The risk of cancer in twins: A report from the childhood cancer survivor study

Abstract

Twin concordance studies help evaluate the contribution of genetic factors in childhood cancers, but previous reports have primarily focused on leukemia because of the rarity of other malignancies. In the current report, a large cohort of childhood cancer survivors was used to: (1) describe twin concordance patterns for a range of cancers, (2) calculate the standardized incidence rates of cancers in twins, and (3) describe clinical features and outcomes of concordant twins. Cancer family history was obtained on the 211 twins participating in the Childhood Cancer Survivor Study (CCSS) (14,352 participants surviving > or =5 years after a malignancy diagnosed at <21 years during January 1, 1970-December 31, 1986) to calculate probandwise twin concordance rates and standardized incidence ratios (SIRs) using Surveillance, Epidemiology, and End-Results data. Seven monozygotic twin pairs were concordant for cancer (six for leukemia, one for non-Hodgkin lymphoma), yielding probandwise concordance rates of 9.5%, 20.7%, and 20.0% for all cancer, leukemia, and non-Hodgkin lymphoma (NHL), respectively. No concordance was observed among dizygotic twins or for dissimilar cancers. The SIR in monozygotic twins was 23.3 (95% CI = 11.1-48.9) for all cancer, 112.4 (95% CI = 50.5-250.1) for leukemia, and 40.5 (5.7-287.5) for NHL. Concordant twins were similar in age at diagnosis and vital status. Twin concordance for cancer is largely restricted to monozygotic twins and hematological malignancies, consistent with in utero malignancy transmission demonstrated by others. Our data support clinical monitoring of the twins of cases with hematological malignancies, and does not contribute evidence for genetic factors in other cancers.