BEACON Researchers at Work: Using evolutionary computation to enhance breast tumor recognition in microwave images

This week’s BEACON Researchers at Work blog post is by MSU graduate student Blair D. Fleet.

Blair Fleet

In 2010, I received my B.S. in Electrical Engineering from Morgan State University. During a college visit for graduate school, I became introduced to evolutionary computation (EC) through a presentation given by Dr. Erik Goodman. As a result of the presentation, I saw the endless possibilities of using EC to solve and optimize a variety of engineering problems. Evolutionary computation is a topic that encompasses using optimization techniques that have a theme governed by evolution and/or nature, such as genetic algorithms, or particle swarm optimization.  From then on, it became a goal of mine to concentrate my future studies in using evolutionary computation algorithms to solve electrical engineering problems. I am currently pursuing my M.S. in Electrical Engineering at Michigan State University with a concentration of Signal Processing and Evolutionary Computation.

I’ve recently decided to pursue my Ph.D. in Electrical Engineering, which was heavily impacted by the opportunity that Dr. Goodman, the director of BEACON, presented to me. He informed me of collaboration with himself; Prof. Meng Yao, the principal investigator (PI) from East China Normal University located in Shanghai, China; and Dr. John Deller, a professor from Michigan State University whose research concentration is in signal and speech processing. The main goal of this collaboration is to use evolutionary computation to enhance and optimize the signal and image processing techniques being used in the BRATUMASS (Breast Tumor Microwave Sensor System) developed by Prof. Yao et al.

I immediately felt this was the perfect research topic for me because I saw and still see how powerful of an impact this research can have on the community. Unfortunately, African American women have the highest mortality rate from breast cancer of any ethnic group in the United States. This is primarily because of the difference in the awareness of and access to screening tests. I can’t stress enough the importance of finding cancer in the earliest of stages. A late diagnosis means a greater probability of the tumor being more aggressive, which leads to a greater chance of dying from the breast cancer.

Another reason why this research is important to me is because of a pivotal moment in my life, and my family’s lives. Five years ago, my mother was diagnosed with lymphoma, which is a type of blood cancer that can affect various places throughout the body. The cancer, which was located in her chest, was found after it had grown to be the approximate size of a tennis ball. She underwent surgery, and chemotherapy, and has been free of cancer ever since. I couldn’t help but think, what if her cancer was caught in the earliest stages? Would one of her vocal cords be paralyzed as result of her intensive surgery, which came from the advanced size of the tumor? Probably not. The key to surviving any type of cancer is to detect the disease at the earliest of stages, but that requires efficient screening and detection technologies.

Diagram showing A) structure of breast tissue and b) schematic of BRATUMASS detecting position

Figure 1. (a) A structure of breast tissue and (b) A schematic of the BRATUMASS detecting position, where the red dot represents the location of the antenna, the green dot represents the location of the metal slice, A is the transmitting antenna, B is the receiving antenna, and C is the center clapboard.

The BRATUMASS gives promise to a new, innovative way to screen for breast tumors.  The device uses ultrawideband microwave signals, which have a power of approximately 6 mW, to detect breast tumors. Note that this power usage from the microwave signals is less than that of the microwave signals emitted from the average cell phone, which is approximately 1 W. This means that radiation emitted from the antenna has no detectable impact on the human body. The ultimate goal of this research is to be able to use this procedure in place of damaging breast examination procedures such as mammograms, which involve ionizing radiation.

During a typical mammogram, the breasts are tugged, pushed and flattened, which leave patients feeling uncomfortable. During the BRATUMASS screening, one simply lays relaxed on one’s back while the specialist takes the transceiver and metal frame and goes around the circumstance of each breast. The transceiver collects data from several different positions around the breast. At each position, a pulsed microwave signal is sent from the transmitter (A) in the direction determined by the metal frame. Information is collected by the receiver (B) about the electric field based on the reflection and scattering of the microwave pulses.

There are notable differences in the microwave returns between normal breast tissue and malignant breast tissue; however, extensive research still needs to be done in that area. For example, normal breast tissue and malignant breast tumor have different dielectric constants. The dielectric constant for normal breast tissue is approximately 10, while it is usually greater than 50 in malignant breast tissue. The values of the constants will differ from person to person, so the techniques developed have to be applied to a wide variety of people. Our research will involve using the collected data to reconstruct an enhanced image of the breast while selecting and classifying tumors based on the artificial microwave features, e.g. dielectric constants, in addition to many other features characterized from using EC techniques. This research has the potential to serve as the safest means of early detection and localization of breast cancer.

For more information about Blair’s work, you can contact her at fleetbla at msu dot edu.

This entry was posted in BEACON Researchers at Work and tagged , , , , , . Bookmark the permalink.

Comments are closed.