Towards Translation of Dedicated Breast Cone-Beam CT for the Detection of Microcalcifications

Abstract

Breast cancer is the most common cause of cancer in women and the second most frequent cause of cancer death in women. In a woman’s lifetime, there is approximately a 1 in 8 chance of her being diagnosed with breast cancer, and an increased chance based on genetic and familial risk factors. There is a clear and present need to reduce the burden of breast cancer through screening and diagnosis for early detection of the disease and to potentially begin treatment earlier. Breast cancer screening using mammography has demonstrated mortality reduction. Currently, the primary breast cancer screening modality is full-field digital mammography (DM) and is being steadily replaced by digital breast tomosynthesis (DBT). DBT provides quasi 3D images along with a synthetic 2D “mammogram”. The 2D nature of DM however suffers from tissue superposition, where structures are overlapped and distinguishing between them is challenging; this problem is exacerbated in dense breast tissue. While DBT provides quasi-3D images, it suffers from out-of-plane artifacts which distort the image, and the information is incomplete due to limited angle acquisition. Dedicated breast CT (BCT) provides a solution to alleviate patient discomfort due to compression, while also providing full 3D images, potentially improving diagnostic capabilities. My project evaluates the detectability of soft tissue mass lesions compared with microcalcification clusters to explore options for optimizing system parameters for the specified imaging tasks. Then, using the same cascaded system analysis, my project aims to optimize the combination of system parameters to maximize microcalcification detection using numerical observer models. The last part of my project involves investigating the imaging performance of a newly developed, upright breast CT clinical prototype system by imaging a breast-mimicking phantom to derive spatial-frequency-dependent quantitative imaging metrics such as MTF and NPS. Additionally, we report subjective interpretation of mass lesion and microcalcification detectability in the phantom. Finally, dosimetry data is reported for a range of scan protocol parameters.