Monday, October 22, 2012

End the war metaphors for cancer screening and treatment

Like millions of cycling fans all over the world, I was very disappointed about revelations that 7-time Tour de France champion and cancer survivor Lance Armstrong had been using banned performance-enhancing substances for most of his career, despite years of increasingly vigorous denials. To many people, Lance transcended his sport through his seemingly indomitable will to win - not only Tours, but a very public fight against advanced testicular cancer. Lance's words of encouragement to other people struggling with cancer were marked with war metaphors: personal battles, wars to be won at all costs against a tenacious and unyielding enemy. And while there's no such thing as cheating in the effort to beat a medical condition, the evidence now shows that Lance's determination to win at all costs in sports led him down a path that has resulted in his disgrace.

On Thursday I will speak about cancer screening at the Johns Hopkins Bloomberg School of Public Health's Fall Policy seminar on "Science and Public Policy in Conflict." One of the messages I hope to leave with the audience is that we need to end, or at least soften, the harsh war metaphors for cancer screening and treatment, which endow screening tests such as mammograms with far more power than they really have to affect patient outcomes, and leads to uninformed advocacy and public policy that ignores the harms of overdiagnosis and overtreatment that inevitably result. As I blogged last June (and have reposted below), there are "no easy victories" in cancer screening when it comes to the evidence.


Nearly forty years ago, President Richard Nixon famously declared a "War on Cancer" by signing the National Cancer Act of 1971. Like the Manhattan Project, the Apollo program that was then landing men on the Moon, and the ongoing (and eventually successful) World Health Organization-led initiative to eradicate smallpox from the face of the Earth, the "War on Cancer" was envisioned as a massive, all-out research and treatment effort. We would bomb cancer in submission with powerful regimens of chemotherapy, experts promised, or, failing that, we would invest in early detection of cancers so that they could be more easily cured at earlier stages.

It was in the spirit of the latter that the National Cancer Institute launched the Prostate, Lung, Colorectal, and Ovarian Cancer (PLCO) Screening trial in 1992. This massive study, which eventually enrolled more than 150,000 men and women between age 55 and 74, was designed to test the widespread belief that screening and early detection of the most common cancers could improve morbidity and mortality in the long term. Not a few influential voices suggested that the many millions of dollars invested in running the trial might be better spent on programs to increase the use of these obviously-effective tests in clinical practice.

They were wrong. As of this week, the PLCO study is 0-for-2.

Miss #1 occurred in March of 2009 when the PLCO study first reported no mortality benefit from annual PSA testing, a test that a majority of men over 50 undergo routinely. Miss #2 occurred over the weekend, when the Journal of the American Medical Association published a landmark paper that ended with the following paragraph:

We conclude that annual screening for ovarian cancer as performed in the PLCO trial with simultaneous CA-125 and transvaginal ultrasound does not reduce disease-specific mortality in women at average risk for ovarian cancer but does increase invasive medical procedures and associated harms.

The lung and colorectal screening components of PLCO have not yet reported mortality data**, and there is reason to believe that at least the latter will likely yield some positive results. Although it has largely been supplanted by colonoscopy and CT colonography (aka "virtual colonoscopy") in the U.S., flexible sigmoidoscopy was already shown to reduce deaths from colorectal cancer in a randomized trial published in the Lancet last year. And PLCO's screening chest x-rays are probably a loser, but a preliminary report from NCI's National Lung Screening Trial suggest that screening CT scans can reduce lung cancer mortality in heavy smokers. (Even after this report is confirmed in a peer-reviewed scientific journal, there will still be plenty of reasons not to rush into lung cancer screening, as I outlined in a previous blog post.)

Still, these are hardly the magic bullets or the resounding victories that many expected from the "War on Cancer." The same can be said for chemoprevention, or the strategy of prescribing medications for healthy adults to prevent cancers from developing at all. The vast majority of "high risk" women have avoided breast cancer chemoprevention with tamoxifen and raloxifene due to their unpleasant side effects (which include hot flashes and life-threatening blood clots), despite a 2002 recommendation from the U.S. Preventive Services Task Force for clinicians to discuss these drugs with their patients. (This recommendation has not been updated since, largely due to politics, not science.) A new study published in the New England Journal of Medicine has reported that the drug exemestane reduces the risk of invasive breast cancer without the other drugs' side effects. But here's the rub: we can't be sure how many of those breast cancers are the ones that inevitably lead to symptoms and death, rather than the 1 in 3 that are thought to be overdiagnosed.

The bottom line from recent research is that there are no easy victories in cancer screening and prevention - just slow, incremental progress. Companies that have a profitable product to push would like you to believe otherwise, but when it comes to cancer prevention, there is no substitute for a healthy lifestyle: Don't Smoke. Drink in Moderation. Exercise. And Eat a Well-Balanced Plate.

** Note: the PLCO trial subsequently reported that chest x-ray screening did not reduce lung cancer deaths, while flexible sigmoidoscopy reduced colorectal cancer deaths by 26 percent. So PLCO's "final score" is 1 in 4 screening tests evaluated showing a positive effect on health.


  1. Hi Kenny,
    I agree that war metaphors are not constructive in discussing cancer research, screening or treatment. Unfortunately, not smoking doesn't seem to help a woman's odds of not getting breast cancer. And while I'd like to think that eating a well-balanced diet does, there's little (no) evidence to support that, either.

    What bearing Lance Armstrong's story has on the value of mammography is not evident to me; I don't know that it's reasonable to invoke his stream of deception when considering the data and potential value of breast cancer screening in women. As for "1 in 3" breast cancers not needing treatment, as best I can tell that's the hypothetical view of one doctor.

    With respect,

  2. Hi Elaine - I agree with you that there evidence supporting exercise and diet for preventing cancer could be stronger. The 1 in 3 figure comes not only from H. Gilbert Welch's book, but a systematic review from the Cochrane Collaboration, published in BMJ: I've updated the reference in this blog post.

    Interesting that you use the term "stream of deception," since I'd argue that's what we (as in the medical profession) have been feeding vulnerable women and men about breast and prostate cancer screening for more than 2 decades. A majority of primary care physicians nationally routinely screen women with mammography at 35, even though no organizations recommend this practice. More than one-third of men over 75 receives PSA testing that will not extend his life but potentially will reduce its quality. (I'd argue that PSA screening in general has no measurable benefits, but it certainly does not in septua- and octo-genarians.) What aggressive screening does result in is more cancers detected, and more money for those who order and perform the tests. And aggressive treatments that "cure" cancer that didn't need to be found in the first place are touted as victories in the war on cancer, regardless of the financial, emotional, and physical costs.

  3. Hi Kenny,
    Thanks for the link to the 2007 BMJ article, which I've read and will review. I see it as a good example of how a meta-analysis can "output" a particular study depending on what the researchers put into it (or cut). You're right that too many 35 year old women get mammograms for no reason (and based on "fear" or something like that, besides misinformation and poorly-informed docs).

    But my main concern is the blending of findings and recommendations for prostate, breast and other cancers. Mammography has improved big-time since 1985 (lower rads, better training of radiologists, ways to evaluate dense breasts by sono before biopsy, more regulation of facilities), but the blood test for PSA hasn't changed much at all. Also, breast cancer tends more often to affect women in the middle of their lives.

    The USPSTF does important work. But I do wish they'd include an oncologist or two. I also wish those discussing the topic of cancer screening would keep these distinctions (details matter!) in mind.

    Best, Elaine

  4. All great points. Thanks for contributing!