ACCIS – Automated Childhood Cancer Information System

In the European Union, about 1% of all cancers occur in children under 15 years of age, corresponding to a crude estimate of 11,000 new cases per year. Seventy-five percent of these now survive five years after diagnosis, although many will have long-term sequelae, in terms of physical, mental or reproductive impairment over their entire future life.

Childhood cancers represent a diverse spectrum of morphologic tumor types. Whilst over 80% of malignancies in adults are carcinomas, in childhood carcinomas account for less than 10% of tumors. The most common childhood tumor types are sarcomas (45%), leukemias (30%) and lymphomas (15%) ^[1,2]) – (Figure 1). Presumably, the different morphology is a reflection of different histological origin and probably also of differences in aetiology.

Figure 1. Proportionate age distribution of ICCC tumour groups. Europe, 1990s.

An important obstacle in studying the aetiology of tumors in children is their rare occurrence. Due to uncertainty about the causes of cancer in childhood, preventive measures are largely nonexistent. Only studies covering large geographic areas or long time periods will result in sufficient numbers of cases being collected to have the necessary power to answer specific hypotheses of causality. It is the ambition of the Auto.mated Childhood Cancer Information System (ACCIS) to provide the tools necessary to elucidate the reasons for geographical and temporal differences in the incidence of and survival from childhood cancer and thus help to reduce the burden of childhood cancer in the population.

The aim of the ACCIS project is to collect, standardize, interpret and disseminate data on indicators of the cancer burden in the childhood population of Europe. These objectives are being achieved by three approaches:

1. Construction of the ACCIS database containing information on incidence and survival of children with cancer in Europe.
2. Creation and maintenance of the Internet site, as a tool for wide distribution of information about the ACCIS project and collected data.
3. Analysis, evaluation, interpretation and dissemination of the collected data.

With some 160,000 cancer cases registered in Europe over the last 30 years and a population at risk of around 2.6 billion person-years, the ACCIS database is the largest database of young cancer patients in the world.

Figure 2. Contributors to the ACCIS database

Eighty population-based cancer registries in 30 European countries have provided their data for the construction of the ACCIS database, which includes all cases of cancer occurring before the age of 20 (Figure 2).

Every registry provided a list of individual records of cancer cases with a standard set of variables. Received data were validated to limit the coding errors, to allow standard interpretation of the results and to evaluate the overall comparability of the datasets ^[3]). The categories of the International Classification of Childhood Cancer (ICCC) ^[1]) are used to classify tumors and present the results. Survival time was calculated for registries with available follow-up ^[4]).

The population data file contained the number of residents in the registration area in each calendar year of the reported period, by gender and age. All registries also completed a specific questionnaire, which provided the coordinating center with information on the registry and its registration and coding practices.

The ACCIS Scientific Committee examined and commented on each dataset. Any issues that could influence the interpretation of the results were documented within the ACCIS database. The majority of the datasets were considered sufficiently comparable for display to the general public on the Internet site.

Using standard methods of data analysis ^[5]), the incidence rates reported below are standardized to the World standard population. Observed five-year survival was calculated using the life-table method ^[4].

The incidence of childhood cancer in Europe in the 1990s reached 140 cases per million for children aged 0-14. Examination of the time trends has shown that the overall incidence has increased over the 30 years of observation, representing an increase of about 10% per decade in the total European rate.

Figure 3. Time trends of cancer incidence in children (age 0-14). Age-standardized rates, Europe.

Figure 3 shows that a rising trend is observed for the majority of tumor types, apart from a relatively stable incidence of retinoblastoma, renal and hepatic tumors. The decreasing trend seen for the “other and unspecified” tumors can be explained by improvements in diagnosis and registration, as a result of which these tumors are classified in more specific diagnostic groups.

Overall five-year survival for unselected childhood cancer patients diagnosed in Europe in the 1990s reaches almost 75%, with a clear difference being seen between the Eastern and Western European countries, in virtually all diagnostic groups (Figure 4).

An increase in survival over time is seen in all parts of Europe, whilst the differences in survival between the East and the West are diminishing relatively slowly. Despite these geo.graphical differences, the overall survival of European children and adolescents has improved in all of the main diagnostic groups defined by ICCC, including the three major diagnostic groups – leukemias, lymphomas and CNS tumors.

The observations described here are based on good quality datasets, uniformly processed and analysed. More refined analyses are necessary in order to describe the observed patterns more precisely and to be able to generate conclusions and recommendations of importance to public health.

An Internet site was developed to present the ACCIS project, display the criteria for contribution, give information about participants in the ACCIS study, provide links with other resources of data on childhood cancer and, most importantly, to disseminate data on childhood cancer incidence and survival and to give help in the interpretation of the results. The user is able to consult incidence rates and survival for childhood cancer by registry, country, ICCC diagnostic group and five-year age group. The tables are displayed in PDF format. The web page can be explored at the address: http://www-dep.iarc.fr/accis.htm

Figure 4. Five-year survival of children (age 0-14) diagnosed with cancer in different parts of Europe in 1990s. (95% confidence interval represeted by line sections).

The ACCIS database was established with a view to being continuously updated. The database and all other aspects of the project will evolve to make full use of this precious data source. A natural development would be to incorporate data from the registries of other countries, especially those with sparse data, which is the case in developing countries. International differences between incidence rates would point to possible risk factors, while comparison of survival data would reflect differences in availability and delivery of health care. Population-based survival data are practically unavailable, to date, for childhood cancer patients in developing countries. The results of a comparison between Europe and developing countries could provide a powerful tool for the development of public health policies pertaining to childhood cancer in the less affluent nations.

Eva Steliarová-Foucher
International Agency for Research on Cancer (IARC), Lyon, France

References:

^[1]) Kramárová E, Stiller CA (1996) The International Classification of Childhood Cancer. Int J Cancer 68: 759-765.