OMIC-12. PREVALENCE AND SPECTRUM OF GERMLINE PATHOGENIC VARIANTS IN CANCER PREDISPOSITION GENES ACROSS THE CHILDREN'S BRAIN TUMOR NETWORK (CBTN). (1st June 2021)

• Record Type: Journal Article
• Authors: Vaksman, Zalman; McQuaid, Shelly; Bornhorst, Miriam; Zhu, Yuankun; Heath, Allison; Waanders, Angela; Cole, Kristina; MacFarland, Suzanne; Diskin, Sharon
• Is Part Of: Neuro-oncology. Volume 23(2021)Supplement 1
• Page Start: i39
• Page End: i40
• Publication Date: 2021-06-01
• Subjects: Brain Neoplasms -- Periodicals; Brain -- Tumors -- Periodicals; Brain -- Cancer -- Periodicals; Nervous system -- Cancer -- Periodicals; 616.99481
• DOI: 10.1093/neuonc/noab090.159
• Languages: English
• ISSNs: 1522-8517

Abstract

Abstract: Germline variants are known to contribute to the pathogenesis of specific central nervous system (CNS) tumor subtypes; however, a large pan-pediatric brain and nervous system cancer germline susceptibility study has not been performed. To define the prevalence and spectrum of pathogenic variants in known cancer predisposition genes (CPGs; n=200), we analyzed whole genome sequencing (WGS) data from 880 pediatric subjects across 19 different cancer types in the Children's Brain Tumor Network (CTBN). Data were aligned using BWA. Variants were called using GATK and annotated with SnpEff and ANNOVAR. After quality control, variants with a minor allele frequency (MAF) < 0.1% in Gnomad 2.11 or ExAC were retained. Pathogenicity was assessed with American College of Medical Genetics (ACMG) guidelines using a lab-developed modification of ClinVar and InterVar. Automated pathogenic/likely pathogenic (P-LP) calls were manually reviewed by two cancer predisposition clinicians and a bioinformatician. Frequency of P-LP variants was assessed and gene burden testing was performed against Gnomad3.1 (without cancer samples) using Fisher's exact test with Bonferroni adjustment. We observed 214 P-LP variants involving 190 unique individuals (21.6% of cohort). As expected, the most frequent variants were observed in NF1, NF2, and TP53 (n=40 variants in 21% of individuals). ATM, TSC2 and CHEK2 variants (n=23) were observed in another 12% of individuals. An increased burden of P-LP variants was observed for 5 of these 6 genes (p = 1.7x10 -25 to 1.4x10 -2, CHEK2 p=5.5x10 -2 ). We also identified 5 variants in BRCA2 (3 in high-grade glioma), 7 in REQC helicases ( BLM, WRN, REQL4 ), and 16 variants in Fanconi anemia genes. Overall, cases harbored increased burden in P-LP variants in CPG genes (p=8.8x -18 ) and the subset of DNA repair genes (p=4.7x10 -4 ). In conclusion we confirmed the association of variants in established predisposition genes while potentially identifying novel variants and genes associated in CNS tumors.