Design of a 3D Mammography System in the Age of Personalized Medicine
Abstract
Rationale: Over the past few years, 3D mammography, or digital breast tomosynthesis, has been adopted by many breast cancer screening centers. This technology offers many advantages over conventional 2D mammography. It has led to the detection of more cancers, particularly invasive cancers. Also, there are fewer “false alarms,” so women are less likely to be recalled for additional testing that ultimately finds no cancer. Although 3D mammography is an important advancement in breast imaging, there are many unresolved challenges to date. For example, out of all women referred for additional testing after a suspicious screening exam, it is still the case that only about 5% are diagnosed with cancer by biopsy. To achieve a more substantial reduction in the number of false alarms, there needs to be major design changes to the technology used in both screening and call-back exams. These are the two tiers of imaging studies that precede a breast biopsy. In the last few years of our research, we have been focusing on resolving the challenges of overdiagnosis and overtesting through a prototype 3D mammography system that we built at the University of Pennsylvania. The system is capable of more complex motion paths for the x-ray source. We have evidence that these new motions offer advantages in the visualization of dense tissue. This allows radiologists to estimate breast density better (for assessment of breast cancer risk), and it is less likely for dense tissue to hide a tumor or lesion. In addition, we designed the system with a new form of detector motion that allows for “super-resolution.” This greatly improves the visualization of small structures, such as calcifications, making them easier for radiologists to interpret. Although our prototype system represents an important step forward, it was not designed for clinical use and there are major design changes that need to be made for clinical translation. Objectives: Using the prototype system, we experimented with different motion paths for the x-ray tube and detector and found that the same motion path will not benefit all women the same way. For this reason, we are now proposing a personalized scan for each woman. The scan will be customized to the unique attributes of each breast (size, shape, and internal composition). This approach differs greatly from clinical screening exams, in which the x-ray tube follows the same motion path for every woman. Through our experience with the prototype system, we also gained insight into practical concerns that need to be addressed for clinical translation. To accommodate the new detector motion, we found that the detector housing needs to be thicker (bulkier), making patient positioning more cumbersome. We are now proposing different motion combinations for the detector and x-ray tube that will allow the detector housing to be much more compact. This will ensure that women remain comfortable during their exam. Overview of Specific Aims: To personalize screening, we propose that first a 2D “scout” image should be acquired at very low radiation dose. This image will be used to estimate the breast outline and internal composition. The motion of the x-ray tube and detector in the 3D acquisition will then be customized around these attributes. We are also proposing to apply similar principles to magnification mammography, a call-back exam arising after a suspicious screening exam. Currently, this exam still relies on 2D imaging methods. To upgrade this technology, we will develop a personalized, 3D imaging platform for magnification mammography. As the new system designs for screening and call-back imaging are developed, they will be evaluated with a virtual clinical trial using realistic, computer breast models. We will analyze the advantages of the personalized designs in terms of breast density visualization, as well as the detection and characterization of lesions. Ultimate Applicability of t
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Oct 29, 2018
- Source ID
- W81XWH1810082
Entities
People
- Raymond J Acciavatti
Organizations
- United States Army
- University of Pennsylvania