Search:
Menopause Live - IMS Updates
InFocus

Date of release: 03 August, 2009

Overdiagnosis of breast cancer due to screening?


A recently published paper has received considerable publicity in the lay press and is worthy of reflection [1]. It addresses the question of the number of cases of breast cancer which would not have been diagnosed without organized screening (overdiagnosis is the detection of asymptomatic cancers that will not cause death or symptoms).


The introduction to the paper recalled, according to previous papers, that 37% of women aged 40–54 years dying from other causes than breast cancer had lesions of invasive and non-invasive breast cancer at autopsy and half of these were visible on mammogram. The importance of overdiagnosis and treatment was quantified from randomized trials and it was shown that mastectomies and lumpectomies increased by 20% and 31%, respectively. But these randomized trials were not of sufficient length to determine exactly the extent of overdiagnosis. This recent paper compared trends in breast cancer incidence before and after screening, taking account of changes in the background incidence and any compensatory decrease in incidence of breast cancer among older, previously screened women.


Five studies were analyzed from the UK, Canada (Manitoba), Australia (South Wales), Sweden and Norway. The incidence reported was compared to the incidence expected from the immediate previous period.


In UK, following the start of screening in 1988, the incidence of breast cancer increased by 41% (up to 57% with in situ carcinoma) above the expected rate in women 50–64 years old, without any decrease in older women. An increase of 7% in 30–49-year-old women was observed for the same period.


In Manitoba, elective screening began in the 1970s. In the invited group, the incidence increased to 35% above the expected rate (59% with in situ carcinoma). The total rate for the age group 70–84 years was 15% below that expected, but for the age group 35–49 years it was 32% below that expected, which suggests that causes other than screening could have contributed to the decrease among previously screened women.


In New South Wales, screening was implemented between 1988 and 1995 and the incidence of breast cancer increased to 55% over the expected range in women 50–69 years. No compensatory decrease was observed in women over 70 years and the incidence increased regularly in young women.


In Sweden, using the pre-screening period as 1971–85 and the period 1998–2006 to estimate the trend after screening, the estimated increase for invasive cancer over expected rates was 35%. A constant increase in incidence was seen among women too young to be screened. A decrease occurred among women aged 70–84 years, but the incidence approached the expected rate at the end of the observation period; 88% of the increase was therefore uncompensated. When carcinoma in situ was included, overdiagnosis was 46%.


In Norway, the incidence increased in screened women to 52% (including in situ carcinoma) above the expected incidence. A decrease was seen in older women but 86% remained uncompensated. At younger ages, the incidence remained stable, with limited data for these ages.


The meta-analysis of overdiagnosis of breast cancer in the five countries provides a rate ratio of 1.52 (95% confidence interval (CI) 1.46–1.58).


In the discussion of the paper, the authors emphasize the fact that the decrease in breast cancer incidence following the large decrease in use of hormone replacement therapy (HRT) has not been seen in every country and, especially in Norway, the ‘effect of screening was separated from that of hormone replacement therapy use, as incidence trends in regions with and without screening could be compared at the same calendar times. Although use of HRT is likely to be similar, a noticeable increase occurred in invasive cancer with the introduction of screening, both in the Akershus, Oslo, Rogaland and Hordaland counties and in the remaining counties of Norway.’

Comment

This paper raises debate on several issues concerning the benefits or harms of breast cancer screening, a debate with pros and cons and unsolved questions.
 
Overdiagnosis
 
This paper argues that the percentage of overdiagnosed breast cancer is 52 (95% confidence interval 46% to 58%) through systematic screening and interestingly shows that incidence varies among various countries and according to age. It has been published with an Editorial in the same BMJ issue by H. G. Welch, who states that ‘the question is no longer whether but how often the overdiagnosis occurs’. The most convincing evidence is a previous randomized trial also published in the BMJ [2] 3 years ago which also demonstrated that rates of overdiagnosis of breast cancer were 1.32 (95% CI 1.14–1.53) during screening, and 1.10 (95% CI 0.99–1.22) at the end of follow-up. 
In this Editorial, the ‘delicate balance’ of national screening is well discussed, highlighting the difficulties of how to choose the right attitude: since ‘mammograms could benefit in some women but hurts some others’.
As a clinician consulted by women above 50 years, it is difficult to refuse to advise them to follow national programs of breast cancer screening even if we know that there is undoubtedly overdiagnosis. Some propositions may be made: to inform patients about the overdiagnosis of breast cancer between 1 in 6 cases in the Malmö paper to 1 in 3 cases in the recent paper; and to apply systematic screening to women with recognized risk factors. However, if a national screening program is active, it is probably recommended to adhere maximally to it in order to evaluate public health policies. 
 
Mortality gain
 
This paper has not discussed the benefits of screening in term of mortality. This is, however, the only endpoint to consider for screening programs. Gain in mortality over time can come either from the benefits from screening program or from more efficient treatments. In the report by Zackrisson and colleagues [2], a ratio of 1 death to 2 women overdiagnosed was reported, whereas, in a recent paper, the ratio was evaluated to be 1 in 10 women [3]. In a paper using modeling techniques to assess the relative and absolute contributions of screening mammography and adjuvant treatment to the reduction in breast cancer mortality in the United States, total reduction of death attributed to screening varied from 28 to 65% (median, 46%) and treatment contributed the rest [4].
 
Gain in quality of life
 
This is an endpoint that is extremely difficult to evaluate! In our opinion, the definite proof of a gain in quality of life from a breast screening program would be to evaluate the rate of chemotherapy cancelled due to diagnosis at a smaller tumor stage; however, we know that screening is not able to diagnose relatively fast-growing tumors and do not know exactly which tumors benefits from screening programs.  
 
In conclusion, the debate remains open but we need to reassess regularly our diagnostic strategies and algorithms in the management of women diagnosed with breast cancers during screening programs.

Comentario

Anne Gompel
Unité de Gynécologie-Endocrinienne, APHP, Hôtel-Dieu Hospital and University Paris Descartes, France

    References

  1. Jørgensen KJ, Gøtzsche PC. Overdiagnosis in publicly organised mammography screening programmes: systematic review of incidence trends. BMJ 2009;339:b2587. doi: 10.1136/bmj.b2587. Published July 9, 2009.
    http://www.ncbi.nlm.nih.gov/pubmed/19589821

  2. Zackrisson S, Andersson I, Janzon L, Manjer J, Garne JP. Rate of over-diagnosis of breast cancer 15 years after end of Malmö mammographic screening trial: follow-up study. BMJ 2006;332:689-92.
    http://www.ncbi.nlm.nih.gov/pubmed/16517548

  3. Gøtzsche PC, Hartling OJ, Nielsen M, Brodersen J, Jørgensen KJ. Breast screening: the facts or maybe not.BMJ 2009;338:b86. doi: 10.1136/bmj.b86.
    http://www.ncbi.nlm.nih.gov/pubmed/19174442

  4. Berry DA, Cronin KA, Plevritis SK, et al.; Cancer Intervention and Surveillance Modeling Network (CISNET) Collaborators. Effect of screening and adjuvant therapy on mortality from breast cancer. N Engl J Med 2005;353:1784-92.
    http://www.ncbi.nlm.nih.gov/pubmed/16251534