Sepsis Detection : astrategic collaboration will integrate GeodAIsics’ AI technology into HORIBA’s hematology systems to improve early identification of sepsis from routine blood tests. First clinical validation is expected in 2025.
GeodAIsics, a Grenoble-based deeptech company specialized in AI-driven medical software, and HORIBA, a global leader in hematology analysis based in France, have signed a strategic partnership to integrate an AI module into a range of hematology analyzers.
The software will generate an interpretable alert score for laboratory professionals based on routine blood tests, aiming to detect early signs of sepsis, septic shock or systemic inflammatory response syndrome (SIRS).
At the core of the solution is Generative Manifold Learning (GML), GeodAIsics’ proprietary unsupervised generative AI approach. GML can identify subtle biomedical signatures using limited datasets without manual labeling, highlighting correlations that support clinical decision-making.
“This partnership strengthens our AI decision-support offering, especially in infectious diseases. HORIBA hematology systems equipped with our module gain earlier and faster detection capabilities, directly benefiting patient care,” said Cédric Sire, Chief Business Officer at GeodAIsics.
“By integrating GeodAIsics’ AI into our analyzers, we take a decisive step in delivering rapid and reliable diagnostics. This collaboration reflects our shared commitment to innovate for healthcare,” added Michael Bruckner, Marketing & Business Development Director at HORIBA.
Early sepsis detection: a major public health challenge
Sepsis remains a critical medical emergency and one of the world’s leading causes of death, with an estimated 50 million cases and more than 11 million deaths in 2017. Earlier detection, made possible through standard laboratory tests, could significantly improve patient management.
Promising results for sepsis detection
Preliminary developments have shown sensitivity and specificity above 80% in sepsis detection. The first software version is expected to undergo clinical validation in 2025. The project also paves the way for expanding the AI platform to other hematology indications.
About HORIBA analyzers
The Yumizen H1500/H2500 and H500 hematology platforms are widely used in hospital and private laboratories. The integration of GeodAIsics’ AI module will allow automated and interpretable scores to be returned directly within the laboratory workflow.
Key points
Integration of a French AI into HORIBA’s hematology systems.- Early detection of sepsis, septic shock and SIRS from routine blood tests.
- > 80% sensitivity and specificity observed in preliminary results.
- Clinical validation expected in 2025.
