PATH-20. METHYLATION ARRAY PROFILING OF PEDIATRIC BRAIN TUMORS; SINGLE CENTRE EXPERIENCE

BACKGROUND: Significant heterogeneity of pediatric brain tumors poses major challenge on diagnostics. Therefore, we aimed to evaluate feasibility of methylation array in the diagnostic process. METHODS: Methylation array (Infinium MethylationEPIC, Illumina) was performed on DNA extracted from fresh frozen tissue from prospective newly diagnosed and selected retrospective patients. Results from Heidelberg classifier (www.molecularneuropathology.org) were compared to the histological diagnosis and further genetic testing was performed to establish integrated morphological/molecular diagnosis. RESULTS: Within years 2018–2019, we performed methylation array profiling of 102 samples consisting mainly of ependymoma, medulloblastoma high-grade and low-grade glioma. High calibrated score (>0.9) was achieved in 62 patients (61%). In 46 cases (74%) with score >0.9, the histological diagnosis matched the methylation class (MC). In the remaining cases (16) that were classified by histopathology mainly as ependymomas, the methylation profiles were classified as novel molecular entities (HGNET_BCOR, HGNET_MN1, etc.) or different tumor type. In 40 cases (39%) with the score <0.9, six were found to have high normal tissue content. Nine cases had no match in the classifier and 25 were assigned MC with score 0.3 to 0.89. In 20 out of 34 cases with low score, the molecular diagnosis could be confirmed based on copy number variants inferred from the methylation array or using additional testing for gene fusions and mutations. CONCLUSIONS: Our experience on the first 100+ cases demonstrated that methylation array could be integral part of diagnostic process in order to establish integrated morphological and molecular diagnosis of pediatric brain tumors.