Abstract
Yankaskas and colleagues report on a large study investigating the performance of first mammography examination in women aged under 40 years based on pooled data from six mammography registries across the USA [1].
Excluding cancers of the skin, breast cancer is the most common cancer among women, accounting for more than 20% of cancers diagnosed in women in most Western countries. The incidence is strongly related to age; although one out of every eight to ten women will develop breast cancer during their lifetimes, the risk of being diagnosed with the disease before the age of 40 years is less than 1%. According to the Surveillance Epidemiology and End-Results data, the incidence in the US population in 2007 was five per 100,000 per year for women aged 20–29 years, and 45 per 100,000 per year for those aged 30–39 years [101]. Despite the relatively small absolute risk, breast cancer is the leading cause of death in women aged 34–44 years. As such, breast cancer has a large psychological and societal impact, and the general wish to reduce breast cancer mortality by early detection is understandable.
During the last four decades, evidence from randomized controlled trials has mounted that suggests that routine mammography screening significantly reduces breast cancer mortality in women aged 50–70 years and the general view is that this benefit does outweigh the possible harms [2]. For women aged 40–49 years, the benefits are less clear, as was demonstrated by a randomized trial undertaken by 23 National Health Service breast screening units in the UK [3]. In this trial, 160,000 women aged 39–41 years were randomly assigned to an intervention group of annual mammography, until the age of 48 years or to a control group of usual medical care. No preselection on risk factors took place. At a mean follow-up of 10.7 years, a reduction in breast cancer mortality was observed in the intervention group, which did not reach statistical significance (relative risk: 0.83 [95% CI: 0.66–1.04]).
Studies on the performance of mammography in women younger than 40 years are very scarce and have mainly been confined to retrospective examinations in patients with a proven diagnosis of breast cancer. In the 1980s, the American Cancer Society recommended a baseline mammogram for women aged 35–40 years who were at average risk to provide a comparison image that would be available when regular screening began at age 40 years or older. Although this recommendation was revoked in 1992 because of lack of evidence of a benefit, many young women continued to have such a baseline screening. In 2005, 29% of the women in the USA aged 30–39 years reported previously having a mammogram [4]. This observation was the main reason for Yankaskas and colleagues to address the following three questions:
Who are these young women who continue to undergo screening mammography?
What is the accuracy of mammography to detect breast cancer in young women?
What are the pathological characteristics of the cancers detected by screening?
Results
The study was based on the pooled data from six mammography registries across the USA from the Breast Cancer Surveillance Consortium. Mammography data from each registry were linked to its regional or Surveillance Epidemiology and End-Results cancer registry to identify those patients in whom cancer was diagnosed within one year after mammography and to retrieve information on cancer type (e.g., invasive or ductal carcinoma
Significance
The low prevalence of breast cancer is leading to high recall rates, a high number of false-positive test results and a low cancer detection rate in women aged under 40 years receiving mammography. Considering these unfavorable figures, Yankaskas and colleagues quite rightly question the appropriateness of mammography in young women without symptoms [1]. For these women, even a test with a specificity of 95% would have resulted in a positive predictive value not exceeding 4%.
The strengths of the study are its representative population-based nature, the large sample size and the prospective follow-up through linkages with regional cancer registries to indentify the women with a diagnosis of breast cancer. As pointed out by the authors themselves, the study also had some limitations, such as the lack of information on the complete family pedigree or BRCA1 and BRCA2 status, which prevented the authors from identifying women at very high risk of breast cancer. In addition, 5% of the women who had undergone screening mammography self-reported a lump at screening. The detection rate dropped from 1.7 to 1.2 per 1000 mammograms when excluding these women with a palpable lump. It was not clear to the authors why these women had screening mammography. One could question whether they should have been part of the screening group in the analysis. However, in every screening program there will be women with symptomatic disease who use the screening test to find an explanation for their complaints, despite the explicit advice not to do so. An issue that is also not addressed specifically in the discussion is the high proportion of
Finally, no information was provided on diagnostic procedures in patients referred to the hospital for additional diagnostic testing to confirm the breast cancer diagnosis. For a more balanced trade-off between the benefits and the costs (both financial and psychological) of screening in young women, more information is needed than referral and detection rates [7]. How many women with a positive screening test were required to undergo additional imaging, cytology, or a surgical procedure, or both, and what is the impact of a false-positive test on quality of life and the willingness to participate in future screening rounds?
Future perspective
The study by Yankaskas and colleagues is a clear illustration that screening healthy people at a low risk of disease is bound to be a poor solution to reduce the cancer burden [8]. Owing to their low risk it will be virtually impossible to demonstrate the impact of screening mammography in women aged under 40 years. Risk prediction models might be helpful to identify young women with a substantial risk of developing breast cancer in the next 10 years, who could serve as a target population for future studies on the effectiveness of mammography screening. Breast density has been established as a key factor in such risk models, along with parity and family history. The use of other imaging techniques, such as ultrasound and MRI, should also be considered in future studies. For women at high familial risk of breast cancer (i.e., those with a strong family history or a high probability of a BRCA1, BRCA2 or TP53 mutation) there is already strong evidence that contrast-enhanced MRI is much more sensitive than mammography for cancer detection [9]. However, as always, one should keep in mind that aiming at a higher sensitivity will result in a drop of the specificity, thereby significantly increasing the number of false-positive screening tests.
Executive summary
Studies on the performance of mammography in women aged under 40 years are scarce. Despite this lack of evidence, many women continue to have a first screening mammogram before the age of 40 years.
An analysis of pooled data from six mammography registries across the USA, including 117,737 women aged 18–39 years who had received their first mammography in the period 1995–2005, showed that the age-adjusted positive predictive value of screening mammography was 1.3% (95% CI: 1.1–1.5%). The overall age-adjusted breast cancer detection rate was 1.7 per 1000 mammograms (95% CI: 1.4–1.9).
Considering the high recall rates, the high number of false-positive test results and the low cancer detection rate the appropriateness of mammography for women younger than 40 years without symptoms is questionable.
Risk prediction models might be helpful to identify young women with a substantial risk of developing breast cancer in the next 10 years, who could serve as a target population for future studies on the effectiveness of mammography screening or other imaging techniques.
Footnotes