Women whose biopsies show atypical hyperplasia with expression of the cyclooxygenase-2 (COX-2) enzyme are at significantly higher risk for eventual progression to breast cancer, according to an article published online March 11 by the Journal of the National Cancer Institute.

Researchers discovered that women whose atypical tissue expressed the COX-2 enzyme were more likely to develop breast cancer subsequently and that risk increased with higher levels of expression. At 20 years after biopsy, 31 percent of women with high levels of COX-2 in their sample had developed breast cancer compared with 14 percent of those with without expression. Among patients whose tissue samples had moderate levels of COX-2, 24 percent had developed breast cancer.

Investigators examined tissue from 235 women who had a breast biopsy following a mammogram at the Mayo Clinic from 1967 through 1991 and were diagnosed with atypical hyperplasia, as well as reviewing their clinical records.

During an average follow-up of 15 years, 41 (17 percent) women developed breast cancer. The average interval for breast cancer to be diagnosed was 11.4 years.

The researchers found that moderate expression was identified in 71 (30 percent) of samples and that 34 (14 percent) of biopsies exhibited strong expression.

The researchers calculated that, among the women in the study group, the absolute risk of developing breast cancer 15 years after diagnosis of atypical hyperplasia was 13 percent in biopsies with little or no COX-2 expression, 19 percent with moderate COX-2 expression, and 25 percent with strong COX-2 expression. In women who had been followed for 20 years or more, the association was more definitive: stratification of risk was 14 percent, 24 percent, and 31 percent, respectively.

The finding that risk of later breast cancer was associated with increasing levels of COX-2 expression was of borderline statistical significance, likely due to small sample size. Harmann added "But higher expression levels were tightly linked to an increased number of abnormal foci in the biopsy, which we had earlier found to be a strong risk factor for breast cancer."

The researchers also assessed the relative risk of developing breast cancer, based on COX-2 expression in atypical hyperplasia, compared with a "control" population of unaffected women, and found that little COX-2 expression was associated with 2.6 times the risk, moderate COX-2 expression increased risk by 3.5 times, and a strong expression of COX-2 elevated risk by more than 5.6 times, compared to women in the general population.

"Based on these findings, COX-2 expression in atypia may be a biomarker of risk of progression to breast cancer," said senior investigator, Mayo Clinic oncologist Lynn Hartmann, MD. "COX-2 is a relevant candidate because it drives a number of malignant features and has been shown to be important in breast cancer.

"Each year in the United States, about 1 million women have a benign breast biopsy; and some of them receive the worrisome news that they have atypical hyperplasia," Hartmann said. If other studies validate these findings, she noted that one strategy to help manage patients may be to test use of a COX-2 inhibitor such as celecoxib or rofecoxib in a clinical trial on prevention of cancer.

In the biopsy profile thus far, the Mayo risk model includes age, number of regions of atypical cells in the biopsy, COX-2 expression, and status of the normal lobules in the breast.

"Our goal is to individualize risk for breast cancer so that we can provide effective care and risk reduction for each of our patients," said Hartmann.

COX-2 is not usually found in normal breast tissue but has been associated with ductal carcinoma in situ and invasive breast cancers.