Survival for eight major cancers and all cancers combined for European adults diagnosed in 1995–99: results of the EUROCARE-4 study

Article (PDF Available)inThe Lancet Oncology 8(9):773-83 · October 2007with155 Reads
DOI: 10.1016/S1470-2045(07)70245-0 · Source: PubMed
Abstract
EUROCARE is the largest population-based cooperative study on survival of patients with cancer. The EUROCARE project aims to regularly monitor, analyse, and explain survival trends and between-country differences in survival. This report (EUROCARE-4) presents survival data for eight selected cancer sites and for all cancers combined, diagnosed in adult (aged >/=15 years) Europeans in 1995-99 and followed up until the end of 2003. We analysed data from 83 cancer registries in 23 European countries on 2 699 086 adult cancer cases that were diagnosed in 1995-99 and followed up to December, 2003. We calculated country-specific and mean-weighted age-adjusted 5-year relative survival for eight major cancers. Additionally, case-mix-adjusted 5-year survival for all cancers combined was calculated by countries ranked by total national expenditure on health (TNEH). Changes to survival were analysed relative to cases diagnosed in 1990-94. Mean age-adjusted 5-year relative survival for colorectal (53.8% [95% CI 53.3-54.1]), lung (12.3% [12.1-12.5]), breast (78.9% [78.6-79.2]), prostate (75.7% [75.2-76.2]), and ovarian (36.3% [35.7-37.0]) cancer was highest in Nordic countries (except Denmark) and central Europe, intermediate in southern Europe, lower in the UK and Ireland, and worst in eastern Europe. Survival for melanoma (81.6% [81.0-82.3]), cancer of the testis (94.2% [93.4-95.0]), and Hodgkin's disease (80.0% [79.0-81.0]) varied little with geography. All-cancer survival correlated with TNEH for most countries. Denmark and UK had lower all-cancer survival than countries with similar TNEH; Finland had high all-cancer survival, but moderate TNEH. Survival increased and intercountry survival differences narrowed between the data for 1990-94 and 1995-99 for, notably, Hodgkin's disease (range 66.1-82.9 [IQR 72.2-78.6] vs 74.0-83.9 [78.6-81.9]), colorectal (29.4-56.7 [45.8-54.1] vs 38.8-59.7 [50.7-57.5]), and breast (61.7-82.7 [72.3-78.3] vs 69.3-87.6 [76.6-82.7]) sites. Increases in survival and decreases in geographic differences over time, which are mainly due to improvements in health-care services in countries with poor survival, might indicate better cancer care. Wealthy countries with high TNEH generally had good cancer outcomes, but those with conspicuously worse outcomes than those with similar TNEH might not be allocating health resources efficiently.
http://oncology.thelancet.com Published online August 21, 2007 DOI:10.1016/S1470-2045(07)70245-0
1
Articles
Survival for eight major cancers and all cancers combined
for European adults diagnosed in 1995–99: results of the
EUROCARE-4 study
Franco Berrino, Roberta De Angelis, Milena Sant, Stefano Rosso, Magdalena B Lasota, Jan W Coebergh, Mariano Santaquilani,
and the EUROCARE Working group*
Summary
Background EUROCARE is the largest population-based cooperative study on survival of patients with cancer. The
EUROCARE project aims to regularly monitor, analyse, and explain survival trends and between-country diff erences
in survival. This report (EUROCARE-4) presents survival data for eight selected cancer sites and for all cancers
combined, diagnosed in adult (aged ≥15 years) Europeans in 1995–99 and followed up until the end of 2003.
Methods We analysed data from 83 cancer registries in 23 European countries on 2 699 086 adult cancer cases that
were diagnosed in 1995–99 and followed up to December, 2003. We calculated country-specifi c and mean-weighted
age-adjusted 5-year relative survival for eight major cancers. Additionally, case-mix-adjusted 5-year survival for all
cancers combined was calculated by countries ranked by total national expenditure on health (TNEH). Changes to
survival were analysed relative to cases diagnosed in 1990–94.
Findings Mean age-adjusted 5-year relative survival for colorectal (53·8% [95% CI 53·3–54·1]), lung (12·3%
[12·1–12·5]), breast (78·9% [78·6–79·2]), prostate (75·7% [75·2–76·2]), and ovarian (36·3% [35·7–37·0]) cancer was
highest in Nordic countries (except Denmark) and central Europe, intermediate in southern Europe, lower in the UK
and Ireland, and worst in eastern Europe. Survival for melanoma (81·6% [81·0–82·3]), cancer of the testis (94·2%
[93·4–95·0]), and Hodgkin’s disease (80·0% [79·0–81·0]) varied little with geography. All-cancer survival correlated
with TNEH for most countries. Denmark and UK had lower all-cancer survival than countries with similar TNEH;
Finland had high all-cancer survival, but moderate TNEH. Survival increased and intercountry survival diff erences
narrowed between the data for 1990–94 and 1995–99 for, notably, Hodgkin’s disease (range 66·1–82·9 [IQR 72·2–78·6]
vs 74·0–83·9 [78·6–81·9]), colorectal (29·4–56·7 [45·8–54·1] vs 38·8–59·7 [50·7–57·5]), and breast (61·7–82·7
[72·3–78·3] vs 69·3–87·6 [76·6–82·7]) sites.
Interpretation Increases in survival and decreases in geographic diff erences over time, which are mainly due to
improvements in health-care services in countries with poor survival, might indicate better cancer care. Wealthy
countries with high TNEH generally had good cancer outcomes, but those with conspicuously worse outcomes than
those with similar TNEH might not be allocating health resources effi ciently.
Introduction
Information necessary for planning health-care responses
to the cancer burden includes: mortality (number of cancer
deaths in a given period in a defi ned population); incidence
(number of new cancer diagnoses in a given period); and
population-based survival (proportion of incident cases
alive at a given time after diagnosis). Population-based
survival, which can be obtained only from population-based
cancer registries, is usually lower than survival calculated
from hospital series or clinical trials because it includes
patients who do not have access to adequate treatment or
who are not eligible for clinical trials. Survival is also
required to estimate the number of patients with a cancer
diagnosis who are alive at any one time (prevalence).
If survival in one country is substantially lower than
that in other countries, especially those of similar wealth,
the health system is probably not functioning as it should.
In-depth studies are advisable to fi nd the reason for this
problem (eg, late diagnosis, inadequate availability, or
poor access to treatment) and to suggest remedies.
Population-based cancer registries have provided
population-based survival statistics since the 1960s.
1
However, the largest coordinated study is EUROCARE,
2–4
a European cancer registry-based project on the survival
and care of patients with cancer; the study aims are to
monitor, analyse, and explain survival trends and
between-country diff erences in survival. EUROCARE
currently includes data from 83 adult and ten childhood
cancer registries in 23 European countries. This report
presents survival data for eight selected cancer sites and
for all cancers combined, in adult (aged ≥15 years)
Europeans diagnosed between 1995 and 1999 and who
were followed up until the end of 2003 (EUROCARE-4).
The eight sites (breast, colorectum, lung, melanoma,
ovary, prostate, testis, and Hodgkin’s disease) were
selected because they are common or are major and
potentially curable diseases.
Previous EUROCARE reports
5,6
showed that survival
correlated with macroeconomic variables. Therefore, in
this report—to assist between-country comparisons—
Published Online
August 21, 2007
DOI:10.1016/S1470-
2045(07)70245-0
See
Online/Refl ection and
Reaction
DOI:10.1016/S1470-
2045(07)70247-4
*Members listed at end of report
Department of Preventive and
Predictive Medicine,
Fondazione Istituto Nazionale
dei Tumori, Milan, Italy
(F Berrino MD, M Sant MD);
National Centre for
Epidemiology, Surveillance and
Health Promotion, Department
of Cancer Epidemiology,
Istituto Superiore di Sanità,
Rome, Italy (R De Angelis BSc,
M Santaquilani); Centre for
Oncological Prevention,
Piedmont Cancer Registry,
Torino, Italy (S Rosso MD);
Independent Unit of
Oncological Education, Maria
Sklodowska-Curie Memorial
Cancer Centre and Institute of
Oncology, Warsaw, Poland
(M B Lasota MD); and
Comprehensive Cancer Centre
South, Eindhoven Cancer
Registry, Zernickestraat,
Netherlands (J W Coebergh MD)
Correspondence to:
Dr Franco Berrino, Department
of Preventive and Predictive
Medicine, Fondazione Istituto
Nazionale dei Tumori, Via
Venezian 1, 20133 Milano, Italy
franco.berrino@istituto
tumori.mi.it
Articles
2
http://oncology.thelancet.com Published online August 21, 2007 DOI:10.1016/S1470-2045(07)70245-0
all-cancer survival has been presented with countries
ranked by mean total national expenditure on health
(TNEH)
7
in 1994–2002.
Methods
Cancer cases
For 13 countries, the entire population was covered by
cancer registration; the other ten countries were
represented by regional cancer registries that covered
variable proportions of the population (table 1). These
countries constituted a large proportion of Europe and
are from now on referred to in this report as Europe. The
important issue of whether mean survival fi gures are
representative of those of Europe as a whole has been
addressed in a previously published report.
8
All cancer registries collected data according to a
standardised protocol
9
and sent them for central analysis
anonymously (month and year of birth, diagnosis, and
death were collected and stored in a dedicated computer
not connected to the internet), so that no ethical approval
for the study was required. The analyses were done (FB,
RDA, MS) centrally at the Istituto Superiore di Sanità,
Rome, and at the Istituto Nazionale dei Tumori, Milan.
Systematic checks were done (FB and MS) to detect
errors, inconsistencies, or unusual combinations of
cancer site, morphology, sex, and age at diagnosis.
Questionable records were sent back to cancer registries
for verifi cation and correction, and non-correctable
records were excluded. Cancer site and morphology were
coded according to the International Classifi cation of
Diseases for Oncology (ICDO), 3rd edition.
10
Only data
from fi rst diagnosed malignancies in any patient were
included.
Statistical analyses
The main survival indicator used by EUROCARE is
relative survival—the ratio of the recorded survival to the
expected survival in the general population of the same
age and sex. Relative survival is used widely to eliminate
the eff ect of competing causes of mortality, and assists
Proportion of national
population covered by
cancer registration in
1998, %
Total
number of
malignant
cancers
Number of
cases
excluded for
major errors
Proportion excluded from survival analyses, % Number of primary
malignant cancers
included in survival
analyses
Proportion
censored after
<5 years’ follow-
up, %
Proportion
microscopically
verifi ed—all
cancer sites*,%
Diagnosed as
second or
subsequent primary
Known by
death
certifi cate only
Diagnosed
incidentally
at autopsy
Austria 100 172 455 187 5 10 0 146 217 19 93
Belgium 58 81 463 56 2 NA† 0·2 79 622 18 87
Czech Republic 8 19 423 0 5 3·4 5·7 16 651 17 87
Denmark 100 110 072 4 6 1·7 0·1 101 349 9 91
Finland 100 102 806 323 7 2·3 1·7 91 135 11 95
France 17 84 289 54 5 NA† 0 80 016 1 96
Germany 1 28 061 17 7 5·2 0 24 658 20 95
Iceland 100 5162 0 11 0·2 1 4541 10 96
Ireland 100 63 347 2 3 3 0·4 59 259 9 86
Italy 28 395 946 139 5 1·8 0·3 365 832 10 85
Malta 100 5953 0 1 1·7 0·2 5757 10 89
Netherlands 34 117 052 1 8 NA† 0·5 107 444 13 95
Norway 100 97 640 285 12 1·1 0·4 84 125 9 92
Poland 9 60 452 268 4 2·9 0·1 56 131 9 79
Portugal 43 32 235 144 3 NA† 0 31 276 23 94
Slovenia 100 33 934 0 3 2·5 1·2 31 655 8 92
Spain 16 103 231 201 5 3·7 0·1 94 306 9 91
Sweden 100 198 271 184 4 NA† 2·4 185 485 11 98
Switzerland 17 42 239 67 8 1 1·2 37 758 10 95
UK England 100‡ 1 031 985 1282 5 6·7 0·4 900 115 9 83
UK Northern Ireland 100 31 167 1 3 1·7 0·6 28 687 8 81
UK Scotland 100 126 234 7 9 1·2 0·1 110 905 8 85
UK Wales 100 71 247 524 8 12·1 0 56 162 8 NA
All cases 3 014 664 3746 6 0·5§ 2 699 086 10§ 87§
NA=not available. *Percentage of microscopically verifi ed cases were calculated only for registries that sent data for all cancer sites; the cancer registry of Herault (France) did not provide information on
microscopic verifi cation and was excluded. †Cancer registries did not have access to or did not use death certifi cates. ‡Data from nine regional cancer registries were used. §Calculated from entire dataset.
Table 1: Summary of characteristics of EUROCARE-4 cancer cases diagnosed in 1995–99, by country
Articles
http://oncology.thelancet.com Published online August 21, 2007 DOI:10.1016/S1470-2045(07)70245-0
3
survival comparisons between countries with diff erent
background mortalities. Relative survival was estimated
by the Hakulinen method
11,12
by use of estimates of
population lifetables for each registry area. Relative
survival at 5 years conditional to surviving 1 year was
calculated as the ratio of the relative survival at 5 years to
that at 1 year.
Age is a major determinant of relative survival.
2–4
To
account for diff erences in the age structure of the
diff erent populations, relative survival was adjusted for
age by the direct method by use of the international
standard for cancer-survival analysis.
13
This international
standard envisages a diff erent standard-age distribution
according to the age pattern of incidence of the cancer:
one for cancers mainly of young adults (eg, cancer of the
testis, non-Hodgkin lymphoma, acute lymphatic
leukaemia), one for cancers whose incidence varies little
with age (eg, cervix uteri, thyroid, and brain cancer), and
one for cancers that mainly aff ect the elderly (all
other
cancers).
For between-country comparisons of survival for all
cancers combined (excluding non-melanoma skin
cancers), adjustment by case-mix was also necessary
to eliminate the confounding eff ect of higher
incidences of more fatal cancers in some countries
compared with other countries. The sex-specifi c case-
mix distribution of all cases diagnosed in 1995–99 by
cancer site was used for this adjustment, and was
applied separately to each sex because case-mix varies
markedly with sex.
To estimate mean survival of cases from Europe,
survival for each cancer site was weighted according to
the contributions that each cancer registry made to the
whole population by use of the European areas defi ned
by the UN:
14
southern Europe (Italy, Spain, Portugal,
Slovenia, and Malta); central Europe (France, Belgium,
Austria, Germany, Netherlands, and Switzerland); eastern
Europe (Poland and Czech Republic); and northern
Europe, subdivided into the Nordic countries (Denmark,
Finland, Norway, Iceland, and Sweden) and UK and
Ireland. The subdivision for northern Europe was
implemented because of large diff erences in survival
between the two northern European areas. Cancer stage
information was not systematically available in the
dataset, and none of the analyses were adjusted by cancer
stage at diagnosis.
To assess changes in between-country survival
diff erences, we compared data from EUROCARE-3 (data
from those diagnosed in 1990–94) and EUROCARE-4
(data from those diagnosed in 1995–99) and used only
cancer registries that provided data for both periods: 60
of the current 83 cancer registries in 18 of the 23 countries.
Data from Ireland, Portugal, and Belgium were excluded
because their data for the EUROCARE-3 period were
unavailable; data from Iceland and Malta were also
excluded to avoid over-dispersion (and consequently large
CIs) due to their small populations sizes.
Role of funding source
The funding source had no role in study design, data
collection, data analysis, data interpretation, in writing
this report, or in the decision to submit for publication.
FB, RDA, MS had access to all of the data. FB and MS
had fi nal responsibility for the decision to submit for
publication.
Results
Table 1 shows the percentages of populations of each
country that were covered by cancer registration; total
numbers of cancer cases diagnosed in 1995–99; numbers
of records excluded from analyses for major errors (mean
0·1%, range 0·0–0·7); and percentages of cases excluded
because they were not fi rst malignancy, were known by
Number of adults Mean 5-year
relative survival (%)
Testis 20 435 93·8
Lip 8083 93·1
Thyroid 23 158 86·5
Hodgkin’s disease 15 323 83·0
Melanoma of skin 75 078 82·0
Breast (women) 414 298 81·1
Corpus uteri 63 012 78·3
Prostate 276 497 77·0
Penis 4714 73·9
Chronic lymphatic leukaemia 25 317 67·2
Cervix uteri 38 430 66·5
Bladder 131 793 65·8
Salivary gland 5741 64·6
Larynx 30 649 63·1
Soft tissue 14 187 59·3
Kidney 70 846 59·2
Bone 4796 58·4
Vagina, vulva 13 457 57·8
Non-Hodgkin lymphoma 91 815 54·6
Colon 231 514 53·9
Rectum 140 307 53·5
All sites* 2699 086 51·9
Nasopharynx 2973 50·2
Oral cavity 15 794 48·5
Nasal cavities 4553 48·2
Tongue 12 437 45·4
Small intestine 7417 43·2
Ovary 61 372 41·6
Oropharynx 9713 39·8
Chronic myeloid leukaemia 9264 39·5
Multiple myeloma 35 550 34·4
Acute lymphatic leukaemia 3497 30·0
Hypopharynx 7181 25·5
Stomach 113 840 24·1
Acute myeloid leukaemia 19 396 19·0
Brain 40 705 18·7
Biliary tract 24 330 14·1
Lung 343 473 12·6
Oesophagus 48 802 12·3
Liver 33 141 8·6
Pleura 11 851 7·5
Pancreas 68 854 5·5
0204060
Mean 5-year relative survival (%)
80 100
Figure 1: Mean European 5-year relative survival by cancer site for men and women
Horizontal bars indicate 95% CI. *Number of all-site cancer cases was higher than the sum of cases in the list
because all-site cancer cases included more sites than listed (myelodysplastic syndrome and cancer of the thymus,
heart, male breast, placenta, eye, meninges, spinal cord, cranial nerves, other endocrine glands, ill-defi ned sites,
and primary site unknown).
Articles
4
http://oncology.thelancet.com Published online August 21, 2007 DOI:10.1016/S1470-2045(07)70245-0
death certifi cate only (DCO), or were diagnosed at
autopsy. Also shown are the numbers of cases included
in the survival analyses, percentages censored after less
than 5 years of follow-up, and percentages of micro-
scopically verifi ed cases. After these exclusions,
2 699 086 cases were eligible for the survival analyses.
These included 41 major cancer sites (shown in fi gure 1),
summing up to 95·4% of total eligible cases. The
remaining 4·6% included the cases with unknown or ill-
defi ned primary site and a few very rare cancers. Overall,
10% of cases were followed up for less than 5 years,
mainly because these cases were diagnosed in 1999 and
the study closed at the end of 2003. Only 1% of cases were
censored before 4 years. The percentage of microscopically
verifi ed cases was highest (>95%) in: France, Iceland,
Netherlands, and Sweden; and lowest (≤85%) in Poland
and the UK registries.
Table 1 also shows that percentage of cases of multiple
tumours (6% overall) was highest in Norway and Iceland
(above 10%) and low in some recently established cancer
registries (Ireland, Portugal, Belgium, and Malta), which
suggested that some cancers diagnosed before the
beginning of registry activity might have been missed
from this study. The percentage of DCO cases was high
(above 5%) in Wales, Austria, England, and Germany.
This indicator was not available for countries where
cancer registries did not have access to or did not use
death certifi cates.
Figure 1 shows mean area-weighted and non-age-
adjusted European 5-year relative survival and 95% CI for
41 cancer sites in men and women together, which ranged
from 93·8% (95% CI 93·1–94·6) for cancer of the testis
to 5·5% (5·2–5·8) for pancreatic cancer. Relative survival
for all cancers combined 5 years after diagnosis was
51·9% (51·8–52·0) overall, 45·7% (45·5–45·8) in men,
and 58·2% (58·1–58·4) in women.
21 of the 41 cancer sites in this study, which accounted
for 61% of all cancers, had survival higher than that for
all cancers combined. Relative survival was 80% or
better for cases of cancer of the testis, lip, thyroid,
Hodgkin’s disease, melanoma, and female breast
cancer. These cancers with good prognosis represented
about 20% of the total, with female breast cancer
making up 15% of the total. Other cancer sites with
fairly good (60–79%) 5-year relative survival accounted
for about a fi fth of all those diagnosed, and included
corpus uteri, prostate, penis, chronic lymphatic
leukaemia, cervix uteri, bladder, salivary gland, and
larynx. Survival in most cancer cases was of 20–60%,
including frequent cancers such as non-Hodgkin
lymphoma (54·6 [95% CI 54·0–55·2]), and those of the
colorectum (53·7% [53·4–54·0]), ovary (41·6%
[41·0–42·3]), and stomach (24·1% [23·7–24·5]).
Cancer sites that had poor (<20%) survival included
acute myeloid leukaemia, biliary tract, lung, oesophagus,
liver, pleura, and pancreas. These cases account for about
20% of the total and are usually diagnosed at advanced
stage or might not be treated with eff ective treatments;
for most of these cases the aetiology is known and
primary prevention is a priority.
Figure 2 shows 5-year relative survival and 5-year
survival conditional to surviving the fi rst year, for eight
cancers. For most of the cancer cases, survival decreased
with increasing age at diagnosis. Notable exceptions
were breast and prostate cancer, for which survival was
worse in the youngest age group. These age patterns
were similar in all countries (data not shown).
Decreasing 5-year survival with age was often due to
decreasing 1-year survival with age, while survival
conditional to surviving the fi rst year changed only
modestly with advancing age (fi gure 2). For colorectal and
ovarian cancers, survival conditional to surviving 1 year
15–44 45–54 55–64
Age at diagnosis (year)
5-year relative survival (%)
65–74 ≥75 15–54 55–64 65–74 75–84 ≥85
15–44 45–54 55–64 65–74 ≥75
100
80
60
40
20
0
100
80
60
40
20
0
100
80
60
40
20
0
100
80
60
40
20
0
Colorectum Lung
Melanoma Breast (women)
Ovary Hodgkin’s disease
Testis Prostate
Conditional survival
Relative survival
Figure 2: 5-year relative survival of unweighted European pool by age at diagnosis, with 5-year survival
conditional on surviving 1 year
Figure 3: Age-adjusted 5-year relative survival of men and women by
country for eight selected cancer sites, with area-weighted mean European
relative survival
Horizontal bars indicate 95% CI. Countries have been grouped into the areas
defi ned by the UN
13
and ordered alphabetically within areas. The bottom bar in
each graph is the mean European age-adjusted relative survival, weighted by
area. *Data from Iceland and Portugal were excluded from the analysis of
testicular cancer, and data from Iceland, Germany, and Malta were excluded
from the analysis of Hodgkin’s disease because the few cases in one or more age
class made age standardisation unstable.
Articles
http://oncology.thelancet.com Published online August 21, 2007 DOI:10.1016/S1470-2045(07)70245-0
5
0 20 40 60 80 100
0 20 40 60 80 100 0 20 40 60 80 100
0 20 40 60 80 100 0 20 40 60 80 100
Denmark
Finland
Iceland
Norway
Sweden
Ireland
UK England
UK Northern Ireland
UK Scotland
UK Wales
Austria
Belgium
France
Germany
Netherlands
Switzerland
Italy
Malta
Portugal
Slovenia
Spain
Czech Republic
Poland
European mean
Denmark
Finland
Iceland
Norway
Sweden
Ireland
UK England
UK Northern Ireland
UK Scotland
UK Wales
Austria
Belgium
France
Germany
Netherlands
Switzerland
Italy
Malta
Portugal
Slovenia
Spain
Czech Republic
Poland
European mean
Denmark
Finland
Norway
Sweden
Ireland
UK England
UK Northern Ireland
UK Scotland
UK Wales
Austria
Belgium
France
Germany
Netherlands
Switzerland
Italy
Malta
Slovenia
Spain
Czech Republic
Poland
European mean
Denmark
Finland
Norway
Sweden
Ireland
UK England
UK Northern Ireland
UK Scotland
UK Wales
Austria
Belgium
France
Netherlands
Switzerland
Italy
Portugal
Slovenia
Spain
Czech Republic
Poland
European mean
Survival (%)
Survival (%)
Colorectum Lung Skin (melanoma)
Breast (women)
Testis (patients aged 15–64 years at diagnosis)*
Hodgkin’s disease
Ovary Prostate
Eastern Europe
Northern Europe
UK and Northern Ireland
Central Europe
Southern Europe
Articles
6
http://oncology.thelancet.com Published online August 21, 2007 DOI:10.1016/S1470-2045(07)70245-0
did not decline, and increased in the oldest age group,
while for melanoma, prostate, and breast cancer, 5-year
conditional (relative) survival remained constant with
increasing age. For lung cancer, both conditional and
unconditional 5-year relative survival declined steadily
with age.
Figure 3 shows age-adjusted 5-year relative survival
and 95% CIs by country for the eight selected cancers.
Mean results were: 53·8% (95% CI 53·3–54·1) for
colorectal cancer; 12·3% (12·1–12·5) for lung cancer;
81·6% (81·0–82·3) for skin melanoma; 78·9%
(78·6–79·2) for breast cancer; 36·3% (35·7–37·0) for
ovarian cancer; 75·7% (75·2–76·2) for prostate cancer;
94·2% (93·4–95·0) for cancer of the testis (in patients
aged 15–64 years); and 80·0% (79·0–81·0) for Hodgkin’s
lymphoma. The common cancers (colorectum, lung,
breast, and prostate) and ovarian cancer had similar
variation in survival with geography: highest in Nordic
countries (except Denmark) and central Europe,
intermediate in southern Europe, lower in UK and
Ireland, and lowest in eastern Europe. Survival for
melanoma, cancer of the testis, and Hodgkin’s disease
varied much less with geography, although survival was
low for Hodgkin’s disease in Ireland and for melanoma
in Poland and Wales.
Figure 4 shows 5-year relative survival and 95% CIs
for all cancers combined by country and separately for
men and women. These data have been adjusted for age
and case-mix, and are those that would be expected if
age distribution and incidence of the diff erent cancers
were the same in all populations. The mean European
area-weighted 5-year relative survival was: 49·6%
(49·5–49·7) overall; 44·8% (44·6–45·0) for men; and
54·6% (54·4–54·8) for women. The countries (UK was
defi ned as one country because TNEH information
was available only for the whole of the UK) were ranked
by decreasing per capita TNEH (US dollars adjusted per
capita purchasing power),
7
and colour-coded into four
TNEH classes. We noted a moderate correlation
between 5-year age-adjusted and case-mix-adjusted
relative survival for all cancers combined and TNEH
(r
2
=0·56 for females, r
2
=0·43 for males). However,
Denmark and UK had lower survival than countries
with similar TNEH. Finland had better survival than
expected from its moderate health expenditure. Spain,
Italy, and Portugal also had better survival than countries
with comparable TNEH, but they were only partially
covered by cancer registration and, therefore, survival
in these covered areas might not represent that of the
whole country. In Italy, for example, cancer registries
were mainly present in the wealthier northern part of
the country.
15
Non-adjusted relative survival for all cancers
combined for men and women together ranged from
57% on average in Nordic countries (excluding Denmark)
to 48% in Denmark and UK and about 45% in eastern
Europe. From these country-specifi c relative survival, we
estimated that out of about 1·5 million patients who died
within 5 years of follow-up, 1·3 million died from cancer.
If all countries had had the same survival as Nordic
countries (57%), we estimated that 150 000 fewer cancer
deaths would have occurred (12% less).
Figure 5 and tables 2 and 3 show the ranges of age-
adjusted 5-year relative survival for the eight cancer sites
in 1990–94 compared with those in 1995–99. For most
cancers, survival increased and between-country
survival diff erences narrowed. Survival diff er ences
narrowed notably for Hodgkin’s disease, and cancers of
Switzerland
Germany
Norway
France
Iceland
Denmark
Belgium
Netherlands
Sweden
Austria
Italy
UK*
Finland
Ireland
Spain
Malta
Portugal
Czech Republic
Slovenia
Poland
European mean
Men Women
0 20 40 60 0 20 40 60
Survival (%)
>US$3000
US$2000–2999
US$1000–1999
<US$999
80
60
40
20
0
100
5-year adjusted relative survival (%)
Melanoma
Colorectum
Lung
Ovary
Breast
Prostate
Testis
Hodgkin’s disease
Cancer site
Eurocare-4 (1995–99)
Eurocare-3 (1990–94)
Figure 4: 5-year relative survival adjusted for age-mix and case-mix by country for all cancers combined, with
area-weighted mean European survival
Horizontal bars indicate 95% CI. Countries are ordered by total national expenditure and colour-coded by TNEH.
TNEH expressed as per capita purchasing power parity. *Weighted mean of age-adjusted 5-year relative survival in
the four UK regions.
Figure 5: Changes in 5-year age-adjusted relative survival from EUROCARE-3 (1990–94) to EUROCARE-4
(1995–99)
Each point (rhombus or square) represents a country. Analysis for cancer of the testis was done on patients
aged 15–64 years at diagnosis. Iceland and Malta were excluded from all analyses, and Germany was excluded
from the analysis of Hodgkin’s disease because the few cases in one or more age class made age standardisation
unstable.
Articles
http://oncology.thelancet.com Published online August 21, 2007 DOI:10.1016/S1470-2045(07)70245-0
7
the testis, breast, and colorectum, mainly because
countries with poor survival in the earlier period had
improved survival in the later period (tables 2 and 3).
For example, for Hodgkin’s disease, survival increased
from 66·1% (60·3–72·4) to 78·4% (73·1–84·1) in Poland,
and from 70·5% (67·1–74·1) to 78·6% (5·4–82·0) in
Scotland; breast cancer from 61·7% (59·0–64·0) to
73·8% (71·8–75·9) in Poland, from 61·9 (58·0–66·0)
to 69·3% (65·7–73·1) in the Czech Republic, from
66·8% (64·3–69·5) to 71·9% (69·7–74·2) in Slovenia,
and from 72·4% (72·0–72·8) to 77·6% (77·2–77·9) in
England; and colorectal cancer from 29·4% (27·7–31·2)
to 38·8% (37·1–40·6) in Poland, and from 35·4%
(33·0–38·0) to 43·9% (41·5–46·4) in the Czech
Republic. For prostate cancer, large survival increases
were in Poland (38·7% [34·6–43·4] to 60·5%
[58·8–64·3]), Finland (62·3% [60·1–64·6] to 80·0%
[78·6–81·4]), and England (54·7% [53·8–55·5] to 69·8%
[69·2–70·4]). Survival for ovarian cancer increased
slightly with time in most countries, with greatest
increases in Finland (32·0% [29·9–34·3] to 39·6%
[37·3–42·0]), and the Czech Republic (25·1% [20·7–30·5]
to 32·3%; [26·4–39·4]). For lung cancer, which had the
poorest prognosis of all the cancers studied, between-
country survival diff erences narrowed, but survival did
not improve from that reached in 1990–94.
Discussion
This European study on the survival of patients with
cancer who were diagnosed in the second half of the 1990s
and followed up to the end of 2003 has shown that—after
correcting for non-cancer mortality, age at diagnosis, and
case-mix—survival for all cancers combined 5 years after
diagnosis was 44·8% (95% CI 44·6–45·0) for men, 54·6%
(54·4–54·8) for women, and 49·6% (49·5–49·7) overall.
An important fi nding is that all-cancer survival increased
notably in countries where all-cancer survival was low in
the previous period, resulting in a narrowing of between-
country survival diff erences. Previous EUROCARE
projects recorded marked between-country and inter-
regional diff erences in survival,
16
and, although survival
increased almost everywhere with time, survival
diff erences persisted and sometimes increased between
projects.
17
Therefore, our new fi ndings showing that the
Colorectal Lung Melanoma Breast
1990–94
relative
survival (SD)
1995–99
relative
survival (SD)
Diff erence* 1990–94
relative
survival (SD)
1995–99
relative
survival (SD)
Diff erence* 1990–94
relative
survival (SD)
1995–99
relative
survival (SD)
Diff erence* 1990–94
relative
survival (SD)
1995–99
relative
survival (SD)
Diff erence*
Austria 54·9 (0·9) 56·7 (0·9) 1·8 14·4 (0·7) 13·9 (0·7) –0·5 83·7 (1·6) 82·1 (1·5) –1·6 76·4 (0·9) 78·5 (0·9) 2·0
Czech
Republic
35·4 (2·5) 43·9 (2·5) 8·5 7·0 (1·4) 8·2 (1·4) 1·2 69·3 (5·9) 76·0 (5·4) 6·6 61·9 (4·0) 69·3 (3·7) 7·4
Denmark 46·1 (1·0) 49·3 (1·0) 3·2 6·5 (0·5) 7·9 (0·5) 1·4 83·1 (1·6) 85·1 (1·5) 2·0 73·2 (1·1) 77·5 (1·0) 4·4
Finland 52·9 (1·3) 57·8 (1·3) 4·8 8·7 (0·7) 9·6 (0·8) 0·9 82·5 (2·0) 84·5 (1·8) 2·0 80·0 (1·2) 83·6 (1·1) 3·7
France 56·7 (1·0) 57·5 (1·2) 0·8 14·0 (0·9) 12·8 (1·0) –1·2 84·4 (2·2) 85·8 (2·6) 1·4 80·9 (1·0) 82·7 (1·2) 1·8
Germany 52·4 (2·0) 57·5 (2·1) 5·1 11·7 (1·5) 13·2 (1·6) 1·6 83·5 (4·9) 83·4 (4·8) 0·0 75·8 (2·4) 78·3 (2·3) 2·6
Iceland 54·0 (6·4) 57·3 (5·6) 3·4 11·2 (3·0) 14·7 (3·3) 3·4 86·1 (9·6) 83·9 (9·6) –2·2 78·7 (6·1) 87·6 (5·3) 8·9
Italy 51·9 (0·7) 57·4 (0·7) 5·6 10·8 (0·4) 12·8 (0·5) 2·0 79·0 (1·6) 85·1 (1·2) 6·1 79·7 (0·7) 83·7 (0·6) 4·0
Malta 48·1 (8·2) 51·2 (4·9) 3·1 6·5 (3·8) 8·7 (2·9) 2·2 73·0 (13·3) 82·7 (12·4) 9·7 75·4 (7·4) 75·9 (4·5) 0·5
Netherlands 54·4 (1·4) 57·0 (1·3) 2·6 12·0 (0·8) 14·3 (0·8) 2·3 83·9 (2·5) 89·8 (2·0) 5·9 78·1 (1·3) 82·7 (1·2) 4·5
Norway 53·8 (1·1) 58·3 (1·1) 4·5 9·7 (0·8) 10·9 (0·8) 1·2 84·9 (1·5) 87·3 (1·4) 2·3 77·1 (1·3) 82·5 (1·2) 5·4
Poland 29·4 (1·8) 38·8 (1·8) 9·4 6·8 (0·8) 9·2 (0·9) 2·4 55·7 (4·6) 63·0 (4·2) 7·3 61·7 (2·3) 73·8 (2·0) 12·1
Slovenia 37·1 (1·9) 44·2 (1·9) 7·1 8·9 (1·1) 8·8 (1·0) 0·0 66·1 (4·4) 79·6 (3·7) 13·5 66·8 (2·6) 71·9 (2·2) 5·0
Spain 50·5 (1·4) 52·5 (1·2) 2·0 11·4 (0·9) 10·8 (0·8) –0·6 81·4 (3·1) 83·7 (2·5) 2·3 76·2 (1·5) 80·8 (1·3) 4·6
Sweden 54·9 (0·9) 58·3 (0·8) 3·4 10·6 (0·7) 13·1 (0·7) 2·4 88·5 (1·1) 90·2 (0·9) 1·7 82·7 (0·7) 84·3 (0·7) 1·6
Switzerland 55·2 (1·7) 59·7 (1·7) 4·6 10·3 (1·1) 13·6 (1·3) 3·3 89·1 (2·5) 90·3 (2·1) 1·1 75·8 (1·7) 82·3 (1·6) 6·4
UK England 45·9 (0·4) 50·8 (0·4) 4·9 7·9 (0·2) 8·6 (0·2) 0·7 81·0 (0·8) 84·6 (0·7) 3·6 72·4 (0·4) 77·6 (0·4) 5·2
UK Northern
Ireland
46·5 (2·9) 51·8 (1·9) 5·3 7·8 (1·6) 10·2 (1·2) 2·4 88·1 (5·2) 92·1 (3·3) 3·9 71·9 (3·4) 77·4 (2·0) 5·4
UK Scotland 45·7 (1·1) 51·5 (1·0) 5·8 7·4 (0·5) 8·0 (0·5) 0·6 86·2 (2·0) 88·4 (1·8) 2·2 71·6 (1·1) 75·0 (1·1) 3·4
UK Wales 43·9 (1·5) 50·6 (1·4) 6·8 8·3 (0·8) 9·0 (0·9) 0·7 74·7 (3·4) 73·6 (3·3) –1·1 72·9 (1·4) 76·9 (1·3) 4·0
EUROCARE
pool†
49·3 (0·2) 53·5 (0·2) 4·2 9·2 (0·1) 10·2 (0·1) 1·0 82·6 (0·4) 85·4 (0·4) 2·8 75·4 (0·2) 79·5 (0·2) 4·2
*Diff erence in 5-year relative survival for cancers diagnosed between 1995–99 and 1990–94. †Unweighted age-adjusted 5-year relative survival of the registries with data available for both 1990–94 and
1995–99 study periods. All relative survival are for 5 years.
Table 2: 5-year age-adjusted relative survival for colorectal, lung, melanoma, and breast cancer for European adults diagnosed in 1995–99, and change in 5-year age-adjusted relative
survival by country for these cancer sites from EUROCARE-3 to EUROCARE-4
Articles
8
http://oncology.thelancet.com Published online August 21, 2007 DOI:10.1016/S1470-2045(07)70245-0
gap in cancer survival in Europe is narrowing, is
encouraging, and suggests substantial improve ment in
cancer care in countries with poor survival.
The underlying causes of the noted increases in
survival are unlikely to be the same for all cancer sites.
For example the increase in breast cancer survival in
the UK
18
and many other European countries
19
was
accompanied by decreased mortality, and probably
represents the establishment of screening and real
improvements in care; the increase in survival for
patients with prostate cancer might be due mainly to
the widespread introduction of testing of prostate-
specifi c antigen, and consequently early diagnosis and
strong lead-time bias.
20
The decrease in mortality for
patients with prostate cancer that was evident in several
European countries might be largely due to improved
treatment.
21
However, survival was still much worse in eastern
than in western Europe, and within western Europe,
survival in the UK and Denmark was still low for several
cancers. Therefore, 5-year breast cancer survival was
around 83% in Finland and Sweden, 82% in France and
Italy, and 81% in the Netherlands covered by cancer
registration, and about 77% in the UK and Denmark.
For colorectal cancer, 5-year survival was above 57% in
the Nordic countries and several central and southern
European populations, but 51% in the UK and 49% in
Denmark. Whether the diff erence in survival of about
6% (ie, 57% vs 51%) between the UK and Denmark and
other western countries can be due to bias or artifacts
in cancer regis tration is an important question. Two
important sources of artifact can bias comparative
survival estimates: incomplete ascertainment of death
in registered patients with cancer; and incomplete
ascertainment of incident cases. Several EUROCARE
cancer registries used active follow-up to check the life
status of apparently living patients, and recorded only
rare instances of such errors.
22
Supposing (in an extreme
example) that 5% of dead patients were misclassifi ed as
alive, breast cancer and colorectal cancer survival would
increase by 1% and 2% respectively.
22
Incomplete incidence can lead to under-reporting of
numbers of long-term survivors which, if not more-or-
less uniform across all cancer registries, could explain
Ovary Prostate Testis Hodgkin’s disease
1990–94
relative
survival (SD)
1995–99
relative
survival (SD)
Diff erence* 1990–94
relative
survival (SD)
1995–99
relative
survival (SD)
Diff erence*
1990–94
relative
survival (SD)
1995–99
relative
survival (SD)
Diff erence*
1990–94
relative
survival (SD)
1995–99
relative
survival (SD)
Diff erence*
Austria 40·0 (1·8) 42·8 (1·8) 2·8 78·0 (1·5) 84·9 (1·0) 6·9 95·8 (1·6) 94·7 (1·7) –1·1 74·9 (3·5) 78·5 (3·5) 3·6
Czech
Republic
25·1 (4·8) 32·3 (6·4) 7·1 53·0 (6·8) 54·4 (5·7) 1·4 88·6 (6·0) 92·3 (7·3) 3·6 79·7 (12·0) 79·7 (9·3) 0·0
Denmark 28·5 (1·9) 32·3 (2·0) 3·8 38·9 (2·2) 47·7 (2·0) 8·8 96·7 (1·4) 94·7 (1·8) –2·0 77·6 (3·6) 79·6 (3·4) 2·0
Finland 32·0 (2·2) 39·6 (2·3) 7·6 62·3 (2·2) 80·0 (1·4) 17·7 93·3 (3·6) 94·9 (2·9) 1·6 76·8 (3·5) 83·6 (3·6) 6·8
France 36·4 (2·6) 36·8 (3·2) 0·4 75·8 (2·0) 79·1 (2·1) 3·3 93·3 (2·6) 94·8 (3·4) 1·5 82·9 (3·8) 79·4 (3·3) -3·6
Germany 37·0 (5·2) 36·9 (4·9) –0·1 77·5 (3·9) 81·6 (3·6) 4·1 93·6 (5·6) 95·7 (5·6) 2·1 80·8 (6·9) NA NA
Iceland 33·0 (10·9) 30·5 (11·5) –2·5 70·4 (8·4) 79·8 (6·7) 9·4 99·1 (4·4) NA NA 67·4 (12·3) NA NA
Italy 33·6 (1·7) 34·9 (1·5) 1·3 66·3 (1·7) 79·6 (1·2) 13·3 93·3 (2·2) 96·0 (1·7) 2·7 77·4 (2·3) 81·6 (2·1) 4·2
Malta 21·9 (9·6) 30·4 (8·9) 8·6 33·6 (9·3) 71·2 (7·9) 37·5 NA 98·3 (6·2) NA 78·4 (10·7) NA NA
Netherlands 31·5 (2·7) 35·7 (2·9) 4·2 70·3 (2·3) 80·9 (2·0) 10·6 96·1 (2·7) 95·7 (2·6) –0·4 76·2 (4·0) 79·7 (4·6) 3·5
Norway 35·2 (2·3) 37·8 (2·3) 2·7 64·6 (2·0) 74·5 (1·4) 10·0 97·0 (1·6) 95·1 (2·0) –1·8 80·1 (4·3) 83·9 (4·2) 3·8
Poland 27·3 (3·0) 31·0 (3·2) 3·8 38·7 (4·4) 60·5 (3·7) 21·7 83·5 (6·2) 92·1 (3·9) 8·7 66·1 (6·0) 78·4 (5·5) 12·3
Slovenia 27·2 (4·1) 33·4 (4·0) 6·2 44·4 (4·7) 58·2 (3·7) 13·8 94·8 (4·4) 95·2 (4·0) 0·4 75·7 (6·8) 82·5 (7·4) 6·8
Spain 35·8 (3·7) 36·5 (3·4) 0·6 57·2 (3·6) 74·7 (2·2) 17·5 89·3 (4·9) 96·0 (2·5) 6·7 72·5 (4·5) 81·8 (4·0) 9·3
Sweden 38·8 (1·7) 42·2 (1·7) 3·4 66·9 (1·4) 77·3 (0·9) 10·4 95·3 (1·9) 97·7 (1·3) 2·5 79·2 (2·8) 83·1 (2·6) 3·9
Switzerland 32·8 (3·6) 39·3 (4·1) 6·5 68·7 (3·3) 82·9 (2·0) 14·2 97·7 (2·0) 94·0 (4·0) –3·7 75·7 (6·9) 81·9 (6·1) 6·2
UK England 29·3 (0·7) 30·3 (0·7) 1·1 54·7 (0·8) 69·8 (0·6) 15·1 94·7 (0·8) 95·6 (0·8) 0·9 74·5 (1·4) 79·1 (1·3) 4·6
UK Northern
Ireland
31·4 (6·0) 34·0 (4·0) 2·7 54·3 (5·6) 60·8 (3·8) 6·5 87·6 (10·8) 95·1 (4·8) 7·5 71·2 (9·4) 74·0 (6·7) 2·8
UK Scotland 28·5 (1·9) 31·6 (2·0) 3·1 56·0 (2·3) 67·8 (1·9) 11·9 93·6 (2·5) 97·8 (1·8) 4·2 70·5 (3·5) 78·6 (3·3) 8·1
UK Wales 29·1 (2·7) 32·3 (2·6) 3·3 49·9 (3·1) 68·7 (2·4) 18·7 91·6 (3·6) 93·3 (3·4) 1·7 71·4 (5·1) 75·4 (4·9) 4·0
EUROCARE
pool†
32·0 (0·5) 34·2 (0·5) 2·2 61·4 (0·5) 73·9 (0·3) 12·5 94·6 (0·5) 95·5 (0·5) 0·9 75·6 (0·8) 80·1 (0·8) 4·5
NA=not available (5-year age-adjusted relative survival not calculated because not enough cases in some age classes in one of the two study periods). *Diff erence in 5-year relative survival for cancers diagnosed
between 1995–99 and 1990–94. †Unweighted age-adjusted 5-year relative survival of the registries with data available for both 1990–94 and 1995–99 study periods. Patients with cancer of the testis were aged
15–64 years at diagnosis. All relative survival are for 5 years.
Table 3: 5-year age-adjusted relative survival for ovarian, prostate, and testicular cancers, and Hodgkin’s disease for European adults diagnosed in 1995–99, and change in 5-year
age-adjusted relative survival by country for these cancer sites from EUROCARE-3 to EUROCARE-4
Articles
http://oncology.thelancet.com Published online August 21, 2007 DOI:10.1016/S1470-2045(07)70245-0
9
part of the noted survival diff erences. However, for
survivor under-reporting to account for the 6%
diff erence in survival for breast or colorectal cancer,
cancer registries of countries with artifactually low
survival would need to have missed 20–30% of 5-year
survivors (without missing any cancer deaths), without
similar errors occurring in cancer registries with high
survival. To show this, consider 100 cases of breast
cancer with 83% actual survival at 5 years; if 30% of the
5-year survivors (25 of 83 cases) were not registered,
but all 17 fatal cases were registered, the 5-year survival
would be calculated from 58 survivors and 17 deaths,
giving 77% (58 of 75 cases) survival. To explain the 6%
(57% vs 51%) diff erence in 5-year survival for colorectal
cancer in 100 cases, 11 (20% of 57%) survivors would
have to be missed and none of the 43 deaths (46 of
89=52%). On the basis of these considerations, we
concluded that the survival diff erences reported by
EUROCARE-4 probably correspond closely to reality.
Increasing age has been shown to have a large negative
eff ect on cancer survival.
2–4
In the EUROCARE-4
population, most of the eff ect of age was due to diff erences
in short-term survival, suggesting that older people might
present with more advanced disease, or might not (or can
not) receive treatments designed to cure or prolong
survival.
A consistent fi nding of all of the EUROCARE studies is
that between-country survival diff erences are small for
relatively uncommon treatable cancers, such as cancer of
the testis and Hodgkin’s disease, and also for cancers
with very poor prognoses. By contrast, for common
cancers with intermediate prognoses—for which early
diagnosis is typically an important determinant of
outcome—survival still diff ers markedly, and probably
unacceptably, between countries in Europe.
Understanding the reasons for these persistent (but
diminishing) diff erences is important for the public-
health response to cancer in Europe.
Survival is a complex indicator of a country’s
performance in managing cancer. Longer survival
might depend on later death or earlier diagnosis.
Earlier diagnosis, however, might not be of benefi t to
the patient if not accompanied by later death. By
contrast, death postponement might derive from more
eff ective conventional treatment due to earlier
diagnosis, or from the availability of better treatment
facilities. Longer survival might also derive from the
inclusion in the patient population of a number of
cancers that would never have progressed to give
clinical signs had they not been incidentally found at
screening.
23
The EUROCARE approach to disentangling these
possible determinants of survival includes high-
resolution studies. These are population-based between-
country comparisons of survival that use information
accessed from clinical records and can, therefore, adjust
for disease stage at diagnosis and assess the
determinants of disease stage.
24,25
The proportions of
patients cured of cancer have also been estimated.
26
High-resolution studies suggest that most of the
survival diff erences for breast and colorectal cancer are
attributable to diff erences in disease stage at
diagnosis,
24,25,27
while survival diff erences for cancer of
the testis seem to be due mostly to diff erences in access
to appropriate treatment.
28
Variation in the distribution
of cancer subsite, histotype, and other biological
features in standard International Classifi cation of
Diseases categories could be another reason for
geographical variation in survival,
29
and should be
explored further in studies at the population level.
Population-based relative survival, representing early
diagnosis and eff ective care, however, is used for
planning national cancer services.
30
EUROCARE-4 included most European Union
countries, and Iceland, Norway, and Switzerland.
Greece and Luxembourg do not have population-based
cancer registration, and cancer registries from Bulgaria,
Cyprus, Hungary, Latvia, and Lithuania did not
contribute to EUROCARE-4. Estonia and Slovakia
provided population-based survival data for
EUROCARE-3, but were unable to systematically access
important status data on patients archived in their
registries; for this reason accurate follow-up data was
not available and, therefore, the cases could not be used
for EUROCARE-4. Expanded coverage of eastern
European populations is a priority for future
EUROCARE studies.
The total number of all-cancer deaths in the fi rst
5 years after diagnosis was about 1·3 million, with an
overall relative 5-year survival of 51·9%. If all
countries attained the mean survival (57%) of Norway,
Sweden, and Finland (countries with high survival and
medium-to-high TNEH), about 12% fewer cancer deaths
(about 150 000) would occur in the 5 years after
diagnosis.
Contributors
FB (EUROCARE project leader) and MS designed the study, wrote the
report, and were involved in data analysis, development of the
algorithms for the quality checks of the entire EUROCARE-4 dataset,
and data checking. RDA did the statistical analyses and contributed to
writing the report. MS prepared the fi les for the data analyses and
developed the statistical programmes for data checking. SR, MBL, and
JWC took part in writing the report. The EUROCARE Working Group
provided data, took part in data checking, and commented on draft
versions of the report.
Confl icts of interest
The authors declared no confl icts of interest.
The EUROCARE-4 Working Group
Austria—M Hackl (Austrian National Cancer Registry, Vienna);
W Oberaigner (Tyrol Cancer Registry, Innsbruck).
Belgium—E Van Eycken (Flemish Cancer Registry, Brussels).
Denmark—H H Storm (Danish Cancer Society, Department of Cancer
Prevention and Documentation, Copenhagen).
Czech Republic—J Holub (West Bohemia Cancer Registry, Prague).
Finland—T Hakulinen (Finnish Cancer Registry, Helsinki).
France—G Hédelin, M Velten (Bas-Rhin Cancer Registry, Strasbourg);
A V Guizard (Calvados General Cancer Registry, Caen); G Launoy
Articles
10
http://oncology.thelancet.com Published online August 21, 2007 DOI:10.1016/S1470-2045(07)70245-0
(Calvados Digestive Cancer Registry, Caen); J Faivre (Côte d’Or Digestive
Cancer Registry, Dijon); P-M Carli and M Maynadie (Côte d’Or
Haematological Malignancies Registry, Dijon); A Danzon (Doubs Cancer
Registry, Besançon); A Buemi (Haut-Rhin Cancer Registry, Mulhouse);
B Tretarre (Hérault Cancer Registry, Montpellier); M Colonna (Isère
Cancer Registry, Meylan), F Molinié (Loire Atlantique Breast and Colon
Cancer Registry, Nantes); S Bara (Manche Cancer Registry, Manche,
Cherbourg); C Schvartz (Marne-Ardennes Thyroid Cancer Registry,
Reims); N Raverdy (Somme Cancer Registry, Amiens); P Grosclaude
(Tarn Cancer Registry, Albi).
Germany—H Brenner (Abteilung Klinische Epidemiologie und
Alternsforschung Deutsches Krebsforschungszentrum, Heidelberg);
H Ziegler (Saarland Cancer Registry, Saarbrucken).
Iceland—L Tryggvadottir (Icelandic Cancer Registry, Reykjavik).
Ireland—H Comber (National Cancer Registry of Ireland, Cork).
Italy—F Berrino (Project Leader), C Allemani, P Baili, L Ciccolallo,
G Gatta, F Lucca, A Micheli, M Sant, S Sowe, G Zigon (Fondazione
IRCCS, Istituto Nazionale dei Tumori, Milan); P Crosignani and
P Contiero (Cancer Registry and Environmental Epidemiology Unit,
Fondazione IRCCS, Istituto Nazionale dei Tumori, Milan); F Bellù
(Registro Tumori Alto Adige–Tumorregister Südtirol, South Tyrol,
Bolzano); A Giacomin (Biella Cancer Registry, Biella); S Ferretti (Ferrara
Cancer Registry, Ferrara); E Paci and E Crocetti (Firenze Cancer Registry,
Firenze); D Serraino, L Dal Maso, M De Dottori, and M Lise (Friuli
Venezia Giulia Cancer Registry, Friuli Venezia Giulia); M Vercelli and
C Casella (Liguria Cancer Registry, Istituto Scientifi co Tumori di Gonova,
University of Genova, Genova); F Pannelli (Macerata Cancer Registry
Camerino and Marche Childhood Cancer Registry, Marche); M Federico
and I Rashid (Modena Cancer Registry, Modena); M Fusco (Napoli
Cancer Registry, Napoli); A Traina (Palermo Breast Cancer Registry,
Palermo); V De Lisi and L Serventi (Parma Cancer Registry, Parma);
R Tumino (Ragusa Cancer Registry, Ragusa); L Mangone (Reggio-Emilia
Cancer Registry, Reggio-Emilia); F Falcini (Romagna Cancer Registry,
Forli); G Senatore and A Iannelli (Salerno Cancer Registry, Salerno);
M Budroni (Sassari Cancer Registry, Sassari); S Piff er (Trento Cancer
Registry, Trento); R Zanetti, S Patriarca, and S Rosso (Torino Cancer
Registry, Torino); F La Rossa (Umbria Cancer Registry, Perugia);
P Zambon and S Guzzinati (Venetian Cancer Registry, Padova);
R Capocaccia, E Carrani, R De Angelis, P Roazzi, M Santaquilani,
E Grande, R Inghelmann, A Simonetti, A. Tavilla, S Francisci, and
A Verdecchia (Centro Nazionale di Epidemiologia, Istituto Superiore di
Sanità, Rome).
Malta—M Dalmas (Malta National Cancer Registry, Valletta).
Norway—F Langmark (Cancer Registry of Norway, Institute of
Population-based Cancer Research, Oslo).
Poland—J Rachtan (Cracow Cancer Registry, Cracow); S Góźdź,
U Siudowska, and R Mężyk (Holycross Cancer Centre, Kielce);
M Bielska-Lasota (Independent Unit of Oncological Education,
M Sklodowska–Curie Cancer Centre, Warsaw); M Zwierko (Warsaw
Cancer Registry, Warsaw).
Portugal—P S Pinheiro (Southern Portugal Cancer Registry, Lisbon).
Slovenia—M Primic-Žakelj (Cancer Registry of Slovenia, Ljubljana).
Spain—A Mateos (Albacete Cancer Registry, Albacete); I Izarzugaza
(Basque Country Cancer Registry, Basque Country, Vitoria-
Gasteiz); A Torrella-Ramos (Castellón Cancer Registry, Valenzia);
R Marcos- Gragera (Girona Cancer Registry, Girona); C Martinez-Garcia
(Granada Cancer Registry, Granada); C Navarro and M D Chirlaque
(Murcia Cancer Registry, Murcia); E Ardanaz and C Moreno (Navarra
Cancer Registry, Navarra); J Galceran (Tarragona Cancer Registry,
Reus).
Sweden—M Talbäck (Centre for Epidemiology, National Board of Health
and Welfare, Stockholm).
Switzerland—G Jundt (Basel Cancer Registry, Basel); M Usel (Geneva
Cancer Registry, Geneva); H Frick (Grisons Cancer Registry, Chur);
S Ess (St Gallen Cancer Registry, St Gallen); A Bordoni (Ticino Cancer
Registry, Locarno); J C Luthi and I Konzelmann (Valais Cancer Registry,
Sion); N Probst (Zurich Cancer Registry, Zurich); J M Lutz and P Pury
(Swiss Association of Cancer Registries, Geneva).
The Netherlands—O Visser (Amsterdam Cancer Registry, Amsterdam);
J W W Coebergh (Eindhoven Cancer Registry, Eindhoven); R Otter
(Comprehensive Cancer Centre, Groningen).
UK—D C Greenberg (Eastern Cancer Registration and Information
Centre, Cambridge); M Quinn and N Cooper (Offi ce for National
Statistics, London); T Moran (North West Cancer Intelligence Service,
Manchester); D Forman (Northern and Yorkshire Cancer Registry and
Information Service, Leeds); M Roche and S Edwards (Oxford Cancer
Intelligence Unit, Oxford), J Verne (South West Cancer Intelligence
Services, Bristol); H Møller, (Thames Cancer Registry); D Meechan and
J Poole (Trent Cancer Registry, Sheffi eld); G Lawrence (West Midlands
Cancer Intelligence Unit, Birmingham); A Gavin (Northern Ireland
Cancer Registry, Belfast); M P Coleman (London School of Hygiene and
Tropical Medicine, London); R J Black and D H Brewster (Scottish
Cancer Registry, Edinburgh); J A Steward (Welsh Cancer Intelligence
and Surveillance Unit, Cardiff ).
Acknowledgments
We thank Donald Ward for the revision of English, Federica Lucca for
preparing the graphs, and Samba Sowe for technical assistance. This
research was supported by the Compagnia di San Paolo, Torino, Italy.
References
1 Cutler SJ, editor. International symposium on end results of cancer
therapy. National Cancer Institute monograph number 15.
Bethesda, MD: US Government Printing Offi ce, 1964.
2 Berrino F, Sant M, Verdecchia A, Capocaccia R, Hakulinen T,
Estève J. Survival of cancer patients in Europe: the EUROCARE
study. IARC Scientifi c Publications number 132. Lyon: IARC press,
1995.
3 Berrino F, Capocaccia R, Estève J, et al. Survival of cancer patients
in Europe: the EUROCARE-2 study. IARC Scientifi c Publications
number 151, Lyon: IARC press, 1999.
4 Berrino F, Capocaccia R, Coleman MP, et al. Survival of cancer
patients in Europe: the EUROCARE-3 study. Ann Oncol 2003;
14 (suppl 5): 1–155.
5 Vercelli M, Lillini R, Capocaccia R, et al. Cancer survival in the
elderly: eff ects of socio-economic factors and health care system
features (ELDCARE project). Eur J Cancer 2006; 2: 234–42.
6 Micheli A, Coebergh JW, Mugno E, et al. European health systems
and cancer care. Ann Oncol 2003; 14 (suppl 5): 41–60.
7 Organisation for economic Co-operation and Development (OECD).
OECD Health Data 2004, 1st edn. Paris: OECD, 2004.
8 Coleman MP, Gatta G, Verdecchia A, et al. EUROCARE-3
summary: cancer survival in Europe at the end of the 20th century.
Ann Oncol 2003; 14 (suppl 5): 128–49.
9 EUROCARE website. http://www.eurocare.it (accessed July 20,
2007).
10 Fritz A, Percy C, Jack A, et al. WHO. International classifi cation of
diseases for oncology, 3rd edn. Geneva: WHO, 1990.
11 Hakulinen T. Cancer survival corrected for heterogeneity in patient
withdrawal. Biometrics 1982; 38: 933–42.
12 US Department of Health and Human Services. National
Cancer Institute, surveillance, epidemiology and end results
program. SEER cancer statistics review, 1973–1999. Bethesda, MD:
US Department of Health and Human Services, 2001.
13 Corazziari I, Quinn M, Capocaccia R. Standard cancer patient
population for age standardising survival ratios. Eur J Cancer 2004;
40: 2307–16.
14 Globocan 2002 database. http://www-dep.iarc.fr/globocan/database.
htm (accessed May 5, 2007).
15 AIRTUM Working Group. Italian cancer fi gures, report 2007:
survival. Epidemiol Prev 2007; 31 (suppl 1).
16 Sant M, Aareleid T, Berrino F, et al. Survival of cancer patients
diagnosed 1990–94—results and commentary. Ann Oncol 2003;
14 (suppl 5): 61–118.
17 Sant M, Capocaccia R, Coleman MP, et al. Cancer survival increases
in Europe, but international diff erences remain wide. Eur J Cancer
2001; 37: 1659–67.
18 Quinn M, Allen E. Changes in incidence of and mortality from
breast cancer in England and Wales since introduction of screening.
United Kingdom Association of Cancer Registries. BMJ 1995;
311: 1391–95.
19 Sant M, Francisci S, Capocaccia R, Verdecchia A, Allemani C,
Berrino F. Time trends of breast cancer survival in Europe in
relation to incidence and mortality. Int J Cancer 2006;
119: 2417–22.
Articles
http://oncology.thelancet.com Published online August 21, 2007 DOI:10.1016/S1470-2045(07)70245-0
11
20 Brenner H, Arndt V. Long-term survival rates of patients with
prostate cancer in the prostate-specifi c antigen screening era:
population-based estimates for the year 2000 by period analysis.
J Clin Oncol 2005; 23: 441–47.
21 Baade PD, Coory MD, Aitken JF. International trends in prostate-
cancer mortality: the decrease is continuing and spreading.
Cancer Causes Control 2004; 15: 237–41.
22 Sant M, Gatta G. The EUROCARE Database: IARC Scientifi c
Publications. Lyon: IARC press, 1995.
23 Berrino F. The EUROCARE Study: strengths, limitations and
perspectives of population-based, comparative survival studies.
Ann Oncol 2003; 14 (suppl 5): 9–13.
24 Sant M, Allemani C, Capocaccia R, et al. Stage at diagnosis is a key
explanation of diff erences in breast cancer survival across Europe.
Int J Cancer 2003; 106: 416–22.
25 Gatta G, Capocaccia R, Sant M, et al. Understanding variations in
survival for colorectal cancer in Europe: a EUROCARE high
resolution study. Gut 2000; 47: 533–38.
26 Verdecchia A, De Angelis R, Capocaccia R, et al. The cure for colon
cancer: results from the EUROCARE study. Int J Cancer 1998;
77: 322–29.
27 Sant M, Allemani C, Berrino F, et al. Breast carcinoma survival in
Europe and the United States. Cancer 2004; 100: 715–22.
28 Sant M, Aareleid T, Artioli ME, et al. Ten-year survival and risk of
relapse for testicular cancer: a EUROCARE high resolution study.
Eur J Cancer 2007; 43: 585–92.
29 Gatta G, Ciccolallo L, Capocaccia R, et al. Diff erences in colorectal
cancer survival between European and US populations: the
importance of sub-site and morphology. Eur J Cancer 2003;
39: 2214–22.
30 UK Department Of Health. NHS cancer plan. London, 2000.
http://www.doh.gov.uk (accessed June 26, 2007).
    • "Although ovarian cancer is considered as an uncommon malignancy, its mortality rate is the highest among gynecologic malignancies in developed countries and is second after cervical cancer in developing world [1][2][3]. Epithelial ovarian cancers (EOCs) are the fifth most common cancer in women who are living in UK and are responsible for more than 4400 death per year in this country [4, 5]. Age and genetics are the most well established risk factors for ovarian cancer [6, 7]. "
    [Show abstract] [Hide abstract] ABSTRACT: Ovarian epithelial cancers are among the most lethal women's cancers. There is no doubt about the preventive role of oral contraceptive pills (OCPs) in development of ovarian cancers. But, there are limited numbers of studies to address the effect of these agents on the number of cortical inclusion cysts (CICs), their epithelial type and suppression of the metaplastic phenomenon by these pills. The aim of this study was to clarify the role of these agents in the prevention of these cyst formation and tubal metaplasia and also examine the mesenchymal-epithelial transition theory in this context by immunohistochemical methods. The representative section(s) of ovarian cortex from a total number of 201 consecutive total abdominal hysterectomy with bilateral or unilateral salpingo-oophorectomy specimens were examined for mean number of CICs and their epithelial type between two groups of the patients. Group A included the patients who were on oral contraceptive pills for more than 5 years. All of the subjects with other contraceptive methods or a history of less than 5 years contraceptive pills usage were stratified in group B. Sections from 20 cases in which more than five inclusion cysts were found, were selected for IHC staining with calretinine and PAX8 as markers for mesothelium and mullerian epithelium respectively. The mean age of the patients was 51.67 years with no significant differences between two groups. The mean number of cysts were 1.27 and 3.23 in group A and B respectively (P =0.0001). Similarly the mean number of CICs, lined by tubal epithelium, was significantly different between two groups (0.65 vs 2.65, P =0.0001). In IHC staining 123 out of 150 CICs (82 %) were PAX+ while only 7 CICs (4.8 %) showed positive reaction for calretinin irrespective of type of epithelium. Our findings showed that the use of OCP for more than five years in women, significantly prevents development of cortical inclusion cysts in the ovaries which lined by tubal (PAX8 positive) type epithelium. These findings may explain the alternative mechanism of oral contraceptive pills or long time use of progesterone in suppression of tubal type overgrowth and subsequently prevention of ovarian epithelial cancers.
    Full-text · Article · Dec 2016
    • "Pancreatic cancer is one of the major causes of cancer death worldwide [1, 2]. Despite advances in diagnostic and therapeutic techniques, the 5-year survival rate of patients with pancreatic cancer remains less than 10% [3]. Autoimmune pancreatitis (AIP) is a rare form of chronic a11111 "
    [Show abstract] [Hide abstract] ABSTRACT: Pancreatic ductal adenocarcinoma (PDAC) is difficult to distinguish from autoimmune pancreatitis (AIP) because of their clinical and radiological similarities, and therefore simple and minimally invasive surrogate markers for differential diagnosis would be useful. In our previous studies, we identified four microRNAs (miRNAs)-miR-7, miR-34a, miR-181d, and miR-193b -as MAPK-associated microRNAs whose expression was altered significantly with upregulation of the MAPK signaling pathway. Recently it has been reported that these miRNAs could be used as biomarkers in serum samples for accurate diagnosis of pancreatic lesions. The aim of the present study was to evaluate whether these MAPK-associated miRNAs in serum could be used as biomarkers for differentiating PDAC from AIP. We enrolled 69 patients with PDAC, 26 with intraductal papillary mucinous neoplasm (IPMN) and 15 with AIP. The expression of MAPK-associated miRNAs in serum was measured by quantitative real-time PCR. The 2-ΔCT method was used to quantify the expression of miRNAs, and the data were normalized using spiked-in synthetic cel-miR-39. Patients with PDAC or IPMN showed significantly higher amounts of serum MAPK-associated miRNAs than those with AIP (p<0.009 for miR-7, p<0.002 for miR-34a, p<0.001 for miR-181d, p<0.002 for miR-193b). ROC curve analysis demonstrated that these miRNAs had an area under the ROC curve (AUC) of 0.723-0.882 for differentiation between PDAC or IPMN from AIP. Furthermore, serum miR-181d was significantly associated with the presence of metastasis in patients with PDA (p = 0.014). Serum MAPK-associated miRNAs could be novel noninvasive biomarkers for differentiation between PDAC or IPMN and AIP.
    Full-text · Article · Jul 2016
    • "Also this study confirms it. However, there have also been some arguments to the contrary [16]. "
    Full-text · Article · Jan 2016 · PLoS ONE
Show more