(Left image) Normal ovarian surface epithelium - the layer of cells "wrapping" around the normal ovary -shows single cells, similar in size and shape, lined up adjacent to one another.
(Right image) In contrast to the benign ovarian surface epithelial cells, these malignant cells from a clear-cell carcinoma tumor show significant variability in size and shape, with overlap and larger (and darker) nuclei.
Images courtesy Dr. Allen Gown, PhenoPath Laboratories.
Ellen, a 53-year old office manager in New York, didn't pay much attention to the upset stomach she'd had for several weeks. But when she experienced abdominal bloating that caused her clothing to become unusually tight, she was concerned enough to pay a visit to her doctor.
Always careful about getting regular checkups, she had undergone a routine annual pelvic exam a few months earlier and complained of some abdominal tightness, but her doctor detected nothing abnormal.
This time the prognosis was stunningly different. Within days of her more recent doctor's visit, she was having her uterus and ovaries removed after being diagnosed with advanced ovarian cancer, a disease that will kill 14,000 women in the U.S. this year and is the leading cause of gynecological death in this country.
Ellen (not her real name) is typical of the 25,000 American women diagnosed each year with ovarian cancer, often called a silent killer because it presents with no specific symptoms - and because there are no reliable methods to test for the disease at an early stage.
But Hutch scientists in the Public Health Sciences Division believe there is a way to change the grim statistics. Their goal - one that will not be quick or easy - is to design a simple blood test that might detect cancer at its earliest stages and could be used to screen women who are at risk for developing the disease later in life.
As part of the Pacific Ovarian Cancer Research Consortium, a multi-site Specialized Program of Research Excellence (SPORE) in Ovarian Cancer funded by $11.5 million from the National Cancer Institute, Hutch scientists are leading two studies to:
Dr. Nicole Urban, SPORE principal investigator, said that earlier detection of ovarian cancer could have an enormous impact on survival rates.
"The average woman diagnosed with advanced ovarian cancer lives three or four years after the course of her treatment is complete," she said.
"In contrast, five-year survival for localized ovarian cancer is about 90 percent, suggesting that earlier detection would result in a significant mortality reduction."
What has hampered studies to identify potential biomarkers and risk factors in the past, said Dr. Garnet Anderson, principal investigator of the risk factor study, is access to blood samples and behavioral data collected from women prior to diagnosis of ovarian cancer.
"But we're in a unique position here at the Hutch because of the Women's Health Initiative," said Anderson, a biostatistician in the Hutch Public Health Sciences Division and an associate professor of biostatistics at the University of Washington.
WHI, coordinated at the Hutch, is a 15-year, nationwide study of women's health issues, including heart disease, osteoporosis and cancer. In addition to completing extensive questionnaires on their family history, diet, use of medications and other behaviors, the 160,000 women enrolled in WHI have provided blood samples multiple times during the study.
Those stored blood samples, said Anderson, who also is co-principal investigator of the WHI coordinating center, may contain clues to the earliest warning signals of ovarian cancer.
Based on the incidence of ovarian cancer in the general population, Anderson predicts that about 240 of the 93,000 women whose data she will analyze will eventually be diagnosed with the disease.
"By analyzing biomarkers in their blood relative to the time of diagnosis, we can ask: Were they elevated? Could we have detected the disease earlier? Were there other factors that influenced their disease risk?"
In addition to evaluating WHI participants, Hutch researchers will use data and samples collected by research collaborators in the Northwest and California.
But the actual biomarkers still need to be identified, an effort being undertaken by another research component of the SPORE that is looking for genes that are expressed at high levels in ovarian tumors relative to normal tissue.
One molecule already known to be elevated in women with ovarian cancer, called CA-125, is used primarily to evaluate a woman's response to treatment and, although discovered in 1983, has not yet proved to be a sensitive, early-detection method.
Evaluating which newly discovered biomarkers might be used to create sensitive tests, as well as refining methods for utilizing CA-125 for early detection, is the goal of a SPORE component led by PHS investigator Dr. Martin McIntosh, who is also a research assistant professor of biostatistics at UW.
McIntosh said the challenge of biomarker development is to find one that can identify most women with the disease while not placing too high a burden on the majority of women without it. While false-negative results mean some cases will go undetected, false-positive results can cause anxiety, as well as lead to considerable expense and additional health risks due to unnecessary surgery, especially for ovarian cancer.
"Women who have false-positive mammograms after screening for breast tumors would follow up with biopsies," he said. "A false-positive result from an ovarian cancer test would be followed with laparoscopic surgery, a much more invasive procedure."
McIntosh and his colleagues are pursuing two avenues to develop biomarker screening and diagnostic tests.
"It may be possible to use a single marker, maybe CA-125 or another molecule, in a better way. Another approach is to use multiple markers, in combination."
A group at Harvard led by Dr. Steven Skates, developed an algorithm that uses CA-125 levels prior to cancer diagnosis to predict disease onset. PHS investigator Dr. Donna Pauler worked on the project before joining the Hutch last summer.
McIntosh said relatively slow increases in CA-125 levels - even before reaching the threshold of 35 U/ml that doctors use to screen for cancer - may be informative for early detection. A study in the United Kingdom is under way to evaluate the potential of such a test. Subtle differences in women's baseline CA-125 levels also could have predictive value, McIntosh said. "We may be able to use a similar approach with other markers, such as HE4," McIntosh said.
HE4 was identified by Dr. Michel Schummer, one of the SPORE collaborators in Dr. Leroy Hood's group at the Institute for Systems Biology.
But McIntosh pointed out that one marker alone may not detect all forms of ovarian cancer, so developing a panel of markers may be the most effective method.
Once sensitive tests are developed, Anderson's goal is to decide which women will benefit most from regular screening.
"Screening is costly," Anderson said, "and ovarian cancer is a relatively rare disease.
"If we can predict a woman's risk of developing cancer, we might be able to tailor a screening program such that low-risk women are tested infrequently while high-risk women are tested every six months or every year."
Besides mutations in the BRCA1 and BRCA2 genes, other known or suspected risk factors for ovarian cancer include reproductive factors. Ovulation frequency and duration over the course of a lifetime appears to increase risk - women who ovulate more have a higher incidence of cancer - while taking birth control pills exerts a protective effect.
Anderson and McIntosh hope reliable risk assessment and screening methods can be developed, but they caution that it is impossible to predict when.
"The ultimate value of a screening test cannot be determined without a randomized trial, and for ovarian cancer, that will involve many women," Anderson said. "Our goal is to develop a test that is worth the effort and has a high chance of delivering on the promise of early detection."
In addition to risk estimation and biomarker development, other components of the ovarian Specialized Program of Research Excellence (SPORE) include: