Predictions of survival up to 10 years after diagnosis for European women with breast cancer in 2000–2002
Article first published online: 7 NOV 2012
Copyright © 2012 UICC
International Journal of Cancer
Volume 132, Issue 10, pages 2404–2412, 15 May 2013
How to Cite
Allemani, C., Minicozzi, P., Berrino, F., Bastiaannet, E., Gavin, A., Galceran, J., Ameijide, A., Siesling, S., Mangone, L., Ardanaz, E., Hédelin, G., Mateos, A., Micheli, A., Sant, M. and the EUROCARE Working Group (2013), Predictions of survival up to 10 years after diagnosis for European women with breast cancer in 2000–2002. Int. J. Cancer, 132: 2404–2412. doi: 10.1002/ijc.27895
- Issue published online: 28 FEB 2013
- Article first published online: 7 NOV 2012
- Accepted manuscript online: 10 OCT 2012 04:15AM EST
- Manuscript Accepted: 3 SEP 2012
- Manuscript Received: 27 JAN 2012
- Compagnia di San Paolo, Torino, Italy; Public Health Agency for Northern Ireland
- breast cancer;
- long-term survival;
Few studies have addressed longer-term survival for breast cancer in European women. We have made predictions of 10-year survival for European women diagnosed with breast cancer in 2000–2002. Data for 114,312 adult women (15–99 years) diagnosed with a first primary malignant cancer of the breast during 2000–2002 were collected in the EUROCARE-4 study from 24 population-based cancer registries in 14 European countries. We estimated relative survival at 1, 5, and 10 years after diagnosis for women who were alive at some point during 2000–2002, using the period approach. We also estimated 10-year survival conditional on survival to 1 and 5 years after diagnosis. Ten-year survival exceeded 70% in most regions, but was only 54% in Eastern Europe, with the highest value in Northern Europe (about 75%). Ten-year survival conditional on survival for 1 year was 2–6% higher than 10-year survival in all European regions, and geographic differences were smaller. Ten-year survival for women who survived at least 5 years was 88% overall, with the lowest figure in Eastern Europe (79%) and the highest in the UK (91%). Women aged 50–69 years had higher overall survival than older and younger women (79%). Six cancer registries had adequate information on stage at diagnosis; in these jurisdictions, 10-year survival was 89% for local, 62% for regional and 10% for metastatic disease. Data on stage are not collected routinely or consistently, yet these data are essential for meaningful comparison of population-based survival, which provides vital information for improving breast cancer control.
Breast cancer is the most common cause of death from cancer in women worldwide1 and it is the most common cancer in European women.2, 3 Over the last 20 years, several population-based studies have reported on 5-year relative survival for breast cancer in European women,4–10 but few studies have addressed longer-term survival for European breast cancer patients11–13 and most of these were concerned with specific populations rather than Europe as a whole. For many cancers, 5-year survival can give an indication of the success of treatment, but women with breast cancer continue to suffer recurrences for 10 years and more after treatment.14 For this reason, long-term survival for breast cancer may provide a better indication of the success of treatment than 5-year survival.
The aim of this study was to assess survival up to 10 years after diagnosis for European women with breast cancer in 2000–2002, the latest period for which Europe-wide data are available. We were also interested in long-term survival in the age range usually targeted by screening programmes (50–69 years), and by stage at diagnosis.
Long-term cancer survival monitoring is essential for providing policy-makers and health-care planners with accurate data for breast cancer control, particularly in the light of improvements in prognosis that occurred during the 1990s.15
Material and Methods
We analyzed population-based data from the EUROCARE study on the survival of women diagnosed with breast cancer in Europe between 1978 and 2002 and followed up to the end of 2003. We used the period approach16, 17 to estimate long-term survival in the absence of complete follow-up. This approach allows us to predict 10-year survival for women diagnosed in 2000–2002, even though most European cancer registries could only provide follow-up data to the end of 2003. The EUROCARE data and the methods used to check data quality and completeness have been described elsewhere.18, 19
We analyzed data from 24 cancer registries in 14 European countries for 114,312 adult women (15–99 years) who were diagnosed during 2000–2002 with a first primary malignant cancer of the breast. The period approach enables us to predict long-term survival for these patients by using the survival experience of patients who were diagnosed in earlier years and who were alive at some point during 2000–2002. Only these 24 registries could provide the high-quality follow-up data from 1991 to 2002 required to enable long-term period survival estimation for cases alive at some point during 2000–2002. We excluded patients whose cancer was registered from a death certificate (DCO, death certificate only) or discovered at autopsy.
The 24 jurisdictions were grouped into five regions: Northern Europe (Finland, Iceland, Norway, Sweden); United Kingdom (England: East Anglia, Northern and Yorkshire, Oxford, West Midlands; Scotland, Wales); Western Europe (The Netherlands: Amsterdam; Switzerland: Geneva, St. Gallen); Eastern Europe (Czech Republic: West Bohemia; Poland: Cracow, Warsaw; Slovakia); and Southern Europe (Italy: Ferrara, Firenze, Modena, Parma, Ragusa, Romagna; Slovenia).
Age at diagnosis was grouped into four categories: 15–39, 40–49, 50–69, and 70–99 years. Women older than 99 years at diagnosis (51 cases, 0.04%) were not included.
We estimated survival by stage at diagnosis only for the six registries in which it was known for at least 85% of patients (Amsterdam, Finland, Geneva, Norway, Slovenia, Warsaw; 29,126 women, Table 1). Stage was grouped into the categories specified in the EUROCARE-4 protocol: localized cancer (confined to breast); regional extension (spread to regional lymph nodes) and metastatic (tumor spread to distant organs). Two hybrid categories of stage were included in the protocol: local or regional cancer (tumor not confined to site of origin, but unspecified whether spread was to immediately adjacent tissues or to regional lymph nodes), and regional or metastatic cancer (not specified whether spread was to regional nodes or distant organs). In the six registries used for stage-specific survival analyses, however, only 31 cases (0.11%) and 256 cases (0.91%), respectively, fell into these two hybrid categories; they were analyzed with patients for whom the stage was unknown.
Unfortunately, despite the 95.3% completeness of breast cancer staging in East Anglia (England), these data had to be excluded from stage-specific analyses because most patients had been categorized in one of the hybrid stage categories. During the period under study, the cancer registry had been using the WHO/UICC TNM staging system (5th edition). The TNM staging system is more detailed and informative than the categorization of extent of disease (local, regional, metastatic) to determine prognosis and assist clinicians in planning treatment, but TNM Stages II and III categories for breast cancer both include tumors with local and regional extent, so it was not possible to convert TNM breast stage to the extent of disease scale.
We estimated relative survival, a measure of the probability of cancer survival after adjustment for competing causes of death. Relative survival is defined as the ratio of the survival observed among the cancer patients and the survival that would have been expected if they had experienced the same death rates as the general population from which they derive. Expected survival was estimated from population life tables by the Ederer II method.20, 21 Each cancer registry provided all-cause mortality data by age, sex and calendar year (or period) for the population resident in its territory. These data were used to construct life tables that were extended and interpolated, as necessary, to obtain the probabilities of death at each year of age (0–99 years), by sex, for each calendar year between 1978 and 2004 (www.eurocare.it). We estimated 1-, 5- and 10-year relative survival for women who were alive at some point during 2000–2002 by the period approach as the product of interval-specific relative survival values for cohorts of patients diagnosed in earlier years.16, 17
We also estimated relative survival at 10 years, conditional on survival for 1 year, as the cumulative 9-year relative survival for women alive and under follow-up at the end of the first follow-up year. Similarly, relative survival at 10 years conditional on survival for 5 years was estimated as the cumulative 5-year relative survival for women alive and under follow-up at the end of the fifth follow-up year.
Overall (all-ages) relative survival estimates were age-standardized with the international cancer survival standard (ICSS),22 which employs standard age distributions that differ with the age-incidence pattern of the cancer.
Data quality was generally good. Only 1.5% (1,726) of all breast cancers were registered from a death certificate only (DCO) and only in Slovakia (9.7%) was the value higher than 5%. Less than 0.1% (99) of cancers were discovered at autopsy. Overall, only 0.1% of women (160) were lost to follow-up, and in most registries this quality indicator did not exceed 1%; the exception was Geneva (3.6%) (data not shown). In most cancer registries, information on stage at diagnosis was available for <85% of cases.
Almost half (49%) of the women were diagnosed in the age range 50–69 years, about one-third (31%) aged 70–99 years, and only about one-fifth under the age of 50 (5% aged 15–39 years, 16% aged 40–49 years) (Table 1). The age distribution varied somewhat both within and between European regions, with 51% aged 50–69 in Northern Europe, and almost 20% aged 40–49 years in Eastern Europe.
Among the six registries for which information on stage was available for at least 85% of cases; only 6.4% of the 29,126 cases were of unknown stage, ranging from 0.1% in Norway to 14.8% in Finland. Overall, half the tumors (49.9%) were confined to the breast, 38.2% with regional spread and only 5.5% were diagnosed with metastases. The proportion of localized tumors was rather low in Warsaw (44.2%). The proportion of tumors with regional spread was above 40% in Norway (42.9%), the Netherlands (41.8%) and Slovenia (40.6%). The proportion of metastatic tumors varied from 4.4% in Finland to 7.6% in Warsaw and 8.8% in Slovenia.
Age-standardized relative survival at 1 year was about 95% in most European areas but lower in Eastern Europe (91%; Table 2). Survival at 5 years was 81% overall, in the range 83–85% in Northern, Western and Southern Europe, 79% in the UK and 69% in Eastern Europe. Survival 10 years after diagnosis was 71% overall, in the range 73–75% in Northern and Western Europe, 71% in the UK and Southern Europe, but only 54% in Eastern Europe.
Ten-year relative survival conditional on surviving 1 year was about 75%, some 2–6% higher than overall 10-year survival in all European areas, and inter-regional differences were narrower. Ten-year survival among women who had survived 5 years was about 88% overall, with 88% in Northern Europe and 91% in the UK, but only 79% in Eastern Europe.
Ten-year relative survival for women alive at some point during 2000–2002 was highest for women aged 50–69 years except in Eastern Europe (Fig. 1a). The overall value was 77.2% (95%CI 76.7–77.7), with 10-year survival above 80% in Finland, Iceland, Sweden and Geneva (Switzerland), and in Firenze, Modena, Parma and Romagna (Italy). Survival in this age group was above 75% in Norway, in the UK registries of Northern and Yorkshire, West Midlands, Scotland and Wales, and in Amsterdam (Netherlands) and Ferrara (Italy) (data not shown).
In Northern, Western and Southern Europe and the UK, survival increased with increasing age from 15–39 to 50–69 years then decreased sharply. In Eastern Europe, survival was about 65% from 15–39 to 40–49 years, declined in the 50–69 year age group, and was exceptionally low (47.0%, 95%CI 42.5–51.6) in the oldest age group. Age-related survival differences were marked in Northern Europe (66.9% at 15–39 years; 82.3% at 50–69 years) and in Western Europe (62.7% at 15–39 years; 77.8% at 50–69 years) but less so in Southern Europe (66.9% at 15–39 years; 78.1% at 50–69 years) and the UK (66.2% at 15–39 years; 76.3% at 50–69 years).
The age patterns for 10-year survival conditional on surviving 1 year were similar to those for overall 10-year survival, but with somewhat smaller differences between European regions (Fig. 1b). These differences were even smaller for 10-year survival conditional on survival to 5 years (Fig. 1c). In each age group, the differences between overall and conditional 10-year survival were greater for women in the UK and Eastern Europe.
Age-standardized 10-year relative survival decreased with advancing stage (Fig. 2): 88.9% (95%CI 87.8–89.9) for cancers confined to the breast, 62.3% (95% CI 61.0–63.6) for cancers with regional spread, and only 9.6% (95%CI 7.9–11.6) for cancers with distant spread when diagnosed. For the small group (6.4%) of women with cancers of unknown stage, 10-year survival was intermediate (67.6%; 95%CI 63.4–71.7).
Slovenia was characterized by the lowest 10-year survival at each stage (80.1% localized, 52.9% regional and 6.7% for metastatic cancer) and Geneva by the highest survival (93.9% local, 74.8% regional and 19.0% for metastatic cancer).
To our knowledge this is the first Europe-wide study to provide data on long-term survival in breast cancer patients according to age classes that are sufficiently narrow to focus on the age group targeted for breast cancer screening. Brenner et al.12 analyzed 10-year period survival in European women in much broader age categories (15–59, 60–74, 75–99 years).
Overall age-standardized 10-year relative survival was above 70% in Northern, Western and Southern Europe and the UK, but considerably lower in Eastern Europe (54.2%). Consistent with previous findings,12 10-year survival was about 10% lower than 5-year survival in all European regions except Eastern Europe, where the absolute difference was about 15%.
The deficit in overall survival in Eastern Europe was even more marked for women aged 70–99 years. Women with breast cancer in Eastern Europe are generally diagnosed at a later stage than in other European regions,25 reducing the chance of cure or long-term survival, as seen for Cracow (Poland) in this study. High proportions of women presenting with advanced disease suggest diagnostic delay, probably reflecting lack of awareness by women and/or physicians about breast cancer, or lack of adequate screening and diagnostic facilities.25, 26
In the late 1990s, breast cancer patients in the Eastern European countries of Slovakia, Estonia and Poland tended to be treated extensively with hormonal therapy even when they did not have oestrogen receptor-positive disease25 and with adjuvant chemotherapy even when they did not have node-positive disease (the major clinical indication for chemotherapy at that time). Breast-conserving surgery plus radiotherapy was also used less than in other European countries for early-stage breast cancer.25 Thus, medication appeared to be the mainstay of treatment in Eastern Europe, perhaps because surgical and radiotherapy facilities were inadequate. Medical treatment may be less expensive than surgery and radiotherapy.25
New high-resolution studies of detailed population-based data on stage and treatment will be required to understand why women diagnosed with breast cancer in Eastern Europe continue to have poorer survival than in other parts of the continent. Perhaps it is simply a question of health expenditure. The proportion of early-stage breast cancer patients treated conservatively in the late 1990s was directly related to total national expenditure on health,25 suggesting that the low expenditure on health in Eastern European countries18 had an important effect on the quality of treatment.
The highest 10-year survival was seen for women aged 50–69 years, the age band targeted for breast cancer screening by many European countries in the 1990s. This finding, together with the concurrent decline in mortality,27 suggests that breast cancer screening may be effective in Europe, even though less than half the minimum recommended numbers of screenings are taking place in the European Union each year (http://www.eubusiness.com/topics/health/cancer-screening.guide/). Survival studies carried out before the screening era usually showed that women aged 40–49 years had the highest survival and that survival decreased progressively from age 50 onward.28 The shift in the age range with the highest long-term survival may be partly attributable to the introduction of breast screening in some countries, but we do not have data on compliance with screening in each territory, or data on whether each woman had been screened.
Ten-year survival was lowest for women aged 70–99 years at diagnosis in all European regions, a finding also reported by a Dutch study for all disease stages.29 Elderly patients are usually diagnosed with more favorable tumor biology, but are likely to present with more advanced disease and are less likely to be treated with standard protocols. Treatment for elderly breast cancer patients is usually less aggressive than for younger women—perhaps reflecting the higher prevalence of comorbidity in the elderly—and this may partly explain their lower survival, even after adjustment for other causes of death. Less aggressive treatment may particularly affect the survival of elderly patients who are fit, and who would have a longer life expectancy if their breast disease were treated adequately. A recent trial suggested that adjuvant chemotherapy can improve survival in fit elderly breast cancer patients and that standard chemotherapy was superior to the oral agent capecitabine.30
In all regions except Eastern Europe, survival for women aged 15–39 years was similar to survival for the elderly. Ten-year survival conditional on surviving 1 year in these young women was lower than in the elderly (again except for Eastern Europe).
Breast cancer is generally more aggressive in younger than older women,31 but screening for women under 50 years is not known to reduce mortality. For women under 40, the public health priority should therefore be to support research for primary prevention, and to offer genetic counselling for women at high risk.
Consistent with previous analyses,7, 12, 32 we found that survival for UK breast cancer patients had increased considerably compared with earlier periods, so that the survival trend by age for women in the UK is similar to that for women in Northern, Western and Southern Europe. Our survival estimates refer to 2000–2002, but since we used the period method to predict survival up to 10 years, the findings also incorporate the longer-term survival of patients who were diagnosed and treated prior to 2000. The observed improvement, therefore, cannot be a consequence of the various cancer plans launched in the UK, even though the earliest, the Calman–Hine report,33 was launched in 1995. The marked increase in breast cancer survival over time for UK breast cancer patients18, 34 accompanied by declining mortality, suggests real improvements in the care of UK breast cancer patients in recent years.35
Notwithstanding these improvements, however, 5-year survival and, to a lesser extent, 10-year survival in the UK was still somewhat lower than in Northern and Western European countries. By contrast, 10-year survival conditional on having survived the first 5 years was higher in the UK than in the other European regions, particularly for the elderly.
This intriguing pattern could suggest that breast cancer recurrences in the UK were sometimes postponed beyond 5 years by more extensive use of chemotherapy, but the evidence suggests that, if anything, chemotherapy was used less frequently in the UK than elsewhere in the 1990s.36 We cannot exclude that this result may be due to women who survive up to 5 years in the UK being really healthy survivors (to a greater extent than in other countries). Less complete follow-up to ascertain vital status in the UK could also be a factor, but >99% of cancer patients are reliably traced for their vital status (http://www.ukacr.org/content/data-quality). In most European cancer registries, follow-up is performed by matching against a death registry, in some cases complemented by active tracing.
Unfortunately, data on stage at diagnosis were available for at least 85% of cases in only six of the 24 participating European registries, so it was not possible to perform Europe-wide analyses of breast cancer survival by stage. In fact stage information is still not routinely collected by many EUROCARE cancer registries as shown by a recent EUROCHIP/ENCR/EUROCOURSE survey (http://www.tumori.net/eurochip/material/WP5/EUROCHIP3_WP5_Report.pdf).
The national cancer registry of Slovenia provided usable information on stage for almost all cases, but survival was low in all stage categories. In principle, this could be due to late stage at diagnosis, less adequate staging procedures or less adequate treatment. In 1996–98, 81% of Slovenian breast cancer patients underwent surgery and 78% received axillary lymphadenectomy, in 89% of which at least 10 lymph nodes were examined pathologically.25 Together with the relatively high proportion of metastatic cases, both in this study (2000–2002) and for women diagnosed in 1996–98,25 this suggests that staging is in fact generally accurate in Slovenia. Implementation of screening, launched only in 2008 (http://eu-cancer.iarc.fr/cancer-13-breast-screening.html), and breast cancer awareness programmes would appear to offer the best approach to reducing breast cancer mortality in Slovenia.
Among the registries with adequate information on stage, the Geneva cancer registry had the highest overall and stage-specific survival. However, this registry was also characterized by a high proportion of patients lost to follow-up, possibly because migrants diagnosed with serious illnesses such as cancer tend to return to their country of origin.37, 38 A high proportion of cases lost to follow-up artificially inflates survival because deaths are not notified and these cases are still considered alive by the cancer registry. Cancer registries should endeavor to improve the quality of follow-up for vital status, especially of long-term survivors.
Our analysis cannot provide strong conclusions about stage distribution and stage-specific survival neither across Europe nor across the six European registries providing data on stage, but should be considered an attempt to describe the availability of this information in the European cancer registries. Furthermore even when stage is specified we cannot be sure about the comparability of stage between countries because the content of each stage category can vary according to staging procedure available in that country to determine the stage (Will Rogers' phenomenon). Data on the staging procedures, such as the number of lymph nodes examined during lymphadenectomy, were not available in this data set thus it was not possible to take into account the stage migration phenomenon.
The collection of information on stage and the procedures used to determine it was not a core requirement for the EUROCARE-4 study, but it is very important to improve the interpretation of survival differences between European regions. High-resolution studies have been useful for interpreting such differences in survival.39
High-resolution studies on large representative samples of patients are still the only practical means of analyzing population-based cancer outcomes in relation to stage, diagnostic procedures and treatment. They can provide vital information for the control of breast cancer. Specific financial support is necessary to enable cancer registries to collect the detailed information required.
This research was supported by the Compagnia di San Paolo, Torino, Italy. The Northern Ireland cancer registry is funded by the Public Health Agency for Northern Ireland. We thank Chiara Margutti for secretarial assistance.
- 2Cancer burden in the year 2000. The global picture. Eur J Cancer 2001; 37 ( Suppl 8): 4–66., , .
- 3Breast Cancer in Europe—ENCR Cancer Fact Sheet. Eur Network Cancer Registries 2002:2; 1–4.Availabl at: http://www.encr.com.fr/, , .
- 4Berrino F, Sant M, Verdecchia A, et al. eds. Survival of cancer patients in Europe: the EUROCARE study (IARC Scientific Publications No. 132). Lyon: International Agency for Research on Cancer, 1995.
- 5Berrino F, Capocaccia R, Estève J, et al., eds. Survival of cancer patients in Europe: the EUROCARE-2 study (IARC Scientific Publications No. 151). Lyon: International Agency for Research on Cancer, 1999.
- 8Cancer survival in Australia, Canada, Denmark, Norway, Sweden, and the UK, 1995–2007 (the International Cancer Benchmarking Partnership): an analysis of population-based cancer registry data. Lancet 2011; 377: 127–38 Available at: http://linkinghub.elsevier.com/retrieve/pii/S0140673610622313, , , et al.
- 9Up-to-date estimates of breast cancer survival for the years 2000–2004 in 11 European countries: the role of screening and a comparison with data from the United States. Eur J Cancer 2010; 46: 3351–7. Available at: http://www.sciencedirect.com/science/article/B6T68–516Y4SV-5/2/620a18b76aaa0db38be362a58b96e23b, , , et al.
- 24StataCorp. STATA statistical software College Station. Texas: Stata Corporation, version 9.1, 2005.
- 33Expert Advisory Group on Cancer. A policy framework for commissioning cancer services (Calman-Hine report). London: Department of Health, 1995.
- 34Cancer trends in England and Wales 1950–1999. Studies on Medical and Population Subjects No. 66. London: Office for National Statistics, 2001., , , et al.
- 38Migration bias in cancer survival rates. J Epid Biostat 1996; 1: 167–73., , , et al.