Photo by Todd McNaught
Fred Hutchinson will lead a $7 million initiative to develop a blood test that can be used with mammography to improve early detection of breast cancer.
The U.S. Department of Defense-funded Breast Cancer Center of Excellence will involve investigators from 10 institutions in Seattle, Houston and Los Angeles.
Dr. Nicole Urban of the Public Health Sciences Division serves as principal investigator of the four-year project, to be announced today. Dr. Nancy Kiviat, professor of pathology at the University of Washington School of Medicine and head of pathology at Seattle's Harborview Medical Center, will serve as co-principal investigator.
"Our goal is to evaluate a variety of breast-cancer biomarkers for their contribution to the early detection of breast cancer," said Urban, also a research professor of health services in the UW School of Public Health and Community Medicine.
"Our vision is that a simple blood test could be used along with mammography to detect breast cancer early in the disease process. We're particularly interested in markers that identify aggressive forms of breast cancer that are missed by mammography."
Mammography indeed saves lives, reducing deaths from breast cancer by up to 30 percent among women over age 50. But it misses about one in five breast cancers overall.
It's least effective in detecting cancer in younger women, whose denser breast tissue is difficult to X-ray. Its other limits in younger women include a high rate of false positives that result in needless biopsies and the inability to distinguish between aggressive, life-threatening cancers versus slow-growing tumors that may not need treatment.
"Despite decades of research on mammography as a tool for early detection, its efficacy remains in question, and some women still die of the disease despite annual screening," Kiviat said.
"We can dramatically improve breast-cancer detection and diagnosis by complementing mammography with molecular-screening tools that provide information about the genetic and cellular traits of their cancer and how it will behave."
The focus of the project, called the Center for the Evaluation of Biomarkers for Early Detection of Breast Cancer, will be to develop and test a panel of serum biomarkers - substances in the blood, such as proteins, or specific genetic sequences - that are detected in higher-than-normal amounts in women with breast cancer.
Blood and tissue samples will be obtained from 1,800 women undergoing mammography, biopsy or breast-cancer surgery during the next four years at Swedish Medical Center and Los Angeles-based Cedars-Sinai Medical Center.
Using biomarkers for cancer screening is not new; single-marker blood tests for prostate and ovarian cancer, for example, have been used widely for more than a decade. An enzyme called PSA, or prostate-specific antigen, is used to detect prostate cancer while it is still localized and treatable, while a protein called CA-125 can flag the recurrence of ovarian cancer. Studies are under way to test the effectiveness of CA-125 in detecting early-stage disease in asymptomatic women. However, no adequate breast-cancer biomarker has yet been identified, perhaps because there is no single pathway to breast-cancer development, Urban said.
"It's unlikely we will ever find a single marker that will perfectly detect and diagnose breast cancer, because breast cancer is not one disease, as not all breast cancers look the same at the molecular level," she said. " We challenge the paradigm that a single marker that detects all breast cancer is even necessary. Instead, we will seek a panel of about 10 markers that can be used with each other."
A breast-cancer biomarker panel could be used in several ways, Urban said.
For example, a woman could have her blood tested in conjunction with her annual screening mammogram. If the mammogram were clearly positive, she would be referred for biopsy. If the mammogram findings were ambiguous, she would be referred for biopsy only if the marker panel suggested a malignancy. If the mammogram were negative but the blood test suggested cancer, she would be referred for additional tests such as MRI, or magnetic-resonance imaging.
Such a marker panel also could be used between annual mammograms to find fast-growing cancers at an earlier stage than they would otherwise be identified, especially in high-risk women.
One factor in developing such a panel is detecting which markers are mutually complementary and best improve sensitivity when used together.
Also key to the success of such a screening tool is interpreting results in a statistically meaningful way, taking into account a woman's personal biomarker behavior, including how these levels may naturally fluctuate over time, instead of a "one-size-fits-all" approach to screening.
Leading this effort is Dr. Martin McIntosh of PHS and a research associate professor of biostatistics at UW. He already has developed such statistical methods for ovarian-cancer screening using new biomarker discoveries.
Change over time
"For early cancer detection, it is much more useful to look at change over time in multiple markers than to take a snapshot of any single marker," McIntosh said.
"For example, in any woman, some biomarkers may consistently fluctuate over time, while others may be quite stable. By giving greater weight to stable markers than unstable ones and by monitoring them over time, only small changes, like those expected in early-stage cancer, are needed to signal a positive screen."
Such an approach to early cancer detection is unique, Urban said.
"Martin is one of the few statisticians in the world to have studied longitudinal use of marker panels, taking into account a woman's biomarker history when determining whether the results are positive or negative," she said.
Who are the collaborators?
Just as synergy among biomarkers is needed to develop an effective screening tool, so too is synergy among investigators key to productive collaboration.
Already collaborating in cancer biomarkers, the group includes researchers from Fred Hutchinson, Group Health Cooperative of Puget Sound, Institute for Systems Biology, Pacific Northwest Research Institute, PhenoPath Laboratories, Swedish Medical Center, University of Washington and Virginia Mason Research Center, all of Seattle.
Others are M.D. Anderson Cancer Center, Houston, and Cedars-Sinai Medical Center, Los Angeles.
Supporting communication among the 10 institutions will be an innovative, Web-based system that will allow investigators to share real-time information.
Instead of linear and message-based like e-mail, the system is idea-based. Each topic or idea is given its own Web page within the system, and researchers then can add their comments and discussion to any of the topics.
"It's a way of communicating among any number of people who can post what they have to say on a Web page and allow others to edit," said Dr. Nicole Urban of PHS. "It's like creating a document collaboratively."
Supporting documents such as summary statistics, timelines, protocols and meeting minutes also can be posted, allowing investigators to jump quickly from discussion to data and references.
The biomarkers to be evaluated
- Circulating breast-tumor cells, which can be identified by the fact that they contain specific genetic instructions for making proteins associated with breast tissue, such as the protein mammaglobin.
- Growth factors detectable in serum or plasma that are associated with angiogenesis, or blood-vessel development, often present in cancer because tumors need a blood supply to survive. Growth factors to be evaluated include vascular endothelial growth factor, or VEGF, which is associated with 80 percent of breast tumors and is an indicator of poor prognosis among women with estrogen-sensitive tumors;
- Lipid markers, or fatty acids, detectable in serum or plasma, which are associated with pathways of cancer development. Preliminary studies with various lipid markers have resulted in breast-cancer detection rates of up to 90 percent with a "false-positive" rate of only 5 percent.
- Antibodies to mutated or over-expressed proteins that are associated with aggressive breast cancer, such as Her2/neu and p53, which are found in 30 percent and 50 percent of breast tumors, respectively; and
- Other promising new markers that may become available from emerging new technologies such as proteomics, which can identify telltale "protein signatures" in the blood that are linked to diseases such as cancer.