Clinical decision support in hematological malignancies using a case-grounded AI agent

Multidisciplinary tumor boards integrate longitudinal treatment histories, molecular profiling and rapidly evolving evidence to guide decisions in hematological malignancies, yet access to this level of subspecialty deliberation is increasingly uneven. Here we develop HemaGuide, a locally deployable, modular large language model agent that converts unstructured clinical documents into structured case representations, autonomously routes cases to specialized decision modes ('guideline', 'advanced' and 'molecular') and grounds recommendations in disease-specific guideline flowcharts and a clinical decision memory of >2,000 real-world tumor board cases. In expert-blinded benchmarking on 45 high-complexity cases across six foundation models, HemaGuide substantially improved concordance with tumor board decisions. A systematic ablation study across 11 layers confirmed that performance gains were routing-type-dependent, with no single component sufficient across case types. Automated classification of 70 clinically relevant missense variants showed high concordance with expert standards; no oncogenic variant was downgraded to benign and the whole workflow was completed under real-time conditions on commodity hardware with a median latency of 39 s rather than the hours typically required for manual molecular board workflows. In a simulated practice study, agent-assisted resident physicians achieved near-senior concordance and partially outperformed senior physicians in their subspecialty. External validation on 555 independent cases from a second academic center yielded 81.8% concordance across 47 entities, and a prospective 1-month silent trial on 64 consecutive, unselected cases achieved 82.8% concordance. Hallucinations occurred in 2 of 664 evaluated cases (0.3%). Together these data provide evidence that locally deployable, case-grounded large language model agents can deliver auditable clinical decision support across hematological malignancies, with concordance maintained across institutions and under real-time conditions on commodity hardware.

Authors

Institutions

Publication Details

Journal
Nature Medicine
Published
2026-06-30
DOI
https://doi.org/10.1038/s41591-026-04494-4
Primary Topic
Artificial Intelligence in Healthcare and Education
Type
article
Field-Weighted Citation Impact
0.00

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Clinical decision support in hematological malignancies using a case-grounded AI agent

Martin Dugas, Michael von Bergwelt‐Baildon, Carsten Müller‐Tidow, Antonia Schach et al.
Nature Medicine
Artificial Intelligence in Healthcare and Education
article

Clinical decision support in hematological malignancies using a case-grounded AI agent

Martin Dugas, Michael von Bergwelt‐Baildon, Carsten Müller‐Tidow, Antonia Schach, Tim Wagner, Martin H. Dreyling, Julian Zoller, Marc S. Raab, Jakob Nikolas Kather, A Metzler, Linus Kruk, Andreas Trumpp, Mirco Friedrich, Peter Dreger, Niklas Kehl, Tim Sauer, Tobias Tix, Silvia Barbosa, Xuewei Wu, Sven Cuntz, René Onken, Julius J. Michel, David Sedloev, Michael Kalz, Jonathan Naboschni, Oliver L. Saldanha, Jan H. Frenking, Cornelius Funk, Kai Rejeski, Alanna Kirschner, Sarah Richter
article en

Abstract

Multidisciplinary tumor boards integrate longitudinal treatment histories, molecular profiling and rapidly evolving evidence to guide decisions in hematological malignancies, yet access to this level of subspecialty deliberation is increasingly uneven. Here we develop HemaGuide, a locally deployable, modular large language model agent that converts unstructured clinical documents into structured case representations, autonomously routes cases to specialized decision modes ('guideline', 'advanced' and 'molecular') and grounds recommendations in disease-specific guideline flowcharts and a clinical decision memory of >2,000 real-world tumor board cases. In expert-blinded benchmarking on 45 high-complexity cases across six foundation models, HemaGuide substantially improved concordance with tumor board decisions. A systematic ablation study across 11 layers confirmed that performance gains were routing-type-dependent, with no single component sufficient across case types. Automated classification of 70 clinically relevant missense variants showed high concordance with expert standards; no oncogenic variant was downgraded to benign and the whole workflow was completed under real-time conditions on commodity hardware with a median latency of 39 s rather than the hours typically required for manual molecular board workflows. In a simulated practice study, agent-assisted resident physicians achieved near-senior concordance and partially outperformed senior physicians in their subspecialty. External validation on 555 independent cases from a second academic center yielded 81.8% concordance across 47 entities, and a prospective 1-month silent trial on 64 consecutive, unselected cases achieved 82.8% concordance. Hallucinations occurred in 2 of 664 evaluated cases (0.3%). Together these data provide evidence that locally deployable, case-grounded large language model agents can deliver auditable clinical decision support across hematological malignancies, with concordance maintained across institutions and under real-time conditions on commodity hardware.

Nature Medicine
University of Leeds (GB), German Cancer Research Center (DE), Heidelberg University (DE), University Hospital Heidelberg (DE), LMU Klinikum (DE), National Center for Tumor Diseases (DE), Heidelberg Institute for Stem Cell Technology and Experimental Medicine (DE), First Affiliated Hospital of Jinan University (CN), Hochschule für Technik und Wirtschaft Dresden – University of Applied Sciences (DE), Deutsches Konsortium für Translationale Krebsforschung (DE), Technische Universität Dresden (DE), Ludwig-Maximilians-Universität München (DE)
Breast Cancer Research Foundation, Deutsches Krebsforschungszentrum, National Institute for Health and Care Research, Department of Health and Social Care, European Commission, Deutscher Akademischer Austauschdienst, Deutsche Forschungsgemeinschaft, Technische Universität Dresden, Else Kröner-Fresenius-Stiftung, Deutsche Krebshilfe, Deutschen Konsortium für Translationale Krebsforschung, Gemeinsame Bundesausschuss, Universitätsklinikum Heidelberg, HORIZON EUROPE Framework Programme, Leeds Biomedical Research Centre
Peace, Justice and strong institutions
Openalex Percentile: Top 11%
Artificial Intelligence in Healthcare and Education
DOIOpen Access PDFOpenAlex
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.