Breast cancer is the most frequent cancer in Europe and the second leading cause of death from cancer in developed countries [2, 3]. Demographic trends indicate a continuing increase in this substantial public health problem [3].

The natural history of breast cancer is not completely understood. Many appear to originate at a preinvasive stage, ductal carcinoma in situ (DCIS), which may subsequently progress to invasive breast cancer. Evidence for this includes:

However, it is unclear why some, but not all, DCIS progresses to invasive cancer and why some invasive breast cancers arise de novo with no DCIS.

Cancer screening aims to detect cancer before symptoms appear. Screening is not useful for all cancers, and for the disease to be amenable to screening, there must be a latent period during which it would be possible to detect the disease before it reaches an advanced stage.

Breast cancer is a heterogeneous disease both with regard to variation among tumours and also the cellular composition within individual tumours. Two events are thought to occur in its natural history and progression: early dissemination and phenotypic progression [10]. If one or both events are related to the size of the breast tumour, then early detection of the tumour may prevent dissemination of tumour cells or the development of a more aggressive tumour that may lead to improved prognosis.

For breast cancer, there is a detectable preclinical phase where tumours are visible on mammograms before they become palpable. Tumours detected are more likely to be non-invasive, and if already invasive less likely to have regional or distant spread.

Screening aims to identify asymptomatic individuals at an early point in their natural history when delivery of early treatment could potentially reduce mortality. Direct comparison of outcomes in women presenting with symptomatic breast cancer is not valid without taking into account three potential biases:

These biases can potentially be removed in randomised trials of population-based screening as described later.

A screening test should detect the majority of women who have cancer (high sensitivity) while eliminating most women who do not have cancer (high specificity). Sensitivity is defined as the proportion of all women with breast cancer who test positive. Specificity is defined as the proportion of all those without breast cancer that test negative. An ideal screening test should have 100% sensitivity and specificity, but in reality no screening test will achieve this. Different modalities for breast cancer screening are described later, but mammography has been the commonly used modality for general population-based screening programmes due to its combined high sensitivity and specificity compared to other screening modalities.

The success of any screening programme will be dependent on the acceptability of the screening test to the population being screened. This will be reflected by the rate of uptake following an invitation to attend for screening. In an overview of mammographic screening programmes operating in Europe for the period 2005–2007 [11], 13 of the 26 programmes achieved the European Union benchmark of acceptable participation (>70%), and 9 achieved the desirable level (>75%) [12]. In England in 2014–2015, 2.11 million women aged 45 and over were screened by the NHS Breast Screening Programme, with an overall uptake of 71.3% [13].

In order to be effective, screening should be repeated at regular intervals. This is a balance between the cost and practicality of screening too frequently and minimising the number of interval cancers that present between screens.

There is variation with regard to the time interval between mammographic screens across EU states, most adopting a policy of biennial mammograms [14], while in the UK women are invited every 3 years.

Growth time or lead time is an important factor with regard to deciding on frequency of screening [15]. Younger women are more likely to have cancers with a shorter growth time potentially increasing interval cancers as opposed to an increased likelihood of slower growing cancers in older women. So theoretically younger women should have a shorter interval between screens compared to their older counterparts. The UK breast cancer frequency trial showed no benefit for women aged 50–62 years of annual mammograms over 3 yearly mammograms [16]. The study did show that tumours diagnosed in women in the study arm were significantly smaller, but there was no significant difference in terms of nodal status, histological grade and predicted mortality between the two groups.

The impact of the introduction of breast screening programmes across Europe has been the development of breast multidisciplinary teams both for the diagnosis and treatment of breast cancer. The use of triple assessment for diagnosis in breast care originated in breast screening practice and has subsequently been widely adopted in symptomatic practice.

Potential physical and psychological hazards of screening using mammography include:

It has been estimated that screening women every 3 years from age 47–73 in the UK breast screening programme may cause 3–6 cancers per 10,000 women screened [17]. Digital mammography, which uses a lower radiation dose, is now more commonly used in screening programmes, and it is likely that this risk is now reduced.

Of the women recalled and found not to have cancer, the majority will only have further imaging (ultrasound, further mammography). Some will have a breast biopsy, usually a core biopsy under local anaesthetic but occasionally a formal open surgical biopsy under general anaesthetic. Numerous studies have assessed the psychological impact of a false-positive result on women with conflicting results. A recent systematic review [19] concluded that, in the population at general risk of breast cancer, a false-positive result can cause breast cancer-specific psychological distress, lasting for up to 3 years.

It is important to continually compare the cost-effectiveness of health screening programmes against other healthcare provisions. This can be expressed as cost per quality-adjusted life year (QALY) gained. This has proved difficult to accurately achieve for breast screening due to ongoing uncertainty regarding the overall benefits and the development of breast screening over time (e.g. one- to two-view mammography, age extension, film to digital conversion, etc.). An assessment of the cost-effectiveness of the UK NHS Breast Screening Programme concluded that there was a moderate probability of breast screening being cost-effective at a willingness-to-pay threshold of £20,000 per QALY [20] but that further research was required.

In summary, breast cancer is a significant health problem, the natural history of which is sufficiently understood for the application of mammographic screening with the aim of early detection and treatment of cancer so as to improve outcome. Mammography as a population-based screening modality is well accepted in Europe and the UK as evidenced by the relatively high participation rate as previously depicted. Mammography is a reliable test (high sensitivity and specificity) in the group undertaking screening, and the subjects are relatively unharmed with the minimal radiation exposure especially with the widespread adoption of digital mammography across Europe.