Roanoke Times
Copyright (c) 1995, Landmark Communications, Inc.
DATE: SUNDAY, April 9, 1995 TAG: 9504100039
SECTION: NATIONAL/INTERNATIONAL PAGE: A-16 EDITION: METRO
SOURCE: The Boston Globe
DATELINE: BOSTON LENGTH: Medium
The researchers at Massachusetts General Hospital and Harvard Medical School said the new development could yield, within a few years, a blood test to identify women who have a high short-term risk of ovarian cancer.
About 80 percent of ovarian cancers caught early can be cured. Because they cause few symptoms, however, 75 percent of ovarian tumors are found late, when only 20 percent are curable.
Ovarian cancer is infrequent, and current detection methods, which include blood tests of a tumor antigen called CA-125, pelvic exams and ultrasound imaging, are so nonspecific that a government panel recently concluded that none was appropriate for widespread screening.
The researchers, however, reported at a scientific meeting in Washington that checking the amount of CA-125 in a woman's blood at intervals over months or years gave a much more reliable indication of her ovarian cancer risk than a single measurement did.
Up to now, the CA-125 test has been deemed inadequate because an estimated 50 percent of Stage I, or early, ovarian tumors don't produce elevated levels of the antigen in the blood. Thus, half of all potential early diagnoses would be missed. In addition, some benign conditions can raise CA-125.
But the researchers, Steven J. Skates, a biostatistician at MGH, and Robert C. Knapp, professor emeritus at Harvard and former chief of gynecology at Brigham and Women's Hospital in Boston, said Friday they have found a new way to analyze CA-125 measurements that makes the test more specific.
The new method involves measuring CA-125 several times and watching for an increase in the antigen, said Skates in a telephone interview. This way, he estimated, the test would yield only 6 or 7 false positives for every actual case of cancer. The women who tested positive could then receive further testing, possibly including surgery, to determine whether they actually have cancer.
This level of accuracy would give the new method a ``positive predictive value'' of 16 percent, Skates added, compared with about 3 percent with current techniques. Researchers say a test for early ovarian cancer would have to have a positive predictive value of 10 percent to be considered cost-effective and to avoid subjecting too many women to anxiety and further invasive testing.
Commenting on the report, Michael G. Teneriello of Allegheny Hospital in Pittsburgh said the improved accuracy the researchers say they can obtain with the test ``is very notable - that sounds like the Holy Grail'' in the search for a more specific ovary cancer detection test.
Teneriello, author of a review of early detection for ovarian cancer published this year, called the research ``very promising'' and said the scientists who carried it out ``are the premier people in the world'' in their field.
by CNB