METB-08. GENOMIC AND MOLECULAR PROFILING OF PEDIATRIC CENTRAL NERVOUS SYSTEM TUMORS: A SINGLE INSTITUTION ANALYSIS

BACKGROUND: Molecular characterization has revolutionized management of pediatric central nervous system (CNS) tumors, with genomic sequencing for diagnosis and determination of actionable variants now largely considered standard of care. We report a single institution’s experience performing genomic sequencing for pediatric CNS tumors over nearly five years. METHODS: Nucleic acid isolated from tumor tissue and paired normal specimens from patients at Children’s Hospital of Philadelphia (CHOP) undergoes paired next generation sequencing (NGS) and copy number analyses of 237 cancer genes and a fusion panel of more than 700 exons of 117 cancer genes (tumor only). All variants are tiered 1-4 according to known clinical significance. Descriptive statistics summarize clinical and molecular data and ensemble-based methods will be used to correlate molecular and clinical characteristics. RESULTS: Between January 1, 2016 and June 30, 2020, 497 patients ≤ 21 years old with median age of 9.5 years (n=243 females; 49%) contributed to 539 cases across 25 CNS tumor subtypes. The most common diagnoses were pilocytic astrocytoma (n=118), medulloblastoma (n=49), and diffuse midline glioma (DMG; n=39). Clinically significant variants (tier 1,2) were detected in 91% of the patients. Most tumors demonstrated tier 1,2 single nucleotide variants (SNV)/insertions/deletions (n=263, 53%), most commonly in TP53>BRAF>H3F3A, and/or CNVs (n=310, 62%), most commonly gains of 7q>7p>1q. Thirty percent of tumors demonstrated tier 1,2 fusions (n=149), most commonly involving BRAF, NTRK2, or FGFR. Average tumor mutation burden in patients without germline predispositions was 3.5/Mbp (standard deviation 3.8). Germline/presumed germline pathogenic variants were found in 18% (90/497) of patients tested with most common germline alterations in TP53 followed by NF1 and APC. CONCLUSION: A large, single site cohort of molecularly characterized CNS tumors demonstrate the clinical and biologic utility of real-time, institution-based genomic sequencing. Molecular findings in correlation with clinical characteristics and outcomes are underway.