Vignesh A Arasu1, Roy Harnish1, Cody McHargue1, Wen Li1, Lisa J Wilmes1, David Newitt1, Ella Jones1, Laura J Esserman2, Bonnie N Joe1, and Nola M Hylton1
1Radiology and Biomedical Imaging, University of California, San Francisco, San Francisco, CA, United States, 2Surgery, University of California, San Francisco, San Francisco, CA, United States
Automated measurements
of whole breast segmentation are becoming an essential process to the
development of quantitative and reproducible imaging biomarkers.
We have developed a method for automated whole breast tissue segmentation and assess its
performance using a test dataset, and found approximately 75% of cases had satisfactory segmentation requiring none to minimal manual modification. The current method can likely provide accurate assessment of mean background parenchymal enhancement, but further refinement of breast-chest wall boundary identification is required for other measurements (e.g. breast density).
