BIOM-47. PREDICTORS OF SEIZURE AT ONSET USING A FUNCTIONAL VARIANT ANALYSIS OF TARGETED NEXT GENERATION SEQUENCING IN GBM

Abstract

Abstract BACKGROUND Adverse events (AE) including seizures cause significant morbidity in patients with GBM. We propose a novel method for assessing genomic predictors of AEs using results from a clinical targeted sequencing platform with variant function analysis. METHODS We identified 1,011 consecutive adult patients with newly diagnosed, histologically confirmed IDH-wildtype GBM with targeted exome NGS (Oncopanel) at Dana-Farber Cancer Institute from 2013-2019. Seizure at presentation was retrospectively identified as an AE. Biologic function (high loss, low loss, neutral, low gain and high gain) was assigned to variants using a three-tiered approach leveraging a genetic variant database (OncoKB), followed by analysis using protein prediction tools (Sift, Polyphen2 and Provean). Univariate logistic regression was performed for each relevant altered gene against the outcome of interest with false-discovery rate correction. Genes associated with seizure at presentation were included iteratively in a multivariate logistic model including other predictors of the outcome. RESULTS Our analysis included 470 GBM patients with 107 genes and 12 whole chromosome or arm level candidate variants covered by all versions of Oncopanel and with >10% alteration. Seizure at presentation occurred in 143/463 patients (31%) and was associated with EGFR amplification (high gain) (OR: 2.76, 95% CI: 1.4-5.3, p = 0.04). In a multivariate analysis (including age, sex, and preoperative tumor volume), EGFR amplification remained statistically significant (OR: 1.5, 95% CI: 1.0-2.2, p = 0.03). CONCLUSION Genomic biomarkers based on functional variant analysis of a routine clinical panel may predict adverse events in GBM. Seizure at presentation was independently associated with EGFR amplification. Our ongoing analysis will look at predictors of myelosuppression, thromboembolism, pseudoprogression and early progression using a similar approach. Identifying molecular risk factors could improve the management of patients through supportive care and consideration of prophylactic therapies.