Asian women at risk for breast and ovarian cancers may not be getting the genetic screening that could save them. The reason: computer models commonly used to assess whether women should be tested for harmful genetic mutations may underestimate the risk in families of Asian descent, according to a new study in the Journal of Clinical Oncology.
To determine which patients would benefit from genetic testing for breast and ovarian malignancies, physicians routinely use the BRCAPRO and Myriad II computer models. By providing a woman's family history of these cancers, including the ages they were diagnosed, the programs calculate a probability that the patient carries a harmful mutation in BRCA1 or BRCA2 (genes involved in controlling malignant cell growth). A woman's risk of breast and ovarian cancer increases five-fold or more if she carries mutations in these genes, which underlie up to 10 percent of breast and 15 percent of ovarian cancers.
Current models were designed using data from Caucasian women, primarily from northern Europe and the U.S., raising the question of their utility in assessing mutation risks for other populations. Recent research showed that the software works relatively well in women of Hispanic and African-American descent, but this new study found that in Asian populations the current scheme missed half of the women carrying the mutation.
For physicians this translates into "missed opportunities to protect patients from bad outcomes," says study leader Allison Kurian, an oncologist at Stanford University School of Medicine. Mayo Clinic geneticist Noralane Lindor says there is no magic number that dictates testing but that physicians are more likely to closely monitor and refer for screening patients flagged by the computer programs as having a higher probability of mutation.
The study, which compared each model's success in Caucasian women with those of Asian descent (Chinese, Japanese, Filipino, Korean and Vietnamese), also raised important questions about the effect of race on cancer development: When Caucasian and Asian patients with similar family histories of breast and ovarian cancer were compared, the Asian women had higher rates of genetic mutation, although the rates of these cancers for Asians have traditionally been lower. This could indicate, Kurian says, that BRCA1 and BRCA2 mutations are not as informative for predicting future cancer in this population as it is in whites.
Kurian notes these differences in cancer outcome may - indicate that Asian women have protective genes that Caucasians lack. Some studies have also suggested that diet (particularly fish and tofu), along with environmental factors may play a role in lower cancer rates—although the incidence of breast and ovarian cancer in Asian populations is currently rising.
Further research will be necessary to determine how often BRCA1 and BRCA2 mutations actually occur in Asian groups and what these mutations really mean for Asian women in terms of cancer risk.
Steven Narod, director of the Familial Breast Cancer Research Unit at Women's College Research Institution in Toronto, believes that cost and not flawed computer programs are to blame for undertesting. The only reason these models are even used, he notes, is because Myriad Genetics, Inc., (in Salt Lake City, Utah) holds a patent for the BRCA1/BRCA2 genetic test and charges patients a hefty $3,120 a pop.
Banu Arun, an oncologist at the University of Texas M.D. Anderson Cancer Center in Houston, says that beyond cost physicians must also consider the psychological impact of testing and women's fears that a BRCA1/2 positive test either could spur insurance companies to drop them or raise their rates through the roof.
For now, though, health care providers will continue to use these computer models, however imperfect they may be, to assess women's genetic risks. Although future research will no doubt focus on fine-tuning current models for nonwhite groups, for now Kurian recommends that Asian women and their doctors err on the side of caution when using them to consider genetic testing.