A turning point for mammography?
Data focus the debate on benefits—and risks—for women in their 40s
From the October 1997 ACP Observer, copyright © 1997 by the American College of Physicians.
By Edward Doyle
When Swedish researchers earlier this year presented data showing that mammography for women in their 40s could reduce mortality rates by as much as 44%, many hoped that one of the thornier issues in women's health—whether to screen women in their 40s for cancer—would finally be resolved.
Instead, the National Institutes of Health (NIH) consensus panel that heard the new data decided against recommending mammography for all women in their 40s. It concluded that because of the procedure's risks like false positives and unnecessary biopsies, the decision of whether to start screening at age 40 should be left to individual women.
Supporters of mammography for younger women—primarily physicians from the radiology and oncology community—were enraged; some compared the decision to a death sentence for women in their 40s. They complained that mammography's risk factors for younger women were exaggerated, and that mammography for 40-somethings is almost as effective as for older women.
This discord immediately reverberated throughout the medical community. The National Cancer Institute (NCI), the group that had asked the NIH to convene the consensus panel, ignored the panel's findings and recommended that women in their 40s undergo mammography every other year. At about the same time, the American Cancer Society released new recommendations calling for women in their 40s to undergo mammography annually.
Even today, the issue remains complicated. While Swedish researchers found that mammography can reduce mortality in younger women by up to 44%, meta analyses including all breast cancer research have found the average mortality reduction to be 18%. (These data have not been published but were considered by the NIH panel in January.) While the numbers from the meta analysis fall short of the Swedish findings, they are statistically significant and help make the case for screening women in their 40s.
These newest data also mark a turning point in the debate, with many experts finally conceding that mammography can actually help save lives of younger women. But controversy over how to balance this benefit with the procedure's downsides continues to cause confusion—and divisiveness—among the medical community. Both supporters and critics of early screening say that data is being manipulated and biases are interfering with science. A rift has developed, one that has left women and their doctors without any firm advice on the question of when to start screening for breast cancer.
The numbers, the risks
To follow the debate, it helps to start with some statistics.
According to Suzanne W. Fletcher, MACP, professor of ambulatory care and prevention at Boston's Harvard Medical School, and breast cancer expert, 16 of every 1,000 women in their 40s will be diagnosed with the disease at the end of 10 years. Of those 16, she said, approximately half will survive without the benefit of mammography or any other screening technology. Of those eight women who can be helped, early mammography can help roughly 20%, or fewer than two.
At the crux of the debate, then, is the question: Should society tell 1,000 women to undergo mammography screening to help save the lives of roughly two?
Skeptics say that mammography's risks, particularly a false positive rate of nearly 30% in younger women, are too high to recommend it to all 40-somethings. Russell Harris, FACP, director of the program on health promotion and disease prevention at the University of North Carolina, said that out of every thousand women screened, 300 will get a good scare from a false positive in exchange for helping extend the lives of two of those women.
"The reports anecdotally are that those women suffer a lot of anxiety and worry until they get the extra study or the biopsy," Dr. Harris said. He also noted that one study found that about a third of those women are still bothered by thoughts of cancer six months later. "You have to stack that up against the number of people we're talking about helping," he said.
But individuals who support mammography for women in their 40s counter that the false positive rate for these women is about the same as that of all women, about 15%, not 30%. And while skeptics say that mammography for women in their 40s misses 25% of cancers—the imaging technique misses about 15% of cancers in all women—supporters say that those are worst-case numbers, not averages.
Both sides agree that older women tend to get more breast cancers and that mammography tends to find more cancers in older women. What they can't agree on, however, is exactly how much more effective mammography is in older women. And that is a source of much of the debate.
For years, many researchers measured the effectiveness of mammography by putting women into two groups: women 40 to 49 years old and women 50 and older. By looking at those data, they concluded that mammography was consistently more effective in older women than in women in their 40s because post-menopausal women generally tend to have less dense breasts, which makes mammography more effective.
Supporters say that when you compare mammography's effectiveness in similar age groups—women in their 40s vs. women in their 50s, for example—it becomes clear that there is no dramatic difference in the procedure's effectiveness on younger women.
Daniel Kopans, MD, director of breast imaging at Boston's Massachusetts General Hospital and one of the more vocal proponents of mammography for younger women, said that comparisons of 40-somethings and all women over 50 are misleading. "They say if there's a benefit, it will only save two women per thousand in their 40s," Dr. Kopans said. "What they don't tell you is that using that same rationale, the benefit would be four women per thousand in their 50s and six women per thousand in their 60s. Two is made to look like a trivial benefit."
Everyone agrees that 50 is an arbitrary dividing line, but skeptics say that because older women tend to get more breast cancers, screening will naturally be much more effective. Dr. Fletcher said that clinical studies show that mammography can reduce mortality rates for women in their 50s by about a third, compared to an 18% reduction for younger women. While the difference between a third and 18% may not seem that big, she said that mammography winds up helping to extend the lives of nearly four times as many women in their 50s as women in their 40s. According to Dr. Fletcher, that's because you're taking one third of a larger group of women with cancer.
Beyond the numbers
While arguments about the numbers can be confusing, what makes the debate over mammography even more difficult to resolve is the very nature of screening tests. As Dr. Harris explained, screening procedures of all kinds are often described as producing a few big winners—individuals whose lives are extended—in exchange for many more small losers—people who are somehow harmed by the measure. In other words, the problem with most screening tests—whether mammography or testing for prostate cancer—is that for every benefit, there is a corresponding downside.
David Atkins, ACP Member, a senior health policy analyst at the Agency for Health Care Policy Research and science advisor to the U.S. Preventive Services Task Force, said that one view is to use the technology until it's proven not to work. But he added that the scientific community has some problems with that approach.
"If you go out and promote screening before you're sure whether it works," Dr. Atkins said, "you're often left in the position of never being able to conduct a trial. It's a tension, because the proponents are saying that you're going to wait 10 years for trial results, during which time we could be saving lives with screening."
That tension is highlighted by an interesting result of a number of clinical studies. Several studies have found that when mammography for women in their 40s reduces mortality, the mortality benefit does not appear until seven to 10 years after screening has begun.
This could be because women who began screening at age 45, for example, aren't found to have breast cancer until they're in their 50s. Skeptics frequently cite this explanation as proof that mammography is more effective in older women.
But there is another explanation, one that has to do with how quickly breast cancer tends to grow in younger women. According to Dr. Fletcher, some breast cancers may grow so quickly in younger women that mammography can't detect them in time to help the woman. By the time these cancers are diagnosed, they have become so big that they kill women quickly, and women in both the study's control group and screening group die at about the same time. In such a scenario, mammography can only help control those smaller cancers that take much longer to kill.
The issue brings to light some basic differences in thought between the two sides. Skeptics wonder if the delayed mortality benefit means that screening could have been postponed until later. Supporters, on the other hand, say rapidly growing cancer in younger women bolsters arguments to screen younger women every year.
Robert Smith, PhD, director for cancer detection science with the American Cancer Society (ACS), said that studies show the sojourn time of breast cancer—the period during which physicians can detect cancer before symptoms develop—is about 1.7 years for women in their 40s, compared to about three years for women over 50. Based on this information, and despite the lack of data from randomized control trials comparing the effectiveness of one-year screening intervals to two-year screening intervals, the ACS earlier this year changed its guidelines and now recommends that women in their 40s be screened each year.
Dr. Smith explained that the ACS recommended the shorter screening interval because of the probability that it would help younger women: "If we have an imperfect understanding of breast cancer and we screen women at an interval too wide because influences of cost effectiveness or a lack of understanding about the disease, that's not a shortcoming of mammography, it's a shortcoming of the clinical protocol."
Until recently, that kind of thinking guided national policy on mammography for younger women. In 1989, a dozen groups that included the ACS and the NCI signed a consensus statement recommending mammography for women in their 40s. At that time, Dr. Smith said, there was no evidence from randomized clinical trials demonstrating a statistically significant benefit, but the experts agreed that finding breast cancer early almost always meant that treatment would be far more successful, so the recommendation was made.
That all changed in 1993. An NCI consensus panel found that clinical studies like those in Sweden had shown that mammography for women in their 40s was effective in extending young women's lives, but the evidence was not statistically significant. Rather than continue to recommend the procedure for all women in their 40s, the NCI decided that the evidence was inconclusive and withdrew its support for mammography for women in their 40s.
Subsequently, other groups including ACP and the U.S. Preventive Services Task Force issued similar recommendations. They cited the lack of randomized control trials showing a statistically significant mortality benefit from screening women in their 40s.
What changed? Those who are skeptical of mammography for younger women point out that organizations like ACP and the U.S. Preventive Services Task Force—groups known for their somewhat conservative stance on screening issues—were never part of the 1989 consensus statement. Instead, it was made up primarily of advocacy groups like the ACS and other professional organizations representing the radiology and oncology community.
But others point out that the thinking about screening has been changing. "We used to think that if we got enough patients together and took a look at something, that was good enough," Dr. Fletcher said. "Then we went to randomized control trials. Then we began to realize that a single effect like mortality reduction wasn't the only thing we should be concerned about, that we've got to see how much and what kind of adverse effects this particular procedure causes. Then we began to realize that it's not even as simple as those two, effectiveness vs. adverse effects, that we also have to look at the cost, because society has made it very clear that it doesn't want to keep raising expenditures for health care."
That view helps explain the 1997 NIH panel's decision that despite statistically significant evidence that mammography can help save young women's lives, it should not be recommended for all women in their 40s. The newest evidence may show a mortality benefit, skeptics say, but it isn't strong enough to recommend mammography for all women in their 40s. "One woman might say I don't want to risk a false positive, I don't want that hassle with this kind of uncertain benefit," Dr. Fletcher said. "Another woman will say I don't care what the false positives are, I'm so scared of breast cancer that I want screening."
Groups like ACP and the U.S. Preventive Services Task Force have not reviewed their mammography guidelines since the newest data were released, but those involved in creating those guidelines have said that the they don't feel the new evidence—the Swedish data or the new meta analysis—would change those recommendations.
Supporters of early mammography, however, remain frustrated by the 1997 NIH panel's decision. They complain that the opposite camp is arbitrarily changing the criteria because they refuse to admit that mammography should be recommended for women in their 40s. They complain that now that there is scientific evidence showing a mortality benefit from mammography for younger women, skeptics have shifted attention to risks like false positives. For their part, skeptics counter that the scientific thought about screening is evolving.
Much of the battle is about disseminating information. While even the most ardent of supporters agree that women should be able to make up their own minds about mammography, as the 1997 NIH panel recommended, they fear U.S. women aren't getting all the facts. For proof, they point to the NIH panel report, which they say is riddled with inaccuracies and errors of omission.
"The report didn't say that we heard new data from Sweden that we disagree with," Dr. Kopans said. "It said we can find no evidence of benefit. At least they could have said that we reviewed the new data from Sweden and we don't believe it is strong evidence. But American women and their physicians didn't hear the data."
But the most stinging criticism comes from supporters who are troubled by what they see as a lack of compassion on the part of the skeptics. "If you look at the opposition," said Dr. Smith from the ACS, "a majority of them are not engaging in much hand-wringing about what we can do to screen women differently so that they could have greater protection from this disease. They don't ask what did we do wrong in these trials, what can we do differently to get better results. You don't see that."
Skeptics say that it's not personal, that medicine in this country is grappling with difficult issues that traditional solutions can't adequately address.
In fact, some predict that the debate over mammography portends the types of problems medicine can expect to see more of. "This is going to be a very interesting problem for our society in the future," said Dr. Harris. "The big bangs in medical science—insulin for diabetes and penicillin for infections—come along rarely, where you get a sudden huge improvement and decreased risk. What we're dealing with a lot in science is a lot of small improvements, and what we're going to find is that with all the small improvements each carries its own downside. Trying to balance the improvement with an uncertain downside is going to be tough."
Politics add a powerful voice to the screening debate
The debate over when to begin mammography screening has caught the attention of more than just physicians and their patients. In the past few years, it has come under the scrutiny of politicians.
While the debate has been ongoing for much longer, many point to 1993 as a turning point for the debate. That was the year that a National Cancer Institute (NCI) panel declared that the evidence regarding mammography for women in their 40s was inconclusive. Despite vocal protests from the medical community, the NCI withdrew its recommendation that all women in their 40s should be screened.
Many claim that the NCI was stacked against screening from the start. For proof, supporters of screening for younger women point to the panel's workshop chair and author: Suzanne W. Fletcher, MACP, then Editor of Annals of Internal Medicine, who had published an article arguing against screening for women in their 40s just before the panel convened. They also point out that while the NCI's advisory board said the agency should continue to recommend mammography for 40-somethings, NCI leadership changed its position anyway. It was the only time that the NCI had ignored a recommendation of its advisory board.
In a scathing report issued in 1994, Congress said that the panel had "tarnished the scientific integrity" of the NCI by including panelists who had clear biases. The report also chastised the panel for excluding evidence from the panel and for wording its recommendations in a way that was "a setback for efforts to encourage early screening and early detection."
Supporters were enraged at what they viewed as the politicization of medical decision-making. That changed in 1997, when an NIH consensus panel concluded that there wasn't strong enough evidence to recommend that all women in their 40s get a mammogram. Before the ink had dried on the recommendations, politicians were tripping over themselves to reverse the decision.
At a press conference following the panel, NCI director Richard Klausner, MD, told reporters that he did not agree with the panel's recommendations. Then the Senate passed a nonbinding resolution supporting mammography for younger women. Soon thereafter, NCI's advisory board—the same group whose recommendations had been ignored in 1993—advised the NCI to recommend mammography to all women.
With the political muscle now on their side, supporters have become quiet about the corrupting influence of politics on the debate. Now it's the skeptics who are complaining.
Internist Archives Quick Links
Earn CME Credits with ACP
ACP offers internists many CME options for the completion of AMA PRA Category 1 CME Credits™. Attend live meetings, work online, or watch course recordings on your own schedule.
Explore our many CME credit options.
The Next-Generation Clinical Information Resource
DynaMed Plus is a collaboration between ACP and EBSCO Health. ACP members enjoy free access to this comprehensive tool that optimizes time to answer for busy clinicians, like you. Get started now!