The human ovary, incubator of the eggs from which we all begin, is hidden deep in the pelvis. This protected location may enhance our chances for survival, but it also is a major reason why ovarian cancer often goes undetected until it is too late to save the patient.

More than 22,000 U.S. women will be diagnosed with ovarian cancer this year, estimates the American Cancer Society.

Epithelial ovarian cancer, the most common type, is usually not discovered until its advanced stages. Ovarian cancer is the fourth-leading cause of cancer-related death in the United States -- three times more lethal than breast cancer despite being one-tenth as common. Because it is so aggressive, a cure often requires catching the epithelial-type cancer before a discrete mass of abnormal cells is formed. Only 15 percent of women found to have such a mass outside the ovary survive for five years. But warning symptoms, such as abdominal bloating or pain, generally don't occur until the tumor has already spread beyond the ovary.

The goal of current research is to improve diagnostic capability and narrow the gap between cancer occurrence and cancer detection. What's dividing researchers is whether to push for a technologically sophisticated screen sensitive enough for use in high-risk women, including women with a family history of breast or ovarian cancer -- or one cheap enough to use in a broad population. What both philosophies share is a sense of urgency.

When I first began practicing medicine, I cared for a woman of 23 who complained of back pain for months before a neurologist ordered an MRI, expecting to find a herniated disc. Instead he found ovarian cancer spread to the spine. (I began seeing this patient after the cancer had done its damage.) The patient, treated with extensive surgery and chemotherapy, was one of the "lucky cures": She survived, though she was left with nerve damage and a loss of sensation that has ruined her sex life.

Though this case made me more sensitive to the risks of this deadly disease, it didn't improve my ability to detect it. Like many clinicians, I wished for an early detection test but didn't have one. In one case where I discovered a case of ovarian cancer early enough to save the patient, it was purely by accident: The diagnosis followed an ultrasound that I ordered because of abnormal liver tests. After this patient's "incidental" tumor was removed and didn't recur, I half-wondered if all women -- despite the great expense -- should be ultrasounded just in case.

David Fishman, a professor of obstetrics and gynecology at New York University School of Medicine, says he has a better idea. Fishman, who directs the National Ovarian Cancer Early Detection Program, uses a multi-tiered approach to identify women most likely to benefit from intensive screening. At present, only 10 percent of ovarian cancer cases are found in women who can be predetermined to be in high-risk groups.

He begins with genetics, identifying women at greatest risk because of a family history of the disease or other risk factors. Then he applies sophisticated techniques, including the latest in protein study (proteomics) and spectrometry, to analyze millions of protein fragments for a pattern mathematically predictive of ovarian cancer. The goal is to identify women with a high likelihood of having cancer before more testing and possibly surgery to remove the ovaries are considered. For a blood test to be useful in a large population, it needs to be almost 100 percent specific for a disease. That is, if the test indicates you have the disease, that prediction must be accurate.

Detecting the behavior of proteins as they begin to act abnormally is groundbreaking work. Says Fishman, "This technique holds promise for many other cancers and chronic diseases, though clinical validation in a large group is crucial and has not yet been achieved."

In the meantime, Fishman aims to establish an early detection program clinicians can use, by combining blood analysis with genetic risk assessment, a specialized ovarian Pap test, and an experimental contrast ultrasound study that uses special nanoparticles to identify the abnormal blood supply unique to cancer before a mass forms .

Early detection centers focus on women who may be at high genetic risk for ovarian cancer -- women like Cara Kealy, a 36-year-old mother of two from Mount Vernon, N.Y. After Kealy developed breast cancer at age 30, she was found to have a BRCA-1 gene mutation, which is associated with a high risk of ovarian cancer. Fishman was thus able to tell her, based on established pathological testing of ovaries removed from other patients, that her ovaries have a 20 percent chance of having premalignant changes. This was enough to make Kealy decide to have her ovaries removed once she no longer wants the option of becoming pregnant again. "I know I'll have them out. It won't be this year, but it will be soon," she said.

Fishman hopes to extend his experimental protein test to women who may be at as great a risk as Kealy and not know it. But broad-scale clinical application will require approval from the Food and Drug Administration, ample financial support from product manufacturers and linking of centers around the world.

The spirit of international cooperation that Fishman envisions for the acceptance of his protein test doesn't yet exist. Critics of his spectrometric blood analysis haven't been able to reproduce his findings of almost 100 percent specificity, which were published in 2000. Fishman says improvements in technology -- he can now look at millions of protein fragments at once -- make his results far more reliable and reproducible. But a group of researchers at Yale remain unconvinced. They also say the state-of-the-art computer required for Fishman's method is too costly to be practical for broad use in a clinical setting.

Focus on Volume

Instead, the Yale team is studying a much smaller number of whole proteins, using more traditional test tube techniques.

Led by Gil Mor and David Ward, the Yale group published a study in the May 10 issue of the Proceedings of the National Academy of Sciences, identifying four proteins -- leptin, prolactin, osteopontin and insulin-like growth factor-II -- that all tend to be abnormal in ovarian cancer more than 95 percent of the time. True, the abnormalities aren't unique to ovarian cancer, acknowledges Mor, director of the Reproductive Immunology Unit at Yale University School of Medicine, but identifying abnormalities in four complete proteins -- not just protein fragments -- is "a good start," he says; he intends to end up with 12 or more in an assay, or blood analysis, that will cost $10 to $20 to run -- instead of thousands of dollars.

"You're not going to get an $18,000 mass spectrometer in every doctor's office," Mor said. "Instead, we need a combination of well-characterized proteins and how they behave in certain abnormal patterns."

Like Fishman, Mor's group combines basic research with clinical experience, targeting women in high-risk groups. He sees the whole-protein blood test as "the red light that something is going on in the ovary. Hopefully it will be like a mammogram for the ovary, useful for women at higher risk of cancer."

Mor's partner in the study, David Ward, recently moved to Las Vegas to become deputy director of the Nevada Cancer Institute. Like many other gamblers in Vegas, he likes to play the odds, only instead of cards or dice he bases his predictions on the biological properties of proteins. While there is not yet "a totally reliable, robust, inexpensive early test for many types of cancer, including ovarian," Ward said, "I think the odds are a 100-1 that in two to three years there will be a test for every woman. It is doable with current technology. We need to get there by sharing and combining our data, by cooperation."

But before full collaboration can occur, there needs to be a resolution in methods. Sam Mok, director of Harvard Medical School's Laboratory of Gynecologic Oncology at Brigham and Women's Hospital, suggests that a combination of techniques may work the best, with Fishman's spectrometry -- an excellent discovery tool as a biomarker as an excellent discovery tool, but still too inaccurate to use in clinical practice. Measuring a small number of biomarkers, following the Yale team method, he says, "is probably more reliable and the more promising approach to use in the clinic."

Mok makes an important distinction between a screening test that is useful for high-risk groups and one that can be applied to the entire population. He said, "Since only 10 percent of ovarian cancers occur in the high-risk group, the impact will be far greater if we can find a blood test that meets the higher bar for the general population."

Like Fishman and Ward, Mok sees the goal of screening the whole country as achievable with current technology. But it would require an international effort and the extension of the multi-tiered approach to larger groups of women. "Studies to date indicate that ultrasonography can increase the chances of finding ovarian cancer by about tenfold, from 2 percent with a blood test alone to about 20 percent with a blood test combined with imaging," he said. "However, we don't know if this is sufficiently sensitive for early stage disease. The surgeon needs to feel confident that he will have at least a 10 percent chance of finding cancer identified by the screening program. We need studies in large groups to prove that we can separate those who require an operation from those who don't."

To get these large clinical studies to occur, it would seem that the big centers will need to do a better job of realizing the value in each other's work. The Food and Drug Administration can broker these deals and make sure that the funding is sufficient to support the necessary research and application to patients in practice. All have the same goal: removing a diseased ovary before it is too late.


Marc Siegel, an associate professor of medicine at the New York University Medical School, last wrote for Health about the treatment of depression during pregnancy. His new book, "False Alarm: The Truth About the Epidemic of Fear," will be published in August.

Ovarian cancer cells, shown in a tinted microscopic image, are hard to detect until they've spread beyond the ovary.