SGO: Ovarian Cancer Screening Test Found Highly Accurate

SAN DIEGO, March 5 -- A highly sensitive blood test for ovarian cancer, based on six biomarkers, could be used to detect the disease while it is still in its curable early stages, researchers said here.

The test takes advantage of six proteins associated with ovarian cancer and a bead-based multiplex assay to measure them simultaneously, according to Aliza Leiser, M.D., of Yale.

In a group of more than 500 volunteers, the six-biomarker test had a sensitivity of 97.5% and a specificity of 99.7%, Dr. Leiser told a plenary session at the annual meeting of the Society of Gynecologic Oncologists.

That level of accuracy, she said, yielded a positive predictive value of 99.3% in the test population and a negative predictive value of 99.2%.

In the general population, where the occurrence of ovarian cancer is one in 2,500, the test would have a positive predictive value of 12.6% -- high enough to make it an "acceptable" screening test, Dr. Leiser said.

She noted that ovarian cancer has a five-year survival rate of 90% for stage I and II. On the other hand, stage III and IV disease has a five-year survival rate of just 20%.

Because of that, she said, "development of a sensitive and specific method of early detection is a priority."

The research builds on earlier work at Yale, in which a team led by Gil Mor, M.D., Ph.D., used a set of four biomarkers - leptin, osteopontin, insulin-like growth factor II, and prolactin - to develop a test that used the well-known ELISA technology. Dr Mor was the senior author of the paper presented here.

A Chinese firm is now commercializing that ELISA test, Yale reported last year.

For the current test, the researchers added two other biomarkers -- CA-125 and macrophage migration inhibitory factor -- and used a commercial bead-based system that needs less than 50 milliliters of serum and can measure up to 100 different proteins at once.

For this study, the researchers first tested the system in 75 patients newly diagnosed with ovarian cancer of all stages and 75 healthy matched controls. Dr. Leiser said the tests incorrectly classed one healthy control as having cancer and one patient as being healthy.

In a validation set, she said, the system misclassified two of the patients as being healthy, but made no mistakes on the 290 healthy controls.

Of the 37 blood samples from patients with stage I and II cancer, four were misclassified as healthy, meaning the system got 89% correct. All 123 stage III and IV samples were correctly classified.

Dr. Leiser concluded that the test has excellent sensitivity and specificity and now needs to be tested prospectively to see how well it applies to the general population.

Yale is enrolling patients in a prospective unblinded study among women at high risk for ovarian cancer and healthy controls, in which at three-month intervals volunteers will be given the six-marker test, a test for CA-125 alone, and trans-vaginal ultrasound.

The study "may represent a significant step toward population-based screening for ovarian cancer," commented Laura Havrilesky, M.D., of Duke Medical School in Durham, N.C..

Dr. Havrilesky, in a formal discussion of the study, noted that development of a screening test for ovarian cancer has been difficult because of its relative rarity.