Menopause Live - IMS Updates

Date of release: 11 January, 2010

Rethinking the impact of breast and prostate cancer screening. Part 2: The comment

Note: the first part of this Commentary, The evidence, was issued on January 4, 2010.


The question of the efficacy of screening and its possible adverse effects, both the emotional impact and the unnecessary procedures due to overtreatment, is regularly discussed. A recent paper, on which a Menopause Live commentary was issued (August 3, 2009), suggested that as many as 50% of the breast cancers diagnosed through screening were over-diagnosed [1]. 
The paper by Esserman and colleagues [2] also provides some arguments to suggest that, especially for prostate cancer but also for breast cancer, a certain amount of disease is over-treated. The paper specially emphasizes the fact that mortality is not impacted to a great extent by screening and suggests alternative strategies. The paper addresses the hormone-dependent tumors as one entity and tends to combine the data for prostate cancer and breast cancer. We suggest that the figures for the two cancers are different. Whereas the mortality is not modified in the USA for prostate cancer and only mildly altered in Europe (1 death for 1410 individuals screened), it is decreased in women with breast cancer. Even if the decrease in very aggressive diseases is not sufficient (is it even possible?), it is not still clear whether there are some cancers for which an early diagnosis may preclude chemotherapy. It is becoming more and more obvious that each case of breast cancer has its own biology and that these characteristics condition the outcome. However, there are so far no routine tests to evaluate this. It is still difficult to quantify the health impact of breast screening, since, in the USA and in most of the European countries, adherence is not complete and national programs are not generalized. There might also be some beneficial aspects of national programs, such as in France where the quality of mammography and radiologic equipment has drastically increased, leading to more accurate diagnosis. It is not clear whether this is the case also in the USA and other countries.
Concerning, however, the emphasis on other strategies to help to decrease mortality, this is definitely an approach that must be promoted, at least in part. Considering the drastic developments in research for predictive markers during the last 5 years and the dramatic increase in therapies targeted at cancer phenotypes, such an approach is supported by many research groups and the pharmaceutical industry. Yet, it is not clear whether molecular profiling of tumors and the tailoring of treatment by specific target therapies could improve not only survival, but also quality of life and the cost effectiveness of breast cancer therapies. 
The additional aims of proposed strategies in Esserman’s paper are the prevention and evaluation of individual risk [2]. Prevention can be promoted by national campaigns for lifestyle improvement and these have been started by several public health authorities in Western countries. Chemical prevention is another potentially promising method, yet not in the focus of clinical research since such trials are too expensive and long-term, which makes them unattractive for pharmaceutical companies. It should therefore be proposed that national or international academic departments may develop and sponsor such programs. 
The development of clinical score systems for the evaluation of individual risk is an ongoing effort. Since the first Gail score was proposed, other scores have been formulated that can be used either for screening patients with a family history of breast cancer or to include clinical factors. Most of the scores developed recently include breast density with other clinical factors; these remain, however, of insufficient sensitivity. The c-statistic measures the reliability of prediction of a scoring system; it evaluates the probability that the system will give a higher score to a woman who develops breast cancer than to the one who does not. If the c-statistic is 0.50, the scoring system has no better predictive value [3]. Most of the scores have a c-statistic ≤ 66%, as reported in the paper from Cummings and colleagues [3]. 
Two recent scores, which have not been evaluated in a large population, could be of some additional interest since they combine genetic risks and clinical data; both are available freely on the internet. The Cancer gene software was developed by D. M. Eurhus in the USA (Texas Southwestern Center for Breast Care, Dallas) and integrates to BRCA probability (BRCAPRO) with risks for other familial cancers such as HNPCC (hereditary non-polyposis colorectal cancer, MMRpro), pancreas (PancPRO) and the Claus and Gail index [4]. IBIS has been developed by Tyrer and Cuzik in the UK. It integrates BRCA1,2 and a third locus frequency but with low penetrance, and personal risk factors, such as age at menarche and menopause, body mass index, atypical hyperplasia, ductal carcinoma in situ, parity and age at first full-term pregnancy [5]. As far as I know, there have not been any comparisons between these different tools.
The development of programs to increase the sensitivity of predicting the risk in individual women and men is obviously one of the goals to decrease cancer mortality. Encouraging preventive strategies may also impact on the outcome of the more aggressive tumors: decreasing obesity, for example, could be an important target since breast cancers can be more aggressive in obese women. However, in the meanwhile, and as clinicians engaged in routine practice, mammograms remain an efficient tool to manage our patients. In addition, if health policies are developed at a national level, despite the remaining uncertainty about the benefits, it seems necessary to encourage adherence to the programs in order to gain statistical power in evaluating their outcomes.


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


  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.

  2. Esserman L, Shie Y, Thompson I. Rethinking screening for breast cancer and prostate cancer. JAMA 2009;302:168592. Published October 21, 2009.

  3. Cummings SR, Tice JA, Bauer S, et al. Prevention of breast cancer in postmenopausal women: approaches to estimating and reducing risk. J Natl Cancer Inst 2009;101:38498.