High-tech spy systems that enable the Central Intelligence Agency to conduct covert, often deadly, operations are being tested at the Medical Center to detect breast cancer early and accurately. Passage of a Senate appropriations bill may mean that clinical trials could begin by the end of the year.

The methods of the intelligence and medical communities are strikingly similar, whether they be tracking the movement of foreign troops or cancer tissue, or conducting landscape reconnaissance necessary for both target and tumor recognition.

Mitchell D. Schnall, M.D., associate professor of radiology and chief of the MRI (magnetic resonance imaging) section, detailed the pilot studies conducted at Penn at a recent news conference. Sen. Arlen Specter , Susan Blumenthal, deputy assistant secretary for women's health and assistant surgeon general, and federal intelligence and information specialists also attended.

Blumenthal, who initiated the project with the CIA and defense experts through the U.S. Public Health Services' Office of Women's Health in 1994, called the combination of medical and intelligence expertise an "innovative collaboration to do battle against a common national enemy." For the past few years, the National Aeronautics and Space Administration has worked with the National Cancer Institute to develop high-resolution, high-contrast digital mammography systems.

According to figures from the Department of Health and Human Services (DHHS), one in eight women will get breast cancer during her lifetime, up from one in 20 women in the 1950s. Since 1960, almost 1 million women in the United States have died from breast cancer, the most-frequently diagnosed cancer in women and the nation's second-deadliest cancer in women. Lung cancer is the first. However, with early detection, the five-year survival rate for breast cancer is 96 percent.

Blumenthal's office hoped to improve current imaging techniques for mammography and enhance newer technologies, such as MRI. To reach that goal, the office turned to the nation's military and intelligence agencies because defense imaging technology is considered to be a decade ahead of medical technology in knowledge. The intention was to extricate new approaches to diagnosis and detection from the "black box" of the CIA's classified intelligence technology.

"I thought, 'If we can see missiles from 18,000 miles away, if the Hubble telescope can see the surface of Mars, and if during the Gulf War we were able to put smart missiles down chimneys,' " Blumenthal said, " 'why can't we more accurately detect small lumps in a woman's breast right in front of us?' "

Mammography, a 40-year-old X-ray technology and the most-popular method of breast imaging, reportedly decreases deaths by 30 percent in women aged 50 and older. But there are obstacles: A mammogram cannot penetrate dense breast tissue. Mammograms miss cancer in 15 percent of cases, many of them in younger women who tend to have denser breast tissue. Mammography also has a high rate of false-positive readings, resulting in unnecessary biopsies, each averaging $2,000, according to DHHS figures. In three out of four of the 500,000 biopsies done yearly on suspicious lesions detected by a mammography, no cancer is found.

At the news conference, several advanced imaging technologies were presented via computer models flashed on overheads. In each case, the intelligence use was described, followed by an explanation of the medical application.

One technique is a two-dimensional image alignment used by the military to study changes in an area captured on aerial-reconnaissance film. The change might be troop movements or the appearance of helicopters; in mammography the change would be growth of cancerous tissue in the breast. By superimposing precisely images taken over a period of time and deleting areas that remain unchanged, subtle alterations can be easily identified.

Another technique is called three-dimensional volume alignment. It is used by the military to evaluate changes in target scenes captured in images taken from a variety of angles. The technology is being tested to enhance MRI, which uses multiple scans to search for differences in tissues; cancer appears as a bright spot. The problem with MRI is that patient movement skews the alignment of the scans, which is already difficult because of the breast's soft tissue. Physicians also might wrongly label bright areas as cancerous. According to Schnall, bright regions indicate cancer only 50 percent of the time. Scans aligned using defense techniques may improve accurate identification.

Physicians also use MRI to examine the rough edges of lesions, which can be useful in differentiating cancerous from benign tumors. But current technology is inadequate. Intelligence experts must distinguish objects--a military craft from a passenger plane, for instance--from miles away. They use a neural network system to take an outline of the object in question and compare it to a known outline logged in a database. Schnall and his colleagues are using this technique to analyze the borders of a lesion, hoping to train the system eventually to recognize a pattern as cancerous or benign.

A similar technique used by the military to detect small targets, such as tanks or buildings, is being studied at the University of Chicago. In this case, the neural network system may reduce the difficulty mammography has with imaging "noise" or surrounding tissue. It might also help reduce the rate of false-positives.

Schnall noted that it might take two years to complete clinical trials on the studies at Penn. Sen. Specter, as chair of the Senate Subcommittee on Intelligence and chair of the Senate Appropriations Subcommittee on Labor, Health and Education, pushed Schnall to name a date when clinical trials could begin "if cost were not an issue."

Cost and administrative issues aside, Schnall replied, "Tomorrow." Specter promised a speedy passage of the Senate's appropriations bill.

Blumenthal stressed that the Food and Drug Administration has promised swift approval for the techniques once clinical trials have been conducted and proven to be successful.

Return to Compass Features for March 26, 1996