Scientists have unveiled a solution to long-standing challenges in molecular data integration for central nervous system (CNS) tumor diagnostics. A Nature Medicine study introduces a nanopore sequencing workflow called Rapid-CNS2 that delivers real-time tumor classification and DNA insights within just 30 minutes during surgery and comprehensive molecular profiling within 24 hours. Complementing it is MNP-Flex, a platform-agnostic methylation classifier that achieved over 99% accuracy across over 78,000 samples worldwide. Together, these tools, developed by researchers at University Hospital Heidelberg, Hopp Children’s Tumor Center (KiTZ), and the German Cancer Research Center (DKFZ), offer advanced speed and accuracy, enabling rapid, actionable insights to guide personalized treatment strategies for CNS tumors.
Advancing brain cancer diagnostics
CNS tumors are notoriously complex, demanding precise and timely diagnostics to guide treatment. Traditional approaches often require weeks to yield actionable insights, delaying critical decisions. In this article, first author Areeba Patel, PhD, and colleagues built on previous research on Rapid-CNS2, which utilizes adaptive nanopore sequencing. Unlike traditional methods, nanopore sequencing uses adaptive sampling to target specific DNA regions in real time, streamlining workflows and reducing the need for specialized equipment.
The system was demonstrated as a proof-of-concept using two Oxford Nanopore Technology (ONT) sequencing devices: the MinION, a small and portable sequencer ideal for on-the-go experiments, and the GridION, a more robust device designed for higher-throughput sequencing tasks. Led by senior authors Martin Sill, PhD, and Felix Sahm, PhD, this innovative workflow was validated in studies conducted at the University Hospital Heidelberg and the University of Nottingham. Rapid-CNS2 consistently delivered high-precision diagnoses, even for small biopsy samples as tiny as 1.5 mm in diameter. Drastically reducing diagnostic timelines from 20 days to an average of 30–40 hours addresses critical global challenges in cancer diagnostics.
Rapid-CNS2 identified over 91% of single nucleotide variants confirmed by next-generation sequencing and correctly classified nearly 93% of CNS tumors to their methylation families. Its ability to assess critical markers, such as MGMT promoter methylation, across entire regions could provide insights that conventional methods often miss. The platform also can detect complex genetic changes with precision. For example, it identified a subclonal deletion in the EGFR gene in a glioblastoma sample—a mutation often overlooked by traditional sequencing methods. These capabilities underscore the advantages of long-read sequencing in providing a comprehensive genetic and epigenetic landscape.
Intraoperative and platform-agnostic applications
Simulated and real-time tests showed that crucial tumor classifications could be made within 15–30 minutes of sequencing, enabling timely treatment decisions. For instance, diffuse gliomas were reliably categorized into isocitrate dehydrogenase (IDH) mutant or wild-type subtypes, critical distinctions for guiding therapy. The platform provides actionable insights in over 70% of intraoperative cases by integrating methylation and copy number variation data. This rapid turnaround has the potential to redefine surgical decision-making and improve patient outcomes.
The methylation classifier MNP-Flex complements Rapid-CNS2 and works across different sequencing platforms and sample types, achieving 99.6% accuracy in identifying tumor families. This compatibility ensures that the tools can be integrated into various clinical workflows.
Rapid-CNS2 and MNP-Flex were designed for users with minimal computational expertise to lower barriers to adoption and make advanced molecular diagnostics available to more healthcare providers and patients. As these tools are scaled for broader clinical use, they could support democratizing access to state-of-the-art diagnostics, particularly in regions with limited resources.
