Elizabeth Suh-Burgmann, MD
The jury is still out on whether use of an algorithm to help classify risk levels in ovarian cancer can improve overall survival (OS), but a company is already marketing the system in Britain and plans to bring it to the United States before the end of the year, and women’s advocacy groups here are working out strategies for handling the interest in early testing that may result.
Results of the UKCTOCS trial of 200,000 postmenopausal women are expected out in mid-December on whether the ROCA (Risk of Ovarian Cancer Algorithm) reduces the number of women who die from the disease. ROCA is based on the idea that changes in CA-125 (glycoprotein) levels when measured on an individual basis can give better information about a woman’s health pattern than if assumptions are made on population characteristics in general.
Nevertheless, gynecologic oncologists say that the CA-125 test does not stand by itself as a fully reliable diagnostic tool and that careful study of UKCTOCS results will be needed to justify any assertions that the ROCA algorithm raises the test to the level of a routine screening tool.
“Previous trials have shown that screening results in excessive unnecessary surgery and associated harms, as well as patient anxiety, so any findings would have to include a careful assessment of this before changes in practice could be adopted,” says Elizabeth Suh-Burgmann, MD, chief of gynecologic oncology at Kaiser Permanente Walnut Creek Medical Center, California.
CA-125 levels are measured in monitoring of ovarian cancer treatment effectiveness, detection of recurrence, and screening for high-risk women. Tests have traditionally relied upon a cutoff point, above which levels of CA-125 may indicate disease progression. The UKCTOCS trial sought to examine whether the CA-125 test could improve survivability by detecting disease earlier in progression. Changes in women’s individual levels of CA-125 were measured and risk levels assigned using ROCA.
Initial UKCTOCS results issued in May of this year indicated that of 296,911 annual screens done over an average of 3 years follow-up, 640 women underwent surgery and 133 were found to have ovarian cancer, leading researchers to conclude that screening methodology had an 85.8% sensitivity for ovarian cancer. For each case of ovarian cancer, 4.8 operations were performed.1
Basing risk on a cutoff level would have identified only half the women with ovarian cancer, the researchers concluded.
Results of the trial due out in December will show whether there was an improvement in OS. However, marketing of ROCA by British company Abcodia is already under way with claims that the algorithm has been “validated” through trials and represents “a powerful aid to clinical decision making for the early detection of ovarian cancer.”
Abcodia is offering the test for postmenopausal women between 50 and 85 and those 35 to 85 who have ≥10% lifetime risk of ovarian cancer.
There is strong interest in a potential early screening tool that would enable physicians to catch the disease in early progression when the prospect of a cure is much greater, but oncologists are concerned about the potential for over-enthusiasm.
Previous trials have shown as many as 85% of women with ovarian cancer have elevated CA-125, and < 1% of woman without ovarian cancer have elevated values. High or low values may have no significance. For example, about 50% of stage I ovarian cancer patients have normal CA-125 values, as do the majority of patients with mucinous ovarian cancers.
“CA-125 has been a great marker with which to follow high grade, serious ovarian cancers, and it is the most common indication of recurrent disease after initial chemotherapy,” says Franco Muggia, professor of medicine (oncology) at New York University Perlmutter Cancer Center, and former chair of the Medical Oncology Committee of the Gynecologic Oncology Group. “However, as a screening tool even in high-risk populations, it has not been proven useful. In particular, a high value in premenopausal women has very low specificity.”
Burgmann agrees that UKCTOCS has shown improved sensitivity and specificity compared with use of a CA-125 benchmark. Strong data on OS could provide a compelling argument for the adoption of screening, Burgmann says, but she worries that objectivity may have been compromised, based on the early marketing of ROCA.
“It was developed in the UK and has been licensed to a private company in the UK who is advertising it already on their website as being clearly beneficial despite the lack of definitive data to this effect. This troubles me as it raises concerns about potential conflicts of interest. As with all trials, any conclusions will need to be reviewed carefully before making changes in patient care,” Burgmann says.