Don’t Count on 23andMe to Detect Most Breast Cancer Risks, Study Warns

23andMe likely saved my life. I got tested because I had severe, unexplained exhaustion. 23andMe didn't find a gene causing my exhaustion, but it did tell me I had a BRCA1 defect. (Yes, it was a defect found mostly in Ashkenazi Jews, and that's my ethnic group.) I had zero family history of breast or ovarian cancer. I've since learned that 1-in-3 women with hereditary breast and ovarian cancers have no family history of breast or ovarian cancer -- they inherited the gene from their father. My PCP insisted I was fine because I had just had a clean mammogram. I fought for six months to get an MRI. The MRI found breast cancer on one side. BRCA1 cancers are exceptionally aggressive, but they got it in time. We now know that at least half of breast cancers diagnosed in women under 60 are hereditary. A high proportion of colon cancers in middle-aged people are hereditary. We have the tools to save millions of middle-aged people who are struck down from nowhere by cancer. What we need is much, much more genetic testing for hereditary cancer. But genetics is hard to understand, and this article makes it worse. 23andMe *does* make clear that only the Jewish variants are tested. The FDA will *not* yet let 23andMe release results on other variants yet. Jews volunteer for genetics testing far more than most other groups, so we have more data on Jews. Please, New York Times. Have Gina Kolata do a major feature on hereditary cancer. She has the science skills to get it right.

My wife who has Parkinson's disease for the lat 18 years was one of the first people to test at 23&Me as it's co owner had the gene for Parkinson's and had a special lower rate fee for PD patients to study their genetic history. The most recent 23&Me test said that she had a lower to no risk for PD ,this from just several months ago and all from a saliver sample that is more then 7 years old...take the 23&Me results all with a grain of salt ,maybe a pound of salt and check with your own physician on any 23&Me medical diagnosis.

About the only thing that is clear is that the limits on the scope and reliability of the information from 23&Me's testing is not clear to the vast majority of the people who get tested. That's a real problem if we're dealing with advertising disclosures for ordinary consumer products. It's a very serious problem with health information, regardless what reasonable alternatives exist.

I just heard during a weekly breast cancer conference that I participate in that genetic testing (the full shebang) for those that walk in off the street and just want to be tested is now $250 for breast cancer at least. This is from a well-reputed cancer society. I don't think that includes the genetic counselling for that price... that would be crazy low. Back when I tested for the BRCA gene, they charged my insurance around $11000. I'm BRCA2+ unfortunately. 23&Me needs to test for all or none. I agree with Dr. King and am afraid the partial results can be dangerously misleading.

Mary-Claire King: “The F.D.A. should not have permitted this out-of-date approach to be used for medical purposes. Misleading, falsely reassuring results from their [23andMe] incomplete testing can cost women’s lives.” Maybe in the 23andMe fine print there was a reference to the test being relevant only to a subset of Ashkenazi Jews? Still, straightforward and correct statements in the spirit of Dr Mary-Claire King's seem depressingly rare. She demonstrates an example others could follow for everyone's benefit.

I signed up for 23 and Me before I knew I was BRCA 1 positive. They don't show me as positive because I'm not Ashkenazi. I'm OK with that. I've learned a good bit about my heritage. They never said their test would reveal all. I've had both breast and ovarian cancer. It's not their fault. I'm doing great!

23 and Me is bogus. I am a psychiatrist and wanted to see how the tests could satisfy my curiosity about medical tests, for fun. 23 and Me test for about 600 000 SNP’s (individual variations) but they only test for about half of the several dozen SNP’s that are of medical interest. They deliberately crippled their test. Now, this article shows that they are happy to say someone is negative for a gene when they don’t bother doing all the tests. They test for so many other SNP’s that there is no excuse for leaving these out. Their documentation is full of warm and fuzzy stuff but not the nitty gritty. Their customer service is reasonably prompt, but they quickly go quiet when you try to dig down.

I wish journalists would stop reporting on studies that have not been peer-reviewed. The results described here may be solid, but until they have undergone peer-review and been published, they should not be reported. Note: I am a biomedical researcher with >160 published papers. I would not issue a press release or allow one to be issued about my results until they have undergone peer review.

This article's unstated crux boils down to the fact that genomics is still very much an evolving field in its early days. As someone who works in a cancer genomics lab with a heavy emphasis on next generation high-through put sequencing, I am well aware that the quality of the data and the interpretation is hugely dependent on a number of variables ranging from the wet lab reagents and protocol optimization to technical sequencing choices to data processing pipeline definition. Many of these choices can be subjective. Almost none of the details are visible to the public. Instead, the public has been presented a highly sanitized, highly simplified picture of what they can expect. It is truly a wild west out there for genetic testing.

Sure, but you can upload your 23andMe or Ancestry or FTDNA results to Promethease for twelve bucks and then get a report that tells you if you have many of the other known variants. I did just that, and now am aware of all of the breast cancer genes I carry that 23andMe didn't warn me about. Promethease also identifies a slew of other genetic disorders and risk factors. Like thousands.

Not specifically related, but perhaps of interest to those commenting or choosing to read them here. The mother of our children succumbed January 1, 2019 to a second incidence of breast cancer (the first was in 1986 followed by removal of that breast). She continued a pattern of annual mammograms on the remaining breast at a well-recognized hospital. The last routine mammography in March, 2018 showed no tumor ~ and with no other symptoms of cancer then. Only in her subsequent final battle was it disclosed to her (or any of the family) that "routine" mammography only scans about 85% of the breast tissue. The portion under the arm pit that is closest to the lymph nodes (infamous for trafficking malignancies) was excluded. In her case, a primary tumor there had spread to her liver, spine, and lungs before any diagnosis of such.

This is reminding me of the Theranos business.

23andME should be used only for entertainment and for curiosity. NO one should use the results for predicting health risk.

So on top of the problem with these companies selling your genetic information to anyone who offers enough money (insurance companies, drug companies etc.), now we learn that the data itself may be significantly inaccurate? That is scary indeed. Careful how you deal with these sleazeball companies, they are basically the "facebook" of healthcare.

If a test that you pay for gets the results wrong 90% of the time, it is criminal to allow that company to continue offering that test, giving false hope (and possibly harming) those who have inherited the dreaded and deadly BRCA gene associated with breast cancer. One of the comments states that people with an inherited gene that causes breast cancer would already know the gene runs in the family and don't need a test to figure it out. I beg to differ. The families of most people who died from breast cancer many years ago, and even today, often don't know their grandmother, great aunt etc. died from breast cancer, and assumed it was a cancer of the part of the body it had spread to (kidneys, liver etc.) that killed them. The same thing happens with prostate cancer. The family thinks their father or grandfather died from stomach or colon cancer, when in fact it was metastatic prostate cancer that had spread to their colon that actually killed them.

I joined 23andMe four years ago. Its ancestry DNA service is brilliant -- I've learned more about my family and ancestors than I ever dreamed possible -- far superior to Ancestry.com et al. However, even though 23andMe's health reports have always been available here in Europe, I take them with a grain of salt. I can't assume they're accurate, in the same way medicine itself is an art and a lot of guesswork.

I subscribe to both GenomeWeb and 23-and-me. In my life, 23andme has had significant importance. My doctors do NOT prescribe genetic tests to the relatively healthy. Do yours? 23andme was cheap, and I signed up in case I could help with the fight against the medical condition I do have-- adult onset asthma. When 23andme was finally able once more to offer genetic information, I learned I carry a gene for 1-alpha antitrypsin deficiency. Not only does that mean a propensity to asthma but it also means I should watch out for my liver's health-- thank you 23andme. Meanwhile, my beloved GenomeWeb is on a crusade to denigrate 23andme and FREQUENTLY CITES my beloved NYTimes articles! I expect to see a precis of this article soon.

My family all did 23andMe for fun years ago and when the BRCA info was emailed out- my mother, sister and I tested positive. It was almost impossible for my sister and me to get the more comprehensive test from our doctors office since our insurance didn't want to cover it. It is a VERY expensive test. We would have never pursued if not for the 3 positives from the 23andMe. I had also paid for an expensive genetic test from a fertility specialist and the results were the same as my 23andMe, both positive Tay-Saches carrier. I know these "Lab" companies all make lots of money off their patents for their gene detection and I think this aspect of making money on the "higher-end" should be considered. At least 23andMe was affordable and I didn't need to jump through hoops with my insurance or pay a lot out of pocket for information I would have never sought out on my own. The picture just seems bigger than this article seems to paint it.

I had the opposite -- an Ashkenazi male, I downloaded my data and ran it through a public database for mutations and it came back positive for BRCA2 mutation. I had further testing which showed I was negative, and it was only 2 years later that there was enough data to show that 23 and Me provides false positives for some BRCA2 mutations. So be wary whether 23 and Me reports you are negative and you have a family history (talk to a doctor!) or if you download the data and run it through a public database and it reports you are positive (talk to a geneticist!).

There are many different mutations in BRCA2 alone - and many at the same site. This isn’t what 23and Me was designed for. However, many take comfort from the results, despite their limitations. When I realized that the familial breast cancer in my family was probably a specific BRCA2 mutation, I contacted people at 23andMe. But with new cancer cases, various members have used Myriad or Invitae for more careful screening. We are actually lucky 23andMe doesn’t do it all - when you find something like that, you want to be in touch with your doctor and a good genetic counselor.

23andMe is a consumer novelty. To use it as a diagnostic tool, without understanding the genetics of what you seek to diagnose, and the capacity of the test to diagnose it are mistakes. The oncologist mentioned likely made both.

I have been a 23andme customer since 2011. My primary interest was genealogic at that time and the family tree functionality they offered back then helped me get started on researching and building a large family tree. However, I found I that the health indicators they were reporting on interesting and an extra benefit. They said that I had a higher than average possibility of having AMD and I do have this. Also a higher possibility of diabetes and I do have pre-diabetes. That being said, if you see something in your test that is a significant health worry, then as in any other medical interaction, it behooves you to get a 2nd or even a 3rd opinion. In this case, that would appear to mean additional DNA tests or at least genetic counseling. DNA companies should be required to inform the person being tested who is doing the actual test, by what company and at what technology level. This will help consumers getting a 2nd test to choose a different platform. With my test, I know that it was done on a model 3 chip. 23andme is now using a model 5 chip. I've not been able to find anything that states what the difference is between chips but it might be important. Perhaps, since 23andme is selling health info, the FDA should require them to retest people on older chips when they upgrade to a newer, better chip?

Let's stop pretending the FDA is here to protect us. While the agency pounces on nutritional supplements and raw milk, it won't ban cigarettes. One, big reason: tax revenues. State and local governments reap about $17 billion a year in tobacco tax revenues while more than $300 billion per year is being spent on medical care and lost productivity due to tobacco use.

So why offer online genetic health testing? Come on! To make money!

This doesn't surprise me in the least. As a genetic counselor I've spoken with many patients who have had 23andme testing and claim they've had "carrier screening." [I must resist rolling my eyes.] Way back before the FDA halted and update 23andme's testing, I ordered my own test to see what the reporting looked like, how the tests were conducted, etc. According to 23andme, I am not a carrier for non-syndromic hearing loss due to a mutation in the GJB2 gene. However, reputable (read: medically-accurate and reliable) carrier screening has shown I do have a mutation in this gene and therefore am a carrier for this condition. Yes, 23andme says they only look for a few mutations, but how many non-medical (or even medical but non-genetic) people can appreciate what that means? I tested negative for the 3 founder mutations in the BRCA genes but know this alone doesn't determine my chance to develop breast (or associated) cancers. Not everyone has the same background I do and, frankly, I don't think 23andme cares whether people taking their test understand the nuances. It's a shame.