Analyse complex measurements using explainable methods and medical expertise.
Biomedical data is high-dimensional, heterogeneous and often limited. Experts need to understand what a classification is based on.
Through interdisciplinary collaboration, IAP develops methods for structure discovery, quality assessment and explainable classification. Examples include FlowXAI, research on inflammatory conditions and the Cinderella method for assessing haemodilution.
Scientifically evaluated methods, Proofs of Concept (PoCs) and research prototypes with documented evaluation status. Development is aligned with medical experts on the intended use and the validation it requires.
Read the technical detail ↓The development objective is a workflow that supports suitable routine cases and identifies complex or uncertain cases for specialist review. Explanations show which features and rules underpin a classification. Actual clinical workflow benefits require evaluation in the intended setting.
FlowXAI examines explainable classification and case-level reliability in flow-cytometry data. The Cinderella work addresses critical haemodilution in bone-marrow samples. Both illustrate why measurement quality and domain interpretation matter alongside the choice of learning method.
The inflammation study examines 80 patients and 38 healthy controls. Transfer to further patient populations, instruments and antibody panels needs assessment for the particular application.
Thrun, M. C., Hoffmann, J., Krause, S. W., Krawitz, P., Stier, Q., Neubauer, A., Brendel, C. & Ultsch, A. (2026). Self-explaining artificial intelligence for the classification of B cell non-Hodgkin lymphoma: A diagnostic decision support study. PLOS Medicine, 23(7), e1004889.
FlowXAI combines assessment of data structures with explainable lymphoma classification and an estimate of reliability for each individual result. This helps domain experts understand results and focus their review on cases requiring closer examination.
Hoffmann, J., Rheude, A., Neubauer, A., Brendel, C. & Thrun, M. C. (2024). Development of an explainable AI system using routine clinical parameters for rapid differentiation of inflammatory conditions. Frontiers in Immunology, 15, 1364954.
Development of an explainable method for differentiating inflammatory conditions using clinical measurements.
Hoffmann*, J., Thrun*, M. C., Röhnert, M. A., von Bonin, M., Oelschlägel, U., Neubauer, A., Ultsch, A. & Brendel, C. (2023; online 2022). Identification of critical hemodilution by artificial intelligence in bone marrow assessed for minimal residual disease analysis in acute myeloid leukemia: The Cinderella method. Cytometry Part A, 103(4), 304–312.
AI for estimating critical blood dilution in bone marrow samples. *J. Hoffmann and M. C. Thrun contributed equally.
Commissioned prototype · related collaborative research
Interpreting flow-cytometry data requires expertise. Rare entities, different measurement conditions and limited training data complicate automated classification.
Project details →
Research project · 2024 study
Different inflammatory conditions can cause similar symptoms. The study examines what additional support can be obtained from blood parameters.
Project details →
Hospital prototype · collaborative research
The quality of blood and bone-marrow samples affects interpretation. Dilution and complex cell populations need to be considered in the analysis.
Project details →Option 1
IAP investigates the problem, develops and evaluates the approach, and hands over the agreed prototype with documentation. Your IT team takes responsibility for further product development and operation.
Option 2
IAP develops the agreed application, including its interface and integrations, connects it to the intended workflow, and completes acceptance and handover. Your organisation or operating partner takes responsibility for subsequent operation.
In a free, non-binding 20-minute introductory discussion, we clarify your objective, starting point and a useful next step.
