Multi-agent system predicts knee osteoarthritis progression with interpretable evidence
OA-MAP: Evidence-Grounded Multi-Agent Multimodal Framework for Interpretable Knee Osteoarthritis Progression
Artificial IntelligenceMultiagent Systems
Summary
Knee osteoarthritis is a disease that worsens over time, and doctors want better ways to predict its progress. The authors created OA-MAP, a computer system that uses different types of medical images and patient information to estimate how the disease might develop. It explains its predictions by showing evidence from scans, clinical tests, and scientific papers, and lets doctors check and adjust the results. Tests with real patient data showed that the system can predict worsening knee damage and pain with reasonable accuracy.
What this means in practice
- •For clinical it teams: Integrate a multi-agent platform to automate knee osteoarthritis progression assessments with transparent evidence support for clinician review.
- •For chronic disease management teams: Use combined imaging and clinical data-driven tools to monitor patient osteoarthritis progression, improving decision-making accuracy.
Authors
Sixu Chen, Mingrui Yang, Qiang Guan, Xiaojuan Li
Abstract
Knee osteoarthritis (KOA) progression prediction can support patient monitoring, requiring the integration of multimodal data and multidomain expertise. Moreover, isolated risk estimates provide limited insight underlying a prediction. To automate the progression assessment workflow and reduce manual effort while providing interpretable findings and supporting evidence, we present OA-MAP, an autonomous multi-agent framework for evidence-grounded assessment of structural and pain progression in KOA. The system incorporates modality-specific agents including MRI, X-ray, and clinical agents, together with a coordinator agent. This framework can autonomously recruit specialist agents, select tools for prediction and analysis, and retrieve literature as external evidence based on user request and available patient information. An uncertainty-informed human-in-the-loop mechanism enables clinicians to review and correct intermediate findings, triggering recomputation of affected results. We evaluate the prediction models using 600 participants from the FNIH Osteoarthritis Biomarkers Consortium cohort. On the test set of 100 participants, the fusion models achieve AUROCs of 0.80 for structural progression and 0.68 for pain progression. A case study illustrates how OA-MAP combines risk estimates with intermediate findings, cross-modal conflicts, literature support, and uncertainty indicators to support interactive review.