ABSTRACT: Breast cancer is a leading cause of cancer and death from cancer among women in the developed and developing world. Detecting and treating breast cancer earlier in its natural history improve prognosis and result in a reduction in breast cancer mortality. There have been eight population-based randomized controlled trials (RCTs) of mammography screening, which individually and collectively provide strong support for the efficacy of breast cancer screening. The evaluation of modern service screening also has shown that modern breast cancer screening is contributing to reductions in breast cancer mortality at a rate as good as or better than that observed in the RCTs. In the last decade, different interpretations of the evidence from the RCTs and observational studies have resulted in different screening guidelines and contentious academic debates over the balance of benefits and potential harms from breast cancer screening. In this paper, the historic and recent evidence supporting the value of breast cancer screening will be described, along with the underpinnings of the current debate over the relative and absolute benefit of regular mammography screening.
Introduction
In the world, breast cancer is the most common cancer diagnosed in women and the most common cause of death from cancer. Globally, an estimated 1.4 million new cases of breast cancer were diagnosed in 2008.[1] In the United States, breast cancer is the most common cancer diagnosed among women, and the second leading cause of death from cancer.[2]
Breast cancer is a progressive disease that becomes systemic as the size of the tumor increases. Tumor size, nodal involvement, and histologic grade all are linked to prognosis, and tumor size is strongly associated with nodal involvement and advanced histologic grade, each of which portends worse survival.[3] When breast cancer is diagnosed while still localized to the breast, 5-year relative survival is 98.6%, compared with 83.8% for regional disease, and 23.3% for distant disease.[4] The effects of these prognostic indicators also are evident in longer-term follow-up, that is, of 20 years or more.[3]
In the early 1970s, experimental evidence from a prospective randomized controlled trial (RCT) demonstrated conclusively that screening with a combination of clinical breast examination (CBE) and mammography using general purpose x-ray equipment could reduce the risk of dying from breast cancer.[5] Later, the Swedish Two-County Trial, utilizing dedicated mammography units, demonstrated that mammography alone resulted in a significant reduction in breast cancer deaths.[6] Since then, eight additional RCTs have been conducted, which together provide firm evidence of the efficacy of screening in reducing breast cancer mortality.[7,8] Moreover, recent evidence from a 29-year follow-up of the Swedish Two-County Trial demonstrates the substantial long-term benefit of invitation to mammography screening in relative and absolute terms.[9]
Modern Breast Cancer Screening
Physical examinations of the breast
Before the introduction of mammography, breast cancer was detected principally by women themselves, either by finding signs of a tumor incidentally or during breast self-examination (BSE), or by a health professional at a CBE. BSE is a formal method of structured self-examination of each breast (usually performed monthly), and is distinct from self-detection of symptoms incidentally during daily activities. The examination protocol for CBE is similar to that for BSE, but CBE is performed by a health professional. The potential contribution of self-detection and CBE to clinical outcomes depends on whether or not mammography screening programs are available. In settings in which women have access to regular mammography screening, self-detection and CBE continue to play important roles in detecting interval cancers; mammography will not detect all breast cancers, because of the heterogeneous nature of the disease and the large variations that exist in breast density. In the Two-County Trial of breast cancer screening, women in the group invited to screening who were diagnosed with an interval cancer had better long-term survival than the control group, which was likely the result of counseling that patients received after the mammography exam, to be alert to breast changes.[3] Among those who are not within the age range of women offered mammography screening, incidental self-detection, BSE, and CBE are the only methods that can lead to an earlier detection of a palpable mass.
Mammography
Examples of mammographic images from 1940 to 2006Mammography is a low-dose x-ray examination of the breasts performed with dedicated imaging equipment to detect abnormalities that may be breast cancer. The image receptor in modern mammography equipment is either screen-film or digital. Full-field digital mammography (FFDM) has largely replaced screen-film units over the past decade, and according to the US Food and Drug Administration (FDA), 88% of all accredited mammography units in the US are FFDM units.[10] Dedicated mammography equipment is specifically designed to produce high-quality images of the breast at a minimum x-ray dose (approximately 3 to 4 mGy per view).[11] In the US, image quality and interpretive skills have also been improved through early efforts by the American College of Radiology’s Mammography Accreditation Program (ACRMAP), and subsequently as a result of the passage of the Mammography Quality Standards Act (MQSA) of 1992, which requires a facility to meet a broad range of technical and personnel standards in order to be certified by the US Food & Drug Administration (FDA).[12] The incremental improvement in image quality is evident in examples of mammograms derived from different mammography imaging technologies from the 1940s to today (Figure 1).
Performance Measures for 1,960,150 Screening Mammography Examinationsa Performed in the US From 2002 to 2006 by Age, Based on Breast Cancer Surveillance Consortium Data as of 2009
A mammography screening examination is performed by specially trained radiology technologists, and involves two views of each breast: a craniocaudal (CC) view and a mediolateral oblique (MLO) view. The recall rate generally is higher at the first exam compared with subsequent exams, but overall it typically ranges from 7% to 10% and is slightly higher in younger women compared with older women.[13] In most instances, the use of a multimodality approach (detailed mammography examination, hand-held ultrasound exam, and magnetic resonance imaging [MRI] in selected cases, combined with the occasional use of interventional methods such as fine needle–aspiration biopsy or larger bore–needle biopsy) provides a preoperative microscopic diagnosis (Figure 2).
The sensitivity and specificity of mammography vary somewhat by age, with sensitivity, specificity, and the positive predictive value (PPV) improving with increasing age (Table 1).[14] Historically, age-specific differences in the accuracy of mammography have been a central issue in the debate over the value of screening women under age 50, and generally mammography performance in women aged 40 to 49 has been compared with that in women 50 years of age and older. This comparison has led to the mistaken impression that performance measures are uniform in postmenopausal women, and measurably poorer in premenopausal women. As shown in Table 1, although there is improvement in sensitivity, specificity, and PPV with increasing age, mammography performance in adjacent decades of life is quite similar. Not shown in Table 1 is PPV as the percentage of biopsies that result in a diagnosis of breast cancer, of which the median percentage is approximately 32%.[15] While an intuitive interpretation of the PPV is that it mainly is influenced by the false-positive rate, in fact the PPV mostly is affected by the underlying prevalence of disease. Therefore, even if screening performance is exactly the same, PPV will be lower in the age group with lower disease prevalence.
