You are on page 1of 11

European Journal of Cancer (2012) 48, 1532– 1542

Available at

journal homepage:

Epidemiology of glial and non-glial brain tumours in Europe

Emanuele Crocetti a,⇑, Annalisa Trama b, Charles Stiller c, Adele Caldarella a,

Riccardo Soffietti d, Jana Jaal e, Damien C. Weber f, Umberto Ricardi g, Jerzy Slowinski h,
Alba Brandes i, RARECARE working group

Clinical and descriptive epidemiology unit, ISPO – Palazzina 28/A Via delle Oblate 2, 50141 Florence, Italy
Evaluative Epidemiology, Department of Preventive and Predictive Medicine Fondazione IRCCS “Istituto Nazionale dei Tumori” Via Venezian
1, 20133 Milano, Italy
Childhood Cancer Research Group, University of Oxford, Richards Building, Old Road Campus, Headington, Oxford OX3 7LG, UK
Department of Neuroscience, University and San Giovanni Battista Hospital Via Cherasco 15, Turin, Italy
Department of Radiotherapy and Oncological Therapy, Haematology and Oncology Clinic, Tartu University Hospital, Vallikraavi str. 10, 51003
Tartu, Estonia
Radiation Oncology Department, Hôpitaux Universitaires de Genève, 4 rue Gabrielle Perret-Gentil, CH-1211 Geneva 14, Switzerland
Radiation Oncology, University of Turin, Azienda Ospedaliero-Universitaria San Giovanni Battista di Torino, Via Genova 3, 10126, Italy
Jerzy Slowinski Dept. of Epidemiology in Bytom,School of Public Health, Medical University of Silesia, Piekarska 18, 41-902 Bytom, Poland
Department of Medical Oncology, Ospedale Bellaria-Maggiore, Azienda ASL Via Altura 3, 40139 Bologna, Italy

Available online 7 January 2012

KEYWORDS Abstract To the central nervous system (CNS) belong a heterogeneous group of glial and
Central nervous system non glial rare cancers.
Glial tumours The aim of the present study was to estimate the burden (incidence, prevalence, survival and
Non-glial brain tumours proportion of cured) for the principal CNS cancers in Europe (EU27) and in European
Rare cancer regions using population-based data from cancer registries participating in the RARECARE
Epidemiology project.
Estimates We analysed 44,947 rare CNS cancers diagnosed from 1995 to 2002 (with follow up at 31st
Incidence December 2003): 86.0% astrocytic (24% low grade, 63% high grade and 13% glioma NOS),
Prevalence 6.4% oligodendroglial (74% low grade), 3.6% ependymal (85% low grade), 4.1% Embryonal
Survival tumours and 0.1% choroid plexus carcinoma. Incidence rates vary widely across European
regions especially for astrocytic tumours ranging from 3/100,000 in Eastern Europe to
5/100,000 in United Kingdom and Ireland. Overall, about 27,700 new rare CNS cancers were
estimated every year in EU27, for an annual incidence rate of 4.8 per 100,000 for astrocytic,
0.4 for oligodendroglial, 0.2 for ependymal and embryonal tumours and less than 0.1 for
choroid plexus carcinoma.
More than 154,000 persons with rare CNS were estimated alive (prevalent cases) in the EU at
the beginning of 2008.

⇑ Corresponding author: Tel.: +39 0557972508; fax: +39 0557972535.

E-mail address: (E. Crocetti).

0959-8049/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
E. Crocetti et al. / European Journal of Cancer 48 (2012) 1532–1542 1533

Five-year relative survival was 14.5% for astrocytic tumours (42.6% for low grade, 4.9% for
high grade and 17.5% for glioma NOS), 54.5% for oligodendroglial (64.9% high grade and
29.6% low grade), 74.2% for ependymal (80.4% low grade and 36.6% high grade), 62.8%
for choroid plexus carcinomas and 56.8% for embryonal tumours. Survival rates for astrocytic
tumours were relatively higher in Northern and Central Europe than in Eastern Europe and in
UK and Ireland. The different availability of diagnostic imaging techniques and/or radiation
therapy equipment across Europe may contribute to explain the reported survival differences.
The estimated proportion of cured patients was 7.9% for the ‘glial’ group to which belong
astrocytic tumours.
Overall results are strongly influenced by astrocytic tumours that are the most common type.
This is the first study to delineate the rare CNS cancer burden in Europe by age, sex and Euro-
pean region.
Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction tumours are included. However, despite the rarity of enti-

ties selected by using this cut-off, there are many sub-enti-
Central Nervous System (CNS) cancers are a group ties for each group (i.e. different histologies and WHO
of different tumour entities anatomically close to each grade of glial tumours) with different prognosis and dif-
other but diverse in terms of morphology, site, molecu- ferent treatment approaches that complicate the picture.
lar biology and clinical behaviour and presumably This means that management of these rare brain tumour
aetiology.1 subgroups has to be considered hyper-specialistic, as well
In Europe, the standardised (World) incidence of pri- as treatment options centralised in large volumes hub
mary CNS cancers ranges from 4.5 to 11.2 cases per centres with recognised expertise in this field. Further-
100,000 men and from 1.6 to 8.5 per 100,000 women.2 more, central pathology review in CNS tumours is partic-
The two most common CNS cancers, high-grade glioma ularly important. The experience of modern clinical trials
and brain metastasis occur more frequently during adult- (i.e. European Organisation for Research and Treatment
hood and especially among the elderly. In Europe, the of Cancer (EORTC) and Radiation Therapy Oncology
peak of incidence is 18.5/100,000 in people aged Group (RTOG) studies on grade III gliomas) showed
P65 years.2 The relative frequency of CNS tumours is very high (up to 50%) inter-observer variability among
however highest during childhood, when they account neuropathologists in the differential diagnosis between
for 23% of all the cancers diagnosed.3 astrocytic and oligodendroglial tumours.9
In adults the 5-year survival rate for the primary CNS The objective of this study was to establish a picture
cancers in Europe was 17% for males and 19% for of incidence, prevalence and survival of rare CNS can-
females (1995–2002),4 with differences across European cers in Europe based on the new RARECARE list of
regions.5 Survivorship is higher for young European tumours.
patients – 63% – than for the elderly ones.6
Statistics on CNS tumours are estimated by grouping 2. Materials and methods/cancer cases
all malignancies arising in all the CNS anatomic sites:
meninges, brain, spinal cord, cranial nerves and other Rare cancers of the CNS presented in this paper are
localisation of CNS (ICD-10 topography codes C70- based on the new list of cancer types provided by
C72).7 RARECARE. The list is organised into three tiers.
However, rare tumours are more appropriately The bottom tier (tier 3) corresponds to the World
defined as a combination of topographical and morpho- Health Organisation (WHO) name of individual cancer
logical characteristics, as defined by the International entities [
Classification of Diseases for Oncology (ICD-O).8 The pat-gen/] and their corresponding ICD-O-3 morphology
Surveillance of Rare Cancers in Europe (RARECARE; and topography codes. Tier 3 entities were grouped into project, is a large collaboration of pop- categories of cancers (tier 2) considered similar from the
ulation-based cancer registries (CRs) across Europe point of view of clinical management and research.
which provides a list of rare cancers on the basis of topog- These categories were further grouped into more general
raphy and morphology. Clinical factors, such as difficul- categories (tier 1), considered to involve the same clini-
ties in achieving diagnosis, in clinical decision making cal expertise and patient referral structure. Accordingly
and in conducting clinical studies and the lack of stand- the following cancer entities were identified and will be
ardised treatment mainly affected the definition of the described in this paper:
type of cancers of the RARECARE list. Under the
threshold proposed by the RARECARE project, inci- Glial tumours (tier 1 entity) which include astrocytic,
dence lower than 6 per 100,000 per year, CNS malignant oligodendroglial, oligoastrocytic and ependymal
1534 E. Crocetti et al. / European Journal of Cancer 48 (2012) 1532–1542

tumours with malignant behaviour only (tier 2 enti- December 2003. For 11 countries CRs covered the entire
ties) and national population (Austria, Iceland, Ireland, Malta,
Non glial tumours of CNS and pineal gland (tier 1 Norway, Slovakia, Slovenia, Sweden, Northern Ireland,
entity) which includes embryonal tumours and cho- Scotland and Wales). Other 10 countries were repre-
roid plexus carcinomas (tier 2 entities). sented by regional CRs, covering variable proportions
of their respective national populations. In the present
Glial tumours were also divided and analysed in low- paper, data from 76 CRs out of the 89 accepting to par-
and high-grade gliomas according to the WHO classifi- ticipate in the RARECARE project were considered.
cation.10 Low grade astrocytic included the ICD-O3 CRs which did not classify cancers according to the
morphology (M) 9382, 9400, 9410, 9411, 9420, 9423 third edition of the ICD-O3 and also those which col-
and M 9424; high grade astrocytic included the ICD- lected data on childhood cancers only were excluded
O3 M 9381, 9401, 9430, 9440-9442. Low grade oligoden- from the analysis.
droglial tumours were M9450 and M 9460; the high The incidence analysis considered case incidents from
grade group included M 9451 only. Finally the low 1995 to 2002 and excluded specialised registries. Thus,
grade ependymal tumours included M9391 and M 64 CRs were included in the incidence analyses. Age-
9393; the high grade group was made by M 9392. standardised incidence rates per 100,000 were computed
Table 1 shows the morphology and topography codes using the European standard population.
of the tumours described in this article. The prevalence per 100,000 was estimated at the
It is worth stressing that pilocytic astrocytoma (grade index date of 1st January 2003. Only data from 22
WHO I), the most frequent brain tumour of children is CRs with data covering the period 1988–2002 were used
excluded, following its downgrading to non-malignant for prevalence estimates. The counting method,11 based
behaviour in ICDO-3. Other primary and secondary on CR incidence and follow-up data, was applied to CR
tumour entities usually included in the CNS (metastasis, data from 1988 to 2002. The completeness index
sarcomas, germ cell tumours, meningiomas, gliomas of method,12 was used to estimate complete prevalence,
optic nerve) are also excluded from this analysis because and involved adding the estimated surviving cases diag-
they are described in other dedicated papers. Unspeci- nosed prior to 1988 to those counted in 1988–2002.
fied malignant neoplasms (8000/3 and 8001/3) are also The expected number of new cases per year and of
excluded from the analysis. prevalent cases in Europe (EU27) was estimated multi-
RARECARE data were extracted from the EURO- plying the crude incidence and prevalence estimates
CARE-4 study dataset covering the period of diagnosis (obtained as described above) to the 2008 European pop-
1978–2002, with status information available up to 31st ulation (EU27 = 497,455,033) provided by EUROSTAT

Table 1
Data quality indicators of rare central nervous system cancers diagnosed 1995–2002 and archived in 76 RARECARE cancer registries.
Tier Entity Cases Data quality indicator ICD-O3a codes
Death Autopsy Microscopic Cases 1995–1998 Topography Morphology
certificate (%) verification censored before five
only (%) (%) years (%)
Rare central nervous
system (CNS)
1 Glial tumour Of CNS 43,125 1.2 0.5 87.0 0.5 C71, C72.0, 9380-9384, 9391-9460
2 Astrocytic tumours 38,653 1.3 0.5 86.1 0.4 C71, C72.0, 9380-9382, 9384,
of CNS C72.8-C72.9 9400-9442
2 Oligodendroglial 2866 0.5 0.4 94.9 1.0 C71, C72.0, 9450-9451,9460
tumours of CNS C72.8-C72.9
2 Ependymal tumours 1606 0.4 0.7 95.2 1.6 C71, C72.0, 9383, 9391-9394
of CNS C72.8-C72.9
1 Non-glial tumour of 1872 0.4 0.2 96.8 2.0 C71, C72.0, 9362, 9390, 9470-
CNS and pineal C75.3 9474, 9490, 9500-
gland 9505, 9508
2 Embryonal tumours 1822 0.3 0.2 97.0 2.0 C71, C72.0, 9362, 9470-9474,
of CNS C75.3 9490, 9500-9505, 9508
2 Choroid plexus 50 4.0 0.0 92.0 0.0 C71, C72.0, 9390
carcinoma of CNS C75.3
The following tumours are excluded: metastasis, unspecified malignant neoplasms (8000/3 and 8001/3), sarcomas, germ cell tumours,
meningiomas, gliomas of optic nerve and pilocytic astrocytoma (grade WHO I).
E. Crocetti et al. / European Journal of Cancer 48 (2012) 1532–1542 1535

( In providing glial and non glial documentation, although the proportion varied among
tumours burden estimates, we assumed that the popula- cancer entities from 86.1% for astrocytic tumours to
tion covered by our CRs was representative of the popu- 95% to ependymal tumours. The 97.0% of non-glial
lation of the EU27 as a whole. We assumed that tumours of CNS and pineal gland was microscopically
incidence, survival and the population for the period verified. The proportion of cases known from autopsy
2003–2008 did not change. A total of 4,302,067 cancer was very low, with a maximum of 0.7% for ependymal
cases were used to produce the prevalence estimates. tumours.
Period survival indicators for the years 2000–2002 Quality indicators for CNS tumours are computed on
were estimated using the Brenner algorithm.13 Period all participating CR (44,947 cases from 76 cancer regis-
analysis provides more up-to-date survival experience tries) while incidence rates are reported for general CR
by considering survival experience in 2000-2002 (sur- only (44,842 cases from 64 cancer registries). The sur-
vival data were collected from 46 European CRs partic- vival analysis was undertaken on 46 CR because only
ipating in the RARECARE study). Survival has been those had data available to perform period survival.
computed as relative survival that is the ratio of absolute
survival to expected age-specific and sex-specific survival 3. Results
of the general population. It is a measure of the excess
mortality for patients with cancer compared with that 3.1. Incidence
for the general population. The proportion of cured
patients for rare CNS tumours was estimated using a Table 2 shows crude, sex and age-specific incidence
mixture or cure model.14 This type of survival model rates recorded in Europe during the period 1995–2002.
assumes that cured cases have the same mortality as Among CNS cancers, astrocytic tumours were the
the general population, while the complementary frac- most common tumours, with a standardised incidence
tion (the fatal cases) has an excess death rate attributed rate of 4.8 (3.0 high grade, 1.2 low grade and 0.6 glioma
to the CNS cancer.14 NOS) per 100,000 per year in Europe, followed by oligo-
This analysis was carried out on 44,947 cases of rare dendroglial tumours (0.4 overall; 0.3 low and 0.1 high
CNS cancers diagnosed in 1995–2002. grade), embryonal (0.2) and ependymal tumours (0.2
The main data quality indicators for the cancer diag- overall; 0.17 low grade and 0.03 high grade). The inci-
noses in 1995–2002 are presented in Table 1. Among glial dence rate of choroid plexus carcinomas was under 0.1
tumours 1.2% of the cases were known from death certif- per 100,000/year.
icate only (DCO); among non glial tumours DCO were Rare CNS tumours were slightly more common in
0.4% ranging from 0.3% (embryonal tumours) to 4.0% men than in women. The male to female crude rates
(choroids plexus carcinomas). About 87% of the glial ratio ranged between 1.5 (embryonal), to 1.4 (astrocytic)
tumour cases included in the analysis had histological and to 1.3 (oligodendroglial), Table 2.

Table 2
Observed cases with crude incidence (rate per 100,000/year) and standard errors (SE) in Europe. Rates and SE by sex and age, with estimated
incident cases in Europe (EU27). Cases diagnosed 1995–2002 in 64 European cancer registries (CRs).
Rare central nervous EU overall Sex Age Estimated cases
system (CNS) cancers in EU27 per year
Male Female 0–19 20–39 40–59 60+
Observed Rate SE Rate SE Rate SE Rate SE Rate SE Rate SE Rate SE N
Cases 1995-
Glial tumour of CNS 43,037 5.4 <0.1 6.3 <0.1 4.5 <0.1 1.2 <0.1 2.5 <0.1 7.0 <0.1 12.1 0.1 26,610
and pineal gland
Astrocytic tumours of 38,588 4.8 <0.1 5.7 <0.1 4.0 <0.1 0.9 <0.1 2.0 <0.1 6.2 <0.1 11.6 0.1 23,859
Oligodendroglial 2,845 0.4 <0.1 0.4 <0.1 0.3 <0.1 0.1 <0.1 0.3 <0.1 0.6 <0.1 0.4 <0.1 1759
tumours of CNS
Ependymal tumours 1604 0.2 <0.1 0.2 <0.1 0.2 <0.1 0.2 <0.1 0.2 <0.1 0.2 <0.1 0.2 <0.1 992
of CNS
Non-glial tumour of 1805 0.2 <0.1 0.3 <0.1 0.2 <0.1 0.6 <0.1 0.2 <0.1 0.1 <0.1 <0.1 <0.1 1116
CNS and pineal
Embryonal tumours 1755 0.2 <0.1 0.3 <0.1 0.2 <0.1 0.6 <0.1 0.2 <0.1 0.1 <0.1 <0.1 <0.1 1085
of CNS
Choroid plexus 50 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 0.0 <0.1 0.0 <0.1 0.0 <0.1 31
carcinoma of CNS
1536 E. Crocetti et al. / European Journal of Cancer 48 (2012) 1532–1542

As regards age-specific incidence rate (per 100,000), age-groups with the exception of the oldest one (60+
incidence rates for glial tumours overall (11.9/100,000/ years) that shows an increasing-incidence trend.
year) and for astrocytic tumours (11.6/100,000/year) Table 3 describes the standardised incidence rates
were highest in the oldest age group (60 years and (1995–2002) for EU in the whole and each European
more). Incidence peaked in the age group 40–59 years Region. For the purposes of the analyses five regions
for oligodendroglial tumours (0.6). The incidence of were identified as follow: Northern Europe (Iceland,
ependymal tumours was quite stable across the age- Norway, Sweden), United Kingdom and Ireland (Eng-
groups while embryonal tumours had their highest inci- land, Scotland, Wales, Northern Ireland, Republic of
dence among younger patients (0–19 years; 0.6/100,000/ Ireland), Central Europe (Belgium, Austria, France,
year) and then incidence fell throughout the age range. Germany, The Netherlands, Switzerland), Eastern Eur-
Table 2 shows the number of new cases of CNS ope (Poland, Slovakia) and Southern Europe (Italy,
tumours estimated in the European Union (EU27) per Malta, Portugal, Slovenia, Spain).
year. There are about 27,700 new diagnosis of CNS can- There was marked geographical variation in inci-
cers per year of which, 26,610 are ‘glial tumours of CNS’ dence for astrocytic tumours, with the highest rate in
(23,859 astrocytic tumours with unspecified gliomas; UK and Ireland (5.1 cases per 100.000) and the lowest
1,759 oligodendroglial tumours and 992 ependymal (3.1) in eastern Europe. The differences in geographical
tumours) and 1,116 are ‘non glial tumours of CNS rates for all the other types of rare CNS tumours were
and pineal gland’ (99% of which are embryonal negligible, Table 3.
Fig. 1 shows incidence trend for the period 1995–2002 3.2. Prevalence
for glioma, as overall and for age-classes. The overall
trend is flat and the same is true for all the analysed Table 4 shows the observed prevalence proportions at
2, 5, 15-year and the estimated complete prevalence in
Europe (index date 1st January 2003).
More than 154,000 persons were alive in EU at the
beginning of the year 2008 with a past diagnosis of rare
CNS cancers, of whom 130,000 had had in their past
clinical history a ‘glial tumour of CNS’ and about
23,000 a ‘non glial tumour of CNS and pineal gland’.
As regards ‘glial tumours of the CNS’, 19% and 32%
were diagnosed within 2 and 5 years before the preva-
lence date, respectively. The difference (13%) between
these two proportions represents the proportion of cases
in the 3rd–5th years after diagnosis, presumably still
undergoing clinical follow-up. The remaining fraction
Fig. 1. Glioma incidence trend by age groups. of 68% represents long-term survivors (surviving more
than 5 years), and 47,000 of these (47% of the total) were

Table 3
Age-standardised incidence rates (per 100,000) for central nervous system (CNS) tumours in 1995–2002, with standard errors (SE) by European
Rare central nervous system (CNS) Regiona EU overall
Northern Central Eastern Southern United
Europe Europe Europe Europe kingdom and
Adj. SE Adj. SE Adj. SE Adj. SE Adj. SE Adj. SE
rate rate rate rate rate rate
Glial tumour of CNS and pineal gland 5.3 0.1 4.8 0.1 3.5 0.1 4.4 0.1 5.7 <0.1 5.0 <0.1
Astrocytic tumours of CNS 4.6 0.1 4.2 0.1 3.1 0.1 3.8 0.1 5.1 <0.1 4.4 <0.1
Oligodendroglial tumours of CNS 0.4 <0.1 0.3 <0.1 0.3 <0.1 0.3 <0.1 0.4 <0.1 0.4 <0.1
Ependymal tumours of CNS 0.3 <0.1 0.2 <0.1 0.2 <0.1 0.2 <0.1 0.2 <0.1 0.2 <0.1
Non-glial tumour of CNS and pineal 0.3 <0.1 0.3 <0.1 0.2 <0.1 0.3 <0.1 0.2 <0.1 0.3 <0.1
Embryonal tumours of CNS 0.3 <0.1 0.3 <0.1 0.2 <0.1 0.3 <0.1 0.2 <0.1 0.3 <0.1
Choroid plexus carcinoma of CNS <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1
Northern Europe (Iceland, Norway, Sweden), Central Europe (Austria, Belgium, France, Germany, the Netherlands, Switzerland), Eastern
Europe (Poland, Slovakia), Southern Europe (Italy, Malta, Slovenia, Portugal, Spain).
E. Crocetti et al. / European Journal of Cancer 48 (2012) 1532–1542 1537

Table 4
Observed prevalence per 100,000 (Prev) with standard errors (SE) by time (2, 5 and 15 years) from diagnosis, with estimated complete prevalence
per 100,000 and estimated total prevalent cases in EU27 of central nervous system (CNS) tumours. Index date 31 December 2003.
Rare central nervous system (CNS) cancers Observed prevalence Estimated prevalence
2 years after 5 years after 15 years after Complete EU27
diagnosis diagnosis diagnosis
Prev. SE Prev. SE Prev. SE Prev. SE No. of cases
Glial tumour of CNS and pineal gland 5.0 0.1 8.3 0.1 14.5 0.2 26.3 0.4 130,764
Astrocytic tumours of CNS 4.1 0.1 6.3 0.1 10.9 0.1 20.4 0.4 101,593
Oligodendroglial tumours of CNS 0.6 <0.1 1.2 <0.1 2.0 0.1 2.7 0.1 13,187
Ependymal tumours of CNS 0.3 <0.1 0.8 <0.1 1.7 0.1 3.8 0.1 19,125
Non-glial tumour of CNS and pineal gland 0.4 <0.1 0.8 <0.1 1.6 0.1 4.7 0.2 23,569
Embryonal tumours of CNS 0.4 <0.1 0.7 <0.1 1.5 <0.1 4.3 0.2 21,470
Choroid plexus carcinoma of CNS <0.1 <0.1 <0.1 <0.1 0.1 <0.1 0.3 0.1 1735

those surviving more than 15 years after diagnosis. (about 13,000) and choroid plexus carcinomas (1,700).
Therefore, the majority of patients alive with a diagnosis Choroid plexus carcinomas, embryonal and astrocytic
of glial tumours of the CNS are cases diagnosed several tumours were those with the highest proportion of very
years before (long survivors). A possible explanation for long term (more than 15 years) survivors (67%, 63% and
this is the high frequency of low grade tumours which 47%, respectively), Table 4.
have a good prognosis.
As regards ‘Non-glial tumours of the CNS and pineal 3.3. Survival
gland’, 9%, 17% and 34% of the prevalent cases were
diagnosed within 2, 5 and 15 years before the prevalence Table 5 shows 1 and 5-year survival of first and sec-
date, respectively. The very long term survivors, those ond tier entities of rare CNS tumours. Survival was bet-
who survived more than 15 years after diagnosis, repre- ter for ‘non glial tumours of CNS and pineal gland’ (601
sented the majority of these cases (66% of the total, cases analysed), than for ‘glial tumours of CNS’ (13,667
about 15,500 subjects). cases). One- and five year survival were 81.0% and
Astrocytic tumours were the most common among 43.8%, and 56.9% and 19.6%, respectively.
prevalent rare CNS cancers (about 100,000 cases), fol- Within glial CNS tumours, ependymal tumours had
lowed by embryonal tumours (about 21,000), ependy- the best 5-year prognosis (74% overall, 80% low grade
mal tumours (about 19,000), oligodendroglial tumours and 37% high grade). Intermediate prognosis was

Table 5
One-year and 5-year relative survival (%) for central nervous system (CNS) tumours with standard errors (SE) by European region for 2000–2002.
Rare central nervous system (CNS) EU overall Regiona
EU EU Northern Central Eastern Southern UK and
1 year 5 years 1 year 5 years 1 year 5 years 1 year 5 years 1 year 5 years 1 year 5 years
(SE) (SE) (SE) (SE) (SE) (SE) (SE) (SE) (SE) (SE) (SE) (SE)
Glial tumour of CNS and pineal gland 44 20 51 23 46 23 46 17 51 19 36 17
(0.5) (0.4) (1.1) (0.9) (0.9) (0.8) (1.7) (1.2) (1.2) (0.9) (0.7) (0.6)
Ependymal tumours of CNS 90 74 94 79 89 69 82 66 89 75 91 78
(1.4) (2.1) (2.5) (4.2) (2.7) (4.2) (6.9) (8.3) (3.7) (5.1) (2.7) (4.1)
Oligodendroglial tumours of CNS 83 55 87 63 82 60 69 34 86 54 81 51
(1.4) (1.8) (2.7) (0.4) (2.9) (3.6) (6.6) (5.8) (3.2) (4.3) (2.3) (3.2)
Astrocytic tumours of CNS 39 15 46 16 42 17 43 14 46 13 31 12
(0.5) (0.3) (1.2) (0.9) (1.0) (0.8) (1.8) (1.2) (1.3) (0.9) (0.7) (0.6)
Non-glial tumour of CNS and pineal 81 57 80 48 87 68 78 50 79 55 79 56
gland (1.7) (2.2) (3.7) (4.7) (2.8) (3.8) (6.3) (7.9) (4.5) (5.5) (3.3) (4.4)

Choroid plexus carcinoma of CNS 78 63 100 50 86 50 50 100 100 57 38
(8.7) (10.7) (25.1) (13.2) (35.4) (35.4) (18.2) (20.0)
Embryonal tumours of CNS 81 57 79 48 87 67 79 49 78 54 80 56
(1.7) (2.2) (3.8) (4.8) (2.9) (3.8) (6.3) (8.2) (4.6) (5.5) (3.3) (4.5)

Statistic could not be calculated.
Northern Europe (Iceland, Norway, Sweden), Central Europe (Austria, Belgium, France, Germany, the Netherlands, Switzerland), Eastern
Europe (Poland, Slovakia), Southern Europe (Italy, Malta, Slovenia, Portugal, Spain).
1538 E. Crocetti et al. / European Journal of Cancer 48 (2012) 1532–1542

reported for oligodendroglial tumours (55% overall; according to their relative frequency (overall < 6/
65% low grade and 30% high grade) and low survival 100,000), astrocytic tumours represent the majority of
was estimated for astrocytic tumours (15%). Astrocytic them accounting for about 86% of all the 27,726 CNS
tumours include aggressive-phenotype tumours as glio- tumours annually estimated in EU and for about 90%
blastoma (ICD-O 9440) but also lower grade astrocyto- of the 23,859 ‘glial tumours of CNS’. Therefore, the
mas (ICD-O 9400) and gliomas with a relatively good overall epidemiological figures of rare CNS tumours in
prognosis. According, the 5-year survival varied from Europe are greatly influenced by astrocytic tumours,15
4.9% for high grade to 43% for low grade tumours. especially by high grade astrocytic that are the most fre-
Within ‘non glial tumours’, a good prognosis was quent (62% of astrocytic tumours) and have the worse
estimated for choroid plexus carcinomas (63%) and an survival. Astrocytic tumours influence also the overall
intermediate survival was observed for embryonal prevalence of rare CNS tumours. In fact, due to the
tumours (57%). poor survival of astrocytic (high grade) tumours, their
For glial tumours five-year survival was slightly relevance is slightly lower for complete prevalence than
higher for women (20.7%; 95% confidence interval for incidence. Astrocytic tumours represent 78% of the
19.6–21.9) than for men (18.7%; 95% CI 17.8–19.7) 154,000 EU complete prevalent cases, and ‘glial
(data not shown). Overall, 5-year survival was quite tumours’ in total represent 85%.
good for children and adolescents (0–19 years; 58.1%), Incidence of astrocytic tumours is higher among men
and for young adults (20–39 years; 52.8.0%), and than women. The 1.4 male excess observed in Europe
decreased steadily for adults (40–59 years; 19.1%), and has also been documented in the US, both in the
especially for older subjects (60+ years; 4.4%) (data SEER16 and in the Central Brain Tumour Registry of
not shown). The decreasing survival with increasing the United States (
age was a common pattern for all rare CNS types with Overall, the median age estimated at diagnosis for
the exception of ependymal tumours for which the CNS tumours is 56 (55 for males and 57 for females)
decrease was less pronounced than for other tumours. years in the United States16 and 53 (53 for males and
The overall EU 5-year survival for glial tumours was 54 for females) years in Europe.
19.6%, with the highest values in northern (22.8%) and The overall incidence increases with age with the
central Europe (22.6%), intermediate value in southern highest values among subjects of 60+ years. This overall
Europe (19.2%) and the lowest survival rates in eastern result is mainly driven by astrocytic tumours, while
Europe (17.2%) and in UK and Ireland (16.6%). The embryonal tumours have the highest incidence rate in
overall geographical difference was mainly driven by dif- the age-group 0–19, oligodendroglial tumours in the
ferences for astrocytic tumours with less evident gradi- age-group 40–59 and ependymal tumours have stable
ent for the other rare CNS types. rates. This confirms that, although rare CNS incidence
There was geographic variation in 1-year survival increases with age, some of the CNS tumours are more
from astrocytic tumours, ranging from 31% in UK common in younger adults. During childhood CNS
and Ireland, to 46% in northern and southern Europe tumours are the most common type of solid tumour
(data not shown). accounting for one-fifth to one-fourth of all the
3.4. Estimated proportion of cured patients The incidence of CNS cancers varies widely across
the world with the highest rates in developed Western
The proportion of patients that did not show an (Europe and North America) and Western-type (Aus-
excess mortality compared with that for the general pop- tralia/New Zealand) countries.2 Also across Europe
ulation (‘cured patients’) varied widely among the differ- there is certain variability in age-standardised incidence
ent groups of rare CNS tumours. The highest rates—the highest values were evidenced in UK, Ire-
proportion of cured, 40.8%, was reported for ‘non glial land and northern Europe and the lowest ones in East-
tumour of CNS and pineal gland’; the lowest (7.9%) for ern Europe. The geographical differences worldwide
‘glial tumours of the CNS’, mainly due to the very poor may be due, at least in part, to the availability of
survival of astrocytic tumours that are the most com- highly advanced imaging technology,17 and this may
mon glial tumours (and with high grade astrocytic which be true, to some extent, also across Europe. The use
represent 62% of all astrocytic tumours). of magnetic resonance imaging (MRI) may also lead
to the diagnosis of unexpected lesions. In a recent sys-
4. Discussion tematic review and meta-analysis of 16 studies, the
prevalence of neoplastic incidental brain findings was
The present study shows that rare CNS cancers rep- 0.70%.18 The prevalence of neoplastic lesions
resent a vast array of tumours with different biological significantly increased with age and it was more likely
characteristics and different epidemiological features. using high resolution MRI sequences than standard
Although all rare CNS tumours belong to rare tumours resolution sequences.18 However, survival data do not
E. Crocetti et al. / European Journal of Cancer 48 (2012) 1532–1542 1539

support a relevant amount of indolent lesions to UK-England (17.6%) and UK-Northern Ireland
explain the differences in Europe, and especially the (15.3%).4 Low survival values were measured in UK
high incidence rates in UK and Ireland.4 and Ireland also in the present study. In a recent paper,
As regards incidence of CNS tumours although the the survival differences between European regions were
aetiology of such tumours is largely unknown there is mainly attributed to the case-mix (morphological type
a major concern for the possible role of cellular phone. of tumours). However, also after the adjustment for
The studies published up to the beginning of 2011 pro- age and morphology there were an excess risk of death
vided conflicting results thus are not useful to interpret of 10% in Southern Europe, 20% in Central Europe,
geographical differences in incidence of CNS across 30% in Eastern Europe and 40% in UK and Ireland in
EU regions. Several studies, including the recent Inter- comparison with Northern Countries.5
phone study, did not demonstrate an increased risk The overall geographical difference was mainly driven
within approximately 10 years of ‘reasonable’ use of by differences for astrocytic tumours with less evident
phones for any tumour of the brain or any other head gradient for the other rare CNS types. Astrocytic
tumour.19,20 However, other studies suggested higher tumours include aggressive-phenotype tumours as glio-
risk for malignant brain tumours in people with ipsilat- blastoma but also tumours with a relatively good prog-
eral mobile phone use and >10 years latency period.21 nosis. In our series glioblastomas were less frequent in
The issue has been recently addressed by a working Eastern (38%) than in the other European regions
group coordinated by the International Agency for (around 50%). This suggests that other prognostic fac-
Research on Cancer specifically aimed at a IARC tors than case mix of the neoplasm contributed to the
Monograph on the potential carcinogenic hazards from survival variation across Europe.
the exposure to radiofrequency electromagnetic fields Similar geographical variation in survival was found
frequency (range 30kHz-300GHz). According to the for adults with any type of brain tumour in EURO-
critical revision of the available evidences IARC classi- CARE-4.4 Survival was quite similar across Europe
fied radiofrequency electromagnetic fields as possibly for children and young adults, though with lower values
carcinogenic to humans (Group 2B). The Monograph in UK.6
is in press,22 a concise summary with the main conclu- Survival was slightly better for women than for men,
sions has been recently published.23 as elsewhere documented2,26 and the strong tendency for
As regards time trends, data from the present study survival rates to worsen with age observed in the present
show stable (or even decreasing) incidence trends in all study is consistent with previous findings from survival
the age-classes analysed with the exception of the age- time-trend analyses in Nordic countries2
group 60+ years which shows an increasing trend. The geographical differences in 1-year survival may
The increased use of technology imaging devices, such be due to a variation in the timeliness of diagnosis and
as computed tomography (CT) and MRI, may have con- therapy which may exert its effect in the short term.
tributed to the more precise diagnosis of such tumours. Brain cancer 1-year survival in 23 European countries
Trends for CNS with and without microscopic confirma- ranged between 34.2% and 48.3%, with high survival
tion showed an increase in incidence for microscopic ver- estimates in Switzerland, France, Sweden, Belgium and
ified cases especially among the elderly. The improved Italy and lower estimates in Poland, Czech Republic,
ability to diagnose brain tumours by stereotactic and Ireland, Denmark and United Kingdom–Northern Ire-
frameless biopsy procedures may have contributed to land.4 Five-year survival conditional on having survived
the increase in microscopically verified cases.24,25 the first year after diagnosis varied less between coun-
The present study showed very different survival tries.4 Also in the present study there were values lower
results for different rare CNS tumours with the worst than the EU average for UK and Ireland for 1-year sur-
prognosis for astrocytic tumours (14.5% survival at vival for astrocytic tumours.
5-years, 5% for high grade astrocytic tumours), In the treatment of brain cancer, radiotherapy is
moderately good values for embryonal (56.8%), oligo- widely used. Therefore, the access to radiotherapy might
dendroglial (54.5%) and choroid plexus carcinomas influence the treatment outcome and consequently sur-
(42.6%) and the best survival for ependymal tumours vival of CNS tumours. The number of linear accelera-
(74.2%). The overall CNS value is strongly influenced tors per million population is far lower in UK and
by the frequency of astrocytic tumours. Recent Ireland, as well as in Eastern Europe compared to Nor-
European estimates – that included all the CNS histolo- dic and Central-Europe countries. ESTRO QUARTS
gies – showed an overall 5-year survival of 19.7% for project has analysed the ratio of actual number of
brain tumours.4 megavoltage therapy units to the evidence-based
This datum was quite similar across Europe with val- required numbers in 13 European countries.27 The
ues higher than the European average for Finland largest gap was seen for Eastern-European countries
(26.8%), Portugal (24.5%) UK-Wales (23.8%) and (Slovenia and Poland) followed by the Czech Republic
Norway (23.5%) and values below the average for and UK. In contrast, in Northern and Central European
1540 E. Crocetti et al. / European Journal of Cancer 48 (2012) 1532–1542

countries (Sweden, France and Belgium), the availability ependymal and embryonal tumours and almost 60 for
of megavoltage therapy units exceeded 90% of the choroid plexus carcinomas.
QUARTS estimate. Rare CNS tumours represent a heterogeneous group
The lower survival rates reported in this study for with great differences in terms of frequency, prognosis
some EU regions could be explained by less accessibility and treatment approaches. Despite remarkable
to Magnetic Resonance Imaging (MRI). EUROSTAT improvement in the ability to diagnose and treat brain
data on medical technology (http://epp.euro- tumours, the prognosis for most of rare CNS cancer showed for the year 2000 a huge var- patients remains poor. The varying availability of highly
iability of the number of MRI units among EU advanced imaging technology and of radiotherapy
countries: from 0.0 per 100,000 persons in Lithuania equipment as well as the lack of evidence based proto-
and 0.1 in Slovakia to 0.8 in Italy. Different availability cols reduces the possibility of getting timely diagnosis
of surgical, radio-therapeutical and chemotherapeutical and treatment. This means that management of these
facilities across Europe can also contribute to explain rare brain tumours has to be considered hyper-specialis-
the lower survival observed in some countries. Further tic, and the treatment has to be centralised in large vol-
data from EUROSTAT showed that in the year 2000, umes hub centres with recognised expertise in this field.
the number of operations on the nervous system (not
only for cancer) varied from 69 in Bulgaria and 74 in Funding
Romania to 1268 per 100,000 inhabitants in Belgium.
Evidence based medicine is, by definition, weaker in rare This research was supported by the European Com-
tumours: this results, in CNS tumours, in a reduced util- mission through the Executive Agency for Health and
isation of more aggressive, still not of high-evidence lev- Consumers (Grant No. 2006113), and the Programma
els, treatments such as total resection in low grade Italia-USA Malattie Rare (Grant No. 526D/42).
gliomas achievable by more modern techniques (i.e.
awake surgery) or less intense chemotherapeutic Conflict of interest statement
approaches in malignant tumours, such as adjuvant
temozolomide longer than the 6 conventional cycles, None declared.
more than one salvage treatment or the use of Gliadel
wafer at recurrence.28,29 Appendix A. The RARECARE Working Group for this
In the present study pilocytic astrocytoma were paper consists of:
excluded from the analyses. We do not know whether
these tumours are represented in the same proportions Austria: N. Zielonk (Austrian National Cancer Reg-
in the glioma NOS and astrocytoma NOS tiers across istry); Belgium: E. Van Eycken (Belgian Cancer Regis-
Countries. This is especially important for the younger try), H. Sundseth, (European Cancer Patient
age groups. In EUROCARE-4, pilocytic astrocytoma Coalition), S. Marreaud (European Organisation for
accounted for 49% of all astrocytomas in children aged Research and Treatment of Cancer), R. Audisio (Euro-
0–14 and 24% of astrocytomas in adolescents and young pean Society of Surgical Oncology); France: G. Hedelin,
adults aged 15–24 years; inclusion of pilocytic astrocy- (Bas-Rhin Cancer Registry); G. Launoy (Calvados
toma increased the 5-year period survival estimates for Digestive Cancer Registry); A.V. Guizard (Calvados
astrocytoma from 63% to 78% in children and from General Cancer Registry); A.M. Bouvier (Côte d’Or
56% to 64% in adolescents and young adults.4 Digestive Cancer Registry); A.S. Woronoff (Doubs Can-
Prevalence estimates are a useful tool for public cer Registry); A. Buemi (Haut-Rhin Cancer Registry);
health strategies and evaluations, as they measure the B. Tretarre (Hérault Cancer Registry); M. Colonna
burden on the health care system.30 Prevalence of CNS (Isère Cancer Registry); S. Bara (Manche Cancer Regis-
tumours was about 30 per 100,000, below the threshold try); O. Ganry (Somme Cancer Registry); P. Grosclaude
of the European definition for rare diseases (50 per (Tarn Cancer Registry); Germany: B. Holleczek (Saar-
100,000) therefore these tumours are rare according also land Cancer Registry); J. Geissler (C.M.L. Advocates
to the European definition on rare diseases31 and may Network; Iceland: L. Tryggvadottir (Icelandic Cancer
profit to the EU Directive on orphan drugs. Prevalence Registry); Ireland: H. Comber (National Cancer Regis-
estimates showed that there are about 154,000 citizens try of Ireland); Italy: F. Bellù (Alto Adige Cancer Reg-
living with rare CNS tumours in EU, 27,000 of them istry); S. Ferretti (Ferrara Cancer Registry); D. Serraino
diagnosed within 2 years and therefore in a period with (Friuli Venezia Giulia Cancer Registry); M. Vercelli
very intensive follow-up. As with most cancers, mortal- (Liguria Cancer Registry c/o IST/UNIGE, Genoa); S.
ity is highest for CNS tumours in the short term. The Vitarelli (Macerata Province Cancer Registry); M. Fede-
ratio between complete prevalence and estimated annual rico (Modena Cancer Registry); M. Fusco (Napoli Can-
number of diagnosed cases is around 4 for astrocytic cer Registry); A. Traina (Palermo Breast Cancer
tumours, 8 for olygodendroglial tumours, 20 for Registry); M. Michiara (Parma Cancer Registry);
E. Crocetti et al. / European Journal of Cancer 48 (2012) 1532–1542 1541

A. Giacomin (Piedmont Cancer Registry, Province of Cancer Registry); M. Roche (Oxford Cancer Intelli-
Biella); R. Tumino (Cancer Registry and Histopathol- gence Unit); J. Verne (South-West Public Health Obser-
ogy Unit, “M.P. Arezzo” Civic Hospital, Ragusa); L. vatory); D. Meechan (Trent Cancer Registry); G.
Mangone (Department of Research Azienda Ospedali- Lawrence (West-Midlands Cancer Intelligence Unit);
era Arcispedale Santa Maria Nuova – IRCCS, Reggio M.P. Coleman (London School of Hygiene and Tropical
Emilia); F. Falcini (Romagna Cancer Registry); A. Ian- Medicine), J. Mackay (University College of London);
nelli (Salerno Cancer Registry); M. Budroni (Sassari UK-Northern Ireland: A. Gavin (Northern Ireland Can-
Cancer Registry); S. Piffer (Trento Cancer Registry); cer Registry); UK-Scotland: D.H. Brewster (Scottish
T. Intrieri (Tuscan Cancer Registry); F. La Rosa Cancer Registry); I. Kunkler (University of Edinburgh);
(Umbria Cancer Registry); P. Contiero (Varese Cancer UK-Wales: J. Steward (Welsh Cancer Intelligence &
Registry); P. Zambon (Veneto Cancer Registry); P.G. Surveillance Unit).
Casali, G. Gatta, A. Gronchi, L. Licitra, M. Ruzza, S.
Sowe, (Fondazione IRCCS Istituto Nazionale dei
Tumori); R. Capocaccia, R. De Angelis, S. Mallone, 1. Wrensch M, Minn Y, Chew T, Bondy M, Berger MS. Epidemi-
A. Tavilla (Centro Nazionale di Epidemiologia, Istituto ology of primary brain tumors: current concepts and review of the
Superiore di Sanità); A.P. Dei Tos (Local Health Unit literature. Neuro Oncol 2002;4:278–99.
No. 9, Region of Veneto); A.A. Brandes (Medical 2. Ferlay J, Shin HR, Bray F, et al. GLOBOCAN 2008, Cancer
Oncology Department, Local Health Unit, Bologna); Incidence and Mortality Worldwide: IARC CancerBase No. 10
[Internet]. Lyon, France: International Agency for Research on
Malta: K. England (Malta National Cancer Registry); Cancer; 2010. Available from:
Norway: G. Ursin (Cancer Registry of Norway); Nether- 3. Kaatsch P. Epidemiology of childhood cancer. Cancer Treat Rev
lands: O. Visser, R. Otter, S. Siesling, J.M. van der Zwan 2010;36:277–85.
(Comprehensive Cancer Centre the Netherlands); 4. Sant M, Allemanni C, Santaquilani M, et al. EUROCARE-4.
J.W.W. Coebergh (Eindhoven Cancer Registry), H. Survival of cancer patients diagnosed in 1995–1999. Results and
commentary. Eur J Cancer 2009;45:931–91.
Schouten (University of Maastricht); Poland: J. Rach- 5. Sant M, Minicozzi P, Lagorio S, et al. Survival of European
tan (Cracow Cancer Registry); S. Gozdz (Kielce Cancer patients with central nervous system tumours. Int J Cancer 2011;
Registry); M. Zwierko (Warsaw Cancer Registry); M. doi:10.1002/ijc.26335. [Epub ahead of print].
Bielska-Lasota (National Institute of Public Health – 6. Gatta G, Zigon G, Capocaccia R, et al. Survival of European
National Institute of Hygiene, Warsaw); J. Slowinski children and young adults with cancer diagnosed 1995–2002. Eur J
Cancer 2009;45:992–1005.
(Department of Neurosurgery in Sosnowiec, Medical 7. World Health Organization. International Classification of Dis-
University of Silesia); Portugal: A. Miranda (Southern eases. 10th revision, Geneva; 1992.
Portugal Cancer Registry); Slovakia: Ch. Safaei Diba 8. Percy C, Shanmugaratnam K, Whelan S, et al. International
(National Cancer Registry of Slovakia); Slovenia: M. Classification of Diseases for Oncology (ICD-O), 3rd ed. Geneva,
Primic-Zakelj (Cancer Registry of Slovenia); Spain: A. Switzerland: World Health Organization; 2000.
9. Kros JM, Gorlia T, Kouwenhoven MC, et al. Panel review of
Mateos (Albacete Cancer Registry); I. Izarzugaza (Bas- anaplastic oligodendroglioma from European Organization For
que Country Cancer Registry); A. Torrella-Ramos Research and Treatment of Cancer Trial 26951: assessment of
(Castillon Cancer Registry); R. Marcos-Gragera (Epide- consensus in diagnosis, influence of 1p/19q loss, and correlations
miology Unit and Girona Cancer Registry, Oncology with outcome. J Neuropathol Exp Neurol 2007;66:545–51.
Coordination Plan, Department of Health and Catalan 10. Louis DN, Ohgaki H, Wiestler OD, et al. The 2007 WHO
classification of tumours of the central nervous system. Acta
Institute of Oncology, Girona, Spain); M.J. Sánchez Neuropathol 2007;114:97–109.
(Granada Cancer Registry); C. Navarro (Department 11. Capocaccia R, Colonna M, Corazziari I, et al. Measuring cancer
of Epidemiology, Murcia Regional Health Authority, prevalence in Europe: the EUROPREVAL Project. Ann Oncol
Murcia, CIBER Epidemiologı́a y Salud Pública (CIB- 2002;13:831–9.
ERESP)); Eva Ardanaz (Navarra Cancer Registry); J. 12. Capocaccia R, De Angelis R. Estimating the completeness of
prevalence based on cancer registry data. Stat Med
Galceran (Tarragona Cancer Registry); C. Martinez- 1997;16:425–40.
Garcia, J.M. Melchor (Escuela Andaluza de Salud Púb- 13. Brenner H, Söderman B, Hakulinen T. Use of period analysis for
lica); Sweden: J. Adolfsson (Stockholm-Gotland Cancer providing more up-to-date estimates of long-term survival rates:
Registry); M. Lambe (Uppsala Regional Cancer Regis- empirical evaluation among 370,000 cancers patients in Finland.
try), T.R. Möller (Lund University Hospital); U. Ring- Int J Epidemiol 2002;31:456–62.
14. Francisci S, Capocaccia R, Grande E, et al. The cure of cancer: a
borg (Karolinska Institute); Switzerland: G. Jundt European perspective. Eur J Cancer 2009;45:1067–79.
(Basel Cancer Registry); M. Usel (Geneva Cancer Reg- 15. Schwartzbaum JA, Fisher JL, Aldape KD, Wrensch M. Epidemi-
istry); S.M. Ess (St. Gallen Cancer Registry); A. Bor- ology and molecular pathology of glioma. Nat Clin Pract Neurol
doni (Ticino Cancer Registry); I. Konzelmann (Valais 2006;2:494–503.
Cancer Registry); J.M. Lutz (National Institute for Can- 16. Altekruse SF, Kosary CL, Krapcho M, et al., editors. SEER
Cancer Statistics Review, 1975–2007, Bethesda, MD: National
cer Epidemiology and Registration); UK-England: D.C. Cancer Institute. Available from:
Greenberg (Eastern Cancer Registration and Informa- 1975_2007/, based on November 2009 SEER data submission,
tion Centre); J. Wilkinson (Northern and Yorkshire posted to the SEER web site, 2010.
1542 E. Crocetti et al. / European Journal of Cancer 48 (2012) 1532–1542

17. Parkin DM, Bray F, Ferlay J, Pisani P. Global Cancer Statistics, 25. Hoffman S, Propp JM, McCarthy BJ. Temporal trends in
2002. CA Cancer. J Clin 2005;55:74–108. incidence of primary brain tumors in the United States, 1985–
18. Morris Z, Whiteley WN, Longstreth Jr WT, et al. Incidental 1991. Neuro-Oncol 2006;8:27–37.
findings on brain magnetic resonance imaging: systematic review 26. Micheli A, Ciampichini R, Oberaigner W, et al. The advantage of
and meta-analysis. BMJ 2009;339:b3016. women in cancer survival: An analysis of EUROCARE-4 data.
19. Ahlbom A, Feychting M, Green A, et al. Standing Eur J Cancer 2009;45:1017–27.
committee on epidemiology. epidemiologic evidence on 27. Bentzen SM, Heeren G, Cottier B, et al. Towards evidence-based
mobile phones and tumor risk. A review. Epidemiology guidelines for radiotherapy infrastructure and staffing needs in
2009;20:639–52. Europe: the ESTRO QUARTS project. Radiother Oncol
20. INTERPHONE Study Group. Brain tumour risk in relation to 2005;75:355–65.
mobile telephone use: results of the INTERPHONE international 28. Brem H, Piantadosi S, et al. Placebo-controlled trial of safety and
case-control study.Int J Epidemiol 2010;39:675–94. efficacy of intraoperative controlled delivery by biodegradable
21. Khurana VG, Teo C, Kundi M, Hardell L, Carlberg M. Cell polymers of chemotherapy for recurrent gliomas. The Polymer-
phones and brain tumors: a review including the long-term brain Tumor Treatment Group.. Lancet 1995;345:1008–12.
epidemiologic data. Surgical Neurol 2009;72:205–15. 29. Mason WP, Maestro RD, Eisenstat D, et al. Canadian recom-
22. IARC. IARC monograph on the evaluation of carcinogenic risks mendations for the treatment of glioblastoma multiforme. Curr
to humans. In: Non-Ionizing radiation, radiofrequency electro- Oncol 2007;14:110–7.
magnetic fields, vol. 102. Lyon: International Agency for Research 30. Porter KR, McCarthy BJ, Freels S, Kim Y, Davis FG. Prevalence
on cancer, in press. estimates for primary brain tumors in the United States by age,
23. Baana R, Grosse Y, Lauby-Secretan B, et al. Carcinogenicity of gender, behavior, and histology. Neuro Oncol 2010;12:520–7.
radiofrequency electromagnetic fields. The Lancet Oncol 31. European Parliament and Council of the European Communities
2011;12:624–6 [published Online: 22 June 2011]. (1999). Decision No 1295/1999/EC of the European Parliament
24. Dammers R, Schouten JW, Haitsma IK, et al. Towards improving and of the Council of 29 April 1999 adopting a programme of
the safety and diagnostic yield of stereotactic biopsy in a single Community action on rare diseases within the framework for
centre. Acta Neurochir . doi: 10.1007/s00701-010-0752-0. action in the field of public health (1999 to 2003).