Mammogram Controversy ‘Artificial,’ Study Says
Article date: December 11, 2013
Standardizing data shows fewer discrepancies in benefits
By Stacy Simon
Many experts are closer to agreement over the benefit of mammograms than they realize, according to research presented today at the San Antonio Breast Cancer Symposium. American Cancer Society Senior Director of Cancer Screening Robert Smith, PhD, and his colleagues say that standardizing mammography data shows regular mammograms do save lives from breast cancer.
“The controversy over the effect of mammographic screening on breast cancer mortality is largely artificial,” they write in the report, which appears in the journal Breast Cancer Management.
The issue of how much good mammograms really do is a longstanding debate within the scientific community about breast cancer screening.
Organizations including the American Cancer Society differ in the way they estimate the benefits and harms of mammograms for breast cancer screening, which leads to different screening recommendations. Recommendations are based on estimates of how many women must be screened to prevent 1 death from breast cancer and range from 90 to 2,000. Smith and his colleagues reviewed the data to figure out why the estimates differ so greatly.
Guidelines for screening mammograms
The American Cancer Society recommends annual mammograms for women beginning at age 40 and continuing as long as they’re in good health. Other organizations recommend starting later (at age 50) and getting mammograms every 1-2 years. Some also specify an age to stop getting mammograms.
The differences are based on interpretations of the benefits and harms of screening. Finding breast cancer early can save a woman’s life. But finding something that has to be investigated – and turns out not to be fatal – can unnecessarily expose a woman to the risks of tests and treatments.
The bottom line for many organizations is how many women must be screened in order to save 1 life from breast cancer. The fewer the women who must be screened to save a life, the lower the risks of screening. There is, however, no magic number. Recommendations are made by experts who review and interpret the evidence.
Crunching the numbers
Smith and his colleagues examined 4 leading reviews of randomized trials of breast cancer screening: the UK Independent Review, the Nordic Cochrane Institute Review, US Preventive Services Task Force (USPSTF) Review, and the EUROSCREEN Review. They found 2 main areas to account for much of the discrepancies.
The first discrepancy lies in the difference between women invited to be screened and women who actually are screened. Some reviewers base their results on the numbers of women asked to make an appointment and get a mammogram, rather than the numbers of women who follow through and get their mammograms. The number of women needed to be invited is always going to be larger, making it seem like more women are needed to be screened to prevent 1 death from breast cancer. In general, only about 77% of the women invited for screening actually come in to get the test.
The second discrepancy lies in differences of follow-up time among reviewers. It can take many years to see the full benefit of screening. Some analyses only looked at the effect after 10 years of follow up, but about twice as many lives are saved if you look after 20 years instead.
Another effect is the age range included in the study. The overall effect of screening depends on how common breast cancer is in the people being screened. Screening women 50 and over, in whom breast cancer is more common, saves more lives overall than screening women under 50. When women 40-49 are included in screening studies, the studies will find higher numbers of women need to be screened to save a life than the studies that are limited to those 50 and older.
For example, the Nordic Cochrane Review employed a screening and follow-up period that encompassed the same 10 years, and calculated that 2,000 women ages 40 – 74 years needed to be invited to screen to prevent 1 death from breast cancer. Smith’s group recalculated the estimate based on women who were actually screened, the average benefit seen in the studies they looked at, and a follow-up period of 20 years. They found that only 600 women in that age group would need to be screened to prevent 1 death. The number drops to 300 when only women age 50-69 are considered.
“It is very important to appreciate that the benefits of screening are a long term investment,” said Smith. “We have long term data to show the benefit steadily accrues.”
The risk of overdiagnosis
In determining guidelines for mammograms, experts must weigh the benefits and risks of screening. One of these risks is overdiagnosis, which occurs when screening finds cancer that if it had gone undetected, would never have caused the woman any problems during her lifetime. Because it’s found, however, the woman must be tested and possibly treated, exposing the woman to potential side effects as well as worry and stress.
Estimates of overdiagnosis differ among reviewers. However, not all studies take into account several factors that complicate the calculations. One of these factors is the trend that more women may be getting breast cancer over time. Another is that some women are being diagnosed with breast cancer earlier due to screening. These cancers would have been found eventually over time, often many years later.
According to Smith’s group, when these factors are taken into consideration, overdiagnosis is in the relatively low rate of 10% or less.
Despite differences in interpretations of screening trials, the evidence demonstrates that regular mammogram screening finds breast cancer earlier when it’s easier to treat, and reduces breast cancer deaths. Read the American Cancer Society’s complete recommendations for early detection of breast cancer.
Citation: Real and artificial controversies in breast cancer screening. Published in the November 2013 issue of Breast Cancer Management. First author Stephen W. Duffy, MSc, Queen Mary University of London.
Reviewed by: Members of the ACS Medical Content Staff
Thank you for your feedback.