Automated segmentation of breast fat-water MR images using empirical analysis

Breast density (BD) has been advocated as a risk factor for the development of breast cancer. BD is typically measured from mammograms. However for longitudinal studies of patients at risk, BD can be better assessed using MRI due to the lack of ionizing radiation and the 3D capabilities of the technique. A fat-water (FW) imaging technique called RAD-GRASE was developed to acquire images of the entire breast in a few minutes and can generate fat-fraction maps, which can be used to assess BD. The time consuming manual segmentation on ~19 slices per exam can be challenging. In this paper, we present a method to automatically segment the breast tissue in FW images and yield FW profiles of the region of interest (ROIs).