Breast cancer screening: Time for rational discourse
Article first published online: 12 JUN 2014
© 2014 American Cancer Society
Volume 120, Issue 18, pages 2800–2802, September 15, 2014
How to Cite
Brawley, O. W. and O'Regan, R. M. (2014), Breast cancer screening: Time for rational discourse. Cancer, 120: 2800–2802. doi: 10.1002/cncr.28788
- Issue published online: 4 SEP 2014
- Article first published online: 12 JUN 2014
- Manuscript Accepted: 23 APR 2014
- Manuscript Received: 4 APR 2014
The controversy over the routine use of screening mammography rages on, fueled by several conflicting studies published within the last 12 months.[1-3] In this issue of Cancer, the distinguished clinical trialist and statistician Donald Berry writes a commentary on breast cancer screening and a failure analysis screening article published in this journal last year, which concluded that most deaths from breast cancer occur in women who do not have regular mammograms.[3, 4] Berry emphasizes the inherent biases involved in screening and uses modeling to support his argument that screening mammography has minimal, if any, real impact on decreasing mortality from breast cancer. On first examination, his commentary might appear to be a harsh criticism of this journal and of the article published in it. After digesting it, the commentary contains elements of constructive criticism that bring a searching perspective to the article and to breast cancer screening in general.
Also in this issue of the journal, Kopans et al were offered the opportunity to respond to Berry's criticism of their article. Their rebuttal is impassioned and makes short shrift of concerns that the value of screening might be exaggerated. They make several important points to support their promammography stance. First, Kopans and colleagues note that the breast cancer death rate changed to a much greater extent among women who participated in screening compared with those who did not, although all of the women had access to the same modern, systemic therapy. Second, the finding that mammography allows the detection of breast cancers at an early stage must translate into a reduction in breast cancer death. Third, the authors state that the harms of mammographically detected false-positives, emphasized by the US Preventive Services Task Force (USPSTF), are largely exaggerated. Fourth, they emphasize the major flaws with almost all of the randomized mammography trials, including but not limited to the reality that the equipment used is now outdated. In addition, they focus on the controversy regarding the optimal age to start screening, stating that there is nothing magical that happens to a woman's breast at age 50 years that would make mammograms so much more effective than in women in their 40s.
These authors and, indeed, both sides of this contentious argument have strong points and strong opinions. The scientific process involves the publication of research articles after an editor makes a decision with the help of peer review. The later publication of letters to the editor and commentaries about the article are an important part of the process as the scientific and medical community determines what an article means and its importance to the practice of medicine. We need frank, direct, and respectful discussion through the entire process.
There are several aspects regarding screening mammography that are generally accepted by most objective scholars. First, screening mammography is not perfect and misses a fairly considerable amount of cancers. Second, mammography increases the detection of small, early stage breast cancers, although its impact on decreasing the incidence of later stage breast cancers is less clear, indicating that a subset of mammographically detected breast cancers are probably overdiagnosed. Third, there are flaws in the design of the majority of the randomized trials, including inappropriate randomization and women in the control groups receiving mammograms. Fourth, as described by Dr. Berry using compelling data, there are inherent biases with all screening tests, including mammography.
The issue of overdiagnosis is important and deserves specific focus. In a recent analysis of the Surveillance, Epidemiology, and End Results database, Bleyer and Welch suggest that mammography leads to “massive overdiagnosis” of breast cancer. They suggest that greater than 30% of breast cancers detected by screening mammography are in fact overdiagnosed. As outlined by Kopans and colleagues in this issue of Cancer, numerically, this would mean that approximately 70,000 women are overdiagnosed with breast cancer on an annual basis, which accounts for almost one-third of breast cancers diagnosed in the United States annually. An important weakness of the randomized mammography trials and these retrospective analyses is the lack of accountability regarding which subtype of breast cancer women are being diagnosed with after screening mammograms. It is certainly possible that a subset of women is diagnosed with small estrogen receptor-positive breast cancers that, analogous to prostate cancer, would never become clinically relevant in their lifetimes. In contrast, small HER2-positive and triple-negative breast cancers almost certainly would become clinically relevant and, without treatment, would likely impact survival. Defining overdiagnosis as simply being diagnosed with breast cancer but not dying from this disease takes minimal account of the advances that have been made in breast cancer therapy and in tailoring treatment to individual breast cancer subtypes.
Unfortunately, the end result of these conflicting publications focused on screening mammography is that the general public gets very mixed messages. The changing news media dynamic tends to hype publications with extreme results on either end of the screening spectrum. One week they are using articles like that by Webb et al to promote screening for everyone, whereas the next week they are citing the recent long-term follow-up of the Canadian Screening trial as a reason not to get mammograms. These days, it is common for research institutions, scientists, journals, and even research funders to issue press releases that exaggerate the importance of findings and usurp the traditional process of scientific evaluation. Indeed, it is amazing how often a press release calls a single study “a game changer” that should affect policy. Science teaches us that it is rare that 1 study or 1 article should change policy.
True legitimate differences of opinion can and should exist, and the press should interview and write about experts with various opinions. Unfortunately, there is an increasing intolerance of diverse opinions, as evidenced by both of the distinguished parties to this debate in these 2 thoughtful, well written, but strongly opinionated pieces. The intolerance is often fueled by arguments with pseudo facts and scare tactics that are inappropriate (although it is not our position that this is the case in either of these commentaries). The end result of the coverage of screening issues by the media creates a situation in which the public can be misinformed.
The press has challenges managing their conflicts of interest on this hot-button topic as well. The tendency of Americans to believe in screening—and the fact there is passionate disagreement among experts—produces an opportunity for the press to exploit the story. We used to say that controversy sells newspapers. The digital age and social media have amplified the power of controversy to attract our focus.
Several organizations publish cancer screening recommendations. These organizations use different processes to assess the scientific literature. There is increasing concern about both financial and emotional conflicts of interest interfering in the generation of guidelines. It is obvious from the Berry and Kopans et al commentaries that the emotional conflicts of interest are actually more difficult to manage than the financial conflicts.
The Institute of Medicine has recently published guidelines for systematic literature reviews and clinical practice guidelines. It is noteworthy that the Institute of Medicine is very concerned about committee membership. They recommend that committees involved in making guidelines should be made up of individuals capable of understanding the issue, but they specifically recommend against including specialists who treat or manage patients with the disease in question on the committee. The lack of specialists on these guidelines groups is one of the major criticisms voiced by advocates of screening regarding the guidelines developed by the USPSTF.
The 2014 American Cancer Society guidelines for breast cancer screening are anxiously awaited. Until then, we note that all of the large screening trials have flaws, as outlined previously. These studies were long-term logistical challenges.
Reasonable experts agree that the body of evidence suggests that there is some benefit to mammography screening among women aged ≥40 years. Most studies suggest that the benefit increases with age, and there are limitations to mammography screening at every age. Although the benefits of mammography screening have almost certainly been exaggerated, this does not mean that it does not save lives or that women should not get it. It means we need to use it with caution, honestly explain the limitations, and realize we need to develop a better test.
No specific funding was disclosed.
CONFLICT OF INTEREST DISCLOSURES
The authors made no disclosures.
- 7Institute of Medicine of the National Academies. Clinical Practice Guidelines We Can Trust. Washington, DC; The National Academies Press; 2011.