Diagnostic impact of DNA methylation classification in adult and pediatric CNS tumors

Over the past decade, the neuropathological diagnosis drastically changed by integratingmorphological features with molecular biomarkers. The molecular era succeeded to refine theneuropathological diagnostic accuracy but yet, a significant number of CNS tumor diagnosesare still challenging, especially in children. DNA methylation classification has emerged as apowerful machine learning approach for clinical decision-making in CNS tumors. The aim ofthis study is to share our experience of using the DNA methylation classification in daily routinepractice with illustration of clinical cases. We used a class system to evaluate the discrepanciesbetween histo-molecular and DNA methylation diagnosis with a specific focus on adult vspaediatric CNS tumors. In our study, we observed 40% of class I, 47% of class II and 13% ofclass III among the match cases. In other words, DNA methylation testing confirmedmorphological diagnoses in 63% and 23% of adult and paediatric cases, respectively. Therefinement of diagnosis was particularly observed for the paediatric population (65% vs 21%for the adult population). We also discussed the cases classified with low calibrated scores. Inconclusion, our study shows that DNA methylation classification represents an added value forall CNS tumor types diagnoses, particularly in paediatric population.