Can texture of tissue surrounding microcalcifications in mammography be used for breast cancer diagnosis?

This study investigates whether texture properties of the tissue surrounding microcalcifications (MCs) can contribute to breast cancer diagnosis. A case sample of 100 MC clusters (46 benign, 54 malignant) from 85 dense mammographic images included in the Digital Database for Screening Mammography, is analyzed. Regions of interest containing clusters are processed using wavelet-based enhancement and individual MCs are segmented by local thresholding. The segmented MCs are removed from original image data and the surrounding tissue area is subjected to texture analysis. The feasibility of four texture feature sets (first-order statistics, gray level co-occurrence matrices, gray level run length matrices and Laws' texture energy measures) in discriminating malignant from benign tissue was investigated using a k-nearest neighbor classifier. Laws' texture energy measures achieved the best classification accuracy 89% (sensitivity 90.74% and specificity 86.96%).