We estimated the risk of subsequent bone cancer among 9170 patients who had survived two or more years after the diagnosis of a cancer in childhood. As compared with the general population, the patients had a relative risk of 133 (95 percent confidence interval, 98 to 176) and a mean (+/- SE) 20-year cumulative risk of 2.8 +/- 0.7 percent. Detailed data on treatment were obtained on 64 patients in whom bone cancer developed after childhood cancer. As compared with 209 matched controls who had survived cancer in childhood but who did not have bone cancer later, patients who had had radiation therapy had a 2.7-fold risk (95 percent confidence interval, 1.0 to 7.7) and a sharp dose-response gradient reaching a 40-fold risk after doses to the bone of more than 6000 rad. The relative dose-response effect among patients who had been treated for retinoblastoma resembled that among patients with all other types of initial tumors, although the cumulative risk of bone cancer in the retinoblastoma group was higher. Similar numbers of patients were treated with orthovoltage and megavoltage; the patterns of risk among categories of doses did not differ according to the type of voltage. After adjustment for radiation therapy, treatment with alkylating agents was also linked to bone cancer (relative risk, 4.7; 95 percent confidence interval, 1.0 to 22.3), with the risk increasing as cumulative drug exposure rose. We conclude that both radiotherapy and chemotherapy with alkylating agents for childhood cancer increase the subsequent risk of bone cancer.

No abstract

Compared with the general population, members of Li-Fraumeni syndrome families have an exceptionally high risk of developing multiple primary cancers. The excess risk of additional primary cancers is mainly for cancers that are characteristic of Li-Fraumeni syndrome, with the highest risk observed for survivors of childhood cancers. Cancer survivors in these families should be closely monitored for early manifestations of new cancers.

Linkage analysis was performed on a large pedigree with an autosomal dominant platelet disorder and a striking propensity in affected family members to develop hematologic malignancy, predominantly acute myelogenous leukemia. We report the linkage of the autosomal dominant platelet disorder to markers on chromosome 21q22. Four genetic markers completely cosegregate with the trait and yield maximum logarithm of difference scores ranging from 4.9 to 10.5 (theta = .001). Two flanking markers, D21S1265 and D21S167, define a critical region for the disease locus of 15.2 centimorgan. Further analysis of this locus may identify a gene product that affects platelet production and function and contributes to the molecular evolution of hematologic malignancy.