EPID-18. THE GENOMIC ARCHITECTURE OF GLIOBLASTOMA PREDISPOSITION IS HIGHLY POLYGENIC IN PATIENTS WITH A POSITIVE FAMILY HISTORY OF GLIOMA

Abstract

Abstract BACKGROUND Genetic predisposition to glioblastoma (GBM) is highly polygenic and typically reflects the additive contributions of many common polymorphisms. However, patients with a positive family history of glioma may have a monogenic cancer predisposition syndrome. Such patients would be expected to have reduced contributions from common risk alleles. METHODS We sought to compare the polygenic risk of GBM among patients with a confirmed family history of glioma to those with a negative family history. Study subjects included 1628 GBM patients from the GICC study, who completed detailed questionnaires regarding personal and family medical history and underwent genome-wide genotyping of blood-derived DNA on the Illumina OncoArray. Polygenic risk scores were generated based on the 17 independent hits discovered by genome-wide association study (GWAS), with variants weighted by their respective effect sizes. RESULTS As expected, polygenic risk scores were significantly higher in GBM patients than in cancer-free control subjects (P=6.6x10-57). A total of 78 GBM patients had a positive family history of glioma, validated by record review. Across strata of family history, we observed that GBM patients with a positive family history of glioma had higher polygenic scores than patients with a negative family history (P=0.039). These results were robust in sensitivity analyses, where each of the 17 contributing variants were excluded from the model, one at a time. CONCLUSIONS Polygenic risk scores for GBM are significantly higher in patients with a positive family history of glioma. This suggests that, while family history can be due to rare, penetrant alleles, a large contingent of familial GBM may arise in families with a heavy polygenic burden of GBM risk alleles. Strategies for rare variant discovery might consider a two-pronged approach of sequencing kindreds with multiple affected members, as well as solitary GBM diagnoses in patients that have low polygenic risk scores.