Publication: Moonlight AI's Breakthrough Detecting MDS from Peripheral Blood
- cruiz56
- Jun 24, 2025
- 1 min read
Updated: 2 days ago
Myelodysplastic syndrome (MDS) diagnosis has traditionally relied on invasive bone marrow biopsies and complex genomic testing with a 4-5 week turnaround time. Unfortunately, the vast majority of these biopsies come back negative for MDS. Moonlight AI has published a novel AI-based approach to detect MDS from neutrophil morphology in peripheral blood. This marks a significant step toward faster, less invasive diagnostics in hematology and oncology.
Read the full article in Annals of Hematology here.
Moonlight AI’s Novel Approach Using Peripheral Blood
Moonlight’s newly developed AI model analyzes neutrophil morphology at the sub-cellular level to extract diagnostic insight directly from peripheral blood smears. This approach offers several advantages:
Non-invasive sampling: Peripheral blood collection is simpler, less painful, and more accessible than bone marrow biopsies.
Early detection: The AI model identified dysplastic features based on known morphological signatures, but too subtle for human observers to detect.
Streamlined diagnostics: The model provided results in under 2 minutes with no additional sample preparation, enabling a first-line diagnostic approach to MDS diagnosis.
Increased access: This technology could enable rapid diagnosis in clinical settings without the need for specialized molecular testing infrastructure.
An International Consortium of Clinical Partners
Moonlight AI published this work with 4 clinical partners: the University of Basel, Universitätsspital Basel, the Universitätsspital Zürich, and Charité Universitätsmedizin in Berlin. The next phase of development will feature a global consortium of clinical partners who are interested in shaping the next generation of MDS diagnostics. These partnerships aim to improve the diagnostic accuracy of the model in a real-world clinical setting and to integrate the solution into routine clinical workflows.
