by ecancer reporter Clare Sansom

In the UK, mammography screening for breast cancer has been offered to all women aged between 50 and 70 since 1988.

During the last few years, however, increasing concerns about the value and risks of screening have been raised in the medical and general media.

This led the UK's National Cancer Director and the chief executive of Cancer Research UK to commission an independent review of breast screening.

The report took just under a year to produce and was published in October 2012. Its findings were presented to delegates at the NCRI conference in a packed plenary session on Tuesday morning.

One of the review's commissioners, CRUK Chief Executive Harpal Kumar, set the scene by describing the rationale for the study and the questions it was asked to explore. It was designed to determine the extent of the benefit of screening in breast cancer deaths avoided and the risks of overdiagnosis. It is useful to highlight the difference between overdiagnosis and a false negative results: overdiagnosis specifically applies to slow-growing cancers that are picked up through screening and removed, but that would not have caused a problem during the lifetime of the women concerned.

The report was carried out by a panel of experts chaired by epidemiologist Professor Sir Michael Marmot None of the panel members had previously published on breast screening. The panel spoke to a wide range of screening experts from the UK and overseas and reviewed the published literature on both randomised controlled trials (RCTs) and observational studies.

As Sir Michael Mermot was unable to attend the meeting, the results of the review were presented by another panellist, Professor David Cameron from the University of Edinburgh. Cameron began by describing the trials that fed into their statistical analysis. The panellists chose to focus on RCTs, which were considered more rigorous, although no trial methodology was flawless.

Eleven trials with follow-up times long enough for survival benefit to be identified were selected. A meta-analysis of these found an odds ratio of death from breast cancer in the screened arm of 0.8 (95% confidence interval 0.72-0;89), indicating a reduction of approximately 20% in the risk of death from the disease. The best estimate for the percentage reduction in the risk of breast cancer death until age 79 was calculated to be 0.43%, indicating that screening 10,000 women should prevent 43 deaths in this age group.

The risk of overdiagnosis was less easy to estimate, as it is clearly impossible to identify retrospectively which cancers are over-diagnosed. Varying the parameters used in the calculation gave estimates of overdiagnosis rates from the same RCTs ranging from 0% to 36%. The panel selected two methodologies that seemed the most informative; these gave an overdiagnosis rate of 11% from a population perspective and 19% from the perspective of an individual woman. Although these rates seem high, and overdiagnosis is certainly a problem, breast cancer is still relatively rare; they estimated that 681 cancers are detected for every 100,000 women screened for 20 years, and that 129 of these would not have needed treatment.

The panel concluded that the balance of the benefit and risk of breast cancer screening came down on the side of benefit, and so recommended that screening should continue. However, they also recommended that the patient information leaflets should be re-written to give women a better idea of this balance in order to make an informed choice.