Investigation of 3D electrical impedance mammography systems for breast cancer detection

Breast cancer is a major disease in women worldwide with a high rate of mortality, second only to lung cancer. Hence, there is considerable interest in developing non-invasive breast cancer detection methods with the aim of identifying breast cancer at an early stage, when it is most treatable. Electrical impedance mammography (EIM) is a relatively new medical imaging method for breast cancer detection. It is a safe, painless, non-invasive, non-ionizing imaging modality, which visualizes the internal conductivity distribution of the breast under investigation. Currently some EIM systems are in clinical trials but not commercialized, as there are still many challenges with sensitivity, spatial resolution and detectability. The research in this thesis aims to enhance and optimize EIM systems in order to address the current challenges. An enhanced image reconstruction algorithm using the duo-mesh method is developed. Both in simulations and real cases of phantoms and patients, the enhanced algorithm has proven more accurate and sensitive than the former algorithm and effective in improving vertical resolution for the EIM system with a planar electrode array. To evaluate the performance of the EIM system and the image reconstruction algorithms, an image processing based error analysis method is developed, which can provide an intuitive and accurate method to evaluate the reconstructed image and outline the shape of the object of interest. Two novel EIM systems are studied, which aim to improve the spatial resolution and the detectability of a tumour deep in the breast volume. These are: rotary planar-electrode-array EIM (RPEIM) system and combined electrode array EIM (CEIM) system. The RPEIM system permits the planar electrode array to rotate in the horizontal plane, which can dramatically increase the number of independent measurements, hence improving the spatial resolution. To support the rotation of the planner electrode array, a synchronous mesh method is developed. The CEIM system has a planar electrode array and a ring electrode array operated independently or together. It has three operational modes. This design provides enhanced detectability of a tumour deep within the tissue, as required for a large volume breast. The studies of the RPEIM system and the CEIM system are based on close-to-realistic digital breast phantoms, which comprise of skin, nipple, ducts, acini, fat and tumour. This approach makes simulations very close to a clinical trial of the technology