Mammogram Screening: The Dartmouth Bombshell

By Dr. Kathleen T. Ruddy

There is no more heated controversy in all of breast cancer research than the arguments surrounding mammogram screening. Women are confused, their doctors are confused, and health policy advisors are caught in the crossfire of scientific debate, public outcry, and the need to control costs. The most recent volley in the battle over mammogram screening was launched from Dartmouth and it was explosive.

Dr. Gilbert Welch, professor of medicine at Dartmouth Institute for Health Policy and Clinical Practice in Hanover, New Hampshire, was concerned that the widespread utilization of mammogram screening over the past thirty years has led to over-diagnosis of breast cancer.  Welch looked at government-derived statistical data on mammogram screening, the overall incidence of breast cancer, and survival rates over a period of three decades to determine if his concern was valid.  Dr. Arche Bleyer of the Department of Radiation Medicine at the Oregon Health and Science University in Portland was Welch’s co-investigator.

In his study design, Welch assumed that the overall burden of breast cancer in the United States has remained stable, despite the perceived rise in the incidence of breast cancer which he believes is associated more with diagnosing early-stage disease via screening than with a true rise in the number of women with life-threatening illness.

Welch examined data on breast cancer trends between 1976 – 2008.  He chose 1976 as a starting point because prior to 1976 mammogram screening was not widely used.  It was only after First Lady Betty Ford’s diagnosis of breast cancer in 1974 that breast cancer became a national issue.  Funding of breast cancer research rose exponentially thereafter, followed by a national campaign for early diagnosis propelled by the joined forces of the profit-driven healthcare industry and breast cancer foundations racing for a cure.

Welch discovered that there was an absolute increase in the incidence of breast cancer, from 112 per 100,000 women in 1976, to 234 per 100,000 in 2008.  Thus, as a direct result of three decades of widespread mammogram screening, an additional 122 women per 100,000 were diagnosed with breast cancer.  But did they live longer as a result of their screening-derived early diagnosis and treatment?  Welch says, “No.”

According to Welch’s analysis, only eight of the 122 women diagnosed with breast cancer would have eventually develop advanced disease.  He concludes in his paper, published in the New England Journal of Medicine, November 29, 2012.  that “over the past 30 years, 1.3 million women were over-diagnosed with breast cancer.”  Welch further estimates that in 2008, breast cancer was “over-diagnosed in more than 70,000 women.”

It’s fair to say that Welch’s use of the term ‘over-diagnosed’ is, at best, misleading and, at worst, inflammatory. The women in Welch’s study were not over-diagnosed with breast cancer, but were correctly diagnosed with breast cancer.  No one was told she had breast cancer who did not, in fact, have breast cancer.  Indeed, it is quite likely that at least one-third of the women diagnosed with breast cancer between 1976 – 2008  (i.e., 1.3 million women) were over-treated, but not over-diagnosed, with a disease that might well have never threatened their lives.

The problem is that, at this time, neither Welch nor any other scientist in the world can accurately identify the 1.3 million breast patients who were over-treated.  Because we cannot determine breast cancer that is potentially lethal from that which is relatively indolent, doctors must, of necessity, treat all women as if their breast cancer is the one that might threaten her life.

Welch concludes his paper by saying, “screening is having, at best, only a small effect on the rate of death from breast cancer”, and he admonishes the medical community to utilize mammograms more sparingly.

I disagree with Welch’s conclusion, and I am not alone in holding the opinion that there is another, better way of viewing the important data he collected in his study.

While it is certainly true that a portion, perhaps even a large portion, of breast cancer is accurately diagnosed in women whose disease is not life-threatening, until we are able to accurately identify those tumors that are destined to kill the patient from those that are not, we ought not to abandon the only screening method we have for our most common female malignancy.  Rather than pull back on mammogram screening, an intervention that unquestionably has saved thousands of lives, we ought to screen more aggressively but with better tools.

There’s no doubt that mammogram screening leaves much to be desired.  We didn’t need another study to tell us that, though Welch’s paper certainly brings into clear view the need to distinguish life-threatening breast cancer from that which can simply be observed.  What we need is a better discriminator, not abdication of the only screening method we currently have for the most common malignancy in women everywhere in the world.

Of course, the best solution would be to find a completely preventive breast cancer vaccine such as the one developed at the Cleveland Clinic by Professor Vincent Tuohy, published in Nature Medicine, May 2010.  Tuohy’s vaccine was 100% effective in preventing breast cancer in three different animal models and is ready for clinical trials to see if it is safe and effective for use in women.  Such a prophylactic breast cancer vaccine would be an excellent way to put an end to the endless, strident arguments over the value of mammogram screening, finally replacing the confusion with much-needed prevention.