- Email: [email protected]
Breast screening policy: Are we heading in the right direction?
Clinical Radiology 66 (2011) 915e919
Contents lists available at ScienceDirect
Clinical Radiology journal homepage: www.elsevierhealth.com/journals/crad
Review
Breast screening policy: Are we heading in the right direction? A. Evans*, P. Whelehan Dundee Cancer Centre, Ninewells Hospital and Medical School, University of Dundee, Dundee, UK
art icl e i nformat ion Article history: Received 11 October 2010 Received in revised form 25 February 2011 Accepted 2 March 2011
There is a large body of evidence supporting 2-yearly screening of women aged 50e69 years. There is good evidence for a mortality reduction from mammographic screening in women aged 40 to 49 years but a 1-year interval is required. The lack of specificity of screening in young women does remain a problem. There is no evidence to suggest that a single screen between the ages of 47 and 50 years within a programme screening at 3-year intervals will reduce mortality; the trials showing a mortality benefit in women in their 40s included multiple screening episodes and shorter screening intervals. There is no randomized, controlled trial evidence to support screening in women aged above 70 years and screening this age group will cause greater harm than in younger women through higher rates of overdiagnosis and consequent over-treatment. The randomized phase of the screening age extension, which at the moment is planned to last only 6 years, should not be immediately followed by general implementation of the policy. Only if and when additional mortality reductions and an acceptable balance between benefit and harms are shown to be achieved by the extra screens should the 2007 Cancer Reform Strategy policy on age extension be implemented. Resources saved by delaying or abandoning the roll-out of the age extension could potentially be redirected towards reducing the current 3 year screening interval to 2 years in women aged 50e69 years. However, reducing the screening interval to 2 years for women aged 50e69 years would require significantly more screening invitations and resources than the proposed age extension. Ó 2011 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.
Introduction Breast screening policy has recently been the subject of heightened debate in the UK following the publication of the previous government’s Cancer Reform Strategy (CRS),1 and in the US following the US Preventive Services Task Force (USPSTF) 2009 recommendations for breast screening.2,3 The recommendations for breast cancer
DOI of original article: 10.1016/j.crad.2011.06.006. * Guarantor and correspondent: A. Evans, Mail box 4, Ninewells Hospital and Medical School, University of Dundee, Dundee DD1 9SY, Scotland, UK. Tel.: þ44 (0) 1382 632196; fax: þ44 (0) 496363. E-mail address: [email protected] (A. Evans).
screening are disconcertingly divergent in the two documents. The different approaches can, in part, be explained by the pre-existing disparity in screening practice between the two countries and perhaps also by the methodologies used to formulate such policies and recommendations.
UK and US policies In the UK the CRS was published in 2007 by the Department of Health and it stated that there would be an extension of routine breast screening to nine rounds between the ages of 47 and 73 years with the guarantee of a first mammogram before age 50 and the concurrent implementation of digital mammography. No change to the
0009-9260/$ e see front matter Ó 2011 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.crad.2011.03.024
916
A. Evans, P. Whelehan / Clinical Radiology 66 (2011) 915e919
3-year screening interval was made. In the US, the 2009 USPSTF recommendations caused a public outcry by recommending against routine provision of a screening service below age 50da change to its previous recommendation and contrary to widespread practice. The USPSTF in 2009 further recommended that biennial screening by mammography in the 50e74 year age group should be provided but that evidence was lacking on screening after age 74.
Rationale for UK policy In the UK, the CRS breast screening policy statements were formulated following consultation with an expert advisory panel. The document does not describe the methodology used to arrive at the recommendations and no direct reference is made to any published evidence underlying the policy statements on breast cancer screening. Following publication, concern regarding the effectiveness of the age extensions led to the adoption of a proposal from the University of Oxford to randomize the extension in order to facilitate more robust future evaluation of screening effectiveness in the new age bands. Invitation batches are, therefore, randomized either to include women aged 47e49 or aged 71e73 years. The portion of each age cohort not being invited constitutes a control group. Five pilot sites began the randomization project and other sites are now joining following demonstration of feasibility.
Rationale for US policy The USPSTF 2009 recommendation to screen with biennial mammography in the 50e69 year age group is based largely on a published study using six mathematical models developed independently within the Cancer Intervention and Surveillance Modeling Network of the US National Institutes of Health (NIH).4 These models evaluated 20 screening strategies in which screening was performed either annually or biennially with starting ages ranging from 40 to 60 years and cessation ages ranging from 69 to 84 years. The work suggested that biennial screening maintained 81% of the benefit of annual screening but reduced the false-positive results by almost half. Screening women aged 50e69 biennially achieved a median breast cancer mortality reduction across the different mathematical models of 16.5%. Initiation of biennial screening at age 40 reduced mortality by an additional 3%. Screening women aged 70e79 years was associated with a 7% additional mortality reduction but with higher rates of over-diagnosis than in the younger age groups. (Over-diagnosis is a term commonly used to describe detection of cancers that would not have become clinically important during a person’s remaining lifespan d an inevitable consequence of screening but nonetheless important to minimize.) When the outcome measure was total life years gained, rather than mortality rate, the benefits were greater for extending screening to younger rather than older women. The 2009 recommendation of the USPSTF not to support routine screening of women in their 40s despite this evidence is
based on their opinion that mortality reduction is a more important endpoint than life years gained. Their recommendation statement was modified the month after issue to make their position less emphatic and to support individualized decision-making on when to begin screening. None of the modelled screening strategies comprised annual screening for younger women and biennial screening for older women. This is surprising since the correct screening interval is determined by the lead time (the time between mammographic detection and when clinical detection would have been possible) achieved by the screening test. It is well established that that the sojourn times (the time between the possibility of detection by mammography and the clinical presentation) and subsequent lead times achieved by screening are longer for older women and shorter for younger women. The reported sojourn time for screening women aged 40e49 is 1.7e2.7 years, for 50e59 is 3.1e3.7 years and for 60e69 is 3.8e4.2 years.5e7
Outstanding controversies Current debate remains focussed on which age groups to screen and how frequently, and the two questions are interrelated. In general, screening policy has been largely based on the findings of the randomized, controlled trials (RCTs) of screening, particularly high-quality trials conducted in Sweden.8e12 However, none of the Swedish trials was designed to examine the mortality benefit of screening in a particular age group. The combined Swedish results stratified in 5 year age bands according to age at randomization show significant mortality reductions in the age bands 55e59 [relative risk (RR) 0.76 (0.59e0.98)], 60e64 [RR 0.68 (0.50e0.92)], and 65e69 [RR 0.69 (0.49e0.96)].11 There is little controversy in the mainstream of opinion about the value of screening women aged 50e69. However, the appropriate screening interval for women of this age group remains open to debate.
Screening interval from age 50 years The lead time for screening is age dependent, and in women from age 50, mammographic screening achieves a lead time of between 3 and 5 years.13,14 Two-yearly mammographic screening is likely to be appropriate in women from age 50, as indicated by a recent analysis of pooled interval cancer data from six European countries. Interval cancer rates of 29% of the expected breast cancer incidence without screening in year 1 and 63% in year 215 indicate that 2-yearly is the maximum screening interval that is appropriate. A recent analysis of interval cancer rates in the NHSBSP gave very similar results.16 In the US the debate on screening intervals is focused on whether 1-yearly or 2-yearly screening is preferable. The apparent continued acceptance in the UK of its 3-year interval is surprising. The UK Co-ordinating Committee on Cancer Research (UKCCCR) breast screening frequency trial17 compared the prognostic features of invasive cancers detected by three annual screens following a prevalent attendance with those in a control group who had
A. Evans, P. Whelehan / Clinical Radiology 66 (2011) 915e919
a screening invitation after the standard 3-year interval. The study compared predicted mortality, not observed mortality. The predicted RR of breast cancer death in the study arm versus the control arm was 95% or 89% using two different prognostic indices but the differences were not statistically significant. The study had a number of important limitations. It was powered to detect a 25% reduction in breast cancer mortality in the study group compared with the control group. This was unrealistic as the breast cancer mortality achieved by screening versus no screening is of this magnitude. Screening was by single view only at a time when screening quality was lower than currently. One of the prognostic indices used (The Nottingham Prognostic Index) was derived from symptomatic patients and was not appropriate for predicting the survival of small screendetected cancers in the frequency trial.18,19 Approximately 20% of the cancers in each arm were of unknown nodal status, making prognostication unreliable. The effect of increased ductal carcinoma in situ (DCIS) detection in the study arm was not taken into account as the analysis was based solely on invasive cancers. These limitations mean the results of the frequency trial should not be used to defend the UK’s 3-year screening interval.
Screening women younger than 50 A meta-analysis by Hendrick et al. in 199720 of women aged 40e49 in all RCTs showed a statistically significant mortality reduction of 18% [RR 0.82 (0.71e0.95)]. The Swedish overview results published in 2003 did not show a statistically significant benefit for women in this same age group.11 This may have been because the results of the Kopparberg portion of the two-county trial were not € and the included in the later analysis. Both the Malmo Gothenberg trials demonstrated a statistically significant mortality reduction in women in this age group [RR 0.64 (0.45e0.89); p ¼ 0.009 and RR 0.56 (0.32e0.98); p ¼ 0.035, respectively)12,21; both studies incorporated short screening intervals (18 months for women with dense breasts) and multiple screening rounds (at least five). A physical examination was not included in either of these trials. Critics of mammographic screening have commented that many of the women in the Swedish RCTs aged 40e49 at the time of randomization received many of their screening € study, 47% of the episodes after age 50. In the Malmo screening episodes of women randomized younger than 50 occurred after age 50. In the Gothenberg study, 31% of the cancers in the study group were diagnosed after age 50. The Gothenberg results have since been re-analysed according to age at cancer diagnosis, suggesting mortality reductions between 31% and 44% in women aged 39e49 diagnosed during the screening period, depending on the precise mortality endpoint.12 The UK Age Trial was the only RCT to perform all screening episodes before age 50.22 Participants aged 39e41 were randomly selected and were screened annually: two views at their first screening and one view at subsequent screens. There was no clinical examination. The RR of breast cancer death in this study group was 0.83 (0.66e1.04). A major strength of the UK Age Trial is that the
917
data are not contaminated by screening episodes after age 50. The limitations of the study include the fact that only one mammographic view was performed after the first round of screening and that film/screen mammography was used and not full-field digital mammography, which has subsequently been found to be more sensitive in denser breasts and younger women.23 Mammographic screening in women in their 40s has long been a point of debate. Critics highlight the fact that women in their 40s have only half the breast cancer incidence of women in their 50s; however, advocates of screening younger women point out that 34% of total life years lost to breast cancer are from cancers diagnosed between ages 40 and 49.24 Although breast cancer incidence increases with age, the proportion of female deaths attributable to breast cancer is highest in women in their 40s and 50s (16 and 14%, respectively).25 Several problems are encountered during screening of women younger than 50, including relatively high numbers of false-positive screening results and screening-provoked benign surgical biopsies.26 There is a modest reduction in mammographic sensitivity when comparing women in their 40s with women in their 50s27; however, the widespread introduction of digital mammography should ameliorate this problem, according to the results of the Digital Mammographic Imaging Screening Trial.23 More serious issues related to screening women younger than 50 are the high rate of interval cancers and the short duration of time between a normal screen and return to normal background incidence.28,29 These findings indicate that the lead time of screening in younger women is short (between 18 and 24 months); therefore, screening of women aged 40e49 requires a short screening interval, ideally 1 year.
Screening women older than 70 Because of the high incidence of breast cancer in women older than 70 and because mammographic sensitivity and specificity are high, screening is attractive in theory. However, there are a number of problems associated with screening for this age group. Competing causes of mortality make breast cancer proportionally a less common cause of death in older compared with younger women.30 Only 4% of all deaths in England and Wales in women aged 70e79 and 2% in women aged 80e89 are due to breast cancer, compared with 16% in women in their 40s.25 The Swedish RCTs included only 5000 women older than 70 and showed a non-significant 18% mortality excess in the screening arm.11 A recent evaluation of service screening in Northern Sweden showed significant mortality reductions for women aged 40e49 [RR 0.64 (0.43e0.97)] and women aged 50e69 [RR 0.70 (0.54e0.91)], but no mortality benefit was seen for women age 70e74 [RR 1.08 (0.58e2.03)].31 The shorter life expectancy and less aggressive tumour profile means that over-diagnosis becomes a larger problem when screening women older than 70 compared with younger women. In a recent trial of extended adjuvant endocrine therapy, 72% of deaths in women older than 70 with breast cancer were non-breast cancer deaths.32 Data from the American
918
A. Evans, P. Whelehan / Clinical Radiology 66 (2011) 915e919
Surveillance Epidemiology and End Results (SEER) programme show that in women aged over 70 diagnosed with localized invasive breast cancer, 86% of deaths at the end of follow-up had resulted from causes other than breast cancer.33 Therefore, it is unclear whether screening women older than 70 is beneficial. A particular issue in this age group is the over treatment of low-risk DCIS lesions, which are unlikely to progress to invasive cancer in a woman’s lifetime. The proposed UK trial of watchful waiting versus conventional treatment is welcomed. A large RCT in this age group has not been conducted.
Conclusions and recommendations There is a large body of evidence supporting 2-yearly screening of women aged 50e69. There is good evidence for a mortality reduction from mammographic screening in women aged 40e49 but a 1-year interval is required. The lack of specificity of screening in young women does remain a problem. Good-quality RCTs have not been performed to assess the benefit of screening for women older than 70. Although screening of women in this age group provides a high cancer yield with good sensitivity and specificity, any mortality benefit is likely to be modest because of competing causes of mortality; therefore, the value of screening women in this age group remains questionable. What of screening policy in the UK? There is no evidence to suggest that a single screen between the ages of 47 and 50 within a programme screening at 3-year intervals will reduce mortality; the trials showing a mortality benefit in women in their 40s included multiple screening episodes and shorter screening intervals. It is true that a stated aim of the lower extension is that women receive their first screen by their 50th birthday. We agree that women receiving their first screen within a year of their 50th birthday would be a good thing. Screening women from age 47 with a 3 year interval is something quite different. Efforts could perhaps have been focussed on ensuring women are first invited between their 49th and 51st birthdays. Although changing organizational systems to achieve this may be difficult, overcoming these difficulties may have been more cost-effective than screening women 3-yearly from age as low as 47. There is no RCT evidence to support screening in women aged above 70 and screening this age group will cause greater harm than in younger women through higher rates of over-diagnosis and consequent over-treatment. The randomized phase of the screening age extension, which at the moment is planned to last only 6 years, should not be immediately followed by general implementation of the policy. Instead, once sufficient accrual has taken place in line with the power calculations for the study, extending the length of the study if necessary to ensure conclusive results, the randomized women should be followed up and the results analysed. Only if and when additional mortality reductions and an acceptable balance between benefit and harms are shown to be achieved by the extra screens should the 2007 CRS policy on age extension be implemented. Resources saved by delaying or abandoning the roll-out of
the age extension could potentially be redirected towards reducing the current 3 year screening interval to 2 years in women aged 50e69. However, reducing the screening interval to 2 years for women aged 50e69 would require significantly more screening invitations and resources than the proposed age extension.
References 1. Department of Health. Cancer reform strategy. London: DoH; 2007. 2. U.S. Preventive Services Task Force. Breast cancer screening. Available from: http://www.uspreventiveservicestaskforce.org/uspstf/uspsbrca.htm; 2009 [accessed 7 October 2010]. 3. Calonge N, Petitti DB, DeWitt TG, et al. Screening for breast cancer: U.S. preventive services task force recommendation statement. Ann Intern Med 2009;151:716e26. 4. Mandelblatt JS, Cronin KA, Bailey S, et al. Effects of mammography screening under different screening schedules: model estimates of potential benefits and harms. Ann Intern Med 2009;151:738e47. 5. Tabar L, Fagerberg G, Chen HH, et al. Efficacy of breast cancer screening by age: new results from the Swedish two-county trial. Cancer 1995;75: 2507e17. 6. Duffy SW, Day NE, Tabar L, et al. Markov models of breast tumor progression: some age-specific results. J Natl Cancer Inst Monogr 1997; 22:93e7. 7. Shen Y, Zelen M. Screening sensitivity and sojourn time from breast cancer early detection clinical trials: mammograms and physical examinations. J Clin Oncol 2001;19:3490e9. 8. Andersson I, Aspegren K, Janzon L, et al. Mammographic screening and mortality from breast cancer: the Malmo mammographic screening trial. BMJ 1988;297:943e8. 9. Tabar L, Fagerberg G, Duffy SW, et al. Update of the Swedish two-county program of mammographic screening for breast cancer. Radiol Clin North Am 1992;30:187e210. 10. Frisell J, Lidbrink E, Hellstrom L, et al. Followup after 11 years d update of mortality results in the Stockholm mammographic screening trial. Breast Cancer Res Treat 1997;45:263e70. 11. Nystrom L, Andersson I, Bjurstam N, et al. Long-term effects of mammography screening: updated overview of the Swedish randomised trials. Lancet 2002;359:909e19. 12. Bjurstam N, Bjorneld L, Warwick J, et al. The Gothenburg Breast Screening Trial. Cancer 2003;97:2387e96. 13. Paci E, Duffy S, Giorgi D, et al. Population-based breast cancer screening programmes: estimates of sensitivity, overdiagnosis and early prediction of the benefit. In: Duffy S, Hill C, Esteve J, editors. Quantitative methods for the evaluation of cancer screening. London: Arnold; 2001. p. 127e35. 14. Duffy SW, Gabe R. What should the detection rates of cancers be in breast screening programmes? Br J Cancer 2005;92:597e600. 15. Tornberg S, Kemetli L, Ascunce N, et al. A pooled analysis of interval cancer rates in six European countries. Eur J Cancer Prevent 2010;19: 87e93. 16. Bennett RL, Sellars SJ, Moss SM. Interval cancers in the NHS breast cancer screening programme in England Wales and Northern Ireland. Br J Cancer 2011;104:571e7. 17. The Breast Screening Frequency Trial Group. The frequency of breast cancer screening: results from the UKCCCR Randomised Trial. Eur J Cancer 2002;38:1458e64. 18. Anderson TJ, Alexander FE, Lamb J, et al. Pathology characteristics that optimize outcome prediction of a breast screening trial. Br J Cancer 2000;83:487e92. 19. Evans AJ, Rakha EA, Pinder SE, et al. Basal phenotype: a powerful prognostic factor in small screen-detected invasive breast cancer with long-term follow-up. J Med Screen 2007;14:210e4. 20. Hendrick RE, Smith RA, Rutledge 3rd JH, et al. Benefit of screening mammography in women aged 40e49: a new meta-analysis of randomized controlled trials. J Natl Cancer Inst Monogr 1997;22:87e92. 21. Andersson I, Janzon L. Reduced breast cancer mortality in women under age 50: updated results from the Malmo Mammographic Screening Program. J Natl Cancer Inst Monogr 1997;22:63e7.
A. Evans, P. Whelehan / Clinical Radiology 66 (2011) 915e919 22. Moss SM, Cuckle H, Evans A, et al. Effect of mammographic screening from age 40 years on breast cancer mortality at 10 years’ follow-up: a randomised controlled trial. Lancet 2006;368: 2053e60. 23. Pisano ED, Gatsonis C, Hendrick E, et al. Diagnostic performance of digital versus film mammography for breast-cancer screening. New Engl J Med 2005;353:1773e83. 24. Moss S. Should women under 50 be screened for breast cancer? Br J Cancer 2004;91:413e7. 25. Office for National Statistics. Mortality statisticsdcause: review of the registrar general on deaths by cause, sex and age. In: England and Wales. London: HMSO; 2005. 2006. 26. Johns LE, Moss SM. Randomized controlled trial of mammographic screening from age 40 (“Age” trial): patterns of screening attendance. J Med Screen 2010;17:37e43. 27. Rosenberg RD, Hunt WC, Williamson MR, et al. Effects of age, breast density, ethnicity, and estrogen replacement therapy on screening mammographic sensitivity and cancer stage at diagnosis: review of 183,134 screening mammograms in Albuquerque, New Mexico. Radiology 1998;209:511e8.
919
28. Tabar L, Faberberg G, Day NE, et al. What is the optimum interval between mammographic screening examinations? An analysis based on the latest results of the Swedish two-county breast cancer screening trial. Br J Cancer 1987;55:547e51. 29. Bucchi L, Ravaioli A, Foca F, et al. Incidence of interval breast cancers after 650,000 negative mammographies in 13 Italian health districts. J Med Screen 2008;15:30e5. 30. Heron M, Hoyert DL, Murphy SL, et al. Deaths: final data for 2006. National vital statistics reports: from the Centers for Disease Control and Prevention, National Center for Health Statistics, National Vital Statistics System 2009;57:1e134. 31. Jonsson H, Bordas P, Wallin H, et al. Service screening with mammography in Northern Sweden: effects on breast cancer mortalitydan update. J Med Screen 2007;14:87e93. 32. Chapman JAW, Meng D, Shepherd L, et al. Competing causes of death from a randomized trial of extended adjuvant endocrine therapy for breast cancer. J Natl Cancer Inst 2008;100:252e60. 33. Schairer C, Mink PJ, Carroll L, et al. Probabilities of death from breast cancer and other causes among female breast cancer patients. J Natl Cancer Inst 2004;96:1311e21.
Contents lists available at ScienceDirect
Clinical Radiology journal homepage: www.elsevierhealth.com/journals/crad
Review
Breast screening policy: Are we heading in the right direction? A. Evans*, P. Whelehan Dundee Cancer Centre, Ninewells Hospital and Medical School, University of Dundee, Dundee, UK
art icl e i nformat ion Article history: Received 11 October 2010 Received in revised form 25 February 2011 Accepted 2 March 2011
There is a large body of evidence supporting 2-yearly screening of women aged 50e69 years. There is good evidence for a mortality reduction from mammographic screening in women aged 40 to 49 years but a 1-year interval is required. The lack of specificity of screening in young women does remain a problem. There is no evidence to suggest that a single screen between the ages of 47 and 50 years within a programme screening at 3-year intervals will reduce mortality; the trials showing a mortality benefit in women in their 40s included multiple screening episodes and shorter screening intervals. There is no randomized, controlled trial evidence to support screening in women aged above 70 years and screening this age group will cause greater harm than in younger women through higher rates of overdiagnosis and consequent over-treatment. The randomized phase of the screening age extension, which at the moment is planned to last only 6 years, should not be immediately followed by general implementation of the policy. Only if and when additional mortality reductions and an acceptable balance between benefit and harms are shown to be achieved by the extra screens should the 2007 Cancer Reform Strategy policy on age extension be implemented. Resources saved by delaying or abandoning the roll-out of the age extension could potentially be redirected towards reducing the current 3 year screening interval to 2 years in women aged 50e69 years. However, reducing the screening interval to 2 years for women aged 50e69 years would require significantly more screening invitations and resources than the proposed age extension. Ó 2011 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.
Introduction Breast screening policy has recently been the subject of heightened debate in the UK following the publication of the previous government’s Cancer Reform Strategy (CRS),1 and in the US following the US Preventive Services Task Force (USPSTF) 2009 recommendations for breast screening.2,3 The recommendations for breast cancer
DOI of original article: 10.1016/j.crad.2011.06.006. * Guarantor and correspondent: A. Evans, Mail box 4, Ninewells Hospital and Medical School, University of Dundee, Dundee DD1 9SY, Scotland, UK. Tel.: þ44 (0) 1382 632196; fax: þ44 (0) 496363. E-mail address: [email protected] (A. Evans).
screening are disconcertingly divergent in the two documents. The different approaches can, in part, be explained by the pre-existing disparity in screening practice between the two countries and perhaps also by the methodologies used to formulate such policies and recommendations.
UK and US policies In the UK the CRS was published in 2007 by the Department of Health and it stated that there would be an extension of routine breast screening to nine rounds between the ages of 47 and 73 years with the guarantee of a first mammogram before age 50 and the concurrent implementation of digital mammography. No change to the
0009-9260/$ e see front matter Ó 2011 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.crad.2011.03.024
916
A. Evans, P. Whelehan / Clinical Radiology 66 (2011) 915e919
3-year screening interval was made. In the US, the 2009 USPSTF recommendations caused a public outcry by recommending against routine provision of a screening service below age 50da change to its previous recommendation and contrary to widespread practice. The USPSTF in 2009 further recommended that biennial screening by mammography in the 50e74 year age group should be provided but that evidence was lacking on screening after age 74.
Rationale for UK policy In the UK, the CRS breast screening policy statements were formulated following consultation with an expert advisory panel. The document does not describe the methodology used to arrive at the recommendations and no direct reference is made to any published evidence underlying the policy statements on breast cancer screening. Following publication, concern regarding the effectiveness of the age extensions led to the adoption of a proposal from the University of Oxford to randomize the extension in order to facilitate more robust future evaluation of screening effectiveness in the new age bands. Invitation batches are, therefore, randomized either to include women aged 47e49 or aged 71e73 years. The portion of each age cohort not being invited constitutes a control group. Five pilot sites began the randomization project and other sites are now joining following demonstration of feasibility.
Rationale for US policy The USPSTF 2009 recommendation to screen with biennial mammography in the 50e69 year age group is based largely on a published study using six mathematical models developed independently within the Cancer Intervention and Surveillance Modeling Network of the US National Institutes of Health (NIH).4 These models evaluated 20 screening strategies in which screening was performed either annually or biennially with starting ages ranging from 40 to 60 years and cessation ages ranging from 69 to 84 years. The work suggested that biennial screening maintained 81% of the benefit of annual screening but reduced the false-positive results by almost half. Screening women aged 50e69 biennially achieved a median breast cancer mortality reduction across the different mathematical models of 16.5%. Initiation of biennial screening at age 40 reduced mortality by an additional 3%. Screening women aged 70e79 years was associated with a 7% additional mortality reduction but with higher rates of over-diagnosis than in the younger age groups. (Over-diagnosis is a term commonly used to describe detection of cancers that would not have become clinically important during a person’s remaining lifespan d an inevitable consequence of screening but nonetheless important to minimize.) When the outcome measure was total life years gained, rather than mortality rate, the benefits were greater for extending screening to younger rather than older women. The 2009 recommendation of the USPSTF not to support routine screening of women in their 40s despite this evidence is
based on their opinion that mortality reduction is a more important endpoint than life years gained. Their recommendation statement was modified the month after issue to make their position less emphatic and to support individualized decision-making on when to begin screening. None of the modelled screening strategies comprised annual screening for younger women and biennial screening for older women. This is surprising since the correct screening interval is determined by the lead time (the time between mammographic detection and when clinical detection would have been possible) achieved by the screening test. It is well established that that the sojourn times (the time between the possibility of detection by mammography and the clinical presentation) and subsequent lead times achieved by screening are longer for older women and shorter for younger women. The reported sojourn time for screening women aged 40e49 is 1.7e2.7 years, for 50e59 is 3.1e3.7 years and for 60e69 is 3.8e4.2 years.5e7
Outstanding controversies Current debate remains focussed on which age groups to screen and how frequently, and the two questions are interrelated. In general, screening policy has been largely based on the findings of the randomized, controlled trials (RCTs) of screening, particularly high-quality trials conducted in Sweden.8e12 However, none of the Swedish trials was designed to examine the mortality benefit of screening in a particular age group. The combined Swedish results stratified in 5 year age bands according to age at randomization show significant mortality reductions in the age bands 55e59 [relative risk (RR) 0.76 (0.59e0.98)], 60e64 [RR 0.68 (0.50e0.92)], and 65e69 [RR 0.69 (0.49e0.96)].11 There is little controversy in the mainstream of opinion about the value of screening women aged 50e69. However, the appropriate screening interval for women of this age group remains open to debate.
Screening interval from age 50 years The lead time for screening is age dependent, and in women from age 50, mammographic screening achieves a lead time of between 3 and 5 years.13,14 Two-yearly mammographic screening is likely to be appropriate in women from age 50, as indicated by a recent analysis of pooled interval cancer data from six European countries. Interval cancer rates of 29% of the expected breast cancer incidence without screening in year 1 and 63% in year 215 indicate that 2-yearly is the maximum screening interval that is appropriate. A recent analysis of interval cancer rates in the NHSBSP gave very similar results.16 In the US the debate on screening intervals is focused on whether 1-yearly or 2-yearly screening is preferable. The apparent continued acceptance in the UK of its 3-year interval is surprising. The UK Co-ordinating Committee on Cancer Research (UKCCCR) breast screening frequency trial17 compared the prognostic features of invasive cancers detected by three annual screens following a prevalent attendance with those in a control group who had
A. Evans, P. Whelehan / Clinical Radiology 66 (2011) 915e919
a screening invitation after the standard 3-year interval. The study compared predicted mortality, not observed mortality. The predicted RR of breast cancer death in the study arm versus the control arm was 95% or 89% using two different prognostic indices but the differences were not statistically significant. The study had a number of important limitations. It was powered to detect a 25% reduction in breast cancer mortality in the study group compared with the control group. This was unrealistic as the breast cancer mortality achieved by screening versus no screening is of this magnitude. Screening was by single view only at a time when screening quality was lower than currently. One of the prognostic indices used (The Nottingham Prognostic Index) was derived from symptomatic patients and was not appropriate for predicting the survival of small screendetected cancers in the frequency trial.18,19 Approximately 20% of the cancers in each arm were of unknown nodal status, making prognostication unreliable. The effect of increased ductal carcinoma in situ (DCIS) detection in the study arm was not taken into account as the analysis was based solely on invasive cancers. These limitations mean the results of the frequency trial should not be used to defend the UK’s 3-year screening interval.
Screening women younger than 50 A meta-analysis by Hendrick et al. in 199720 of women aged 40e49 in all RCTs showed a statistically significant mortality reduction of 18% [RR 0.82 (0.71e0.95)]. The Swedish overview results published in 2003 did not show a statistically significant benefit for women in this same age group.11 This may have been because the results of the Kopparberg portion of the two-county trial were not € and the included in the later analysis. Both the Malmo Gothenberg trials demonstrated a statistically significant mortality reduction in women in this age group [RR 0.64 (0.45e0.89); p ¼ 0.009 and RR 0.56 (0.32e0.98); p ¼ 0.035, respectively)12,21; both studies incorporated short screening intervals (18 months for women with dense breasts) and multiple screening rounds (at least five). A physical examination was not included in either of these trials. Critics of mammographic screening have commented that many of the women in the Swedish RCTs aged 40e49 at the time of randomization received many of their screening € study, 47% of the episodes after age 50. In the Malmo screening episodes of women randomized younger than 50 occurred after age 50. In the Gothenberg study, 31% of the cancers in the study group were diagnosed after age 50. The Gothenberg results have since been re-analysed according to age at cancer diagnosis, suggesting mortality reductions between 31% and 44% in women aged 39e49 diagnosed during the screening period, depending on the precise mortality endpoint.12 The UK Age Trial was the only RCT to perform all screening episodes before age 50.22 Participants aged 39e41 were randomly selected and were screened annually: two views at their first screening and one view at subsequent screens. There was no clinical examination. The RR of breast cancer death in this study group was 0.83 (0.66e1.04). A major strength of the UK Age Trial is that the
917
data are not contaminated by screening episodes after age 50. The limitations of the study include the fact that only one mammographic view was performed after the first round of screening and that film/screen mammography was used and not full-field digital mammography, which has subsequently been found to be more sensitive in denser breasts and younger women.23 Mammographic screening in women in their 40s has long been a point of debate. Critics highlight the fact that women in their 40s have only half the breast cancer incidence of women in their 50s; however, advocates of screening younger women point out that 34% of total life years lost to breast cancer are from cancers diagnosed between ages 40 and 49.24 Although breast cancer incidence increases with age, the proportion of female deaths attributable to breast cancer is highest in women in their 40s and 50s (16 and 14%, respectively).25 Several problems are encountered during screening of women younger than 50, including relatively high numbers of false-positive screening results and screening-provoked benign surgical biopsies.26 There is a modest reduction in mammographic sensitivity when comparing women in their 40s with women in their 50s27; however, the widespread introduction of digital mammography should ameliorate this problem, according to the results of the Digital Mammographic Imaging Screening Trial.23 More serious issues related to screening women younger than 50 are the high rate of interval cancers and the short duration of time between a normal screen and return to normal background incidence.28,29 These findings indicate that the lead time of screening in younger women is short (between 18 and 24 months); therefore, screening of women aged 40e49 requires a short screening interval, ideally 1 year.
Screening women older than 70 Because of the high incidence of breast cancer in women older than 70 and because mammographic sensitivity and specificity are high, screening is attractive in theory. However, there are a number of problems associated with screening for this age group. Competing causes of mortality make breast cancer proportionally a less common cause of death in older compared with younger women.30 Only 4% of all deaths in England and Wales in women aged 70e79 and 2% in women aged 80e89 are due to breast cancer, compared with 16% in women in their 40s.25 The Swedish RCTs included only 5000 women older than 70 and showed a non-significant 18% mortality excess in the screening arm.11 A recent evaluation of service screening in Northern Sweden showed significant mortality reductions for women aged 40e49 [RR 0.64 (0.43e0.97)] and women aged 50e69 [RR 0.70 (0.54e0.91)], but no mortality benefit was seen for women age 70e74 [RR 1.08 (0.58e2.03)].31 The shorter life expectancy and less aggressive tumour profile means that over-diagnosis becomes a larger problem when screening women older than 70 compared with younger women. In a recent trial of extended adjuvant endocrine therapy, 72% of deaths in women older than 70 with breast cancer were non-breast cancer deaths.32 Data from the American
918
A. Evans, P. Whelehan / Clinical Radiology 66 (2011) 915e919
Surveillance Epidemiology and End Results (SEER) programme show that in women aged over 70 diagnosed with localized invasive breast cancer, 86% of deaths at the end of follow-up had resulted from causes other than breast cancer.33 Therefore, it is unclear whether screening women older than 70 is beneficial. A particular issue in this age group is the over treatment of low-risk DCIS lesions, which are unlikely to progress to invasive cancer in a woman’s lifetime. The proposed UK trial of watchful waiting versus conventional treatment is welcomed. A large RCT in this age group has not been conducted.
Conclusions and recommendations There is a large body of evidence supporting 2-yearly screening of women aged 50e69. There is good evidence for a mortality reduction from mammographic screening in women aged 40e49 but a 1-year interval is required. The lack of specificity of screening in young women does remain a problem. Good-quality RCTs have not been performed to assess the benefit of screening for women older than 70. Although screening of women in this age group provides a high cancer yield with good sensitivity and specificity, any mortality benefit is likely to be modest because of competing causes of mortality; therefore, the value of screening women in this age group remains questionable. What of screening policy in the UK? There is no evidence to suggest that a single screen between the ages of 47 and 50 within a programme screening at 3-year intervals will reduce mortality; the trials showing a mortality benefit in women in their 40s included multiple screening episodes and shorter screening intervals. It is true that a stated aim of the lower extension is that women receive their first screen by their 50th birthday. We agree that women receiving their first screen within a year of their 50th birthday would be a good thing. Screening women from age 47 with a 3 year interval is something quite different. Efforts could perhaps have been focussed on ensuring women are first invited between their 49th and 51st birthdays. Although changing organizational systems to achieve this may be difficult, overcoming these difficulties may have been more cost-effective than screening women 3-yearly from age as low as 47. There is no RCT evidence to support screening in women aged above 70 and screening this age group will cause greater harm than in younger women through higher rates of over-diagnosis and consequent over-treatment. The randomized phase of the screening age extension, which at the moment is planned to last only 6 years, should not be immediately followed by general implementation of the policy. Instead, once sufficient accrual has taken place in line with the power calculations for the study, extending the length of the study if necessary to ensure conclusive results, the randomized women should be followed up and the results analysed. Only if and when additional mortality reductions and an acceptable balance between benefit and harms are shown to be achieved by the extra screens should the 2007 CRS policy on age extension be implemented. Resources saved by delaying or abandoning the roll-out of
the age extension could potentially be redirected towards reducing the current 3 year screening interval to 2 years in women aged 50e69. However, reducing the screening interval to 2 years for women aged 50e69 would require significantly more screening invitations and resources than the proposed age extension.
References 1. Department of Health. Cancer reform strategy. London: DoH; 2007. 2. U.S. Preventive Services Task Force. Breast cancer screening. Available from: http://www.uspreventiveservicestaskforce.org/uspstf/uspsbrca.htm; 2009 [accessed 7 October 2010]. 3. Calonge N, Petitti DB, DeWitt TG, et al. Screening for breast cancer: U.S. preventive services task force recommendation statement. Ann Intern Med 2009;151:716e26. 4. Mandelblatt JS, Cronin KA, Bailey S, et al. Effects of mammography screening under different screening schedules: model estimates of potential benefits and harms. Ann Intern Med 2009;151:738e47. 5. Tabar L, Fagerberg G, Chen HH, et al. Efficacy of breast cancer screening by age: new results from the Swedish two-county trial. Cancer 1995;75: 2507e17. 6. Duffy SW, Day NE, Tabar L, et al. Markov models of breast tumor progression: some age-specific results. J Natl Cancer Inst Monogr 1997; 22:93e7. 7. Shen Y, Zelen M. Screening sensitivity and sojourn time from breast cancer early detection clinical trials: mammograms and physical examinations. J Clin Oncol 2001;19:3490e9. 8. Andersson I, Aspegren K, Janzon L, et al. Mammographic screening and mortality from breast cancer: the Malmo mammographic screening trial. BMJ 1988;297:943e8. 9. Tabar L, Fagerberg G, Duffy SW, et al. Update of the Swedish two-county program of mammographic screening for breast cancer. Radiol Clin North Am 1992;30:187e210. 10. Frisell J, Lidbrink E, Hellstrom L, et al. Followup after 11 years d update of mortality results in the Stockholm mammographic screening trial. Breast Cancer Res Treat 1997;45:263e70. 11. Nystrom L, Andersson I, Bjurstam N, et al. Long-term effects of mammography screening: updated overview of the Swedish randomised trials. Lancet 2002;359:909e19. 12. Bjurstam N, Bjorneld L, Warwick J, et al. The Gothenburg Breast Screening Trial. Cancer 2003;97:2387e96. 13. Paci E, Duffy S, Giorgi D, et al. Population-based breast cancer screening programmes: estimates of sensitivity, overdiagnosis and early prediction of the benefit. In: Duffy S, Hill C, Esteve J, editors. Quantitative methods for the evaluation of cancer screening. London: Arnold; 2001. p. 127e35. 14. Duffy SW, Gabe R. What should the detection rates of cancers be in breast screening programmes? Br J Cancer 2005;92:597e600. 15. Tornberg S, Kemetli L, Ascunce N, et al. A pooled analysis of interval cancer rates in six European countries. Eur J Cancer Prevent 2010;19: 87e93. 16. Bennett RL, Sellars SJ, Moss SM. Interval cancers in the NHS breast cancer screening programme in England Wales and Northern Ireland. Br J Cancer 2011;104:571e7. 17. The Breast Screening Frequency Trial Group. The frequency of breast cancer screening: results from the UKCCCR Randomised Trial. Eur J Cancer 2002;38:1458e64. 18. Anderson TJ, Alexander FE, Lamb J, et al. Pathology characteristics that optimize outcome prediction of a breast screening trial. Br J Cancer 2000;83:487e92. 19. Evans AJ, Rakha EA, Pinder SE, et al. Basal phenotype: a powerful prognostic factor in small screen-detected invasive breast cancer with long-term follow-up. J Med Screen 2007;14:210e4. 20. Hendrick RE, Smith RA, Rutledge 3rd JH, et al. Benefit of screening mammography in women aged 40e49: a new meta-analysis of randomized controlled trials. J Natl Cancer Inst Monogr 1997;22:87e92. 21. Andersson I, Janzon L. Reduced breast cancer mortality in women under age 50: updated results from the Malmo Mammographic Screening Program. J Natl Cancer Inst Monogr 1997;22:63e7.
A. Evans, P. Whelehan / Clinical Radiology 66 (2011) 915e919 22. Moss SM, Cuckle H, Evans A, et al. Effect of mammographic screening from age 40 years on breast cancer mortality at 10 years’ follow-up: a randomised controlled trial. Lancet 2006;368: 2053e60. 23. Pisano ED, Gatsonis C, Hendrick E, et al. Diagnostic performance of digital versus film mammography for breast-cancer screening. New Engl J Med 2005;353:1773e83. 24. Moss S. Should women under 50 be screened for breast cancer? Br J Cancer 2004;91:413e7. 25. Office for National Statistics. Mortality statisticsdcause: review of the registrar general on deaths by cause, sex and age. In: England and Wales. London: HMSO; 2005. 2006. 26. Johns LE, Moss SM. Randomized controlled trial of mammographic screening from age 40 (“Age” trial): patterns of screening attendance. J Med Screen 2010;17:37e43. 27. Rosenberg RD, Hunt WC, Williamson MR, et al. Effects of age, breast density, ethnicity, and estrogen replacement therapy on screening mammographic sensitivity and cancer stage at diagnosis: review of 183,134 screening mammograms in Albuquerque, New Mexico. Radiology 1998;209:511e8.
919
28. Tabar L, Faberberg G, Day NE, et al. What is the optimum interval between mammographic screening examinations? An analysis based on the latest results of the Swedish two-county breast cancer screening trial. Br J Cancer 1987;55:547e51. 29. Bucchi L, Ravaioli A, Foca F, et al. Incidence of interval breast cancers after 650,000 negative mammographies in 13 Italian health districts. J Med Screen 2008;15:30e5. 30. Heron M, Hoyert DL, Murphy SL, et al. Deaths: final data for 2006. National vital statistics reports: from the Centers for Disease Control and Prevention, National Center for Health Statistics, National Vital Statistics System 2009;57:1e134. 31. Jonsson H, Bordas P, Wallin H, et al. Service screening with mammography in Northern Sweden: effects on breast cancer mortalitydan update. J Med Screen 2007;14:87e93. 32. Chapman JAW, Meng D, Shepherd L, et al. Competing causes of death from a randomized trial of extended adjuvant endocrine therapy for breast cancer. J Natl Cancer Inst 2008;100:252e60. 33. Schairer C, Mink PJ, Carroll L, et al. Probabilities of death from breast cancer and other causes among female breast cancer patients. J Natl Cancer Inst 2004;96:1311e21.