The data of the Australian Paediatric Cancer Registry on childhood cancer incidence in Australia for the 10-year period 1982-1 99 I are presented. The crude average annual incidence of cancer in children under the age of 15 years was 13.8 per IOO.000. The incidence of childhood cancer in Australia is rising.
Significant increases were seen in acute non-lymphoblastic leukaemia, astrocytoma and melanoma. The age-standardised incidence of 14.4 per 100,000 is about 34% higher than in the UK. Most types of cancer had a higher incidence in Australia than in the UK, and the difference was significant for acute lymphoblastic leukaemia, astrocytoma and melanoma. Of particular interest is malignant melanoma, whose incidence in Australia is more than 5 times that in the UK, as a result of excessive UV exposure. Australia has a higher incidence of Ewing's turnour than osteosarcoma, nearly twice that of the UK. International comparative studies may help to elucidate the aetiology of these tumours.
o 1996 Wiley-Liss, Inc.
Although cancer is one of the commonest potentially fatal diseases in childhood, little is known about its aetiology. Descriptive epidemiology is one way of looking for possible aetiological factors. Establishment of a cancer registry to produce descriptive epidemiological data also provides a data base which can be utilised in further epidemiological or clinical studies.
The Australian Paediatric Cancer Registry (APCR) was established in 1977 to provide data on the incidence of childhood cancer in Australia. At that time, there were no such data available as cancer registration was not in existence in most states. There are now population-based cancer registries in each of the states and in the Northern Territory.
Childhood cancer forms only a small proportion of the total number of cases, so it does not feature separately in the reports of most cancer registries. Consequently, it is appropriate to establish a separate childhood cancer registry for the study of the epidemiology of childhood cancer. The relatively small numbers make it more feasible to determine survival and to collect information on the clinical characteristics of childhood cancer than is the case with adult cancer.