Bhagirathi Halalli and Aziz Makandar
ABSTRACT
Breast cancer is the second-most diagnosed cancer in women (after skin cancer) and can often be found through self-examinations before any symptoms appear. According to the America Cancer Society, about 1 in 8 women will be diagnosed with breast cancer in their lifetime. Early detection is critical to improve outcomes, survival rate, and successful treatment of each patient. Computer-aided diagnosis (CAD) technology is used to improve the quality of mammograms, which can assist the radiologist with interpreting the images correctly and identify possible abnormalities that are undetectable by the human eye. While not all breast cancers can be detected through a self-examination or a mammogram, regular screenings can greatly increase the chances of early detection, successful treatment, and even a cure. The use of CAD in mammography can increase sensitivity of mammograms and plays a supporting and final interpretive role in diagnosis. This continuing education course will outline the process of analyzing mammograms utilizing a CAD system and present an updated system that introduces advanced techniques used to overcome the difficulty of identifying subtle signs of breast cancer and its stage. Segmentation, or the separation of the abnormal part (mass) from normal breast tissue on an image, will be presented with current approaches and practices as well as modifications for mammograms. In addition, feature extraction techniques and classification methods, the 2 most important steps in designing CAD systems, will be examined and compared.