BIOM-44. GENOMIC PREDICTORS OF ADVERSE EVENTS IN NEWLY DIAGNOSED IDH-WILDTYPE GLIOBLASTOMA

Abstract BACKGROUND Adverse outcomes including lymphopenia, thromboembolism (VTE), pseudoprogression and seizures cause significant morbidity in glioblastoma (GBM) patients. The association between genomic biomarkers and adverse outcomes in IDH-wildtype (IDH-wt) GBM needs to be further validated. METHODS We identified 1,011 consecutive adult patients with histologically confirmed GBM with OncoPanel testing (capture-based next-generation sequencing of 447 cancer-associated genes) at Dana-Farber Cancer Institute from 2013–2019. IDH-mutant patients and those without consistent follow-up at DFCI were excluded. Seizure at presentation, lymphopenia (absolute lymphocyte count < 1.0x109/L) at the beginning and end of chemoradiation, VTE, radiographic pseudoprogression (< 6 months of RT end) and early progression (< 6 months since completing RT) were identified retrospectively as adverse events. Single nucleotide variants and indels of relevant genes (Tier 1 or 2 mutations) and copy number analysis were derived. Variables were compared using Wilcoxon rank-sum test or Fisher’s Exact test with p< 0.05 indicating statistical significance. RESULTS Among 557 patients included, 198/508 (39%) had MGMT-methylated GBMs. Seizure at presentation occurred in 176/542 (32%) and was more common in TP53- or RB1-mutant GBM. Pre-RT 83/338 (25%), but not post-RT lymphopenia (249/367, 68%), was associated with PIK3CA mutation. VTE was detected in 74/509 (15%) during follow-up and was associated with mutations in PIK3CA and POLE. Pseudoprogression was noted in 43/457 (9%) of patients and was not associated with MGMT status or genomic alterations. Early progressive disease occurred in 33% of patients and was associated with unmethylated MGMT, PIK3CA, STAG2 and PTPN11 mutations. Median overall survival for the entire cohort was 17.5 months. CONCLUSION PIK3CA mutations were associated with baseline lymphopenia, risk of VTE and early progression in this cohort. Genomic biomarkers may help identify patients at higher risk of select adverse events in IDH-wt GBM. This could benefit patients through tailoring of supportive care and prophylactic therapies.