Dramatic Mammography Development: Radiologist Calls for More Radiology

Andrew J. Vickers, PhD

Disclosures

October 14, 2010

In other news, local plumber Vincent Fontino says that more homeowners should have routine plumbing checks: "It's a great way to prevent potentially costly problems building up."

So, another mammography study, another opinion proffered. Although this was an interesting one: according to Daniel Kopans, professor of radiology at Harvard Medical School, a study from Sweden should "end the debate" about mammography in younger women.[1] The study concerned was a nonrandomized comparison of breast cancer deaths comparing areas in Sweden that did and did not offer mammography to women aged 40-49.[2] I guess the findings are interesting in their own way, but it is hard to understand how an observational study could "end the debate" when so many randomized trials have failed to do so.

The reaction to the study, however, does nicely illustrate what I see as several pernicious attitudes to cancer research.

  • The study methodology is rigorous if the conclusions support my prior beliefs. Dr Kopans, the Harvard radiologist, claims that the Swedish study is "robust."[1] It is interesting to compare this with his description of a negative Canadian randomized trial of mammography in young women as "badly designed."[3] The message appears to be that that you shouldn't trust a randomized trial if you don't like the results, but you can throw your weight behind an observational study if you agree with its conclusions. Kopans also described the quality of the mammograms in the Canadian study as "poor." Clearly, radiology in rural Sweden in the 1970s was far advanced of that available in Toronto 10 years later.

  • Whoever has the last laugh wins. A trial is reported showing no effect of a drug, but then a second study is published with statistically significant results. A common reaction is "oh, it turns out that the drug does work." The reverse situation, a positive initial trial, a negative replication, is scientifically identical (it doesn't make a difference to the stats if one trial is published a few months before another). Yet the typical reaction becomes: "no, they proved it was ineffective actually." All of which is to say that we should avoid conclusions on the basis of single studies, if there are other pertinent data. And of course, in breast cancer, there are plenty of other data. What we need is not to cherry pick this or that study, but to look at the big picture of all studies combined. On which point...

  • Three golden rules of scientific critique: demonize, demonize, demonize. The group that has looked at all the data systematically on mammography is the US Preventive Services Task Force.[4] But when the group made a mild change in the recommendations -- instead of advising all women 40-49 years of age to get screened, they said that this was a personal decision -- there was outrage. Dr Kopans described the recommendations as "an ego thing" because the Task Force "didn't want input from experts." He went on to suggest that the Task Force had been hijacked by individuals long opposed to mammography who "are willing to let women die."[5]

  • Motherhood, apple pie, and breast cancer. The reaction to the mammography study is also a good example of the "pinkwashing" of medical research: if it is breast cancer, something should be done; any other cancer, well, interesting, but hardly a major priority. A large randomized study comes out showing that prostate cancer screening saves lives,[6,7] and most of what we hear about is overdiagnosis.[8]Yet the number of cancer deaths compared with the number of cancer diagnoses is similar for prostate (~32,000; ~215,000) compared with breast cancer (~40,000; 210,000).

In short, judge research on its merits, not whether you agree with its findings; look at the big picture, not individual studies; make nice with those you disagree with. A cancer death is a cancer death, regardless of whether a 250-lb linebacker wears pink to raise your awareness of it.

Comments

3090D553-9492-4563-8681-AD288FA52ACE
Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as:

processing....