Abstract 2486: African ancestry is associated with high-risk disease and event-free survival in children with neuroblastoma

Abstract

Abstract Background In a cohort of 3539 children with neuroblastoma, we have previously reported that self-reported black race is significantly associated with high-risk disease (p < 0.001) and late relapse or death (p = 0.04). To investigate if the component of genomic variation that co-segregates with African ancestry is associated with risk group and survival in neuroblastoma, we interrogated genome-wide single nucleotide polymorphism (SNP) genotypes against those phenotypes in 2790 neuroblastoma patients. Methods Genome-wide SNP genotypes from 3404 children with neuroblastoma were obtained from 3 Illumina platforms: HumanHap550 v1, HumanHap550 v3 and HumanQuad 610. Quality filtering removed 17 patients who were genotyped twice, 15 patients with ambiguous sex, 72 patients with some degree of relatedness, 120 patients with poor genotype call rates (defined as >5% total SNPs interrogated without a genotype call) and 390 patients with missing clinical data. Genotypes from the remaining 2790 patients were included in the analysis. Using SNPs common to each of the genotyping platforms, ancestral reference populations from the International HapMap Project (Caucasian - CEU, African - YRI and Asian - CHB/JPT) and the EIGENSTRAT method, a principal components analysis was performed and an ancestry map of all patients was created. Principal Component 1 (PC1), which separated black neuroblastoma patients and the YRI HapMap samples from all other ethnic groups, was tested as a continuous variable in a Cox Proportional Hazard Regression model of event-free survival (EFS) and in an ordinal logistic regression of risk group. Using a k-means clustering algorithm, patients were grouped into 1 of 5 ethnic clusters;EFS of clusters was compared using a log-rank test, and Kaplan-Meier curves were plotted. Results PC1 was correlated with neuroblastoma risk group classification (p=0.0063). PC1 was also correlated with EFS (p=0.044). Compared to Caucasian patients, patients with African ancestry had the poorest EFS (log rank p = 0.0257). Discussion Here we show that the component of genomic variation that co-segregates with African ancestry is associated with high-risk disease and event-free survival. Patients with African ancestry had inferior outcomes to patients of Caucasian ancestry. We hypothesize that germline genetic variants associated with both the development of high-risk neuroblastoma and event-free survival after chemotherapy will have different allelic frequencies in Caucasian and African populations and that patients of African ancestry may also harbor unique genetic variants associated with high-risk disease and event-free survival not found in Caucasian patients. Efforts to identify these genetic variants are underway. These results underscore the need to perform association studies from a variety of ancestries, as each group may yield novel associations. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 2486. doi:1538-7445.AM2012-2486