Mammograms Reduce Breast Cancer Deaths, But Increase Overdiagnosis

The good news is that screening mammography does reduce breast cancer mortality. The review found that women offered screening mammograms are 15 percent less likely to die of breast cancer than women who are not offered mammograms.

The bad news is that women in a screened population are 30 percent more likely to be diagnosed and treated for a cancer that, in the absence of screening, would never have posed a threat to their health.

"It is likely that screening mammography reduces breast cancer mortality, but the other side of the coin is the major harm of overdiagnosis and overtreatment," said lead author Peter Gøtzsche, M.D., director of the Nordic Cochrane Centre in Copenhagen, Denmark.

The review appears in the current issue of The Cochrane Library, a publication of The Cochrane Collaboration, an international organization that evaluates research in all aspects of health care. Systematic reviews draw evidence-based conclusions about medical practice after considering both the content and quality of existing trials on a topic.

Mammograms can detect tumors that are too small to be felt as a lump, theoretically identifying cancer at its earliest and most treatable stage. In the United States, women are recommended to have regular mammograms beginning at age 40. While most European countries also recommend screening mammography, their guidelines advise that women begin screening at age 50.

To determine whether offering mammograms to women with no history or symptoms of breast cancer actually reduces breast cancer mortality, the reviewers pooled together findings from six randomized controlled trials involving half a million women. The researchers excluded a seventh trial from analysis because it was considered too biased to provide reliable data.

Most of the trials enrolled women ages 45 to 64, although one Canadian trial included women ages 40 to 49.

After seven years, women in the screening group were 20 percent less likely to have died of breast cancer compared to women in the control group. The same reduction in breast cancer mortality was seen at 13 years.

According to the reviewers, however, this 20 percent reduction isn't the most accurate estimate, because not all six trials were of equal quality. The four trials that they judged to be of poorest quality yielded the greatest benefit for screening mammography -- a 29 percent reduction in the risk of breast cancer mortality after seven years and a 25 percent reduction after 13 years.

In contrast, in the two trials that they considered to be of highest quality, they found no significant benefit from screening. After seven years, the relative risk of death was slightly higher in the screened population; after 13 years, it was slightly lower. At neither time was the difference statistically significant.

"We believe more in the good trials," said Gøtzsche, "but we have taken the other trials into account as well and reached a compromise of 15 percent. We don't find it likely that it's higher than this."

According to Stephen Taplin, M.D., senior scientist at the National Cancer Institute, evaluating the quality of these trials -- some of which began 30 or 40 years ago -- necessarily involves judgment and a bit of second-guessing.

"Ultimately it's a subjective interpretation," he said. "The most important question is, are there so many fundamental flaws that it makes the conclusions invalid?"

With the exception of the trial that the reviewers excluded from analysis, Taplin doesn't think so.

William Barlow, Ph.D., senior biostatistician at Cancer Research and Biostatistics in Seattle, also considered the flaws identified by the reviewers to be, for the most part, minor issues.

"But even if the risk reduction is 15 percent, that to me is a significant reduction in breast cancer mortality," he said. "So then the real issue becomes the trade-off between a reduction in breast cancer mortality and the adverse events caused by having detected breast cancers that may not have gone anywhere."

The reviewers estimated that for every 2,000 women who are invited to get mammograms for 10 years, one woman's life will be prolonged as a result of detecting and treating a potentially lethal cancer.

However, another 10 healthy women will be transformed into cancer patients and undergo treatment needlessly.

An additional 200 will have the anxiety-inducing experience of a false positive -- being told about a suspicious finding on a mammogram that further testing reveals to be noncancerous.

"Screening mammography is clearly a double-edged sword," said Lisa Schwartz, M.D., co-director of the Veterans Administration Outcomes Group in White River Junction, Vermont. "Regular screening will save some lives but will cause even more women to be harmed through the unnecessary diagnosis and treatment of cancers that would never have affected their health, were it not for screening."

Many of those cancers are of a type called ductal carcinoma in situ, or DCIS. DCIS is too small to be felt as a lump and is almost always detected with mammography. About one in five cancers picked up on a mammogram -- 60,000 cases a year in the United States -- is of this type.

In most cases, these abnormal cells, found in the milk ducts, will never invade the surrounding breast tissue. The problem is that researchers don't know which of these cancers will progress and which ones won't. As a result, all women who are diagnosed with DCIS are advised to have the lesions removed.

Barlow doesn't think the fault lies with mammography.

"If we were better judges of the cancers, then there wouldn't be a problem with overdiagnosis," he said. "I don't blame screening for that. I blame our inability to determine which cancers are life threatening and which ones are not. So the challenge is to do a better job of determining the prognostic potential of that tumor rather than necessarily blaming screening for detecting it."

Better predictive tests may some day be available. For now, said Schwartz, women are faced with a difficult choice -- and, she adds, it is a choice.

"Our approach to breast cancer screening has fostered a climate where women are seen as irresponsible if they do not undergo screening. But screening has important trade-offs," she said. "We need to make sure that women understand this is a real decision because it has real consequences in both directions."