Although relatively rare in comparison with cancer in older adults, cancer is the second leading cause of death in persons aged 1-14 years. The epidemiology of cancer among children differs markedly from that of adults, both in the patterns of anatomic sites involved and the predominant histologic types. Most notably, the tumors diagnosed in children frequently involve the hematopoietic and central nervous systems or are of mesenchymal origin. In contrast, malignancies of epithelial tissues, which are predominant in adults, are uncommon in children. Similar to adult cancers, the etiology of many childhood cancers remains unclear.

The Cancer Data Registry of Idaho receives several requests per year from physicians and others for data on pediatric cancer incidence for the state of Idaho. In 1997, the North American Association of Central Cancer Registries (NAACCR) included Childhood Cancer Incidence in North America, 1988-1992 as a section in Cancer in North America, 1989-1993.1 Modeled after the NAACCR paper, this report describes the incidence of pediatric cancers in Idaho, with comparisons to U.S. data (NAACCR) and comparisons among health districts and counties in Idaho.

METHODS

The data analyzed for this report include cancers diagnosed between 1982 and 1996 among Idaho residents under 15 years of age. Cases were grouped according to the International Classification of Childhood Cancer (ICCC), which is based on histologic features of childhood cancers.2

A total of 618 cases of cancer were diagnosed among Idaho resident children under the age of 15 between 1982 and 1996. It was not possible to assign a group code of the ICCC system to 12 cases (1.9%). In comparison, approximately one percent (0.8%) of cases were excluded for this reason from NAACCR data for the report, Childhood Cancer Incidence in North America, 1988-1992.1 An additional 4 Idaho cases were excluded due to missing county code, resulting in a total of 602 cases included in the study. The following data elements were extracted for each case: age, sex, site, histology, behavior, grade, date of birth, diagnosis date, and county code. Among the 602 cases, 7 (1.2%, compared to 6.1% in Childhood Cancer Incidence in North America, 1988-1992) had unlikely age/site/morphology combinations, but were included in the analyses. County code was defined as county of residence at time of diagnosis. Health district was assigned based upon county code (see Appendix).

All rates presented were calculated per million population, and are fifteen-year person-year averages for the period 1982 through 1996 (rates per million, rather than per 100,000, are commonly used for pediatric cancers). Age-adjustment for state rates was performed using the direct method to the 1970 U.S. population. Pediatric cancer groupings with 100 or more cases statewide from 1982-1996 were selected for analysis by health district and county.

Expected cases and standardized incidence ratios were calculated for individual health districts and counties by applying age- and sex-specific rates for the state of Idaho to population data from the health districts and counties. Adjusted incidence rates for these geographic areas were calculated as the products of the standardized incidence ratios and the state age-adjusted rate. Statistical tests comparing observed and expected numbers of cases were performed using the Poisson probability distribution, two-tailed.

RESULTS

A total of 602 cases that met the study criteria were diagnosed among Idaho residents aged less than 15 years between 1982 and 1996, yielding an overall age-adjusted rate of 147.2 cases per million population. Approximately 45% of pediatric cancers were diagnosed among children aged 0-4 years, 25% were diagnosed among children aged 5-9 years, and 31% were diagnosed among children aged 10-14 years.

As indicated in Table 1, 24% of cases were lymphocytic leukemias. Astrocytomas, the next most frequently diagnosed cancer and the most predominant histologic type reported within the central nervous system and miscellaneous intracranial and intraspinal neoplasms grouping, comprised 11% of all childhood cancer cases reported. Wilms' tumor, the principal histologic type within the renal tumors site group, accounted for 7% of the total cases, and neuroblastomas and ganglioneuroblastomas accounted for 6%. Together, these four cancer types accounted for almost half of the reported cases in Idaho.

The distribution of pediatric cancers by ICCC grouping was very similar in the U.S. and Idaho. In addition, for no ICCC grouping was there a statistically significant difference between the observed number of cases in Idaho and the expected number of cases based upon U.S. rates for whites. Age-specific rates differed markedly across age category depending on ICCC grouping, but were similar for Idaho and the U.S.

Table 2 shows Idaho pediatric cancer incidence statistics for all cancers combined, by health district and county. For both sexes combined, the numbers of observed and expected cases were not statistically significantly different for any health district. For cases among females, Health District 1 had significantly fewer cases than expected. For cases among males, Health District 2 had significantly fewer cases than expected, and Health District 3 had significantly more cases than expected. The excess in observed cases among males in Health District 3 was partially attributable to an excess in observed cases in Gem County (11 cases observed and 3.8 cases expected). Overall, and stratified by sex, several additional counties had statistically significantly fewer cases of pediatric cancer than expected, but all such results were based upon fewer than 10 observed cases, causing unstable rates and making interpretation difficult.

Table 3 shows Idaho pediatric cancer incidence statistics for leukemias, by health district and county. For both sexes combined, the numbers of observed and expected cases were not statistically significantly different for any health district. For cases among males, Health District 2 had significantly fewer cases than expected. Stratified by sex and overall, Bonneville County had statistically significantly fewer cases than expected, but the results were based upon fewer than 10 observed cases, causing unstable rates and making interpretation difficult.

Table 4 shows Idaho pediatric cancer incidence statistics for lymphocytic leukemia, by health district and county. Health District 2 had significantly fewer cases than expected among males. Bannock County, part of Health District 6, had significantly more cases than expected. Kootenai County had statistically significantly fewer cases than expected among females, but the result was based upon fewer than 10 observed cases, causing unstable rates and making interpretation difficult.

Table 5 shows Idaho pediatric cancer incidence statistics for central nervous system and miscellaneous intracranial and intraspinal neoplasms, by health district and county. For no health district were the numbers of observed and expected cases statistically significantly different among males, females, or overall. Stratified by sex and overall, several counties had statistically significantly more cases than expected, but all such results were based upon fewer than 10 observed cases, causing unstable rates and making interpretation difficult.

DISCUSSION

These data demonstrate a strong similarity between pediatric cancer incidence in Idaho and the U.S. For no ICCC grouping was there a statistically significant difference between the observed number of cases in Idaho and the expected number of cases based upon U.S. rates for whites. With few exceptions, there is little evidence of differences in pediatric cancer incidence by health district or county within Idaho. However, the small numbers of cases for many ICCC groupings, particularly when examined at the health district or county levels, make spatial trends difficult to interpret. Based upon the results in this report, the Cancer Data Registry of Idaho will perform additional descriptive epidemiologic analyses for, and continue to monitor, pediatric cancers among males in Gem County, and lymphocytic leukemia in Bannock County.

REFERENCES

1. Carozza SE et al., Childhood Cancer Incidence in North America, 1988-1992. In Howe HL, Lehnherr M (Eds). Cancer in North America, 1989-1993. Volume One: Incidence. Sacramento, CA: North American Association of Central Cancer Registries, 1997.
2. Kramàrovà E , Stiller CA, Ferlay J, Parkin DM, Draper GJ, Michaelis J, Neglia J, Qureshi S (Eds). International Classification of Childhood Cancer 1996, IARC Technical Report No. 29. Lyon: International Agency for Research on Cancer, 1996.