Abstract General Information

Title

IMPLEMENTATION OF DNA METHYLATION PROFILING FOR BRAIN TUMOR DIAGNOSIS: THE FIRST BRAZILIAN EXPERIENCE

Introduction, Objectives, Methods, Results, and Conclusion.

Introduction: DNA methylation profiling has revolutionized diagnostic neuropathology since its implementation in Heidelberg. It associates a high throughput methylation array data with a machine learning based classifier in order to predict the accurate diagnosis of several brain tumor entities. This approach allows not only appropriate classification of difficult-to-diagnose cases, but also to subclassification within tumor classes, such as Medulloblastomas and Ependymomas. Despite its use became widespread in several countries, methylation profiling for tumor classification has never been performed in Brazil.

Objective: To implement and validate a DNA methylation array and the Heidelberg Brain Tumor classifier to aid in the diagnosis of brain tumors in Brazil.

Methods: Eight cases with known or highly suspected diagnosis and a normal brain cortex tissue samples were selected for this validation. DNA samples from FFPE specimens were evaluated using the BeadChip HumanMethylation450 (8 samples) or Infinium MethylationEPIC 850 (1 sample) arrays. The raw data for each sample were uploaded in the Heidelberg Neuropathology Website and evaluated in the classifier versions 11b4 and 12.3. The results were interpreted together with all other clinical, histopathological, immunohistochemical and molecular data in order to render an integrated diagnosis.

Results: Two samples were not suitable for classification due to bad fixation and poor DNA quality. Six cases received a high calibrated score for known methylation classes, matching with the original or expected diagnosis. One specimen had a low calibrated score for methylation class Glioblastoma, RTK2 subtype, but this information was considered for the final diagnosis of Glioblastoma, IDH-wildtype. The integrated diagnosis for the other cases were Astroblastoma, MN1 altered; Diffuse hemispheric glioma, H3 G34-mutant; Anaplastic Pleomorphic Xanthoastrocytoma; Diffuse high grade glioma, H3 K27-altered, Medulloblastoma, SHH-activated, subclass 4 and Control tissue, hemispheric cortex.

Conclusion: DNA methylation profiling was able to confirm and refine the diagnosis of the majority of this first set of cases, with a relatively simple workflow and interpretation. Poor fixation may interfere in test accuracy, so it is crucial to maintain good practices in specimen handling. Applying the array to a higher number of cases will allow us to get more experience with this tool and benefit more patients with a precise diagnosis.

Keywords (separated by comma on a single line)

Neuropathology; Methylation; Brain Tumor; Classification