What's the Point of a Biomarker? A Look at BRCA.

I've ranted a bit before about how confusing is the steady onslaught of genetic associations to disease. Every day there seems to be a new genetic tag for another disease. Some diseases, though, have very clear and well-established genetic associations. Among the most researched is the connection between breast cancer and mutated BRCA genes (typically called BRCA1 and BRCA2). The generally accepted statistic is that between 5 and 10 percent of women with breast cancer will carry an altered BRCA gene. According to the National Cancer Institute, women with an altered BRCA gene have three to seven times higher chance of developing breast cancer than those without (altered BRCA genes are also associated with other cancers, including male prostate cancer). Women who have BRCA1 or BRCA2 gene changes have between a 36% and 85% chance of developing breast cancer and between a 16% and 60% chance of developing ovarian cancer during their lifetimes.

These strong associations have made BRCA among the most researched genetic associations out there - a PubMed search for BRCA lists 776 papers. In turn, these associations means that BRCA tests are some of the most popular genetic tests out there - they're even being pitched directly to consumers. Indeed, they are for many the first taste of genetic profiling. The BRCA test, in many ways, is a peek at our medical futures, how we will, soon enough, be tested for a wide swath of genetic mutations (or presence/absence of genes, or specific combinations of genes, etc), that could indicate certain probabilities of developing certain diseases.

But for all that research and for however pervasive the tests may be, there's still a huge grey area around BRCA. The New England Journal of Medicine has a new paper that shows that the presence of an altered BRCA test does not, in fact, increase mortality among women with breast cancer. In other words, if you have breast cancer, just because you have a buggy BRCA gene doesn't mean that cancer is more lethal. It is, admittedly, a rather fine point, and may be small solace to women with breast cancer - it doesn't as far as I can tell devalue the core assocation between the gene and the disease.

But almost simultaneously, there's a recommendation published in the Journal of the American Medical Association to increase the use of BRCA genetic testing among women with a higher likelihood of developing breast cancer. I'm all for increased use of screening tools - provided that the accuracy (sensitivity/specificity) of those tools doesn't lead to a rash of false positives (or worse, false negatives). But an important part of the moral bargain of screening is that, if it detects disease, that it improve outcomes - that is, we should screen earlier when earlier detection leads to less disease. If knowledge doesn't afford better treatment, then we're creating a very ambiguous arena where people may know they face disease, but can't do anything about it. And ultimately, that's just making people worse off, not better.

Update: EyeonDNA has a nice riff on the ethics & utility of screening for BRCA here.