Ongoing development of microwave breast imaging system components

We present ongoing work at the University of Michigan toward the goal of developing a microwave inverse scattering system for breast cancer imaging. The algorithm is built around the Born Iterative Method with a modified cost function allowing inclusion of a priori information. The calibration is based on a new formulation for S-parameter measurements. The liquid matching medium is an oil-water emulsion with high dielectric constant and moderate loss. Last, we designed a wide-band, linear-phase antenna for both frequency and time-domain applications. Finally, we present numeric studies of a CW microwave breast cancer therapy system.

1. Introduction

Much work has been done in the field of microwave breast imaging. It is based around the assumption that benign and malignant tissues have different dielectric properties, and, if imaged, could provide additional diagnostic information. The two approaches to imaging are time-domain focusing, similar to traditional ultrasound beamforming, and inverse scattering, also called tomography. Beamforming provides fast, qualitative maps of tissue radar backscatter, however, these algorithms assume homogenous backgrounds and rely on the quality of the time domain pulses for resolution. Inverse scattering can provide quantitative maps of tissue permittivity and conductivity, but is computationally expensive, requires full antenna modeling, and are limited by the peak contrast of the object. Our work addresses some of the experimental aspects of breast imaging.