Letters To the Journal Science Raise Concerns Over 2003 Breast Cancer Study
A "landmark" 2003 study that linked mutated forms of two genes with an increased risk for breast cancer might have contained "worrisome flaws" that led to "a significant overestimate" of the risk, according to letters from two separate teams of scientists to the journal Science, which published the study, the Wall Street Journal reports (Begley, Wall Street Journal, 1/7). The study found that women who have inherited mutated versions of either of two genes linked to breast and ovarian cancers have a higher risk for one of the diseases, regardless of whether they have a family history of the diseases.
In the study, researchers examined 1,008 Ashkenazi Jewish women who received treatment at 12 New York medical centers for breast or ovarian cancer between 1996 and 2001. Researchers extracted genetic material from blood samples of the women and found that 104 had a mutation in either the BRCA1 or BRCA2 gene. The mutations are linked to about 5% to 10% of the more than 200,000 new cases of breast cancer diagnosed annually. About one in 400 individuals have one of the gene mutations, but about one in 40 Ashkenazi Jews have one of the mutations. The study found that the risk for breast cancer for women with one of the mutations was 20% by age 40, 55% by age 60 and 82% by age 80; the average U.S. woman has about a 10% risk for breast cancer in her lifetime. The results of the study led to a significant increase in requests among women for genetic testing (California Healthline, 11/11/03).
The letters -- sent by scientists at the National Cancer Institute, the University of Southern California-Los Angeles and the University of Cambridge-England -- said that 22 similar studies have found lower risk estimates for women who have the mutations: 65% for BRCA1 and 45% for BRCA2, on average. Although neither of the letters said that researchers in the 2003 study "deliberately manipulated the data in an unethical way," they criticized the participants selected and the methods used to calculate the breast cancer rates of family members.
According to the letters, the study did not include women who carried the mutations but did not have breast cancer, which could have inflated the calculated risk, and researchers relied on patients to recall which family members had the disease. Sholom Wacholder of NCI said, "If I had a relative with breast cancer, I'd be more likely to mention her (to researchers) because I'd have an idea of what the scientist was looking for." Robert Haile of USC said, "In our opinion, (their) estimates of cancer risk are biased up and are higher than virtually all other published studies."
Wacholder added, "This is not a small disagreement, such as whether the true risk is 90% or 95%. The disagreement is whether the true risk is closer to 50% or to 80%. That's the difference between a coin flip and virtually sure thing. ... These are life-changing decisions, so we want the best information to be out there." Although the letters were sent within weeks of publication of the study, a Science spokesperson said that the magazine delayed their publication to have them reviewed by outside experts -- an "unusual step," according to the Journal. The Science spokesperson added, "We could have done a better job moving these letters more quickly."
According to the Journal, "the true cancer risk posed by the mutation is no mere academic dispute," as some women who learn through genetic testing that they have the mutations decide to have their breasts or ovaries removed to avoid breast cancer. However, according to genetic counselors, "it is too soon to tell how the controversy will affect women who carry BRCA mutations," the Journal reports.
"For many women, when they learn they have the mutation, they figure they're done for," breast cancer expert Susan Love said, adding, "They panic and say do 'everything you can,' which often means prophylactic mastectomy and ovariectomy."
Mary-Claire King, a geneticist at the University of Washington who led the study, wrote in a letter to Science, "We do not think (our) estimates are too high." She added in an interview, "The precise level of risk is much less important than conveying the reality that the risk of breast and ovarian cancer is very, very high. We hope this reality will not be lost in the discussion of optimal statistical approaches" (Wall Street Journal, 1/7).