抄録
Study was made of mortality rates for 5 leading childhood cancers over 2 time intervals in the U.S. nd 3 in Japan. The U.S. rates changed very little over time, and, though the Japanese rates increased progressively, they still were well below those for the U.S. for each site considered. A peak in leukemia mortality appears to be emerging at 3 years of age in Japan as it did 25 years ago among U.S. white children. Kidney neoplasia (principally Wilms' tumor) appears from mortality data to be much less frequent and to have an earlier peak in Japan than in the U. S. The explanation of this difference awaits the accumulation of additional information through Japanese childhood cancer registries. Preliminary results from these registries suggest that the large differences between the U.S.A.nd Japan in mortality from neuroblastoma may be a consequence of dissimilar diagnostic or classification practice.