FREQUENTLY ASKED QUESTIONS
Are the GLOBOCAN estimates for previous years (e.g. 2002) still available?
No. The increasing availability of incidence data from cancer registries and mortality data from vital statistics offices, coupled with the fine-tuning of the estimation methods means that the current incidence and mortality estimates for 2008 cannot be directly compared to estimates from previous versions (e.g. GLOBOCAN 2002). To avoid potential confusion among users, we prefer to provide only the latest and most reliable statistics on the current global cancer incidence and mortality burden (for 2008). If there are specific requirements to view GLOBOCAN 2002, please email a request to cin@iarc.fr.
What are the major differences between GLOBOCAN 2008 and GLOBOCAN 2002?
There are several differences between GLOBOCAN 2002 and GLOBOCAN 2008 which relate to availability of incidence and/or mortality data, methods used to estimate rates in the absence of data and methods used to project available data for 2008.
The methodology for GLOBOCAN 2008 is described in detail in Ferlay et al, Int J Cancer. 2010; 127(12):2893-2917. A summary is also provided using the Data sources and methods menu option.
I used to have my own copy of GLOBOCAN 2002, supplied on CD Rom. Will this be available for GLOBOCAN 2008?
GLOBOCAN 2008 uses a different database structure from GLOBOCAN 2002 and we have not yet been able to prioritise the development of a CD Rom version. However, the online only implementation does allow us to make amendments to the underlying dataset (when, for example errors are notified to us) and to the functionality without the need to redistribute updates to CD Roms.
I have seen other estimates for the global burden of cancer, e.g. from the WHO Global Burden of Disease (GBD) project. Are they comparable with these ones?
IARC and WHO produce different estimates of the cancer burden, and this is an unfortunate source of confusion. WHO (through the GBD project) aims at producing estimates of the burden of 150 diseases (including 28 cancers), using mortality data as the first source of information together with modelling procedures. IARC (through the GLOBOCAN project) uses cancer registry data as the first source of information: when necessary, mortality is derived from incidence estimates and survival probabilities.
These two different approaches have historically produced different estimates of the global burden of cancer. The results were very similar for regions with good vital registration data (Europe, the Americas and some Asian Pacific countries). For other regions, the estimates of cancer mortality (and, subsequently, cancer incidence) differed, particularly in Africa and India. For these new 2008 estimates, IARC and WHO worked closely in order to produce comparable estimates. For countries without vital registration system, the GLOBOCAN 2008 mortality estimates are now based on preliminary results from the WHO, except for Africa for which IARC still relies upon available cancer registry data rather than on model-based approach (see Data sources and methods menu option).
What are the differences between mortality data available in GLOBOCAN 2008 and those available with the WHO mortality databank?
The mortality data presented in GLOBOCAN 2008 are estimates based on recorded data available through the WHO mortality databank for earlier periods together with prediction methods. In addition, the WHO mortality data may have been corrected to take into account some degree of incompleteness or under coverage. These corrections are incorporated into GLOBOCAN 2008.
What is the relationship between GLOBOCAN and Cancer Incidence in Five Continents (CI5)?
The series of CI5 volumes present detailed information on the incidence of cancer recorded by cancer registries (regional or national) worldwide during a specific period of time (generally 5-year). It is the major source of information to compute the country-specific cancer incidence estimates presented in GLOBOCAN. National incidence rates published in CI5 may not correspond exactly to those published in GLOBOCAN 2008 because they are presented for a different time period, and computed using a different set of populations.
I am interested in childhood cancers. Can GLOBOCAN provide any statistics for this group of important cancers?
Childhood cancers make use of a different classification system, including morphological sub-types. The cancer sites available through GLOBOCAN 2008 are based on ICD-10 codes (see The cancer dictionary menu option), which are not satisfactory for childhood cancer studies, but data for 0-14 year age group are available from GLOBOCAN 2008.
How frequently will GLOBOCAN be updated?
We plan to update the GLOBOCAN 2008 estimates every year with available recorded data for 2008 (see below). A new set of estimates for 2012 will be made available in 2013 together with publication of CI5 Vol. X.
We will also, periodically, make minor modifications to the underlying data within GLOBOCAN 2008 in order to correct identified errors. The current release version of GLOBOCAN 2008 will be identified on the website and an error correction log maintained.
Are there specific warnings we should note when using GLOBOCAN 2008?
• Estimates for 2008 provided in GLOBOCAN 2008 may not always correspond to recorded data which are now becoming available. Regular updates of GLOBOCAN 2008 are planed when sufficient recorded national incidence and mortality data for 2008 are available. However, this will have little impact on the estimation of the overall burden of cancer.
• The estimated numbers of cancer cases or deaths for a region may not correspond to the sum of the estimated numbers of cancer cases or deaths of the individual countries: the population for the region may include some small country populations for which no estimates are provided.
• Prevalence estimates are provided for the total adult population only (age 15 and over, no breakdown by age-groups).
• Care should be taken when using the ONLINE ANALYSIS Incidence/Mortality > Prediction option. The expected number of new cancer cases or deaths in a country or region in 2010,...,2030 is computed by multiplying the age-specific incidence/mortality rates estimated for 2008, by the corresponding expected population for 2010,...,2030. These expected populations differ from that of 2008, which is used to compute the estimated rates, in term of age structure and size. Therefore, the sum of the expected numbers of new cancer cases or deaths in 2010,..,2030 in different countries or regions will not correspond to the expected number of cancer cases or deaths in 2010,...,2030 if these were computed as a single area. For example, the expected number of deaths from cervical cancer worldwide in 2030 is different when computed using the world mortality rates and the expected world population for 2030, than the sum of the corresponding estimates for 2030 obtained from the less and more developed regions.
• Kaposi sarcoma (KS) is estimated for sub-Saharan countries only. KS is also provided for sub-Saharan, AFRO and Africa build-in regions. KS has been included in the Africa build-in group although KS is not available for Northern Africa (where it is a rather rare disease).