Merging Clinical Knowledge into Large Language Models for Medical Research and Applications: A Survey

2025-02-28

AI Summary

Current medical AI development lacks systematic construction guidelines, particularly suffering from paradigmatic gaps and inconsistent evaluation criteria for integrating clinical knowledge into large language models (LLMs). Method: This study presents the first comprehensive review of academic and industrial practices in building medical AI systems, proposing an integrated framework encompassing clinical databases (e.g., MIMIC, UMLS), biomedical knowledge graphs, instruction tuning, retrieval-augmented generation (RAG), and multi-granularity evaluation. We establish a six-dimensional comparative taxonomy and distill insights from representative systems—including DoctorGPT and Pangu-Drug. Contribution/Results: The work identifies key knowledge-enhancement pathways and clinical deployment bottlenecks, highlighting interpretability, data privacy, and clinical alignment as critical challenges. It provides both theoretical foundations and practical guidance for the standardized development and trustworthy deployment of medical LLMs.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval
• Cognitive Modeling & Cognitive Systems: Agent Architectures

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Large language models for search

Abstract

Clinical knowledge is the collection of information learned from studies on the causes, prognosis, diagnosis, and treatment of diseases. This type of knowledge can improve curing performances, and promote physical health. With the emergence of large language models (LLMs), medical artificial intelligence (medical AI), which aims to apply academic medical AI systems to real-world medical scenarios, has entered a new age of development, resulting in excellent works such as DoctorGPT and Pangu-Drug from academic and industrial researches. However, the field lacks a comprehensive compendium and comparison of building medical AI systems from academia and industry. Therefore, this survey focuses on the building paradigms of medical AI systems including the use of clinical databases, datasets, training pipelines, integrating medical knowledge graphs, system applications, and evaluation systems. We hope that this survey can help relevant practical researchers understand the current performance of academic models in various fields of healthcare, as well as the potential problems and future directions for implementing these scientific achievements.

Problem

• Lack of comprehensive review on medical AI systems development.
• Need for integrating clinical knowledge into large language models.
• Challenges in applying academic medical AI to real-world scenarios.

Innovation

• Integrates clinical databases into AI training pipelines
• Utilizes medical knowledge graphs for enhanced accuracy
• Compares academic and industrial medical AI systems