Pharmacogenetic and Pharmacogenomic Studies in Pediatric Leukemia
Although current chemotherapy protocols have substantially improved the survival of children with acute leukemia, not all children are cured and many suffer treatment related toxicities. We have sought to improve the outcome of children with leukemia through pharmacogenomic and pharmacoepidemiology studies that are aimed at better defining patient populations at risk for adverse outcome and improving the utilization of supportive care resources. Specifically, we have demonstrated that group sequential methods (GSM) can be used to conserve valuable biosamples in pharmacogenomic studies in the case of a null result. We have also used the Children's Oncology Group (COG) clinical trials as a platform to demonstrate that CYP3A genotypes do not modify relapse risk but may modify the risk of vincristine associated neurological toxicities. Finally, we have merged data from COG with the Pediatric Health Information Systems (PHIS) administrative data set for patients enrolled on a Phase III trial of treatment for de novo acute myeloid leukemia. This merged data may be used not only for analysis of resource utilization, but may also be useful for monitoring Phase III clinical trials prospectively for treatment related toxicities. Ongoing studies are examining the concordance between toxicities reported in COG and PHIS data, the establishment of a chart-review based “gold standard” for determination of the sensitivity, specificity, positive predictive value, and negative predictive value of toxicities defined by COG and PHIS data, and the analysis of resource utilization for patient enrolled on COG AML studies.