There’s a particular kind of pressure in a neurosurgical theatre. A surgeon has just removed a small piece of brain tissue — a biopsy, perhaps only a millimetre or two across — and what happens next will shape the entire course of treatment. Historically, that “next” has meant waiting. Sometimes for days. Sometimes for weeks.
A study published in Nature Medicine in early 2025 suggests that wait might soon be a thing of the past.
The problem with brain tumour diagnosis
Central nervous system (CNS) tumours are notoriously difficult to classify. There are hundreds of subtypes, and getting the diagnosis right matters enormously — different tumour types respond to completely different treatments. The gold standard has long been a combination of histology (looking at the tissue under a microscope) and molecular profiling (reading the DNA). Together, these can take up to three weeks to return results. For a patient with a brain tumour, three weeks is a long time to wait before anyone can meaningfully plan their care.
Enter nanopore sequencing
Nanopore sequencing is a method of reading DNA that works by threading individual strands through tiny protein pores and measuring changes in electrical current as each base passes through. It’s faster and more portable than traditional sequencing methods, and crucially, it can work in real time — generating results as the sequencing happens, rather than after it’s complete.
Researchers at University Hospital Heidelberg, the Hopp Children’s Tumour Center, and the German Cancer Research Center developed a workflow called Rapid-CNS2, which uses this technology to profile brain tumour DNA during surgery. The headline figure: a preliminary classification can be returned within 30 minutes of sequencing starting. A comprehensive molecular profile — covering genetic mutations, methylation patterns, and chromosomal changes — follows within 24 hours.
That’s a reduction from roughly 20 days to under a day.
What it can actually detect
The system identified over 91% of single nucleotide variants (individual letter changes in the DNA sequence) confirmed by more established methods, and correctly classified nearly 93% of tumours into their molecular family. It also demonstrated the ability to catch subtle mutations that conventional short-read sequencing can miss — including a subclonal deletion in the EGFR gene in a glioblastoma sample, which is the kind of finding that can directly influence treatment choice.
Complementing Rapid-CNS2 is a tool called MNP-Flex, a methylation classifier that achieved 99.6% accuracy across more than 78,000 samples worldwide — and works across different sequencing platforms, meaning it doesn’t require any one specific machine.
Why this matters beyond speed
Speed matters, obviously. But the more interesting implication here is what faster diagnostics could mean for equity in cancer care. These tools were specifically designed to be accessible to users without deep computational expertise, and the sequencing hardware they run on — including a small portable device called the MinION — doesn’t require a fully equipped genomics laboratory. In settings where advanced diagnostics are currently out of reach, that portability is significant.
It’s early. Broader clinical validation is still needed before this becomes standard practice. But the proof-of-concept is compelling: real-time molecular diagnostics during brain surgery is no longer hypothetical.
Sources
- Patel, A. et al. (2025). Rapid-CNS2: comprehensive real-time CNS tumour profiling during surgery. Nature Medicine. nature.com
- Oxford Nanopore Technologies — MinION device overview: nanoporetech.com
- WHO Classification of Tumours of the Central Nervous System, 5th Edition (2021)
- German Cancer Research Center (DKFZ): dkfz.de
- Hopp Children’s Tumour Center (KiTZ): kitz-heidelberg.de
