X-ray mammography is the most widely used method to screen asymptomatic women for early detection of breast cancer. The large number of mammograms generated by population screening must be interpreted and diagnosed by relatively few radiologists. It is considered that the use of computerized mammographic analysis will make a vital contribution to easing the increasing workload and assisting in the detection of breast cancer. These authors contend that before the digitized mammogram is analyzed by computer, it must be segmented into its representative anatomical regions. Three anatomical landmarks have to be first extracted automatically: they are the breast border, the nipple, and the pectoral muscle. In this paper, a method for automatically segmenting the pectoral muscle on mediolateral oblique (MLO) view mammogram is proposed.
When the MLO view is properly imaged, the pectoral muscle should always appear as a high-intensity, triangular region across the upper posterior margin of the image. The craniocaudal (CC) view is not considered in this paper because the pectoral muscle is only seen in about 30-40% of CC images. Several factors complicate the segmentation of the pectoral muscle. Depending on anatomy and patient positioning during image acquisition, the pectoral muscle could occupy as much as half of the breast region, or as little as a few percent of it. The curvature of the muscle edge is usually convex, but it can also be concave or a mixture of both. Although the pectoral muscle boundary is perceived to be visually continuous by humans, there are large variations in edge strength and texture. In many cases, the upper part of the boundary is a sharp-intensity edge, while the lower part is more likely to be a texture edge, due to the fact that it is overlapped by fibroglandular tissue. In addition, the muscle edge may be obscured by artifacts on the digitized mammogram, such as sticky tapes. Because of all these factors, automatic segmentation of the pectoral muscle by computer is a demanding task.
Automatic pectoral muscle segmentation is useful in many areas of mammographic analysis. The work of Gupta and Undrill indicates that mammographic parenchyma and the pectoral region may have similar texture characteristics, causing a high number of false positives when detecting suspicious masses. In other words, the pectoral muscle could interfere with automated detection of cancers. Also, the area overlying the pectoral muscle is a common area for cancers to develop, and is particularly checked by radiologists to reduce false negatives. It is therefore necessary to segment out the pectoral muscle before lesion detection, as stated in Ref. 6. Similarly, exclusion of the pectoral muscle is required for automatic breast tissue density quantification.
Online access to SPIE eBooks is limited to subscribing institutions.