EMBR-12. IMPROVED DIAGNOSTIC ALGORITHM FOR DIFFERENTIAL DIAGNOSTICS OF CNS EMBRYONAL TUMORS (FORMER CNS-PNET) BY NEUROPATHOLOGICAL RE-EVALUATION OF 256 CASES AND CROSSVALIDATION BY METHYLATION CLASSIFICATION

• Published in: Neuro-oncology (Charlottesville, Va.) Va.), 2018-06, Vol.20 (suppl_2), p.i71-i71
• Main Authors: Pietsch, Torsten, Figarella-Branger, Dominique, Giangaspero, Felice, Gessi, Marco, Hawkins, Cynthia, Jacques, Thomas, Korshunov, Andrey, Eberhart, Charles, Burger, Peter, von Hoff, Katja, Kool, Marcel, Haberler, Christine
• Format: Article
• Language: English
• Subjects: Abstracts
• ISSN: 1522-8517; 1523-5866
• DOI: 10.1093/neuonc/noy059.196

Abstract Epigenetic profiling has shown that a proportion of cases diagnosed as CNS PNET in the past can be assigned to other tumour entities with similar morphological appearance. In an international effort to re-analyze CNS-PNET aiming for disease-specific re-evaluation of survival data and the development of diagnostic guidelines providing the basis for improved therapeutic approaches, 256 tumours diagnosed and treated as CNS-PNET in the last two decades in 17 countries were reviewed by a panel of neuropathologists according to today’s standards of clinical neuropathological diagnostics including immunohistochemical and molecular pathological assays. The majority of cases were also independently analyzed by methylation array hybridization and classified by random forest algorithm. In this unique cohort, we identified 20 different tumor entities including frequent high grade gliomas. 41% of cases were confirmed as CNS-PNET (now termed CNS-embryonal tumors (CNS-ET) according to the revised WHO-classification) representing two main entities: ETMR displayed typical histopathological features, LIN28A expression and/or C19MC alteration. The other represented a group of tumors with variable degrees of differentiation along neuroblastic/ganglionic lines, corresponding to the WHO diagnoses CNS-(Ganglio)-neuroblastoma or CNS-ET, NOS. The vast majority of these tumors could be assigned to the FOXR2 CNS-NB group by methylation array-based classification. Crossvalidation of neuropathological and epigenetic classification proved that methylation classification represents a useful complimentary tool in the differential diagnosis of CNS-ET. Re-evaluation of prototypic tumors and cases with discrepant diagnoses enabled us to develop an optimized diagnostic algorithm to securely delineate this tumor type from other entities with largely divergent clinical and biological behaviour.