Related Experiment Video
Updated: Jul 2, 2026
05:35
A Point-of-Care Method with Integrated Decision Support Tool to Estimate Anemia at Population Level
Published on: January 19, 2024
Clinical decision support in hematological malignancies using a case-grounded AI agent
Julian Zoller1,2,3,4, Michael Kalz1,2,3,4, Xuewei Wu5,6
1JRG Hematology and Immune Engineering, German Cancer Research Center (DKFZ), Heidelberg, Germany.
Nature Medicine
|July 1, 2026
Summary
HemaGuide, an AI agent for hematologic malignancies, improves decision-making by structuring cases and grounding recommendations in guidelines and past cases. It offers real-time support with high accuracy, enhancing clinical concordance and physician performance.
Area of Science:
- Artificial Intelligence in Medicine
- Computational Hematology
- Clinical Decision Support Systems
Background:
- Multidisciplinary tumor boards are crucial for treating hematologic malignancies, integrating complex data for optimal patient care.
- Uneven access to specialized subspecialty deliberation challenges timely and effective treatment planning.
Purpose of the Study:
- To develop and evaluate HemaGuide, a locally deployable large language model agent for clinical decision support in hematologic malignancies.
- To assess HemaGuide's ability to structure clinical data, route cases to specialized decision modes, and ground recommendations in evidence.
Main Methods:
- HemaGuide, a modular large language model agent, converts unstructured clinical notes into structured case representations.
- Cases are autonomously routed to 'guideline,' 'advanced,' or 'molecular' decision modes, grounded in disease-specific flowcharts and a memory of >2,000 tumor board cases.
- Benchmarking involved expert-blinded evaluation on 45 high-complexity cases, automated variant classification, simulated practice studies, and external validation on 555 independent cases.
Main Results:
- HemaGuide significantly improved concordance with tumor board decisions in expert-blinded benchmarking.
- Automated classification of 70 missense variants showed high concordance with expert standards, with no oncogenic variants downgraded.
- Agent-assisted resident physicians achieved near-senior concordance, and external validation showed 81.8% concordance across 47 entities.
Conclusions:
- Locally deployable, case-grounded large language model agents like HemaGuide can provide auditable clinical decision support for hematologic malignancies.
- HemaGuide demonstrates high concordance across institutions and under real-time conditions on commodity hardware, with a very low hallucination rate (0.3%).
- The system enhances clinical decision-making efficiency and accuracy, potentially improving patient outcomes in hematologic oncology.
Related Concept Videos
02:50Combination Therapies and Personalized Medicine
Combining two or more treatment methods increases the life span of cancer patients while reducing damage to vital organs or tissue from the overuse of a single treatment. Combination therapy also targets different cancer-inducing pathways, thus reducing the chances of developing resistance to treatment.
The combination of the drug acetazolamide and sulforaphane is a good example of combination therapy to treat cancer. The cells in the interior of a large tumor often die due to the hypoxic and...
The combination of the drug acetazolamide and sulforaphane is a good example of combination therapy to treat cancer. The cells in the interior of a large tumor often die due to the hypoxic and...
01:21Cancer Survival Analysis
Cancer survival analysis focuses on quantifying and interpreting the time from a key starting point, such as diagnosis or the initiation of treatment, to a specific endpoint, such as remission or death. This analysis provides critical insights into treatment effectiveness and factors that influence patient outcomes, helping to shape clinical decisions and guide prognostic evaluations. A cornerstone of oncology research, survival analysis tackles the challenges of skewed, non-normally...