P10.22.A UNRAVELLING MOLECULAR HETEROGENEITY IN DIFFUSE HIGH-GRADE GLIOMAS OF PEDIATRIC, ADOLESCENT, AND YOUNG ADULTS THROUGH DNA METHYLATION PROFILING

Abstract BACKGROUND High-grade gliomas (HGGs) are among the most aggressive CNS tumors in pediatric, adolescent, and young adult (AYA) populations, marked by rapid progression and poor survival. These tumors show considerable histopathological and molecular heterogeneity, complicating diagnosis and treatment. Although WHO CNS5 distinguishes adult and pediatric HGGs, the pediatric and AYA subgroup remains a heterogeneous group with overlapping features and considerable molecular variability. MATERIAL AND METHODS A total of 61 pediatric and AYA cases of diffuse HGG, CNS WHO Grade-4, NOS (median age: 19 years; M: F = 1.6: 1), were included in the study. Whole-genome DNA methylation profiling was performed for all cases along with eight controls and classified on the online available Epignostix brain tumor classifier (v12.8), and Methylscape. Additionally, next-generation sequencing (NGS) and RNA fusion analysis were conducted where appropriate, based on insights from methylation-based classification. RESULTS Tumors initially labeled as diffuse HGG, CNS WHO Grade 4, NOS were reclassified into distinct methylation classes. Pediatric HGGs, H3-WT/IDH-WT accounted for 20% (12/61), with 42% (5/12) showing H3K27me3 loss, deviating from WHO criteria. RTK1 subtype was enriched in MMR-deficient and radiation-induced gliomas. Glioblastoma, IDH-wildtype comprised 18% (11/61), including 45% (5/11) pediatric cases. Adult-type glioblastoma, subtype B (5%, 3/61) showed MAPK pathway alterations of therapeutic relevance. Diffuse hemispheric glioma, H3K27M-altered was identified in 6.5% (4/61) of cases, including one located in the temporal lobe—challenging conventional anatomy-based classification. Pleomorphic xanthoastrocytoma (PXA) was seen in 8.2% (5/61), all of which had poor survival outcomes. Other reclassified entities included DHG, H3G34-mutant (n=2), pilocytic astrocytoma (n=3), HGAP (n=1), MPNST (n=2), CNS-BCOR ITD (n=1), neuroepithelial tumor with PATZ1-fusion (n=1), and DLGNT (n=1). Methylation profiling refined diagnoses in 48%, changed them in 16%, confirmed in 15%, while 12% were misdiagnosed and 10% remained unclassified. CONCLUSION Methylation profiling revealed substantial molecular diversity in pediatric and AYA diffuse HGGs, enabling refined classification and identification of rare, clinically relevant subtypes. Deviations from WHO criteria highlight the need to update diagnostic frameworks to better capture subgroup variability. Expanding classifier datasets to include diagnostically challenging cases may enhance prognostic accuracy and clinical relevance. This comprehensive approach advances molecular understanding, diagnostic precision, and better informs targeted therapeutic strategies.