P20.12.b dna methylation profiling of adult spinal glial tumours; results from a large single centre study

Abstract BACKGROUND Glial tumours can occur in any central nervous system location, including the spinal cord. Although rare, spinal gliomas can represent a significant diagnostic challenge and are recognised as separate tumour types in the WHO classification. MATERIAL AND METHODS This study used DNA methylation profiling to characterise adult spinal glial tumours from a large neurosurgical centre, exploring relative frequencies, and the value of this technique in diagnostic workflows. A search in our laboratory information management system (2006-2021) identified 116 tumours, from patients aged 17.2-79.2 years. Diagnoses included ependymomas (all grades/subtypes) and other high- or low-grade gliomas (HGG/LGG). Extrinsic tumours (metastases, meningiomas, schwannomas) were excluded. DNA underwent methylation profiling (Illumina EPIC BeadArrays, n=116). iDat files were processed using the brain tumour classifier (MNP v12.5 R package), classifying cases against a reference cohort (n=2,800). Longitudinal (n=8) and poor QC cases (n=3) were excluded. RESULTS 83% (n=87) classified as an ependymoma (myxopapillary ependymoma (EPN_MPE, n=44), spinal ependymoma (n=43), spinal subependymoma (n=1)). We identified HGGs that were not diagnosed histologically (diffuse midline glioma, H3 K27-altered (n=2), high-grade astrocytoma with piloid features (n=1)) and 41% of cases (n=18/44) diagnosed histologically as an ependymoma, CNS WHO grade 2, classified as EPN_MPE. LGGs were less frequent (pilocytic astrocytoma (n=3), pleomorphic xanthoastrocytoma (n=1), rosette-forming glioneuronal tumour (n=1)). 9% (n=9) did not classify with existing methylation classes. Copy number analysis revealed subtype-specific changes, including prognostic alterations. CONCLUSION In conclusion, this single-centre study highlights the crucial role of DNA methylation profiling as part of diagnostic workflows for spinal glial tumours.