Meta-analyses of RCTs in women 40 to 69 years of age have reported a collective mortality reduction ranging from 20% to 40%.20-22 The data in women 40 to 49 years of age are far more limited and essentially consist of subgroup analyses of larger trials and observational results. The American Cancer Society (ACS) commissioned a systematic review of the breast cancer screening literature to address questions related to the frequency of screening mammography3; synthesis of the literature revealed that screening mammography in women 40 to 69 years of age was associated with a reduction in breast cancer deaths across a range of study designs, with inferential evidence supporting breast cancer screening for healthy women 70 years of age and older.

Hellquist et al published landmark data on the effectiveness of national service screening with mammography in more than 600,000 Swedish women 40 to 49 years of age. After an average follow-up of 16 years, a 26% mortality reduction was observed in women invited to undergo screening, and a 29% decrease in women who ultimately were screened.23 Similarly, the UK Age Trial, involving more than 160,000 women 39 to 41 years of age and with 10 years of follow-up, ascertained a 24% reduction in mortality in women who were screened, but these results did not reach statistical significance.24 Alternatively, the Canadian National Breast Screening Study (CNBSS) documented that in its group of nearly 50,000 women 40 to 49 years of age, no reduction in breast cancer mortality was associated with screening mammography.25 However, the CNBSS data have been largely discredited because of flawed (ie, nonblinded) randomization, as well as inferior mammographic technique and interpretation. Problems with randomization were also evidenced by the fact that there were significantly more women with node-positive advanced cancers in the mammography (experimental) group than in the control group.

Continue Reading

Evolving technology. In addition to study design limitations, the application of outmoded screening data to current guidelines is further confounded by the constant evolution of mammographic technology. Most of the randomized studies used in developing guidelines were conducted in the 1970s and 1980s. Thereafter, digital mammography, computer-aided detection, and more recently digital breast tomosynthesis have been incorporated into routine practice, vastly improving performance compared with the mammography of decades past.26 This point is fundamentally important when the value of current screening mammography is being assessed but is rarely accounted for during the development of breast cancer screening guidelines.

Nonmortality benefits. When the utility of mammography is considered, substantial emphasis is placed on decreasing the mortality rates of breast cancer. Conversely, scant focus is placed on the value of the ancillary benefits of screening—namely, less extensive (ie, breast-conserving) surgery, decreased use of radiation therapy, and reduced use of chemotherapy as a result of detection at an earlier stage. 


Mammography has proved to mitigate a woman’s risk for dying of breast cancer.2,3 However, several potential risks and limitations attributed to screening mammography are worth considering, including financial cost, overdiagnosis, radiation exposure, and quality-of-life issues.

Financial cost. Financial cost is an essential issue in mammography. The aggregate cost in 2010 for population screening ranged from $2.6 billion to $10.1 billion, depending on screening frequency, age at screening, and participation constraints.27 Ultimately, these figures should be viewed with caution when cost per quality-adjusted life-year is evaluated because they are based on fluctuating estimates of mortality reduction that are widely variable.28

Overdiagnosis. Overdiagnosis of breast cancer is a consequence of screening and has recently become a very controversial topic. In the context of breast cancer, overdiagnosis refers to the potential for screening mammography to detect a cancer that would not otherwise have been identified or considered life-threatening, especially if the patient had not undergone mammography.

For evaluating breast cancer overdiagnosis, the literature is inconclusive, primarily because overdiagnosis cannot be measured directly. Clearly, it is both unethical and untenable to conduct a study of breast cancer mortality by randomizing women with a diagnosis of breast cancer to a nontreatment arm. Instead, rates of overdiagnosis are estimated by comparing breast cancer incidence rates in screened and unscreened populations, either contemporaneously or at different endpoints, while an attempt is made to account statistically for lead time bias and differences in breast cancer risk.29