer-aided diagnosis rules; we generated 10 rules for predicting the classification of abnormal lesions and 11 rules for classifying the tumor severity. These rules can help clinicians detect and identify breast cancer efficiency from mammograms and improve medical care quality.

Keywords

• computer-aided diagnosis system
• mammograms
• breast cancer classification