ACUTE leukaemia is an aggressive blood cancer that requires swift and precise molecular classification to inform urgent treatment decisions. Unfortunately, traditional diagnostic methods are often slow and do not fully capture the complex diversity of leukaemia subtypes. A recent breakthrough study introduces a novel framework based on genome-wide DNA methylation profiling that substantially improves classification accuracy and significantly accelerates diagnosis time.
Overview of Study
A breakthrough study has introduced an innovative framework using genome-wide DNA methylation profiling, which enhances classification accuracy and significantly expedites diagnosis. The researchers first built a comprehensive reference cohort of 2,540 samples to define 38 distinct methylation classes. This epigenomic perspective allows clinicians to gain a deeper understanding of the heterogeneity in leukaemia, which has important implications for treatment planning and prognosis.
Methylation- and AI-Guided Rapid Leukemia Subtype Inference (MARLIN)
Building on this, researchers then developed MARLIN (methylation- and AI-guided rapid leukemia subtype inference), a neural network capable of classifying acute leukaemia subtypes from sparse DNA methylation profiles. MARLIN demonstrated exceptional accuracy, concordant with standard diagnostic results in retrospective analyses and ongoing clinical use. Impressively, the system provides real-time classification within two hours of sample receipt.
The approach uses nanopore sequencing technology paired with artificial intelligence, offering a scalable, affordable, and easily implemented diagnostic tool. This advancement promises to complement and enhance current leukaemia diagnostics by facilitating rapid, comprehensive molecular classification that can guide immediate treatment strategies.
Future Outlooks
By revolutionising acute leukaemia diagnosis through AI and methylation profiling, MARLIN stands to improve patient outcomes by providing oncologists with timely, detailed insights into disease subtype and heterogeneity, ultimately supporting more precise and effective care, advancing hope for faster, personalised, and truly life-saving treatment decisions worldwide.
Reference
Steinicke TL et al. Rapid epigenomic classification of acute leukemia. Nat Genet. 2025; https://doi.org/10.1038/s41588-025-02321-z.
