Karsten Jørgensen, of the Nordic Cochrane Center in Copenhagen, talked about the perils derived by over-diagnosis with mammography screening at the third day of the ESMO conference in Milano. Jorgensen opened his talk with a provokative quotation taken by the paper of JA Muir Gray and June Austoker, namely: "All screening programmes do harm; some also do good. The responsibility of the policy-maker is to decide which programmes do more good than harm at reasonable cost and then introduce them, once they are confident that the screening programme could and will reach the standard of quality required for success. The ratio of benefit to harm is not, however, constant and this relationship demonstrates a shifting balance"¹, and posed another similarly provokative question to the audience, namely whether mammography screening for breast cancer can be actually numbered among the programmes that also do good.

To start out, Jorgensen gave some examples of cases where over-diagnosis led to abandonment of the screening programmes: lung cancer with chest x-ray was abandoned as a screeening programme when 29-50 % more cases were found in the screened group. The same happened with prostate cancer screening with PSA (20-70 % extra cases in screened group) and screening for neuroblastoma in children with urine test: 100 % over-diagnosis in two independent trials. Not to talk about the final screening test: autopsy, where practically every single one of us is found to have thyroid lesions!

Two fundamental concepts need to be spelled out and properly defined for every screening programme that we undertake:

1) Over-diagnosis, defined as the phenomenon for which, if screening were detecting breast cancer that would otherwise not have been diagnosed, it would be expected that in controlled trials there would be a persistent excess number of breast cancers in the screened group compared with the control group (the definition is taken from the 1986 Forrest Report, which formed the basis for the breast cancer screening programme in the UK, and that can be read here.

2) Over-treatment: an obvious source of harm associated with any screening programme is unnecessary treatment of cancers that were not going to cause death or symptoms (from the 2003 IARC Handbooks of Cancer Prevention, Volume 7: Breast Cancer Screening, the electronic version can be found here.

As stressed by Jørgensen, "It's not over-diagnosis the problem we need to tackle in breast cancer treatments, it is overtreatment! Think how you would explain to a woman that the surgery, radiotherapy or chemotherapy she underwent, together with all the worries and preoccupation that she and her family went through, were unnecessary? That is the real problem!"

The Nordic Cochrane Center (NCC), where Jorgensen works, is an independent research and information centre that is part of the Cochrane Collaboration (an international network of individuals and institutions committed to preparing and disseminating systematic reviews of the effects of health care). A review about breast cancer screening with mammography was compiled in 2009, and is available upon download here on their website. In this review, eight eligible trials were identified, and 600,000 women were included in the analyses. Three trials with adequate randomisation did not show a significant reduction in breast cancer mortality at 13 years (relative risk (RR) 0.90); four trials with suboptimal randomisation showed a significant reduction in breast cancer mortality with an RR of 0.75. Numbers of lumpectomies and mastectomies were significantly larger in the screened groups (RR 1.31, 95% CI 1.22 to 1.42) for the two adequately randomised trials that measured this outcome; the use of radiotherapy was similarly increased. The authors of the review, Peter Gotzsche and Margrethe Nielsen², conclude that, although screening is likely to reduce breast cancer mortality, with a reasonable estimate of 15% reduction (corresponding to an absolute risk reduction of 0.05%), it also led to 30% over-diagnosis and overtreatment, or an absolute risk increase of 0.5%.

To put the results in other figures, this means that for every 2000 women invited for screening throughout ten years, one will have her life prolonged and ten healthy women, who would not have been diagnosed if there had not been screening, will be treated unnecessarily! Furthermore, more than two hundred women will go through psychological distress due to false positive results. Gotzsche and Nielsen conclude that it is thus not clear whether screening does more good than harm. Jørgensen agrees with them and, as he started out his talk with a provocative question, he concluded it with another thought-provoking message. He showed an old print campaign advertisement by the American Cancer Society, which reads: "If you haven't had a mammogram, you need more than your breasts examined", conveying therefore the message that women would be stupid if they did not undergo mammography screening. But, given current data, such a message needs to be seriously re-evaluated.

References

1. Gray JA and Austoker J (1998) Quality assurance in screening programmes Br Med Bull. 54(4) 983-92

2. Jørgensen KJ and Gøtzsche PC (2009) Overdiagnosis in publicly organised mammography screening programmes: systematic review of incidence trends BMJ 339:b2587.

Additional reading

Boer R, Warmerdam P, de Koning H, et al (1994) Extra incidence caused by mammographic screening Lancet 343(8903) 979.

Esserman L and Thompson I (2010) Solving the overdiagnosis dilemma J Natl Cancer Inst 102(9):582-3

Veronesi A and Serraino D (2009) Screening: is breast cancer overdiagnosed? Nat Rev Clin Oncol. 6(12):682-3

Welch HG and Black WC (2010) Overdiagnosis in cancer J Natl Cancer Inst 102(9):605-13