Respectful Insolence

I’m beginning to understand why evolutionary biologists are so sensitive about how creationists abuse and twist any research that they think can be used to cast doubt upon evolution. Whenever there is research that changes the way we look at evolution or suggest aspects of it that we didn’t appreciate before, where scientists get excited because they see an opportunity for better understanding of evolution, creationists see a chance to use it to launch specious and fallacious attacks against evolution. Sadly, there is no shortage of creationists willing to make fools of themselves to try to prove their pseudoscience. Indeed, creationists have Casey Luskin, Michael Behe, Dr. Michael Egnor, and any number of gnats nipping away at evolution.

I have Mike Adams.

There’s a difference, though. As important as evolution is to biology, as big a threat that creationists are to science literacy and education, they don’t have the immediate and direct affects on life and limb that boosters of quackery like Mike Adams potentially have. This time around, where I see an interesting, if preliminary, study that may end up changing how we calculate the risk-benefit ratio of mammographic screening for breast cancer (or may not), Mike Adams sees yet more “evidence” that mammography is useless when it is anything but.

I’m referring to a study that was presented a couple of days ago at the Radiological Society of North America’s annual meeting. It was done at University Medical Center Groningen in the Netherlands and has been reported widely in the news. Here’s a Reuters report:

CHICAGO (Reuters) – Low-dose radiation from mammograms and chest X-rays may increase the risk of breast cancer in young women who are already at high risk because of family history or genetic susceptibility, Dutch researchers said on Tuesday.

They said high-risk women, especially those under 30, may want to consider switching to an alternative screening method such as magnetic resonance imaging, or MRI, which does not involve exposure to radiation.

“Our findings suggest that low-dose radiation increases breast cancer risk among these young, high-risk women, and a careful approach is warranted,” said Marijke Jansen-van der Weide of the University Medical Center Groningen in the Netherlands.

The timing of this study is almost as bad as the timing of the announcement of the new USPSTF recommendations regarding screening mammography right in the middle of the health care reform debate, coming as it does on the heels of, well, the announcement of the new USPSTF recommendations regarding screening mammography. The reason, of course, is that the new USPSTF recommendations suggested that for women under 50 routine screening mammography provides marginal benefit compared to the risk of false positives, overdiagnosis, and overtreatment. Despite those recommendations, it was still considered advisable for young women at high risk for breast cancer due to family history or known genetic mutations predisposing for cancer to begin mammography no later than age 40 or ten years before the youngest age at which a first degree relative developed cancer, whichever is younger.

I decided to go to the source and look up the abstract, Mammography Screening and Radiation-induced Breast Cancer among Women with a Familial or Genetic Predisposition: A Metaanalysis. The first thing I noted is that this is not a study per se, but rather a meta-analysis, and if there’s one thing that I’ve always said about meta-analyses, it’s “garbage in, garbage out.” No, I’m not saying that this study is garbage; what I am saying is that any meta-analysis is completely dependent on the studies analyzed, how tight the inclusion criteria are for the studies examined, and how high quality the studies are. The problem with an abstract presented at a meeting in a 10 minute talk that I didn’t attend is that it’s really thin gruel for me to analyze. It’s virtually impossible for me to look at the abstract and figure out if the authors did it right. What the authors did in this case was to do a systematic search in Pubmed and EMBASE/Medline using a search strategy of “Breast neoplasms AND Mass screening OR Mammography OR neoplasms, radiation-induced,” combined with text words focusing on high-risk women.

These were their results:

In total 34 articles were found of which 5 were selected by two reviewers. Four studies examined the effect of exposure to low-dose radiation among mutation carriers. One study researched the effect of radiation among women with and without a family history of breast cancer. Pooled OR revealed an increased risk of breast cancer among high-risk women due to low-dose radiation exposure (OR=1.63, 95% CI: 0.92- 2.90). Exposure before the age of 20 (OR=2.0, 95% CI: 1.4-3.0) or five or more exposures were associated with a higher radiation-induced breast cancer risk (OR=1.95, 95% CI: 1.5-2.6).

Even as meta-analyses go this appears to be rather thin gruel. It’s suggestive that there may be increased risk of cancer in young women at high risk for breast cancer who either (1) undergo their first mammogram before the age of 20 or have more than five mammograms. The doses reported were actually quite low, from 0.3 to 24 mSv. (For comparison, typical radiation exposure from normal background radiation is around 2.5 mSv/year.) Another thing to remember is that the population being studied is a distinct minority. Four studies looked at women with documented mutations in BRCA1 and BRCA2 that predispose to familial breast cancer, and only one looked at women with a strong family history.

A Medscape report explains why such women might be at a higher risk of cancer due to low dose radiation:

“We know that the BRCA gene is a radiation repair gene, and when you are missing that, you have a decreased capacity for radiation repair, which will make you more susceptible to the low dose of radiation with mammography. We think this might be the biological basis for what this study found.”

In the discussion following the presentation, it was also noted that the young breast is more susceptible to radiation, which has long been known.

So what does this study mean to physicians and surgeons who take care of patients with breast cancer, as well as primary care physicians? It’s hard to say. It was only a few studies in the analysis, and I have no way of knowing if the authors adequately corrected for overdiagnosis, which could certainly occur in young women who undergo more mammography than women who don’t. As I’ve explained before, diagnosis by screening mammography is a very complex issue, and these sorts of factors are not trivial to control for in retrospective studies. In addition, as was pointed out in various news reports, no single study demonstrated a clearly increased risk of breast cancer due to low dose radiation exposure from mammograms. So, given that the relative risk isn’t enormous, certainly a large prospective study is probably indicated to determine whether these results hold up or are spurious. Certainly, given what we know about radiation and BRCA mutations, the existence of increased risk in young women due to low level radiation is plausible. What would be even more important, however, would be to quantify the magnitude of the increased risk due to starting mammography early, so that the risk-benefit ratio could be recalibrated based on that information and new recommendations based on the new data formulated.

What this study also points to is an acute need to develop better screening tests, preferably one that doesn’t involve ionizing radiation, particularly for young women. As I pointed out before, MRI is too sensitive, not to mention way too expensive, to be used in women at normal risk. It’s that very sensitivity that would exacerbate the problems of overdiagnosis and overtreatment, which in mammography led the USPSTF to rethink recommendations for screening mammography in women under 50. However, we’re not talking about women at normal risk; this study addresses women who are already at very high risk for developing breast cancer and who are very young. As the lead investigator put it:

She emphasized that screening mammography is not a problem for high-risk women 30 years and older. “Screening these high-risk women is still very important, but we must think about alternative nonionizing techniques at younger ages,” she said.

“From our study results, we are talking about women below the age of 20, which is very young. We are also talking about 5 or more exposures. So if a woman starts at 25 and then gets 5 or more mammograms by the time she is 30, her risk will be increased. Up to the age of 30, women should be careful. After 30, perhaps she could do mammography screening every other year; that is one idea. We only used 6 studies in our analysis and we need more prospective studies to find out more about this.”

All of these sound like potentially reasonable ideas to minimize the risk from beginning mammographic screening so early. After all, MRI screening has already been officially been recommended for women at high risk, such as BRCA carriers, and that recommendation was made two and a half years ago. Again, we have to consider the risk-benefit ratio:

The average woman has a 10% chance of developing breast cancer during her lifetime. In comparison, women who are carriers of the BRCA1 gene mutation have a 57% chance, and BRCA2 carriers have a 49% chance of developing breast cancer. Screening these women must start at an early age, since many will get breast cancer in their 30s or 40s, Dr. Jansen-van der Weide said.

Against such high risks, a slightly higher risk of harm from screening would probably still leave the risk-benefit ratio strongly in favor of starting mammogarphy early:

But he said no individual study has shown that mammography increases the risk of breast cancer in women at average or high risk of breast cancer, and estimating how much mammography might increase the risk can vary widely.

“For now, the unavoidable conclusion is that the benefits of early breast cancer detection in women at very high risk outweigh the low possibility of a radiation-induced breast cancer,” he said.

Jansen-van der Weide said the findings are based on a small sample, but when data from the studies were pooled they did reach statistical significance.

Which is what meta-analyses do. They take several studies with inadequate power to find statistically significant differences and try to find such differences by pooling the data fromt he studies. Sometimes their results are correct, and sometimes they are spurious. “Garbage in, garbage out” is a cautionary principle that haunts every meta-analysis ever done, leaving the inescapable conclusion that in this case prospective studies are required.

So what we are left with is a provocative study that hasn’t even been published in full in the peer-reviewed literature yet. It is suggestive of a potentially increased risk for breast cancer in young women with BRCA mutations due to too much low dose radiation, but it is not definitive. In the meantime, it suggests an important area where more research is needed and that we should probably be more cautious in screening young women with BRCA mutations or perhaps even change over to MRI screening of these women.

Not Mike Adams. After a breathless article entitled Mammograms cause breast cancer, groundbreaking new research declares by one of his woo minions named S. L. Baker, Adams decides to show his skill at nuance:

Mammogram pushers now have nothing left to stand on. The complete and utter hoax of mammography has now been wholly discredited through a flurry of groundbreaking studies performed by conventional medicine researchers! Yes, even the industry’s own former advocates now admit mammography harms far more women than it helps.

Why? Because mammography causes the very disease it claims to “detect”. It’s much like a clever sleight-of-hand magician’s trick where they reach for your ear and suddenly produce a coin that was presumably hidden there. But as everybody knows, they put it there themselves! Mammograms offer a similar kind of sleight-of-hand trick (or sleight-of-breast, as the case may be) by actually generating the very disease they claim to find. If so many women hadn’t already been harmed by mammography, the whole thing would be quite hysterical.

“Early detection saves lives,” they say. Except they stupidly forget to tell women the other side of the story: “Mammograms cause cancer.” And if you’re gullible enough to actually irradiate your breasts every year, don’t be surprised — shocked! — if they someday find tumors in them.

No sober weighing of risks and benefits for Adams! Oh, no! If a “conventional” medical modality is not perfect or if there is the slightest risk to it, suddently it’s evil itself. It kills women. Meanwhile, his woo minion so grossly misinterprets a study that it’s evidence of either an ignorance that stinks so bad that it would, as my father was fond of saying, knock a buzzard off a dung wagon (he’d use a different four letter word), or it’s rank dishonesty:

And it may not be only women with a familial risk for breast cancer who are at extra risk from mammography radiation. As NaturalNews covered last year, a report published in the American Medical Association’s Archives of Internal Medicine found breast cancer rates increased significantly in four Norwegian counties after women there began getting mammograms every two years. In fact, the start of screening mammography programs throughout Europe has been linked to an increased incidence of breast cancer.

I discussed the very study Baker is referencing. What Baker is saying about this study (and the present study) and what the studies actually conclude are related only by coincidence. Moreover, as is the usual case, there is no nuance and no discussion of risks versus benefits. To Adams and his crew, all science-based medicine is evil. It’s classic black and white thinking. It’s either “natural” (good) or scientific or “conventional” (evil). There is no middle ground.

Such black and white thinking is clearly another major defining characteristic of a crank or pseudoscientist. We see it in abundance in the anti-vaccine movement, without a doubt. It’s also very much in evidence in quackery supporters like Adams, as it is in creationists, particularly young earth creationists, the latter of whom are usually fundamentalists. Science-based medicine has no room for such thinking, nor does any other field of science. Indeed, science-based medicine is all about the application of science to the nuanced weighing of benefits and risks in order to determine what treatments and medical interventions can be used to maximize benefit and minimize harm. To achieve this balance, an accurate estimation of the risks and benefits of an intervention is required. The study under discussion, while by no means close to definitive, is a step in that direction when it comes to the early detection of breast cancer in young women carrying BRCA mutations.

Comments

  1. #1 Uncle Glenny
    December 3, 2009

    The political gamesmanship also begins on the recommendations. “Analysis” by wonkette, and from Vitter’s web site,

    … “Last month, this task force came out and publicly stated that women under the age of 50 should no longer get mammograms – a recommendation which flies in the face of everything that all women have been taught for decades,” said Vitter. “They also advised against annual mammograms for those over 50. These recommendations could have a drastic and negative impact on hundreds of thousands of women, and they should not be implemented – period. I held a roundtable in Louisiana on this very topic last month and heard from a number of survivors who credited annual mammograms for the early detection of their breast cancer.”…

    Wonkette notes that Barbar Mikulski has already introduced some sort of amendment.

  2. #2 Scott
    December 3, 2009

    Pooled OR revealed an increased risk of breast cancer among high-risk women due to low-dose radiation exposure (OR=1.63, 95% CI: 0.92- 2.90).

    This statement seems quite dubious to me. Since the 95% CI includes 1, my conclusion would be that the data did not support a conclusion of increased risk in this overall group. Sufficiently suggestive to warrant further study, I’d certainly buy – 1 is likely outside the 90% CI. But if a researcher is going to draw actual conclusions based on a larger CI, why quote the 95% and not the one the conclusion was based on? (My worry would be that the answer would be because 95% is the standard, and the author doesn’t recognize the inherent problem that implies.)

    Orac, I’d be very interested to know your thoughts on this bit in particular. I notice that in the post you only concluded that

    It’s suggestive that there may be increased risk of cancer in young women at high risk for breast cancer who either (1) undergo their first mammogram before the age of 20 or have more than five mammograms.

    which are the groups where the 95% CI did not include 1, but you didn’t explicitly address the prior sentence.

    Personally, if I were reading a physics paper that made such a comment in the abstract, it would make me very worried that they hadn’t done the rest of the statistics right – imprecision in conclusions tends to be linked with imprecision in the math.

  3. #3 Party Cactus
    December 3, 2009

    Sometimes I wonder if quacks pull stuff like that just to be contradictory. You know, like that guy in high school that always said any popular music or movie was terrible just just for the sake of being different.

    Maybe you should start referring to that site as Natural News [sic].

  4. #4 Denice Walter
    December 3, 2009

    “Meta-analysis,Schmeta-analysis! I don’t need no stinkin’analysis! Mikey doesn’t like it, so it’s BAD!” Seriously though,I have tried in vain to find out Adams’ educational background(his bio doesn’t give a clue)so I might venture a guess that it doesn’t include much bio,chem,stat,psych,physio,pharma,econo,etc..He describes himself as a former “software executive”- which could mean just about anything(his parents gave him money to startup a small company?).Concerning the “black-and-white thinking” and inability to weigh risk/benefit sounds like he hasn’t yet achieved “formal operational thought”(and he *is* nearly 40!),or has “other problems”.Adams often quotes Null, who reciprocates, making the results “Woo square”.Both of these charlatans sponsor gathering places(pseudo-science *salons*?) for woo-slingers of all stripes.Most recently, besides posting garbage about Orac’s area of expertise,Adams has also publicized an AIDS denialist film (“House of Numbers”) and AIDS denialist “research”(sic).Thus, he has expanded the range of people at risk for serious illness who he can possibly endanger.So what’s next? “The NaturalNews Low-down on Cardio-vascular Disease:How Greedy Cardiologists Will Take Your Money AND Kill You Too.”

  5. #5 MikeMa
    December 3, 2009

    It is a wonder that the human body is resilient enough, in many cases, to survive the natural non-cures Adams et al are pushing in place of real medical care. If not, he’d have no clientele, they all having died from lack of positive care.

    It is a good system for the woo-meisters though: Bilk the marks while they are alive, use their stories to drum up new business, and very few clients actually sue him for damages after they’ve died.

  6. #6 Pablo
    December 3, 2009

    I held a roundtable in Louisiana on this very topic last month and heard from a number of survivors who credited annual mammograms for the early detection of their breast cancer.”…

    Jeez, why bother sitting a panel of actual experts in the fields and stuff like that when we can just have David Vitter hold roundtables for us?

    of course, Vitter was against healthcare reform because I’m sure he is against having the government get involved in making medical decisions for you…

  7. #7 ms physics
    December 3, 2009

    As a medical physicist, I really don’t have words for conveying my anger at things like this. Thank you for providing some of your own.

    Also, how old were the data for the analysis? Pre-MQSA, I understand dose from an average mammography was quite different.

  8. #8 katydid13
    December 4, 2009

    Access to mammograms (and mammogram quality) has been Sen. Barbara Mikuslski’s signature issue(mostly because when she started in the Senate you didn’t need all the fingers on one hand to count the women). Honestly, I don’t think there is any amount of science that could change her mind.

  9. #9 Isis the Scientist
    December 7, 2009

    Ping, Orac (specifically in the comments section). These are some general meanderings I am having today about health care disparity. Any thoughts from you?

  10. #10 Orac
    December 7, 2009

    There may be, as I happen to work in a cancer center that is in a highly urban area with a large African-American population.

    However, not before Wednesday probably, as I’ve already written Tuesday’s post, and to do the topic justice requires some thought (whereas taking down the anti-vaccine movement often does not, given how nonsensical their arguments are).

    This is a complicated issue, and there is much to discuss. Having read your post I will say, however, that endocrine disruptors are not likely to be the culprit and that Kristoff’s claims about how fast breast cancer incidence rates are going up are dubious at best and deceptive at worst (I’m being polite). In fact, teasing environmental influences in breast cancer has been incredibly difficult.

  11. #11 ozdere
    December 10, 2009

    As a medical physicist, I really don’t have words for conveying my anger at things like this. Thank you for providing some of your own.

  12. #12 Realist
    December 11, 2009

    Until ALL research studies are initiated and implemented by researchers devoid of a commercial interest in the outcome (like the USPSTF one), we won’t know the real risk/benefit results of mammography or any other test that uses low-level radiation. It is always amazing to me how many studies we just DON’T do in this country. If you’re a corporation, or a medical research group that relies on funding from these corporations – why look for trouble? We will continue to leave it to countries with socialized medicine to ask the pertinent questions regarding the real health of our citizens. The profit-motive, which has insinuated its way into almost all forms of the research in this country, make a non-biased discussion like this impossible.

  13. #13 janet
    January 1, 2010

    Doctors over diagnose and push mamograms too much and of course it has some effect which is not good. I had my breast turned patches of red and pimple on the side after my last mamo and I will not do it again until I find a lump. It is harmful or it would not turn red.

The site is undergoing maintenance presently. Commenting has been disabled. Please check back later!