Study: Screening Mammography Likely Leads to Overdiagnosis

Authors Say Ability to Detect Ever-smaller Tumors Carries Its Own Risks

October 19, 2016 04:15 pm Chris Crawford
[Young woman getting a mammogram]

A study published Oct. 13 in the New England Journal of Medicine (NEJM)( looked at how mammography has affected reported breast cancer incidence, tumor size at diagnosis, related mortality and the potential for overdiagnosis.

Researchers found that the proportion of tumors detected that were classified as large fell after the introduction of screening mammography, as would be expected, whereas the relative proportion of those classified as small grew -- a shift in size distribution the authors said was mostly due to "substantial increases in the detection of small tumors."

Further analysis of the relative magnitude of that shift, however, and associated case fatality rates suggested that women were more likely to have breast cancer that was overdiagnosed than to have earlier detection of a tumor that would grow to be large. Additionally, researchers said the overall decrease in breast cancer mortality over the study period more likely reflects improvements in cancer treatment than the benefits of earlier detection of tumors.

Story highlights
  • A New England Journal of Medicine study looked at how mammography has affected reported breast cancer incidence, tumor size at diagnosis, related mortality and the potential for overdiagnosis.
  • Researchers found that the proportion of tumors detected that were classified as large fell after the introduction of screening mammography, as would be expected, whereas the relative proportion of those classified as small grew. Further analysis suggested that breast cancer was more likely to be overdiagnosed and an observed overall decrease in breast cancer mortality probably reflects improvements in cancer treatment rather than the benefits of earlier detection of tumors.

Study Methods, Results

Researchers examined data gathered by the National Cancer Institute's Surveillance, Epidemiology and End Results (SEER) Program from 1975-2012 to establish tumor size distribution and size-specific incidence of breast cancer among women age 40 and older. The group then established a baseline period before screening mammography became widespread (1975-1979), as well as a more recent period (2000-2002) that would allow for 10 years of followup data analysis.

Tumor size data from SEER was more accurately tracked in the later years, the researchers noted, with only 5 percent of tumors being recorded as of unknown size from 2008-2012 compared to 33 percent from 1975-1979. The missing data were accounted for using inverse-probability weighting, said the authors, because "If we had directly calculated size-specific incidence by excluding tumors of unknown size, this decreasing frequency of missing data on size would have produced a spuriously low baseline incidence followed by a spuriously large increase, which would have led us to overestimate overdiagnosis and underestimate the contribution of screening to lowering mortality." Overall, they found that from 1975 to 2012, the proportion of breast tumors classified as small increased from 36 percent to 68 percent, while the proportion of large tumors decreased from 64 percent to 32 percent.

Size-specific 10-year risk of death( from breast cancer also was calculated using the baseline period and the followup period. For this analysis, said the authors, the reduced incidence of large tumors was attributed to screening, and this reduction was assumed to translate directly to a reduction in mortality. Therefore, the reduction in mortality from screening mammography was about 12 deaths per 100,000 women, which improved to about 8 deaths per 100,000 with the advent of more recent therapies.

However, the estimated reduction in mortality from treatment alone (i.e., if screening mammography had not been performed) was about 17 deaths per 100,000 women. Thus, said the researchers, improved treatment -- not mammographic screening -- accounted for at least two-thirds of the reduction in breast cancer mortality.

Family Physician's Take

AAFP Commission on Health of the Public and Science member Kenneth Lin, M.D., M.P.H., of Washington, D.C., who also is an associate professor of family medicine at the Georgetown University School of Medicine, told AAFP News that the value of screening mammography has historically been based on the principle that smaller breast tumors are more likely to be treatable than larger ones. However, more recent evidence suggests it isn't always this simple.

"Unfortunately, characteristics of breast cancer other than tumor size, such as the tumor's responsiveness to estrogen, also influence how well individual cancers will respond to treatment," he explained.

Lin said it's important to note that the NEJM study suggested breast cancer screening in the United States has detected many small tumors that would not have progressed to large tumors or caused health problems.

"These tumors are said to be 'overdiagnosed' because they most likely led to unnecessary treatment and side effects of those treatments," he said.

Lin, who also is associate deputy editor for American Family Physician (AFP) online, wrote an editorial published in the April 15 issue of AFP that said although screening mammography does reduce breast cancer-specific mortality, women tend to overestimate its benefit.

"The bad news is the benefit of screening mammography has often been exaggerated while minimizing the harms," Lin said. His editorial noted that for every 1,000 women who start a lifetime program of screening mammography at age 40 instead of age 50, one woman will live 30 years longer, but nearly 600 women will have a false-positive test result, about 70 will have a biopsy that ends up being normal, and two will be treated for a cancer that never would have caused symptoms or problems.

Conflicting Recommendations

The U.S. Preventive Services Task Force (USPSTF) released its final breast cancer screening recommendations( (which the AAFP recommendations mirrored) back in January that say to screen women ages 50-74 using mammography every two years but proceed on a case-by-case basis for women ages 40-49. The group also said evidence was insufficient to recommend screening patients age 75 and older.

However, other organizations such as the American College of Obstetricians and Gynecologists( and the American College of Radiology( recommend commencing screening at age 40, and the American Cancer Society (ACS)( recommends that screening begin at age 45, with a transition to screening every two years at age 55. These varied recommendations can be confusing to patients.

To muddle things further, the Consolidated Appropriations Act of 2016( passed by Congress last December says insurance coverage of screening under the Patient Protection and Affordable Care Act, with no cost-sharing for women, must begin at age 40.

To help patients in their 40s cut through this noise, Lin said it's important for family physicians to initiate a conversation letting them know they have a decision to make. "Are you ready to talk about when you want to start getting mammograms?" he suggested as an opening line.

"I mention that guidelines disagree about the starting age because the decision is a 'close call' and requires their input on how concerned they are about breast cancer and false-positives," Lin said. "I then present the data on each of the outcomes."

Additionally, in his editorial, Lin suggested that if a physician's practice uses a patient portal or contacts patients by email before visits, women could be referred to a site that helps them make an informed decision before their visit, such as Breast Screening Decisions.( "These practical strategies can help family physicians and women in their 40s make more informed decisions about breast cancer screening," he said.

According to Lin, the silver lining in the NEJM study is that it showed the effectiveness of treatment for early breast cancer has improved, "which is a good thing," he noted.

"What may be less intuitive is that the better treatment becomes, the less valuable screening is, since smaller breast tumors don't need to be detected," Lin added. "That's one reason the USPSTF recommends that women be screened every two years after age 50 rather than annually (and the ACS partially concurs) -- it reduces the number of false-positives while still preserving most of the benefits of screening."

Related AAFP News Coverage
USPSTF, AAFP Issue Final Breast Cancer Screening Recommendations
Mammography for Women in Their 40s Remains Individual Decision


Help Patients Confused by ACS' Breast Cancer Screening Guidance
USPSTF, AAFP Recommend Mammography for Older Patient Group


USPSTF Still Recommends Mammography for Women 50-74
Screening for Women in Their 40s Remains Based on Risk Factors


More From AAFP
American Family Physician: Putting Prevention Into Practice -- an Evidence-based Approach: Screening for Breast Cancer