Automatic Approaches for CE-MRI Examination of the Breast: A Survey

Abstract--Breast cancer is the most frequently diagnosed non-skin cancer in women, the second leading cause of death among women. Early detection of a breast cancer is fundamental for ensuring high survival rate. Imaging techniques are used to identify suspicious modifications of breast tissue. Among these, contrast-enhanced magnetic resonance imaging (CE-MRI) is particularly interesting for its lack of exposure to radiation and its ability to highlight differences in vascularisation, typical of cancer lesions. Automatic or semi-automatic methods are especially useful with this technique, due to the high quantity of data, in the form of 4D images (3D space + time), to be analysed in each test. This survey describes approaches to fully automatic computer-aided detection/diagnosis of breast lesions with CE-MRI, with particular emphasis on computational intelligence techniques.