Contributions to cost reduction and sensitivity improvement of microwave breast cancer detection

This dissertation focuses on technology development that contributes to cost reduction and sensitivity improvement in microwave breast cancer detection systems.;A microwave reflectometer measures scattered signals from breast tissue, which allows subsequent signal processing to obtain breast images for screening and diagnosis purposes. In current systems, a commercial vector network analyzer (VNA) functions as a reflectometer, which connects to either a single antenna or an antenna array via switches. An ideal configuration would be a fully-populated reflectometer-antenna array with each antenna connected to its own reflectometer. The VNA is too bulky and expensive for this approach. To address this problem, a low-cost mixer-based reflectometer-antenna subsystem is developed. A free-space calibration method is developed to eliminate the need for connectorized mechanical standards, compensate for antenna reverberation, and allow flexibility in the choice of reference plane position in measurements. Measurement results using the prototype compare favorably with those using a VNA. This free-space calibrated reflectometer-antenna subsystem offers potential for cost reduction in a full-array based microwave breast imaging system.;Heterogeneity in the dielectric properties of normal breast tissue poses a challenge for current microwave techniques to distinguish malignant from normal fibroglandular tissue. Hence, an acoustic and microwave hybrid modality that exploits both dielectric and elastic properties contrasts in breast tissue is proposed. Two methods of inducing tissue deformation in breast---global excitation with a compression plate and local excitation with a focused acoustic transducer---are investigated. 3-D numerical techniques are developed to efficiently model this multi-physics problem. An acoustic beamformer is designed and the finite-element-method based acoustic-structural simulation is conducted to compute the induced tissue deformation. The finite-difference time-domain based numerical method with the use of sheet boundary conditions is adopted and extended to 3-D to simulate the microwave scattering response of the acoustically excited tissue. The numerical study demonstrates a larger and inverted microwave Doppler scattering contrast between the malignant and normal fibroglandular tissues in comparison with their fundamental scattering contrast of a pure microwave detection scheme. This reveals the potential of the hybrid modality for improved sensitivity in detecting a malignant tumor in a heterogeneously dense breast environment.