lmproving diagnosis of central nervous system tumours using genetic and epigenetic tools

Abstract: Brain tumor diagnostics has traditionally been based on histopathology stains. The introduction of immunohistochemistry stains resulted in improved ability to classify these often devastating tumors. The understanding of molecular markers in central nervous system tumors has improved substantially in recent decades. Demonstration of specific genetic changes such as mutations can have crucial impact on selection of therapy. Genetic changes also play an increasingly important role when it comes to classification of these tumors. Correct classification and grading are important to be able to give correct prognostic and predictive information and are of fundamental importance for efficient clinical patient handling. In this thesis, we use several molecular techniques to improve tumor diagnosis and tumor classification and investigate the utility of these methods to give a deeper understanding of these neoplastic processes. In Paper I, we investigated the DNA methylation profiling method for molecular classification of diffuse lower grade gliomas. We showed that DNA methylation profiling not only gave correct diagnostic and prognostic information but also were able to give reliable molecular information enabling molecular classification of the tumors according to the World Health Organization classification system. In Paper II, we assessed changes in DNA methylation pattern over time in diffuse IDH-mutant gliomas. We showed that tumors accumulated methylation alterations during progression, but that the overall methylation patterns most often were maintained upon recurrence. In Paper III, we explored if a proposed immunohistochemistry-based investigation of phenotype predicted survival and tumor recurrence in a clinical cohort of diffuse low-grade gliomas that were reclassified according to the 2021 WHO criteria. In Paper IV, we describe a 16-month-old patient with a tumor in the third ventricle with a relapse two years after diagnosis. The tumor was initially classified as a low-grade glioma but was after methylation profiling reclassified as an infant-type hemispheric glioma. To search for druggable targets and for further refinement of the molecular background both whole genome sequencing and whole transcriptome sequencing were performed. A novel TPR::ROS1 fusion gene was detected activating the MAPK-, PI3K- and JAK/STAT- pathways. In Paper V, we present a cystic pilocytic astrocytoma with KIAA1549::BRAF fusion in a 16-year-old patient. The tumor showed ganglion cell morphology and different vascularization in a nodular component. With extended molecular examination we were able to prove that the cells with ganglion cell morphology were of neoplastic origin. In conclusion, we further demonstrate the importance of adding molecular investigation in the histopathological diagnostic work-up. We also present arguments for the importance of evaluating molecular findings in correlation with the histomorphology picture.