Summary Accurate pathological diagnosis is crucial for optimal management of cancer patients. For the ~100 known central nervous system (CNS) tumour entities, standardization of the diagnostic process has been shown to be particularly challenging - with substantial inter-observer variability in the histopathological diagnosis of many tumour types. We herein present the development of a comprehensive approach for DNA methylation-based CNS tumour classification across all entities and age groups, and demonstrate its application in a routine diagnostic setting. We show that availability of this method may have substantial impact on diagnostic precision compared with standard methods, resulting in a change of diagnosis in up to 12% of prospective cases. For broader accessibility we have designed a free online classifier tool (www.molecularneuropathology.org) requiring no additional onsite data processing. Our results provide a blueprint for the generation of machine learning-based tumour classifiers across other cancer entities, with the potential to fundamentally transform tumour pathology.

The median survival of glioblastoma patients is approximately 12 months. However, 3-5% of the patients survives for more than 3 years and are referred to as long-term survivors. The clinical and molecular factors that contribute to long-term survival are still unknown. To identify specific parameters that might be associated with this phenomenon, we performed a detailed clinical and molecular analysis of 55 primary glioblastoma long-term survivors recruited at the six clinical centres of the German Glioma Network and one associated centre. An evaluation form was developed and used to document demographic, clinical and treatment-associated parameters. In addition, environmental risk factors, associated diseases and occupational risks were assessed. These patients were characterized by young age at diagnosis and a good initial Karnofsky performance score (KPS). None of the evaluated socioeconomic, environmental and occupational factors were associated with long-term survival. Molecular analyses revealed MGMT hypermethylation in 28 of 36 tumours (74%) investigated. TP53 mutations were found in 9 of 31 tumours (29%) and EGFR amplification in 10 of 38 tumours (26%). Only 2 of 32 tumours (6%) carried combined 1p and 19q deletions. Comparison of these data with results from an independent series of 141 consecutive unselected glioblastoma patients registered in the German Glioma Network revealed significantly more frequent MGMT hypermethylation in the long-term survivor group. Taken together, our findings underline the association of glioblastoma long-term survival with prognostically favourable clinical factors, in particular young age and good initial performance score, as well as MGMT promoter hypermethylation.