Microwave imaging may yield better breast cancer detection
A radar technology used to detect anti-personnel land mines may find a promising application in the campaign for early detection of breast cancer.
Susan C. Hagness, an assistant professor of electrical and computer engineering and biomedical engineering, is researching the use of microwave radar imaging for breast cancer detection, as a complement to the standard use of X-ray mammography.
The widespread availability of X-ray mammography has been a life-saving success story, providing an inexpensive and simple approach to early detection. Despite progress, the technology still produces a relatively high number of false negative and false positive diagnoses, Hagness says.
Hagness is researching whether images from microwaves - the same microwaves used to communicate with digital cellular phones, only at lower power - will offer the sensitivity to solve those problems.
"Early detection is the key to survival, so we always want to do a better job of identifying tumors as soon as possible," Hagness says.
Hagness received a boost in September with a $207,000 grant from the The Whitaker Foundation, a non-profit organization dedicated to engineering research that improves medical care. She also started a research partnership with Frederick Kelcz, associate professor of medicine and expert on breast cancer detection.
Although clinical trials are down the road, the team is conducting microwave measurements on breast tissue excised during biopsies at UW Hospital and Clinics. They are also creating computer simulations that test microwave sensor designs in realistic environments.
The work seeks to answer two key questions, Hagness says. First, can microwave imaging detect extremely small tumors, only a few millimeters in size, to improve early detection? And second, can it differentiate between a malignant and a benign tumor, thus eliminating the trauma of unnecessary biopsies for women?
According to Kelcz, mammograms are known to miss a significant number of breast cancers, especially in younger women. False positives are even more widespread. Of the women who follow up with biopsies of an abnormality, about 70 percent of them do not have cancer.
Microwaves are a non-ionizing form of electromagnetic waves that interact with tissue according to water content. In Hagness' research, low-power microwaves are used like sonar to bounce a signal off the tissue. Since tumors have a higher water content compared to normal tissue, they will back-scatter the microwaves and produce an echo. The measurements of those echoes enable Hagness to develop a 3-D image of the tumor.
Sub-surface radar imaging has a number of applications to help reveal the invisible. In addition to land mine detection, they can be used to identify weaknesses in bridge structures, detect the thickness of glaciers or identify archaeological sites.
To date, Hagness is getting some promising images in her computer simulations of the microwave sensor. The ongoing research will help them learn how to distinguish cancerous tumors from harmless ones.
"We want to highlight those differences and determine the classic signatures to the microwave echoes," she says. "This could help us narrow down biopsy cases."
Kelcz says finding alternative approaches is important because early detection saves lives. He notes that when cancers are detected less than a centimeter in size, there is a 95 percent survival rate. But at 2-3 centimeters, the cancer often moves into lymph nodes and survival rates begin to decline.
"We need technologies with the sensitivity to pick up abnormalities early enough to make a clinical difference," he says.
Hagness adds that microwave devices are inexpensive enough to make them widely accessible in health care. Microwave imaging also carries no danger of radiation exposure, and would not require breast compression.
She began the work with staff of the Chicago company Interstitial, Inc., while finishing her doctorate at Northwestern University. Interstitial was recently awarded three patents for this technology; Hagness and her colleagues have two patents pending.