CHICAGO—In the continuing quest for an effective screening method to detect ovarian cancer at an early stage, a new approach using an algorithm has shown promise in early testing as reported here at the ASCO Annual Meeting and featured beforehand in a teleconference by the society.
The Risk of Ovarian Cancer Algorithm (ROCA) is based on patient age and fluctuations in CA-125 blood levels, and in a study of 3,238 postmenopausal women, the approach proved feasible, with a specificity of 99.9%, said lead author Karen H. Lu, MD, Professor of Gynecologic Oncology at the University of Texas MD Anderson Cancer Center.
The approach failed to detect two borderline cases, but no invasive ovarian cancers were missed.
“Could these borderline cases become invasive later? It's possible, but biologically, they tend to act differently,” she said.
Although years ago, researchers discovered that blood levels of the CA-125 (cancer antigen-125) protein are elevated in women with ovarian cancer and it is used to check to see if treatment is working or if cancer has returned, by itself, however, CA-125 did not prove useful as an early marker of cancer, since it misses some tumors and falsely identifies many healthy women as having cancer, Dr. Lu explained.
So recently, researchers started looking at the change in CA-125 levels over time rather than at CA-125 as a single value. “We and others theorized that if someone has low levels and that is doubled, that it could be a sign that something is happening even if the levels are still low. Alternately, if a woman has a high value and it stays there, [we theorized that] it's not likely to be ovarian cancer,” she said.
The purpose of the current study was to assess the specificity and positive predictive value of the ROCA screening strategy for healthy postmenopausal women, using changes in CA-125 over time and age to estimate risk for ovarian cancer.
As part of the strategy, woman with high ROCA scores are referred to transvaginal sonography (TVS), clinical evaluation with a gynecologic oncologist, and, if needed, exploratory surgery.
The prospective study of 3,238 postmenopausal women aged 50 to 74 with no significant family history of breast or ovarian cancer, enrolled over a course of nine years.
Based on ROCA results, women are placed into three risk groups:
- Women at low risk are told to return for annual CA-125 blood screening.
- Women at intermediate risk have their CA-125 test repeated in three months.
- Women at high risk are referred for TVS and clinical evaluation.
Each year, less than 1% of the women were categorized as high risk of ovarian cancer based on their ROCA results and referred for ultrasound.
“Over the nine-year period, only 2.6% of women were ever triaged to a transvaginal ultrasound, and eight women underwent surgery,” Dr. Lu said.
A total of 6.8% of women annually were referred for three-month CA-125 screenings, and cumulatively, the rate was 15.4%. “So the vast majority of women were low risk and just had to come back for CA-125 screening annually,” she said.
Among the eight women who underwent surgery, five ovarian cancers were detected.
Three were early-stage invasive ovarian cancers—“the kind that I as a doctor really want to pick up early,” Dr. Lu said. The other two borderline ovarian tumors were also diagnosed at an early stage.
“Currently we will perform about 10 operations to detect one ovarian cancer. In this study we needed to perform three surgeries to detect one invasive cancer. This was very reassuring to us.”
The three invasive ovarian cancers were all high-grade epithelial tumors; two were Stage 1C and one was IIB.
“All three women with invasive ovarian cancers had at least three years of low-risk, annual CA-125 values prior to a rising CA-125,” Dr. Lu noted.
The other three women who underwent surgery had benign lesions, although one woman was later determined to have endometrial cancer.
Overall, the ROCA had a positive predictive value of 37.5%. The specificity of ROCA followed by TVS for referral to surgery was 99.9%.
Await Results of UK Trial
“The beauty of the algorithm is that each individual woman ultimately established her own baseline, based on her prior CA-125 values,” Dr. Lu said.
She emphasized, though, that the results are still early and can no means be considered “practice changing at this time,” and reminded people that at last year's AACR meeting, for example, about five research groups presented results of their best candidates for early detection of ovarian cancer and none has since panned out.
“We await the results of the definitive trial that examines mortality as an endpoint, which is currently ongoing in the United Kingdom,” she said.
Results of that study, which involves more than 200,000 women, should be available in four or five years, Dr. Lu said.
Commenting on the findings, ASCO 2009–2010 President Douglas W. Blayney, MD, said, “The ROCA represents yet another example of personalized medicine. It also represents a more refined application of known technologies that are widely available.”
The results remain to be validated, he also stressed—“the test is not yet ready for prime time.”