Rethinking Screening Mammography

February 24, 2014

Rethinking Screening Mammography

  1. Andrew M. Kaunitz, MD

Long-term Canadian study confirms overdiagnosis and lack of mortality benefit — where should we go from here?

  1. Andrew M. Kaunitz, MD

Randomized trials conducted from the 1960s to the 1980s indicated that screening mammography reduced breast cancer mortality (Cochrane Database Syst Rev 2013; 6:CD001877), but these studies were conducted in an era when breast cancer treatments were less effective and women's awareness of breast cancer less keen.

Now, Canadian investigators report findings from 25 years of follow-up in a screening mammography trial that was initiated in 1980 and involved some 90,000 women aged 40 to 59. All participants underwent baseline clinical breast examinations by trained clinicians. Women aged 40 to 49 were randomized to 5 annual mammograms (plus annual breast examinations) or to usual care. Women aged 50 to 59 were randomized to 5 annual mammograms (plus annual breast examinations) or to annual breast examinations only. Outcomes were tracked during the 5-year screening period and subsequent follow-up through 2005.

During the screening period, 484 of 666 cancers in the mammography group were screen detected. During the entire study period, breast cancer was diagnosed in 3250 and 3133 women in the mammography and no-mammography groups, respectively; 500 and 505 women died from breast cancer. The 25-year cumulative mortality from breast cancer was similar among women in the mammography and no-mammography groups, and these findings did not vary with age. Within 15 years of baseline, an excess of 106 cases of breast cancer were identified with screening mammography; thus, 22% of screen-detected cancers (106/484) represented overdiagnosed breast tumors.


These findings echo those of other recent studies of breast cancer screening (NEJM JW Womens Health Sep 22 2010 and NEJM JW Womens Health Aug 11 2011). Mammograms are expensive (NEJM JW Womens Health Feb 10 2014) and have high rates of false-positive findings (NEJM JW Womens Health Jan 17 2014). For decades, we have marched to the drumbeat “Mammograms Save Lives.” Annual screens have become an easy recommendation to make — and for our patients, the reassurance that accompanies a normal mammogram is comforting. Many women will be perplexed by this new information; others may view it with suspicion. While we await updated guidance from professional societies, my approach is to encourage patients to follow the 2009 U.S. Preventive Services Task Force guidelines: In average-risk women, screen every 2 years beginning at age 50.

Editor Disclosures at Time of Publication

  • Disclosures for Andrew M. Kaunitz, MD at time of publication Consultant / Advisory board Agile; Bayer; Merck Equity Vertex Royalties UpToDate Grant / research support Agile; Bayer; Endoceutics; Medical Diagnostic Laboratories; Noven; Teva Editorial boards Contraception; MedScape; Menopause; OBG Management Leadership positions in professional societies North American Menopause Society (Board of Trustees)


Reader Comments (9)

ashok vijh Other

If ,as implied,the women are being over-diagnosed and overtreated, why are women still dying of breast cancer--that would indicate that,on the contrary,at least some women are not being properly diagnosed or treated timely. I think lack of biomarkers of high reliability make this area more of a guess work,with medical economists pushing their cost-benefit analyses,without much rigorous science.
Dr. Ashok Vijh,Ph.D., Electrochemical physicist,also publishing on oncology.

RAVI RAO Physician, Oncology, Pvt Practice

Interesting article. I like the conclusion of the paper (worth reading): if we have access to good adjuvant therapy, mammography may not be worth it. There is a definite overdiagnosis of breast cancer. The question, of course, is who is going to tell that to the patient ? Recently, when the USPTF came out against mammograms in 40-50 year olds, I remember the indignation expressed by everyone (including the oncology and radiology colleagues I work with). Mammograms are going to become like pap smears - with marginal utility, expensive, paid for by all of us, and will make us all feel warm and fuzzy that we are doing something.

DONALD STANLEY Physician, Pathology, Maine Medical Center

Prevalence of disease is an input that few physicians use productively. The result of any screening test, to include sensitivity and specificity depends mostly on the prevalence of the disease; or on the clinician's prior probability.

This requires more emphasis in teaching physicians how to use any test.

Bruce Bodner Physician, Surgery, General, Taunton, MA

Is it really that tough, in the USA, to enroll 2000 women, all aged 40, who have never had a mammogram, split them into two groups and screen half them every year with a good digital mammogram, and the other half just follow. Do the same thing with 2000 fifty year olds. report after 10, 15 and 20 years. It would answer some questions.

James Stark, MD Physician, Oncology, Office

As hinted at by Dr. Vassar above, the period after the five years was very long. I read the original paper in the BMJ line by line at least three times and was forced to conclude that the design was fatally flawed. For those cancers diagnosed during the five-year screening period the women in the screening group did significantly better. That improvement was diluted by screening in neither group for the next 20 years. I am very surprised that more reviewers didn't pick this up, that they accepted the methodology at face value. Even studies published in good journals can be very misleading, as was this one.

TOKUNBO AJAYI Resident, Internal Medicine, Salem, MGH

I am concern about generalizing this study because of unequal access to healthcare in lower social economic groups and countries.
Also, it would informative to know patients quality of life after treatment in those that were diagnosed early( less chemo, less radiation and probable better quality of life) as compare to those diagnosed later though both had similar mortality rate.

Carol Vassar, MD Physician, Internal Medicine, pvt practice

The Dartmouth study looking at cancer registries 1976(?) c/w 2008 showed a 50% increase in the dx of breast cancer, rather than 22% ( over 28 years rather than simultaneous, but still a large discrepancy). In the Dartmouth study, would the change of DCIS from cancer to risk factor then allow us to get the 8% reduction in advanced cancers? In the Canadian study, what was the screening in the 25 year follow-up. That follow-up period seems too long. What were the results at 5 years after the end of the study and and 10 years?

BRUNO HAMMERSCHLAG Physician, Pulmonary Medicine, Panama City, PANAMA

The decision we have to take mainly based on risk-benefit ratio for the patient who sits in front of us in not always well served by the cost-efffectiveness decision making process that prevails in public health policies held por populations.

These studies are not done using the average practitioner that takes care for women attending a medical clinic but by doctors whose accuracy detecting a breast nodule have previously been assesed and is significantly above average.

Besides, the advent of new diagnostic tools like tomosynthesis may change these reality sooner than new recommendations will be available. Meanwhile, a huge amount of women may be biassed not using mammograms that could detect cancer on time, when their doctors may not be able to do it.

ADAM CLEMENTS Resident, Hospital Medicine

I'm not completely sure that this study generalizes well to current practice. I've read that the machines and methods used for mammography have changed quite a bit from the 60's, 70's and even the 80's. It would be nice to see newer data with current methods, even if the N value was smaller.

Your Comment

(will not be published)

Filtered HTML

  • Allowed HTML tags: <a> <em> <strong> <cite> <blockquote> <code> <ul> <ol> <li> <dl> <dt> <dd>
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.
Do you have any conflict of interest to disclose?
This question is for testing whether you are a human visitor and to prevent automated spam submissions.

Vertical Tabs

* Required

Reader comments are intended to encourage lively discussion of clinical topics with your peers in the medical community. We ask that you keep your remarks to a reasonable length, and we reserve the right to withhold publication of remarks that do not meet this standard.

PRIVACY: We will not use your email address, submitted for a comment, for any other purpose nor sell, rent, or share your e-mail address with any third parties. Please see our Privacy Policy.