Breast cancer screening

Purpose and design: Three breast cancer screening methods are commonly employed: mammography, breast self examination (BSE), and physical examination by trained personnel (PE). Case-control, retrospective, and prospective studies have examined the efficacy of these screening modalities in reducing breast cancer mortality. However, there are three biases pertinent to many of these studies: lead-time, length, and selection biases. The best way to exclude these biases is to compare screened and unscreened women in a randomized controlled trial with breast cancer mortality as the end point. Eight trials have examined the effect of mammographic screening on breast cancer mortality and two have examined the impact of screening with BSE. In addition, a large trial will soon be initiated in India to assess the impact of screening by PE on breast cancer mortality. This article reviews these trials and discusses the implications of the studies. Results: The overall results of the randomized controlled trials indicate that mammographic screening in women over age 50 can reduce breast cancer mortality by about 25%. However, its efficacy in women between the ages of 40 and 49 is disputed, and another large trial has been initiated in the United Kingdom to resolve this controversy. Preliminary results of two trials indicate that BSE has no impact on breast cancer mortality. However, longer follow-up of these trials is necessary before drawing any conclusions regarding BSE. Conclusions: Mammographic screening in postmenopausal women is an effective means of reducing breast cancer mortality. However, the impact of mammographic screening on breast cancer mortality in premenopausal women is disputed. At least four potentially harmful consequences of mammographic screening merit consideration: lead time effect, radiation exposure, false-positives, and overdiagnosis. Thus, women between the ages of 40 and 49, in particular, should be informed of the potential for benefit and harm prior to mammographic screening.