Intermediate magnification micrograph of a low malignant potential (LMP) mucinous ovarian tumour. H&E stain.Changes in CA125 blood levels—not one-shot tests—may be the key to one of oncology’s most elusive goals: identifying ovarian cancer at a treatable stage, says a recent Cancer study.

By measuring such changes, the study, conducted by MD Anderson Cancer Center, identified women who needed intensive monitoring, and those who don’t.

“The MD Anderson study confirms, albeit on a smaller scale, our previously published data,” says Ian Foster, vice president of the University of Manchester. He was not involved with the MD Anderson study, but he is director and principal investigator of the UK Collaborative Trial of Ovarian Cancer Screening        (UKCTOCSstudy, which is utilizing a similar approach on a larger scale.

“I am optimistic.”

Notoriously, more than 70% of ovarian cancers are discovered at a stage where cure is highly unlikely. Ovarian cancer's early symptoms are hidden, often discovered in early stages only after use of surgical methods. So a belief has grown it is best to identify a smaller group of at-risk women to follow more aggressively.

The biomarker used by MD Anderson—CA125 (carbohydrate antigen 125)—has long been thought a key sign of ovarian cancer. But one-shot takes have not unmasked high-risk patients.

Karen H. Lu, MDSo the MD Anderson group, led by Karen Lu, professor of gynecologic oncology at the University of Texas, looked for alterations in CA125 levels among 4,051 post-menopausal women. (Ovarian cancer risk rises sharply after menopause.) Each woman was tested annually between 2001 and 2013.

Based in part on Foster’s research, the authors predicted a certain leap in a woman's CA125 level would make her higher-risk, warranting a CA125 level check every three months.

An even higher leap meant a woman was highest risk. She would receive a transvaginal ultrasound to seek out designated cysts or masses of possibly cancerous cells that might call for exploratory surgery.

The results: 83.4% of the women received CA125 testing no more than once a year. Another 13.7% qualified for more testing, but didn’t generate further CA125 results calling for an ultrasound. Some 2.9%—117 of the 4,051 women—were deemed to need a transvaginal ultrasound.

The ultrasounds of ten of those women led to surgery. Four of those were found to have early-stage high grade ovarian cancers. One had endometrial cancer. Three had benign cysts. Two had ovarian tumors that were not likely to go malignant.

All the women with invasive ovarian cancers were treated and cancer free from four to 42 months after treatment.

“This MD Anderson study utilized the ROC (Risk of Ovarian Cancer) algorithm, which I developed with Steven Skates (a biostatistics expert with the Dana Farber/Harvard Cancer Center) in the 1990's,” says Foster, who was contacted by email. “The algorithm has since been tested in very large trials in the UK (with over 200,000 participants). It has been shown (in a 2009 Lancet Oncology paper) to have high specificity, positive predictive value and sensitivity in ovarian cancer screening as well as achieving a shift in cancer stage. We await the final results of the UKCTOCS study which will provide data on the mortality reduction achieved by screening.”

Foster says the MD Anderson study confirms “our previously published data on specificity and positive predictive value in a USA population, and is valuable to that extent. It is too small to add to our data on sensitivity. Given that it was not a randomized study, it cannot contribute on the key issue of mortality benefit. So whether ovarian cancer screening can save lives awaits the outcome of UKCTOCS in 2015. I am optimistic. Because the sensitivity is high and we have achieved a stage shift. But I am naturally an optimist; you need to be to work in this field over many years. Others are more skeptical or cautious. Time will tell.”

The early symptoms of ovarian cancer include bloating, pelvic pain, difficulty eating and frequent urination. They are often overlooked or confused with other illnesses.

Ovarian cancer accounts for three percent of cancers among women.