| Summary: | Breast cancer is one of the most common and alarming diseases for women. With early detection and diagnosis, the chances of successful treatment and survival would improve. The diagnosis includes classification or staging of the breast cancer, which plays an important role in the prognosis of the disease hence determining the best treatment for the disease. Breast cancer staging is a way of describing the size of a cancer and how far it has grown. One of the staging numbering systems is the Tumour-Node-Metastasis (TNM) system, which categories cancer into several stages. Most types of cancer have four stages, numbered from I to IV. Normally, staging is determined by doctors by examining the pathology information which describes the spreading area of the cancer. In this study, a series of computer algorithms is applied to produce the area information of the cancerous region. The methodology consists of four phases which are image gathering; image pre-processing using Median filter and PCA; image segmentation using FCM; and staging using the area of the primary tumour as a representation of size in the TNM system. From thirty-five randomly selected mammography images of Malaysian women with malignant tumour, 14.3% is in stage I, 5.7% is in stage II and 80% in stage III. Experimental results also shows that 96.4% accuracy is obtained for stage III but higher errors occur at the boundaries of the staging scale. © Published under licence by IOP Publishing Ltd.
|
|---|