A self-correcting Agentic Graph RAG for clinical decision support in hepatology
Самокорректирующийся агентный Graph RAG для поддержки принятия клинических решений в гепатологии
2025-12-16
SCID: 54.1/nq8frmrf
Discuss with AI
Agentic Graph RAGLLM hallucinationsclinical decision supporthepatology knowledge graphself-correcting retrieve-evaluate-refine loop
Figures from the paper
Abstract (AI)
Introduction: Clinical decision-making in hepatology is currently challenged by the rapid expansion of medical knowledge and the limitations of Large Language Models (LLMs), specifically their unreliability and tendency to hallucinate. Furthermore, standard Retrieval-Augmented Generation (RAG) paradigms often fail to effectively leverage complex medical knowledge structures. Methods: To address these issues, we propose an Agentic Graph RAG framework built upon a clinically-verified hepatology knowledge graph. Our approach utilizes a state-driven agentic system employing a self-correcting "retrieve-evaluate-refine" loop. Within this workflow, agents dynamically generate, semantically validate, assess, and iteratively optimize graph search strategies to construct a comprehensive and accurate context, which is then used by an LLM to generate reliable responses. Results: The framework was evaluated on a custom dataset of clinical questions. It significantly outperformed baseline models (including GPT-4, standard RAG, and Graph RAG) across all evaluation metrics. Specifically, our model achieved superior scores in faithfulness (0.94), context recall (0.92), and answer relevancy (0.91). Discussion: This agentic approach effectively mitigates LLM hallucinations and provides accurate, interpretable answers. These findings demonstrate the framework's potential as a robust, next-generation intelligent clinical decision support tool for hepatology.
Key Findings
1
Achieved faithfulness of 0.94, context recall of 0.92, and answer relevancy of 0.91.
2
Introduces an Agentic Graph RAG framework built on a clinically verified hepatology knowledge graph for clinical decision support.
3
The approach mitigates LLM hallucinations while producing accurate and interpretable hepatology answers, supporting its potential as a robust clinical decision-support tool.
4
The framework significantly outperformed GPT-4, standard RAG, and Graph RAG across all evaluation metrics on a custom clinical-question dataset.
5
Uses a state-driven self-correcting “retrieve-evaluate-refine” loop to generate, validate, assess, and optimize graph-search strategies.
Research Object
Agentic Graph RAG framework for hepatology clinical decision support
Research Subject
Self-correcting retrieval, evaluation, and refinement performance for generating faithful, relevant, and accurate clinical answers
Publication Details
Publication Date
2025-12-16
Journal
Publisher
ISSN
Cited by
9
Open access PDF
Access Type
Author Information
Download PDF
Subscribe to digest
References available in scid.ai8
A Survey on Evaluation of Large Language Models2024
Large language models encode clinical knowledge2023
Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks2020
Redefining Radiology: A Review of Artificial Intelligence Integration in Medical Imaging2023
Benchmarking Retrieval-Augmented Generation for Medicine2024
RAGAs: Automated Evaluation of Retrieval Augmented Generation2024
Large language models (LLMs): survey, technical frameworks, and future challenges2024
From LLM Reasoning to Autonomous AI Agents: A Comprehensive Review2026