SGO: Ovarian Cancer Marker Shows Prognostic Value

SAN DIEGO, March 5 -- Low levels of the tumor marker CA-125 -- measured before adjuvant chemotherapy for epithelial ovarian cancer -- appear to predict a longer time before the cancer recurs, researchers said here.

CA-125 -- the tumor marker most closely associated with ovarian cancer -- is commonly used to measure response to chemotherapy, according to Kristin Zorn, M.D., of the University of Pittsburgh Medical Center.

But a pooled analysis of seven large phase III trials shows that pre-treatment serum CA-125 may also have a prognostic value, Dr. Zorn told a plenary session of the Society of Gynecologic Oncologists meeting here.

Indeed, at the highest level of more than 1,000 U/mL of serum, the risk of progression was 71% higher for any given interval than for the normal level of 35 U/ml, which was statistically significant, Dr. Zorn said. Several intermediate levels were also associated with significantly elevate risk.

Dr. Zorn and colleagues analyzed data from seven studies conducted by the Gynecologic Oncology Group, in which the adjuvant treatment arm consisted of a standard regimen of intravenous cisplatin and paclitaxel, after surgical debulking. Patients in the studies had stage III or IV disease.

All told, Dr. Zorn said, the analysis included 1,299 patients for whom a pre-treatment CA-125 value was available.

The researchers found:

• The median pretreatment CA-125 level was 246 U/ml.
• 7.6% of patients had a normal level of less than 35 U/ml and 20% had a level above 1,000 U/ml.
• The lowest median CA-125 levels were observed in the mucinous subtype (99 U/ml) and in patients with stage III disease that had only microscopic residual disease (117 U/ml).
• 65% of patients with a normal CA-125 level had serous tumors, or about 7% of all patients with serous tumors, which is the most common form of epithelial ovarian cancer.
• On the other hand, 31% of patients with a mucinous tumor had normal CA-125.

The study also showed that a doubling of the pre-treatment level of CA-125 was associated with a 9% increase in the risk of progression, which was significant at P

The study's strengths includes large numbers, prospective data collection and uniform treatment arms, Dr. Secord said, the design of the seven trials means the data can be generalized outside the study setting.

On the other hand, she said, the study is a meta-analysis, which raises the statistical problem of heterogeneity, since there was no central lab to conduct the CA-125 measurements, and there were a limited number of patients with various rare subtypes.

Dr. Zorn said the findings are "reassuring" in that they suggest that a low pre-treatment CA-125 level is probably a good sign even if the small numbers made it impossible to demonstrate definitively.

Because of the difficulty of performing randomized trials, "this is the best data we're ever going to have on this group of patients," she said.