Breast Cancer Detection Simulation Based On UWB Antenna And MRI-derived Breast Model

Breast cancer has become the biggest threat to women's health. Early detection can urge patients to get prompt treatment and thus will help improve the survival rates. Nowadays, X-ray mammography is most commonly used to detect breast tumor, as it has a relatively high accuracy. However, there are some fatal flaws such as exposure to ionizing radiation which may increase the risk of having cancer, compressing of breast which makes women feel uncomfortable. Therefore, this method cannot be used frequently. During the past decade, the ultra-wide band(UWB) microwave imaging for breast cancer detection is presented and further studied. As a complementary technique, it has many advantages such as safe, inexpensive, easily integrated, and portable, which can be used as a regular self-examination method.At present, there are numbers of research groups are engaging in developing this technique. In many early numerical studies, the detection experiment is modeled without the consideration of antenna structure. Instead, the point source is regarded as the antenna for simplicity. In addition, the breast model used is simple cubic and hemisphere. However, the breast outline and the distribution of breast tissues are more complex. Therefore, in order to more closely approximate the microwave propagation in the real environment, the antenna structure and anatomically realistic breast model should be added into the simulation system.In this dissertation, the electromagnetic simulation is completed based on finite-difference time-domain(FDTD) algorithm. Then, UWB antenna modeling is completed by adding the resistive voltage source to the antenna feed port with the lumped element technique. In total, two kinds of antennas are modeled. One operates in the air; the other operates with contacting the skin. Sequently, the second antenna is employed to form an array, adding into the breast tumor detection system simulation. The signals obtained from simulation are processed by the imaging algorithm to verify the feasibility of the imaging algorithms.Three kinds of breast model is derived from clinical MRI(magnetic resonant imaging) images. The first one is a tissue segmented model. Based on this model, the influence of the gland amount on imaging quality is studied and a method to detect multiple tumors is proposed. The second one is a two-dimensional model which will be applied to FDTD simulation. This model is much more anatomically realistic, which approximate the breast outline and the tissue distribution. The third one is a three-dimensional model which is developed by interpolating the 2D MRI slices. Finally, both 2D MRI-FDTD model and 3D MRI-FDTD are integrated into the FDTD program and verify the feasibility of the imaging algorithms through simulation experiments.