RexDrug: Reliable Multi-Drug Combination Extraction through Reasoning-Enhanced LLMs

2026-03-09

AI Summary

Existing approaches struggle to accurately extract variable-length, multi-drug combinations from biomedical literature due to complex compatibility logic and fragmented evidence. This work proposes RexDrug, a novel framework that, for the first time, leverages reasoning-enhanced large language models for this task. RexDrug employs a multi-agent collaboration strategy to generate expert-level reasoning trajectories for supervised fine-tuning and introduces a multidimensional reward function tailored for drug combination extraction to optimize model performance through reinforcement learning. Evaluated on the DrugComb dataset, RexDrug significantly outperforms current state-of-the-art methods and demonstrates strong generalization capability on DDI13. Both manual and automated evaluations confirm that RexDrug produces coherent medical reasoning and precisely identifies complex therapeutic regimens.

Technology Category

• Knowledge Representation and Reasoning: Computational Complexity of Reasoning
• Natural Language Processing: Information Extraction
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Automated Drug Combination Extraction (DCE) from large-scale biomedical literature is crucial for advancing precision medicine and pharmacological research. However, existing relation extraction methods primarily focus on binary interactions and struggle to model variable-length n-ary drug combinations, where complex compatibility logic and distributed evidence need to be considered. To address these limitations, we propose RexDrug, an end-to-end reasoning-enhanced relation extraction framework for n-ary drug combination extraction based on large language models. RexDrug adopts a two-stage training strategy. First, a multi-agent collaborative mechanism is utilized to automatically generate high-quality expert-like reasoning traces for supervised fine-tuning. Second, reinforcement learning with a multi-dimensional reward function specifically tailored for DCE is applied to further refine reasoning quality and extraction accuracy. Extensive experiments on the DrugComb dataset show that RexDrug consistently outperforms state-of-the-art baselines for n-ary extraction. Additional evaluation on the DDI13 corpus confirms its generalizability to binary drugdrug interaction tasks. Human expert assessment and automatic reasoning metrics further indicates that RexDrug produces coherent medical reasoning while accurately identifying complex therapeutic regimens. These results establish RexDrug as a scalable and reliable solution for complex biomedical relation extraction from unstructured text. The source code and data are available at https://github.com/DUTIR-BioNLP/RexDrug

Problem

• Drug Combination Extraction
• n-ary Relation Extraction
• Biomedical Literature
• Precision Medicine
• Pharmacological Research

Innovation

• reasoning-enhanced LLMs
• n-ary relation extraction
• multi-agent collaborative reasoning
• reinforcement learning with multi-dimensional rewards
• drug combination extraction