Nature: A new tool to classify tumors of the central nervous system

To find the optimal treatment options, it is important to know exactly which type of tumor a patient has, but this is difficult with current techniques. The study was coordinated by the University of Heidelberg, and is published in the journal Nature.

The method is based on the methylation of the tumor DNA – a characteristic that is defined by both the cell of origin and somatic changes that occur during tumorigenesis. DNA methylation profiling is a powerful and reproducible technique useful for subclassification of CNS tumors that were until now considered as homogenous diseases. This study uses Chip Arrays to generate genome-wide DNA methylation profiles. It aims to provide CNS tumor classification for all entities and age groups.

Researchers started by developing a reference cohort consisting of 2,801 samples. An unsupervised computer algorithm clustered these samples into 91 CNS tumor methylation classes. The utility of the method was then evaluated with 1,155 diagnostic samples. For about ¾ of the samples, the methylation class matched the result of histopathological analysis. For up to 12 % of cases, however, the methylation-profile approach made a big difference: it caused the sample to be assigned to a different class and thus changed the diagnosis.

The authors have designed a free-online tool, where clinicians can submit DNA-methylation profiles and request a classification of the samples to compliment the classical diagnostic toolbox.