Automated Deep Learning Method for Whole-Breast Segmentation in Synthetic MRI

Motivation: In quantitative breast MRI studies, segmenting the whole-breast region is a key initial step for several clinically relevant applications. Goal(s): To develop a deep learning segmentation method utilizing the nnU-Net architecture for fully automated whole-breast segmentation based on diffusion-weighted imaging (DWI) and synthetic MRI (SyMRI) images. Approach: The U-Net and nnU-Net deep learning algorithms were employed to segment the whole-breast on DWI and SyMRI images. Results: The nnU-Net outperformed the U-Net and the PD of SyMRI exhibited better performance than DWI. Impact: The nnU-Net exhibited exceptional segmentation performance for fully automated breast segmentation of contrast-free quantitative images.