Gene scans that search a person's DNA for disease risks are now being sold on the Internet, but David Hunter thinks they are not ready for prime time. He should know; he uses this powerful new gene-scanning technology in his work, searching out genes responsible for common forms of breast cancer. U.S. News recently reported on the gene scans being sold by companies like 23andme.com, deCODEme.com, and Navigenics.com. In tomorrow's New England Journal of Medicine, Hunter, an epidemiologist at Brigham and Women's Hospital and a professor at the Harvard School of Public Health, explains why direct-to-consumer gene scanning is not yet useful for making personal health decisions. U.S. News interviewed Hunter, and excerpts follow:
Why are you warning people against buying DNA scans on the Internet?
We thought that the idea that these very, very large scale genomewide analyses were suitable for direct-to-consumer marketing needed to be examined and challenged. How are you using these scans, which look at hundreds of thousands of gene variations in individuals, in your cancer research?
We have found this to be a fabulous new research tool. In the past 12 to 18 months it has led to a blizzard of new discoveries that we all hope will have relevance in the future—to patient counseling, as well as in improving our understanding of the genetic basis of diseases. In our own work, we discovered a genetic variation in the FGFR2 gene that is associated with breast cancer risk. Sixty percent of Caucasian women carry one or two copies of the gene variant. It's an example of the whole new field of relatively common genetic variants that we can now associate with risks for diseases. How is that gene different from the BRCA genes, which also predispose women to breast cancer?
The big difference is that the risks associated with being found to carry a BRCA variant are many, many fold higher—a greater than 20-fold risk of breast cancer at a young age, compared with 1.6 fold with the FGFR2 gene variant, even if you inherit both copies. It's a much, much different scale of risk. I'm Caucasian, so the odds are that I have the FGFR2 variation. What do I do about that?
All this information is so new that I don't think we've established [an answer]. The research is just not yet sufficiently mature for most diseases and conditions to really understand how predictive the currently available genetic markers are and whether there's anything that individuals can do about altering their risk. Things are moving incredibly fast, but realistically, we're some years away from having the studies complete. The information is preliminary, premature, and sketchy. What are your specific concerns about consumers buying gene scans?
Somebody who is shown not to carry any of the relative risk variants that we do know about may be falsely reassured that their risk is low. That may lead people to be less vigorous about following current screening recommendations because they have a false sense of security. At present, our experience with counseling about risk and disease tends to be disease-specific, and we counsel about one disease at a time. Consumers are being presented here with information about dozens of diseases [simultaneously]. I don't think we know how to deal with something that is this complex and this new.
Thirdly, the one thing that is almost guaranteed for anybody who signed up for one of these tests and paid for it—and the companies actually tell people this—is that the information you get will need to be updated constantly over months and years. So why not wait until the information is more mature and isn't changing all the time?
Again, we think that this information may be routine in medical practice in 10, 15, or 20 years' time, but there's still quite a way to go to understand the science. Then there's this other extra area of how to responsibly communicate the information. That needs to be thought through very carefully so we don't do more harm than good.
If genome scans aren't yet useful for individuals, what can I do if I have a family history of a disease like breast cancer or prostate cancer?