Parsing the Pros and Cons of Breast Cancer Screening

You might think that screening for breast cancer by mammography is a slam-dunk. It was not the last time I wrote about it; and it is not now. 

I don’t want there to be doubt about my position until the end of this column, however: I favor and recommend mammography, albeit with less conviction than colon cancer or cervical cancer screening. In contrast, I do not specifically recommend prostate cancer screening, although I think that’s a close call. I practice what I preach, of course: I do get colonoscopies, but do not undergo prostate cancer screening, and the women in my family get mammograms.

Ordinarily, my weekly topics are provoked by some new study I think is particularly important, or especially prone to media misrepresentation and misunderstanding. That is not the case today. I just searched for mammography and associated mortality in both the media and the peer-reviewed literature. There are always new studies, of course, and I’ll come back to the most recent before I’m done- but nothing truly immediate or very controversial emerged.

Rather, the provocation this time is simply the Internet. Opinions that at one time might have been quite insular- of, by, and for a gathering of contrarians and iconoclasts, for instance- now reverberate widely, and perhaps forever (time will tell). So it is that a colleague recently sent me a YouTube video of an expert opining on the utility of mammography to ask me what I think.

In the video, the head of the Nordic Cochrane Center makes an emphatic case against breast cancer screening. This is important, because Cochrane is among the leading sources of medical evidence assessment and synthesis in the world, and those who run their centers are highly qualified to judge such evidence.   On the other hand, this particular expert is a rather inveterate contrarian, opposing much of what “Big Medicine” does. That doesn’t make him wrong, but it does highlight his penchant for staking out and defending extreme positions.

The support for this position on mammography is clear enough: some large population studies show no mortality benefit of screening for breast cancer versus no screening at all. Cancer screening, of every kind, certainly has potential to do harm- related to treatment that may be unnecessary, or follow-up testing that may be invasive and dangerous, and needed to determine whether or not cancer is truly present. The harms of such testing are particularly hard to condone when false positives (i.e., the test suggests cancer but there really isn’t any) significantly outnumber true positives (i.e., follow-up testing confirms the presence of cancer), as is true for mammography.

Yet, by way of reminder, I have come to praise (well, defend) mammography, not bury it. I support it, as does the US Preventive Services Task Force, which assigns a “B” grade in support of mammography for women age 50 to 74. A “B” grade on their scale means: “there is high certainty that the net benefit is moderate or there is moderate certainty that the net benefit is moderate to substantial.” By way of comparison, colorectal cancer screening gets an “A” grade, indicating clear evidence of benefit, while prostate cancer screening gets a “C,” indicating uncertainty about the balance of benefit and harms. For those of us in Preventive Medicine, the USPSTF recommendations are something of a bible; you can access them yourself online.

As noted at the start, there have been recent studies on mammography and mortality, and they lend support to the benefits of screening, but with interesting provisos. A study from Canada, and another from New Zealand, both suggest that screening can save lives, but that benefit varies considerably with age and other characteristics.

There are a number of discrete reasons why the benefits of mammography might actually be, or seem to be, less than we might hope. Here’s my list, along with what I consider the main implications in each case.

1) Mammography is far from 100% accurate in general. This is true of most screening tests and simply something we must tolerate in a world where “perfect is the enemy of good.” The implications are that the technology should be improved or replaced to produce better images, and radiologists (and pathologists) should be ever more highly trained to interpret subtleties. Both of those are in fact on-going processes, so the images and their interpretation improve over time.

2) Breast tissue is not all created equal. I fully support the campaign to raise awareness about dense breast tissue, and the need for alternatives to mammography when those images are unreliable. Every time a woman with breast tissue too dense for reliable mammography is screened with a mammogram, it adds to the “evidence” that mammography is not helpful. The reality is that mammograms can work well for some women, while others are much better off with alternative screening tests.

3) Many early stage breast cancers found by screening may not require specific treatment, such as chemotherapy.  Genetic tests are evolving to help determine who needs what treatment. This is crucial to the value proposition of screening, because each time someone receives therapy they don’t really need, it narrows the gap between benefits and risks. 

4) The behavior of breast cancer varies with age of onset. Breast cancer is generally more aggressive before menopause and less so afterward, although this is a rule with many exceptions. Still, it has implications for optimal screening. Younger women might actually benefit from screening more often than yearly, because aggressive cancers can advance too much in a single year for annual screens to defend reliably against their spread. Older women might benefit more from less frequent screens. Such considerations figure in the on-going efforts to optimize screening. My advice here is that you discuss your situation, and family history, with your doctor, and work to customize the frequency of your screening accordingly.

5) Cancer treatment is improving rapidly, and mortality is a very blunt measure. This combination is really the clincher in my view. Cancer mortality rates are declining overall because of advances in treatment. This means that even breast cancer found later, without screening, may be treatable in ways that forestall death. But as noted, death is a very blunt measure, and often the one used to judge the merits of mammography. But the treatment for late stage, metastatic cancer is apt to be much, much tougher and more toxic to the body than the localized treatment for an early stage cancer found through screening. Perhaps the “death” rate would be comparable between them, but the “duress” rate certainly would not be. Paradoxical though it may seem, advances in cancer treatment, especially treatment for advanced cancers, may obscure the advantages of screening because lives can be “saved” with or without early detection. But that does not rule out the desirability of early detection, and less extensive treatment.

Cancer screening is, in the most literal sense, the art and science of looking for trouble. If not considered carefully, applied appropriately, and customized suitably, trouble is just what it can produce. One of the more reliable provisos in all of medicine is: just because we can, doesn’t mean we should.

But I believe we should screen for breast cancer, even though that case is far from the slam dunk many might suppose. We should also work to improve the screening tools, their interpretation, and their more personalized application; you may take comfort in knowing that’s just what the profession is doing, continuously. Reports of the limitations of current screening approaches are part of that process.

Breast cancer screening approaches are not perfect, but I- along with the US Preventive Services Task Force- remain convinced they confer net benefit, and do good. Let’s pursue better methods, but not make perfect the enemy of the good our current methods can do.