U.S. Preventive Services Task Force Recommendations on screening for breast cancer
A Closer Look
Does Mammography Sometimes Detect Too Much Breast Cancer?
The argument for screening women for breast cancer with mammography sounds simple: Mammography can detect cancers before they begin to cause symptoms. Clinical trials have shown that fewer women die of breast cancer when they are screened with mammography than when they are not.
But what if some of the very early stage breast cancers found by mammography are not destined to grow and become potentially lethal? Although the idea of a harmless cancer may seem counterintuitive, such cancers do exist.
Some cancers never grow, or they grow so slowly that they never become clinically detectable. Autopsy studies have found breast tumors in women between age 40 and 70 who died of unrelated causes. And a 2008 study made the provocative suggestion that some very early stage invasive breast tumors may regress spontaneously.
The identification of tumors that would never have become a clinical problem in the patient’s lifetime is known as overdiagnosis.
“With overdiagnosis, we often end up ‘curing’ cancers that didn’t need to be cured in the first place,” said Dr. Barry Kramer, director of the NIH Office of Disease Prevention and an expert on cancer screening.
Because doctors can’t distinguish breast cancers that will progress from those that will not, these “overdiagnosed” tumors will in most cases lead to surgery, possibly with radiation, chemotherapy, or hormone therapy. Some people, therefore, are receiving unnecessary treatment that is of no benefit to them, can be toxic, and often comes with significant side effects.
Studies strongly suggest that overdiagnosis occurs with screening for several cancers, including prostate and lung cancer, as well as breast cancer, said Dr. Kramer.
The precise extent of breast cancer overdiagnosis is unclear. In a paper published in July in the British Medical Journal, Danish researchers estimated that one out of every three tumors detected by screening mammography was overdiagnosed. Their estimate was based on an analysis of breast cancer rates before and after the implementation of government-run mammography screening programs in Australia, Canada, Norway, Sweden, and the United Kingdom.
A 2006 study that was based on 25 years of follow-up with women who were screened with mammography in Sweden concluded that one in six breast cancers was overdiagnosed. Other estimates of overdiagnosis have ranged from 5 percent to 32 percent.
“Pulling Cases Out of the Future”
Several factors can lead to varying estimates of breast cancer overdiagnosis. One is whether researchers count both early stage invasive tumors and ductal carcinoma in situ (DCIS) detected by mammography, or only invasive tumors. “If you count only invasive cancers, estimates will be lower. If you also count DCIS, they will generally be higher,” said Dr. Kramer.
Another factor is what assumptions researchers make about time trends in cancer rates. “In the absence of overdiagnosis, [mammography] should be pulling cases out of the future into the present,” said Dr. Kramer. In other words, every breast cancer detected early by mammography should equal one less cancer detected later on.
This is what the Danish researchers assumed would happen, explained Dr. Kathleen Cronin of NCI’s Statistical Research and Applications Branch. “They assumed that the number of new breast cancer cases diagnosed each year after mammography screening began would have followed the same trend seen before mammography was introduced,” she said.
But in fact, from the 1980s through 2001, a span of time roughly corresponding to the period during which mammography screening programs became established, the number of newly diagnosed breast cancer cases increased year by year. And, as noted in this week’s Journal of the American Medical Association, the incidence of aggressive or later-stage disease has not decreased at a rate that would be explained by an increase in earlier diagnoses.
Changes in breast cancer risk factors––more women delaying childbirth or using postmenopausal hormone therapy, for example––may have contributed to the increase in breast cancer cases, said Dr. Kramer. Overdiagnosis may have, too.
The bottom line? Many other researchers think the Danish investigators may have overestimated the level of breast cancer overdiagnosis. Most agree, however, that mammography does diagnose some breast cancers that would never have needed to be treated.
Recognizing that current estimates of the extent of the problem are unsatisfactory, researchers are striving to develop more accurate ways of estimating just how much overdiagnosis of breast cancer results from mammography in the hopes of sparing some women from unnecessary and often toxic treatments. Dr. Cronin is scientific coordinator for an NCI-supported consortium of researchers who are working to better understand how mammography, among other factors, influences breast cancer rates.
Overdiagnosis can be identified at a population level by comparing cancer rates in groups of people over time. Relevant data are provided by the Breast Cancer Surveillance Consortium (BCSC) and NCI’s Medicare SEER database. Longitudinal data are a critical aspect of these efforts. Over its 15-year existence, the BCSC has collected information on 2,017,869 women, including 86,700 with cancer.
Researchers agree that one solution to the problem of breast cancer overdiagnosis may ultimately emerge from advances in molecular medicine that will enable doctors to distinguish mammography-detected tumors likely to progress from those unlikely to do so.
“If we can identify tumors that are not destined to become a problem, we can avoid treating patients unnecessarily,” said Dr. Sudhir Srivastava, chief of NCI’s Cancer Biomarkers Research Group. “And differentiating progressive and nonprogressive tumors on the basis of molecular features is currently an active area of research within the NCI-supported Early Detection Research Network,” he explained.
In the meantime, for women aged 40 and over who may be wondering if they should continue having regular mammograms, Dr. Stephen Taplin, a breast cancer screening specialist with NCI’s Applied Research Program, offered this advice:
“Today, regardless of age, women are less likely to die of breast cancer than they were 20 years ago because of improvements in treatment and wider use of screening. There is no simple way for a woman to measure the benefit of mammography to her as an individual, but research is under way to improve our ability to estimate a woman’s risk. What we do know is that, on average, for women aged 50 to 75, mammography reduces the likelihood of dying from breast cancer by about 20 percent. For women aged 40 to 49, the reduction in mortality is smaller, although still real, and the balance of benefits versus harms may be more uncertain.
“The balance of benefits and harms is something that is best done by each woman with her health care provider,” Dr. Taplin continued. “Think about what is important to you. Think about how you would feel if you had an abnormal mammogram. Then make the decision that is right for you.”