Featured Meetings and Events A calendar of scientific meetings and events sponsored by the National Institutes of Health (NIH) is available at http://calendar.nih.gov.

Epithelial ovarian cancer is known as the "deadly disease that whispers." Deadly because less than 30 percent of these cancers are detected in stage I when they are confined to the ovaries, highly treatable, and curable in up to 90 percent of patients; whispering because its early symptoms can be ambiguous or nonexistent.

For most ovarian cancer patients, however, the disease has begun to spread by the time it is detected, and only one in four patients with advanced ovarian cancer will be alive 5 years after diagnosis.

Ovarian cancer also provides the context for another story: the role of biomarkers in cancer research. A cancer biomarker is a substance in the body that - by its presence, pattern, or behavior - alerts clinicians to hidden biologic activity that might be associated with some aspect of the precancerous or neoplastic process.

For many years, a protein known as CA125 has been a prominent blip on researchers' radar screens. The history of CA125 illustrates the promise and the challenges involved in the search for a biomarker that could lead to meaningful clinical progress in fighting ovarian cancer.

The Early Years

After the development of monoclonal antibody technology in 1975, many scientists began the search for antibodies that would react with specific tumor tissue types. One of those researchers was Dr. Robert C. Bast, who would go on to become vice president for translational research and the principal investigator for the Ovarian SPORE at the University of Texas M.D. Anderson Cancer Center.

In the late 1970s, Dr. Bast and colleagues used the then new monoclonal technology to develop antibodies against human ovarian cancer, and their 125th hybridoma produced an antibody that would bind to antigens expressed by about 80 percent of epithelial ovarian cancers (which comprise about 90 percent of all ovarian cancers). They named it OC-125 and by 1983 Dr. Bast's group reported the first clinically usable, monoclonal antibody radioimmunoassay to monitor the response of epithelial ovarian cancer to treatment, known as the CA125 test.

Equipped with a test for an antigen that epithelial ovarian cancer cells shed into the blood, researchers wondered if they had a potential screening test.

"To stand alone as a screen for ovarian cancer, a biomarker would have to detect significantly more cases at the early, highly curable stage than the 28 percent we find now. We need a sensitivity of at least 80 percent," says Dr. Beth Karlan.

Dr. Karlan is director of the Gilda Radner Hereditary Cancer Detection Program at Cedars Sinai's Women's Cancer Research Institute in Los Angeles. Since a positive test could lead to abdominal surgery, its positive predictive value needs to be at least 10 percent. "You also want to avoid false-positives in a relatively rare disease, and so the specificity should exceed 99 percent," she explains.

The CA125 test did not meet this threshold, in part because elevated CA125 levels were also found in patients with other types of cancer, and also in several benign conditions; only about 3 percent of patients with elevated CA125 actually had previously undetected ovarian cancer. "Nonetheless we knew we had a potential tool, it was just a matter of learning how to use it," notes Dr. Bast.

Coming of Age

"A prognostic biomarker is something we can measure that correlates to disease outcome," explains Dr. Gary Kelloff, special advisor to NCI's Cancer Imaging Program in the Division of Cancer Treatment and Diagnosis. "If a cellular protein like CA125 can stand in as an effective surrogate for the standard way we measure how ovarian cancer responds to chemotherapy, it could economize and hasten the process of testing new drugs."

In advanced ovarian cancer, surgery to remove all detectable disease is the first-line treatment, usually followed by chemotherapy administered directly into the peritoneum. Unfortunately, most patients eventually see their disease recur and progress. The standard way to mark such recurrence is the Response Evaluation Criteria in Solid Tumors (RECIST) criteria, which rely on imaging to detect a visible mass.

"But when ovarian cancer does recur, CA125 sends us a message about 3 months before we could detect it in any other way," says Dr. Bast.

Since the late 1980s, Dr. Gordon J.S. Rustin has been conducting a series of clinical trials on how to translate this early warning signal most accurately, since many clinicians were monitoring CA125 levels but interpreting them differently. The Gynecologic Cancer Intergroup (GCIG), formed in 1997, joined many of the cooperative groups into a common organization to foster collaboration and consensus.

By 2006, GCIG had agreed on definitions for measuring response and progression. While there may still be a few final issues to resolve, Dr. Rustin believes it is crucial to have the most accurate early date of progression established uniformly in clinical trials. He and others are working to validate the GCIG standard. "A comprehensive analysis of emerging data on CA125 may lead regulatory agencies to adopt it into their thinking about drug development and drug approval," says Dr. Kelloff.

A Crucial Tool

The CA125 assay has twice been refined, and now serves as a versatile tool for ovarian cancer clinicians. It fills several important roles, not to mention its revived prospects as part of a multistep screening approach.

Transvaginal ultrasound (TVU) and color power Doppler imaging can sometimes distinguish a benign from a malignant pelvic mass. A new model incorporates CA125 levels into an imaging approach that has a specificity of greater than 99.6 percent and a sensitivity of more than 70 percent in detecting malignancy, providing the required 10 percent predictive value.

Tumor response and progression are also markers of a patient's prognosis that CA125 testing can measure. The half-life of CA125 is about 14 days following initial surgical resection and the beginning of chemotherapy. Persistent disease that may not be detected by imaging, however, extends this time to more than 20 days, correlating with a poor prognosis.

Dr. Kelloff suggests that CA125 could play an important role in clinical trial design. For example, when used in all arms of a randomized trial, CA125 could signal an earlier time to stop ineffective drug combinations.

"The potential value is great," adds Dr. Rustin. "Patients identified by CA125 can be included in phase II trials, even though they would not have qualified using RECIST. The go/no-go decisions to keep drugs moving through the pipeline can be reached more quickly and economically in many cases."

And back to screening, where it all began for CA125. "Currently, we combine CA125 and TVU and focus on women with a strong family history of ovarian cancer or early-onset breast cancer," says Dr. Karlan. NCI's large PLCO clinical trial is investigating these two screening tests in a general population of older women.

Molecular and genetic research is starting to uncover more specific signatures of the early stages, says Dr. Karlan, who credits NCI's Early Detection Research Network for useful guidelines to develop new biomarkers for screening.

"I expect an effective screen for ovarian cancer to emerge by 2015," predicts Dr. Bast. "And when it does, CA125 may still be one of its integral features."

By Addison Greenwood