Sixth in a series on the long-term experiment we know as annual screening mammography, this post reviews the National Cancer Institute’s most recent information for physicians on the hazards of mammography screening. Whereas less than 1% of women’s lives can be said to have been prolonged because of breast cancer screening, between 48% and 59% of women, depending on their ages, suffered unnecessary treatment as a result of false-positive mammograms. Mammography may even have led to an increase in the incidence of breast cancer over time.

Mammography’s Shadows, VI: The National Cancer Institute Weighs In

Crab Image

Like all federal agencies the National Cancer Institute is subject to indirect political pressure through congressional and executive branch control of agency budgets.  Such pressure can and does influence programmatic decisions. That had been the case with the Breast Cancer Demonstration Project of the 1970s (see this blog’s post for May 25, 2015, “Mammography’s Shadows: The Campaign,”).

But thanks to federal commercialization policy, which prohibits government agencies from competing in the market for commercially available products and services, neither the institutional survival of federal mission or regulatory agencies, nor their employees’ job security, depends on revenue streams from medical procedures, equipment, and pharmaceuticals. Moreover, the science intensive agencies, such as the National Institutes of Health, supplement their internal scientific capabilities with advisory groups and consensus conferences of outside researchers, while those agencies reserve to themselves the duty of final arbiter of disputed research findings.

Thus it should not be surprising that the National Cancer Institute’s own positions on the benefits and risks of screening mammography might vary from those taken by the American Cancer Society and Memorial Sloan Kettering Cancer Center.  When reporting on screening mammography in the framework of its parent agency’s (National Institutes of Health) goals for its “Healthy People Program”[1], a programmatic document,  the NCI cautiously endorsed widespread screening mammography.

But it did so with subtle differences, noted here in italics: In 2012 the institute wrote  biennial “mammography screening allows for the early detection of breast cancer, which may help reduce mortality from breast cancer, especially among women aged 50 to 69 years.”  For these women “there is solid evidence that screening may lower this risk by up to 30 percent.  For women in their 40s, the [mortality] risk may be reduced by about 17 percent.  For women aged 70 and older, mammography may be helpful, although firm evidence is lacking.”[2]

In the NCI’s subsequent breast cancer screening guidance for health professionals, however, the federal cancer agency explicitly relied on the comprehensive ‘physicians data query (PDQ®)’ research database.  This evidence-based foundation, in contrast to the policy goals of the NIH, supported a much different construction. Well over half of the 38 research reports referenced in the guidance were published after 2000, none was published prior to 1990, and missing are citations to the “large randomized trials conducted a generation or more ago” upon which Memorial Sloan Kettering claimed it based its own mammography screening guidance.

Noticeable by its repeated appearance among the published research reports referenced by the NCI is the work of H. Gilbert Welch, Professor of Medicine at Dartmouth’s Geisel School of Medicine, and Peter C. Gøtzsche of the Nordic Cochrane Center, and their co-investigators.  Since the early 1990s Welch in particular has been calling attention to screening mammography’s hazards of over-diagnoses and over-treatment.[3]

Peter C. Gøtzsche, meanwhile, has been a thorn in the side of routine mammography screening advocates, because the Cochrane group’s rigorous meta-analyses of randomized controlled trials in medicine have failed to support the claim that screening mammography reduces breast cancer mortality, while also suggesting it results in significant over-diagnoses and over-treatment.[4]  In a debate about the health of women’s bodies dominated overwhelmingly by men, it’s noteworthy that two laudatory forwards to Gøtzsche’s 2013 book were written by women, the President of the National Breast Cancer Coalition (Fran Vesco) and President of the Royal College of General  Practitioners (Iona Heath).

The NCI evaluated screening mammography research studies on the basis of both their internal validity (reliability of cause-effect relationship, elimination of confounding variables) and external validity (generalizability beyond populations studied).[5]  In the February 2015 update to its December 2014 health professionals’ review of breast cancer screening, the NCI analyzed research findings as they would apply to 10,000 women who undergo annual screening mammography (a) for women in their 40s, 50s and 60s, and (b) extending over two time periods, 10 years and 15 years.  Given cancer latency which can reach beyond 10 years, the longer time periods are important.

The institute’s research analysts found that the best estimates of cancer deaths averted over the next 15 years for women in their 40s attributable to screening mammography did not exceed .16% (or 16 out of 10,000).  That percentage increased slightly for women in their 50s (at most .32% or 32 out of 10,000) and again for women in their 60’s (at most .49% or 42 out of 10,000).  At no time did the estimates approach 1%.

At the same time, the best estimates of the false-positive results during a ten year period reached minimums of 59% (5,940 out of 10,000) for women in their 40s, 58% for women in their 50s, and 48% for women in their 60s.[6]  False-negative results occur as well, and no doubt the fear of that result accounts for what the NCI concluded was the “biggest risk factor for having a false positive mammogram,” namely “the individual radiologist’s tendency to read mammograms as abnormal.”[7]

Especially valuable was the NCI’s effort to wrestle with the question of over-diagnosis, or the diagnosis and treatment of cancers over the 10 year period that would never have become clinically significant.  Since “cancers that will cause illness and/or death cannot be confidently distinguished from those that will remain occult . . . all cancers are treated.”  But an effort to assess the extent and circumstances of over-treatment is critically important to those women suffering through treatment unnecessarily, albeit generating revenue for those furnishing it.

The NCI approached this question by examining studies comparing breast cancer incidence before the introduction of screening mammography, and afterwards.  If mammography screening has an appreciable effect, the incidence of cancer detected in a population should increase during the early years of screening, but then the incidence should decline in later years.  This would result from cancers having been found earlier that would have been palpable only in later years.

What the NCI analysis of population-based screening data from studies in Norway, Sweden, the United Kingdom, and the United States found was that “breast cancer incidence rates increase at the initiation of screening without a compensatory drop in later years.”

One study in 11 rural Swedish counties showed a persistent increase in breast cancer incidence following the advent of screening.  A population-based study showed increases in invasive breast cancer incidence of 54% in Norway and 45% in Sweden in women aged 50 to 69 years, following the advent of screening.  No corresponding decline in incidence in women older than 69 years was ever seen.[8]

While couched in terms of evidence of possible over-diagnosis, the NCI’s analysis of trends in breast cancer incidence before and after the introduction of screening mammography is an important step toward liberating the debate over mammography screening from the very narrow epistemic and conceptual confines within which it has been conducted over the past half-century.

First, measuring the value of screening mammography entirely in terms of “early treatment saves lives,”  limits the question to mortality data, which seems certain and definite.  But the reporting of deaths and determination of their causes is not an exact or detached science, especially when applied to older persons.[9]  Secondly, while male physicians naturally incline to the heroic aspiration to preserve life (postpone death), those female patients who must suffer through ultimately unsuccessful breast cancer treatment may prefer quality of life over its duration.  But those patients are unlikely to know at the onset how long their treatment must be endured and whether it will succeed.

Secondly, the mantra “early detection saves lives” assumes that we know enough about the causes and optimal treatments for various breast cancers to fulfill the promise of “early detection.”  This assumption becomes critical to the validity of all statistically based arguments for screening mammography:  it requires that breast cancer treatment for all women is a constant and thus neutral factor, when it In fact is a critical and unstable variable.

One of the few comments responding to the Canadian National Breast Cancer Study published in the British Medical Journal (2014) that did not get ensnared in numbers generated by randomized control trials came from Donald L. Benjamin, a researcher at Australia’s Cancer & Support Society.  Benjamin’s own research into the efficacy of surgery for breast, bowel, lung, prostate and ovarian cancer (research also based on randomized controlled trials) led him to conclude that breast cancer is a systemic disease, and its tumors are symptoms rather than its causes.[10]  While this is not a new insight, its truth confounds efforts to reduce standard modes of breast cancer detection and mortality to simple comparative numbers.

The appreciation of breast cancer’s systemic character has had a substantial effect not only on determining risk (e.g., the presence of biomarkers such as mutations of the BRCA1 and BRCA2 genes in families experiencing unusual levels of breast and ovarian cancers[11]).   Our increased understanding of the role of estrogen receptors (ER) in breast cancer has led to successful endocrine therapies using tamoxifen, an ER suppressor, and aromatase inhibitors that block the enzymes that convert androgens into estrogens.[12]  How effective these therapies are, and their side-effects, continue to be examined, as they should be.

In short, understanding how breast cancer works is an ongoing process, and breast cancer therapies are not fixed in time, nor are their results for all women constant or uniform. Thus the most important variable in efforts to link screening mammography and breast cancer mortality is the etiology of breast cancer in individual women.  Treating breast cancer morbidity as a neutral constant, as one attempts to establish a linear relationship between tumor phase at detection and breast cancer mortality, is the most serious weakness in all quantitative justifications for screening mammography.

Thirdly, by assessing methods of breast cancer detection or treatment solely within the limited epistemic and conceptual confines of breast cancer itself, we neglect the single most compelling constant of all:  the irradiation of women’s breasts, with inevitable x-ray scatter to the torso during mammography screening, diagnostic mammography, and radiation treatment.

This has turned out to be a very high risk to take, all in order to spare women of any age from breast cancer deaths at the rate of less than on-half of one percent per 10,000.

(Published June 18, 2015.  Our next and final post in the “Mammography’s Shadows” series looks at the now apparent consequences for untold women of their cumulative annual exposures to breast x-rays.  It also notes that it was women, not only because of the nature of their disease but because of cultural attitudes, who were persuaded to participate in this long-term and hazardous medical experiment.)


[1] See:

[2] The federal Healthy People 2010, 2020 target is biennial mammography screening for 70% of all women over 40, a target that was met only briefly in 2003. National Cancer Institute, “Cancer Trends Progress Report-2011-2012 Update: Breast Cancer Screening.” doc_detail.asp?pid=1&did=2011&chid=102&coid=1016&mid=. Downloaded March 9, 2015.

[3] See, for example, H. Gilbert Welch, Should I be Tested for Cancer?: Maybe Not, and Here’s Why (University of California Press, 2006), and H. Gilbert Welch, “Breast Cancer Screenings: What We Still Don’t Know,” The New York Times (Dece,ber 29, 2013).

[4] See Donald G. McNeil, Jr., “Confronting Cancer: Scientist at Work — Peter Gotzsche: A Career That Bristles With Against-the-Grain Conclusions,” The New York Times (April 9, 2002), and Gøtzsche, Peter C. Mammography Screening: Truth, Lies and Controversy. (Radcliffe Publishing Ltd., 2012).

[5] National Cancer Institute, “Breast Cancer Screening (PDQ ®): Health Professional Version,” (December 5, 2014). cancertopics/pdq/screeningbreast/healthprofesional/.Downloaded December 20, 2014).

[6] National Cancer Institute, “Breast Cancer Screening (PDQ®): Harms of Screening Mammography,” Update: February 6, 2015. Downloaded February 25, 2015.

[7] National Cancer Institute, “Breast Cancer Screening (PDQ®): Harms of Screening Mammography,” Update: February 6, 2015. cancertopics/pdq/screening/breast/healthprofessional/page8. Downloaded February 25, 2015.

[8] National Cancer Institute, “Breast Cancer Screening (PDQ®): Harms of Screening Mammography: Overdiagnosis,” Update: February 6, 2015. Downloaded February 25, 2015.

[9] See for example, Kathryn Schulz, “Final Forms:  What Death Certificates Tell us, and What They Don’t,” The New Yorker (April 7, 2014), pp. 32-37.

[10] Donald L. Benjamin, “Comment re. Twenty five year follow-up for breast cancer incidence and mortality of the Canadian National Breast Screening Study: Randomised screening trial,” British Medical Journal (11 February 2014),, Downloaded March 2, 2015; DonaldJ. Benjamin, “The Efficacy of Surgical Treatment of Cancer-20 Years Later,” Medical Hypotheses Vol. 82 (2014), pp. 412-420.

[11] BRAC1 and BRCA2 are human genes and their proteins that participate in the repair of damage to the chromosomes and DNA, and the destruction of cells with damaged DNA.

[12] Monica Morrow, MD, “50 Years in Breast Cancer: Dramatic Progress in Treatment Based On an Improved Understanding of Biology,”  Annual Meeting of the American Society of Clinical Oncology, 2014. Downloaded  January 3, 2015.

Post to Twitter Tweet This Post

No Comments

Leave a Reply

Your email is never shared.Required fields are marked *