Digital Imaging Based Screening and Detection of Breast Cancer
Breast cancer affects thousands of (primarily) women and is best treated when detected early. Currently, the primary form of (successful) breast cancer detection is manual palpation, which detects tumors through their high contrast stiffness relative to the surrounding tissue. The DIET system (Digital Image-based Elasto-Tomography) exploits this contrast to create a novel approach to breast cancer detection based on material property reconstruction of harmonic excitation data from the surface of the breast. These motions are measured using digital imaging in combination with a variety of novel imaging algorithms and approaches. The result is a conceptual technology approach to breast cancer screening that is small, potentially low cost, scalable, and does not involve X-rays, thus opening the opportunity to offer screening at any age with potentially greater compliance than found with mammography. Relying on silicon technologies it is also highly scalable as these technologies improve (regularly) in resolution and capability.
This seminar presents the overall DIET concept and the research successes and "difficulties" to date at the University of Canterbury in New Zealand. A particular focus is paid to providing an overview of the multi-disciplinary methods, technologies and algorithms required to apply material property reconstruction from a strictly optical (digital imaging and computer vision) based sensing system in this Bio-Engineering application. Hence, the talk covers all the major technological modalities required to see this form of breast cancer screening emerge. An overview of methods is given and results are shown for proof of concept simulation, all major imaging and image processing aspects, and recent proof of concept silicone phantom studies on a pre-pre-prototype experimental system.