Advancing the Clinical Potential of Carbon Nanotube-enabled stationary 3D Mammography

Scope and purpose. 3D imaging has revolutionized medicine. Digital breast tomosynthesis (DBT), also recognized as 3D mammography, is a relatively recent example. stationary DBT (sDBT) is an experimental technology in which the single moving x-ray source of conventional DBT has been replaced by a fixed array of carbon nanotube (CNT)-enabled sources. Given the potential for a higher spatial and temporal resolution compared to commercially-available, moving-source DBT devices, it was hypothesized that sDBT would provide a valuable tool for breast imaging. As such, the purpose of this work was to explore the clinical potential of sDBT. To accomplish this purpose, three broad Aims were set forth: (1) study the challenges of scatter and artifact with sDBT, (2) assess the performance of sDBT relative to standard mammographic screening approaches, and (3) develop a synthetic mammography capability for sDBT. Throughout the work, developing image processing approaches to maximize the diagnostic value of the information presented to readers remained a specific goal. Data sources and methodology. Sitting at the intersection of development and clinical application, this work involved both basic experimentation and human study. Quantitative measures of image quality as well as reader preference and accuracy were used to assess the performance of sDBT. These studies imaged breast-mimicking phantoms, lumpectomy specimens, and human subjects on IRB-approved study protocols, often using standard 2D and conventional 3D mammography for reference. Key findings. Characterizing scatter and artifact allowed the development of new processing approaches to improve image quality. Additionally, comparing the performance of sDBT to standard breast imaging technologies helped identify opportunities for improvement through processing. This line of research culminated in the incorporation of a synthetic mammography capability into sDBT, yielding images that have the potential to improve the diagnostic value of sDBT. Implications. This work advanced the evolution of CNT-enabled sDBT toward a viable clinical tool by incorporating key image processing functionality and characterizing the performance of sDBT relative to standard breast imaging techniques. The findings confirmed the clinical utility of sDBT while also suggesting promising paths for future research and development with this unique approach to breast imaging.Doctor of Philosoph