AI method classifies brain tumours using spinal fluid DNA

An international research team involving the Medical University of Vienna has developed a new AI-based approach that can accurately classify brain tumours using tiny fragments of tumour DNA found in cerebrospinal fluid.

The method, called M-PACT, analyses cell-free DNA methylation patterns released by tumour cells into the fluid surrounding the brain and spinal cord. These molecular signatures allow different types of paediatric brain tumours to be identified with high precision, even when only minimal genetic material is available.

Until now, brain tumour diagnosis has largely depended on surgically obtained tissue samples, which are not always feasible or may carry significant risk. The new approach offers a less invasive alternative, potentially enabling earlier diagnosis and more frequent monitoring without the need for surgery.

In addition to tumour classification, the AI system can track molecular changes over time, offering a way to monitor disease progression, treatment response and relapse through simple fluid sampling.

The study [1], published in Nature Cancer, was conducted in collaboration with international partners including St. Jude Children’s Research Hospital and the Hopp Children’s Cancer Centre in Heidelberg. Researchers say further clinical studies will be needed before the method can be introduced into routine practice, but it represents a significant step towards non-invasive brain tumour diagnostics.

More information online

1. M-PACT leverages cell-free DNA methylomes to achieve robust classification of paediatric brain tumours published in Nature Cancer