5355

With the exception of certain predisposing syndromes and in utero exposure to X-rays, there are few confirmed risk factors for childhood cancer. A few studies have found an excess of minor congenital malformations in children with cancer. It has been suggested that these “mild errors of morphogenesis” (MEM), which can be detected at birth, represent general genetic instability that could increase the risk of malignancy. One study noted a positive association between birthmarks, considered to be MEMs, and acute lymphoblastic leukemia (ALL) in children who were not documented as having neurofibromatosis type I, a genetic disease characterized by café au lait spots and an increased risk for ALL. We examined whether the frequency of birthmarks is increased in children with cancer using data from the Collaborative Perinatal Project (CPP), a prospective cohort study comprised of 54,795 U.S. children born between 1959 and 1968. The CPP systematically collected data related to all aspects of the overall health of its participants at multiple time points, including detailed information on the pregnancy, delivery, and growth and development of the child up to the age of 7 or 8 years. In addition, extensive information was collected on social and demographic characteristics of the child’s family. There were 51 incident cases of childhood cancer identified over 7-8 years of follow-up including: 17 leukemias, 8 central nervous system tumors, 6 neuroblastomas, 7 Wilms’ tumors, 5 lymphomas, 3 retinoblastomas, and 5 other cancers. Birthmarks were documented during the first year through history or medical exams that occurred at birth, 4 months and 1 year and included: hemangiomas (port-wine, strawberry, or cavernous), pigmented nevi, lymphangiomas, and café au lait spots. In the CPP, 2505 (4.6%) of subjects had a documented birthmark, including 7 (14%) children with cancer and 2498 (4.6%) children without cancer. In a Cox proportional hazards regression model, birthmarks were associated with a significant increase in the risk of cancer (hazard ratio (HR) = 3.0; 95% confidence interval (CI)=1.35-6.68). Adjusting for gender, race, socioeconomic status, gestational age, or birth weight did not materially change the risk estimate. There was a slight attenuation of the risk estimate when cases diagnosed in the first year of life were excluded (HR=2.6; 95%CI=1.02-6.61). No specific childhood malignancies were notably affected by birthmarks, nor did any particular type of birthmark occur more frequently in children with cancer. In conclusion, although this study was based on a small number of cases, we found birthmarks to be in excess in children who were diagnosed with cancer. These findings provide additional support for the possibility of a shared etiology of MEMs and childhood cancer. This work was supported by the University of Minnesota Children’s Cancer Research Fund.

[Proc Amer Assoc Cancer Res, Volume 47, 2006]