The Shape of Breast Cancer
*Brook K. Byrd1, *Venkataramanan Krishnaswamy2, *Timothy B. Rooney3, *Rebecca A. Zuurbier3, Richard J. Barth Jr.4
1Thayer School of Engineering, Dartmouth College, Hanover, NH;2CairnSurgical, Inc., Lebanon, NH;3Department of Radiology, Dartmouth-Hitchcock Medical Center, Lebanon, NH;4Department of Surgery, Dartmouth-Hitchcock Medical Center, Lebanon, NH
Objective: Little is known about the shape of breast cancer. Many surgeons assume cancers are spherical and base their excisions, especially when guided by a point source, on that assumption. We obtained pre-operative supine MRI images and derived tumor shapes by outlining their edges on successive MRI slices.
Design: Retrospective database review.
Setting: Academic medical center.
Patients: 65 patients undergoing partial mastectomy: 57 invasive cancer, 8 DCIS.
Interventions: Supine MRI.
Main outcome measures: Breast shape: 3-dimensional tumor models were categorized into four categories (spherical, discoidal, segmental and irregular) based on geometrically defined parameters. Ideal resection volumes were determined by adding 1 cm in every dimension to the actual tumor volume and were compared to spherical and actual resection volumes.
Results: 35% of tumors were segmental, 32% discoidal, 21% spherical and 12% irregular. No DCIS tumors were spherical. When the smallest spherical volumes possible were fit to the ideal resection shapes, the volume of tissue defined was 2.2 and 2.7 times greater than the ideal resection volume for discoidal and segmental shaped tumors, respectively. The actual volume excised by surgeons who could see the 3D tumor shape during surgery was 1.4 and 1.7 times the ideal resection volume for discoidal and segmental tumors, respectively.
Conclusions: Information from supine MRIs can be used to display and classify breast tumor shapes. Most breast cancers have segmental or discoidal shapes, few are spherical. Knowledge of tumor shape may facilitate more precise lumpectomies. 
Back to 2019 Abstracts
