ID #658 Multi-omic analysis of archived formalin-fixed paraffin embedded brain tumours

Abstract The study of rare paediatric brain tumours is significantly constrained by limited sample availability, hindering large-scale molecular characterisation and advances in diagnosis and treatment. Long-term biobanking offers a powerful solution by enabling access to archival tumour material collected over decades, thereby expanding cohort size and diversity. Formalin-fixed paraffin-embedded (FFPE) tissue represents the most abundant and stable source of paediatric brain tumour specimens and remains a critical resource for generating genomic and epigenomic data. Here, we leverage a unique FFPE brain tumour biobank comprising specimens collected over the past 80 years at the Women’s and Children’s Hospital (South Australia). This collection includes a broad spectrum of paediatric brain tumours, spanning low-grade to high-grade malignancies and multiple rare tumour entities. However, the age and variable preservation quality of archival FFPE samples present significant technical challenges for modern molecular profiling. This study focuses on optimising multi-omics workflows for archival FFPE tissue, including genomic DNA extraction, DNA library preparation, and downstream analysis. We demonstrate that sample-specific optimisation of extraction and library preparation protocols is essential to achieve sufficient DNA yield and quality, particularly for older FFPE specimens. Pooling DNA extracted from multiple tumour regions within individual cases was critical for meeting minimal input requirements in older samples. Using DNA methylation profiling, we show that archival tumours can be reclassified with greater molecular precision compared to historical histopathological diagnoses. In several cases, methylation-based classification refined tumour subtypes or identified previously unclassifiable tumours, highlighting discrepancies between morphology-based and molecular diagnoses. Our findings demonstrate that advanced multi-omics approaches can be successfully applied to decades-old FFPE paediatric brain tumour samples, unlocking valuable biological and diagnostic insights. This work underscores the importance of archival biobanks in paediatric neuro-oncology and further supports the integration of molecular profiling into diagnostic frameworks to improve tumour classification, risk stratification, and ultimately patient care.