Breast Cancer Screening: Can We Think Too Pink?

The experts at the American Cancer Society (ACS) recommend screening for breast cancer by mammography every year for women over 40.  Wow.  That's a lot of testing.  Judging by conversations I've heard, and overheard, and innumerable cartoons I've seen, mammography is not high on any woman's list of fun things to do.  So maybe the recommendations of the U.S. Preventive Services Task Force (USPSTF) seem more appealing: every other year, starting at age 50, going through age 74.

[Before I go on, allow me a momentary digression into one of my pet peeves in the use of terminology.  This is not cancer prevention.  Cancer screening does not prevent disease.  It may detect it early and make it possible to cure it, thereby preventing a cancer death.  We don't know very much about preventing cancer.]

So how are we doing with mammography?  Are we saving lives?

Eighteen months ago (5-28-2011) I wrote an essay for this blog on the general subject of preventive medicine and touched briefly on screening mammography. I mentioned a book by Welch and colleagues (Overdiagnosed: Making People Sick in the Pursuit of Health).  Now Dr. H. Gilbert Welch is the second author (first author Archie Bleyer) of a paper published in the New England Journal of Medicine, one of the world's leading English-language medical journals.  Bleyer and Welch posed this very question.  They looked at three decades of data and found a substantial increase in detection, via mammography, of early breast cancer.  They did not, however, find a corresponding reduction in the diagnosis of late-stage breast cancer.  Specifically, cases detected early more than doubled, while cases diagnosed late declined by about 8%.

Why is that important?

If we've been advocating screening mammography for women over 40, and we say there has been a 28% decline in breast cancer deaths in this group, we might be inclined to put two and two together and say it's working.  But if early detection isn't substantially reducing cases not diagnosed until more advanced stages, the logical conclusion would be that improved treatment, not earlier diagnosis, accounts for most of the reduction in mortality.

According to the ACS, two to four out of 1,000 mammograms lead to a diagnosis of cancer.  Let's take the middle number (three) and ask what happens to the other 997 patients.  They all get a note saying, "Negative again, thanks for choosing Pink Mammography Services, see you next year."  Right?  Well, not exactly.  Some of them have findings that are not so straightforward.  Some of them wind up getting additional tests, like ultrasound examination of the breast or MRI.  Some of them undergo surgical biopsies.  They spend a lot of time worrying about whether they are harboring a life-threatening malignancy before being told, ultimately, that the conclusion is a benign one.

What about the ones who are diagnosed with early breast cancer?  Well, they all get treatment (assuming they follow their doctors' advice and recommendations). And each such case represents a breast cancer death prevented.  Right? Assuming, that is, that the long-term outcome is that the woman dies from something else.  (After all, that is one of the facts of life in a human body: the long-term mortality rate is 100%, sometimes quoted as one per person.)

Well, to be completely honest, we don't know.  It is entirely possible that some of these early cases were never going to progress to advanced disease and eventually cause death.  And the uncertainty about that was what led Bleyer and Welch to examine more than thirty years' worth of data.

These studies in the realm of epidemiology, public health, and the effects of medical interventions on large populations are difficult to do and more difficult to interpret.  But again, the central finding by the authors was that a large increase (137%) in the detection of cases of early breast cancer was accompanied by a decline of only 8% in the rate of detection of late-stage breast cancer.  And this suggests that other factors, such as more effective treatment of cases detected at later stages, are playing a substantial role in the reduction of the mortality rate from breast cancer.

One of the things we know about the behavior of some cancers is that there are people who harbor these diseases for many years and ultimately die from something else.  Thus it is reasonable to surmise that some women with breast cancer that can be detected by mammography when they have no symptoms would, if never diagnosed, live many more years and go on to die from an unrelated cause.  How common is that?

The answer from Bleyer and Welch:

After excluding the transient excess incidence associated with hormone-replacement therapy and adjusting for trends in the incidence of breast cancer among women younger than 40 years of age, we estimated that breast cancer was overdiagnosed (i.e., tumors were detected on screening that would never have led to clinical symptoms) in 1.3 million U.S. women in the past 30 years. We estimated that in 2008, breast cancer was overdiagnosed in more than 70,000 women; this accounted for 31% of all breast cancers diagnosed.

Now we have some numbers to ponder.  Even if the 1.3 million over three decades, the 70,000 in the year 2008, and the 31% are not exactly right, they should give us pause.  At the very least, they should tell us that we need to know much more about how to figure out which early breast cancers really need treatment and which cases may not.

So what do you do as an individual woman?  First, and especially if you are between 40 and 50, you have to decide whether to follow the ACS or USPSTF recommendations.  Talk to your doctor, and hope he or she really understands the science well enough to answer your questions.  If your mammogram is abnormal, no one is going to tell you that you should just wait and see.  And these data don't tell us that's a good idea, because we cannot tell which cases detected early will ultimately be a threat to life and which ones will not.

But when we look at these numbers on a population scale, we should ask ourselves what impact screening for early detection is really having on outcomes. After all, everything we do in health care costs money, and the supply is limited. We must always ask ourselves the kinds of tough questions that are answered by what policy wonks call a cost-effectiveness analysis, where the amount of money devoted to something is looked at in terms of dollars per "quality-adjusted life-year" (QALY).

If the patient is you, or a loved one, no amount of money seems too much to save a life.  But as a society, we must recognize that resources are finite and figure out where the money will do the most good for the most people.  If we're wondering - and we always should be - about the value of screening tests - Bleyer and Welch have given us more to think about.