Abstract Background: Racial and ethnic disparities in incidence and outcomes for childhood B-cell acute lymphoblastic leukemia (B-ALL) are well established. While genome-wide association studies (GWASs) among children of European (EUR) genetic ancestry have identified robust risk loci, comparable studies in children of African (AFR) genetic ancestry do not exist. Here we report on results from the first GWAS of B-ALL exclusive to AFR children. Methods: A total of 840 AFR cases (diagnosis age <21y) and 3360 sex-/ancestry-matched controls (from ADMIRAL and ZOE 2.0, a childhood caries study) were evaluated. We used 656 cases from the Children’s Oncology Group (COG) ALL frontline trials for discovery and 184 non-COG cases for replication. Samples were genotyped with the Illumina Global Diversity Array and imputed with the TOPMed reference panel. Ancestry was ascertained with RFMix-inferred global ancestry values and principal components analysis. Logistic regression models adjusted for cryptic population substructure evaluated variant risk associations. Discovery and replication results were meta-analyzed using an inverse variance-weighted fixed effects model. Results: We identified 7 genome-wide significant (GWS, P<5 × 10−8) variants, including 2 novel B-ALL risk loci with replication evidence (P<0.05) and striking per allele effect sizes: CNTN4 (meta-analysis odds ratio [OR]=2.09, P=1.4 × 10−11) and FAM174A (OR=2.94, P=1.2 × 10−15). Established B-ALL EUR risk loci were corroborated in these data (ARIDB5, IKZF1; P<5 × 10−8). Five additional novel loci were GWS after meta-analysis including at SLC37A4 (OR=2.02, P=1.1 × 10−8) and CBWD1 (OR=2.69, P=2.7 × 10−9), with suggestive associations in discovery data (P<5 × 10−6) and nominal replication (P<0.05). Notably, all novel risk alleles were low-frequency in reference AFR populations and AFR-specific (minor allele frequencies of 2-9% in 1000 Genomes AFR and <0.1% in non-AFR). Top variants had relevant in silico annotations; variants at CNTN4 and SLC37A4 had high predicted functional probability (>0.8, RegulomeDb). For example, annotations for a top CNTN4 signal suggest a potential leukemogenesis mechanism, overlapping a ChIP-seq peak for CEBPA in transdifferentiated B cells. In our replication data, a polygenic risk score (PRS) with 5 novel AFR-specific GWS variants from discovery had an appreciably stronger B-ALL risk association (top 25% vs rest: OR=2.03, P=1.2 × 10−4) compared to a published B-ALL PRS from a GWAS in EUR children (top 25% vs rest: OR=1.59, P=0.014). Conclusions: Despite the modest number of AFR B-ALL cases, we successfully detected and replicated genotype risk associations with considerable effect sizes. These results demonstrate the critical importance of increasing diversity in genomic studies of childhood ALL to promote discovery and reduce disparities. Citation Format: Cindy Im, Andrew Raduski, Lauren J. Mills, Rebecca A. Johnson, Andrew DeWan, Xiaomei Ma, Joseph L. Wiemels, Catherine Metayer, Jun J. Yang, Heather H. Nelson, Tianzhong Yang, Saonli Basu, Lucie M. Turcotte, Nathan Pankratz, Michael E. Scheurer, Logan G. Spector. Novel genetic risk loci for B-cell acute lymphoblastic leukemia in African American children: Findings from the Admixture and Risk of Acute Leukemia (ADMIRAL) Study [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 2845.
Read full abstract