文献速递牙体牙髓病口腔修复与种植

基于检索增强生成的根管治疗后牙体修复临床决策支持系统

[A retrieval-augmented generation-based clinical decision support system for the restoration of endodontically treated teeth].
PubMed 文献 · Zhonghua Kou Qiang Yi Xue Za Zhi · 2026-01-01

研究构建了基于检索增强生成(RAG)的临床决策支持系统CDSS-RAG-Pro,纳入《牙体缺损、牙列缺损与缺失修复诊疗指南(2022年版)》等资料构建知识库,对DeepSeek、通义千问、ChatGPT 3种大语言模型进行优化。选取30份典型临床病例,由专家小组从决策准确性、逻辑合理性、专业可读性三维度评分,并用ROUGE-L F1和BERTScore F1评估语义相似度。结果显示RAG模式下所有模型各项指标均显著提高,BERTScore F1最高达0.91(DeepSeek)。

要点

  • RAG模式下3种大语言模型各项评价指标均显著优于无RAG模式。
  • RAG模式下BERTScore F1最高为0.91,由DeepSeek模型取得。
  • 模型类型与RAG模式在决策准确性评分和ROUGE-L F1上存在显著交互效应。

核心结论

  • 基于RAG的CDSS-RAG-Pro可提升通用大语言模型在根管治疗后牙体修复决策中的准确性和可解释性。
  • RAG技术作为面向小样本量的轻量化大语言模型配置策略,在牙体缺损修复智能决策方面具有一定优势。

研究设计与样本

  • 系统检索并纳入《牙体缺损、牙列缺损与缺失修复诊疗指南(2022年版)》等口腔修复专业资料构建结构化知识库。
  • 开发RAG模块并集成至DeepSeek-V3、Qwen 2.5-72B-Instruct和GPT-3.5-turbo三种大语言模型,形成CDSS-RAG-Pro系统。
  • 自中山大学附属口腔医院口腔修复科2024年5月至2025年12月临床诊疗记录中筛选30份典型临床病例资料。
  • 由专家小组采用李克特量表从决策准确性、逻辑合理性、专业可读性三维度评分,并用ROUGE-L F1和BERTScore F1评估语义相似度,采用广义估计方程分析差异。

主要发现

  • 与无RAG模式相比,RAG模式下所有大语言模型各项评价指标得分均显著提高(均P<0.05)。
  • RAG模式下BERTScore F1最高达0.91(DeepSeek)。
  • 所有评价指标层面各大语言模型间差异均有统计学意义(均P<0.05)。
  • 模型与RAG模式在决策准确性评分(交互效应χ2=10.11,P<0.001)与ROUGE-L F1值(交互效应χ2=8.57,P<0.05)上存在显著交互效应。

关键数值与推荐

  • 知识库构建纳入《牙体缺损、牙列缺损与缺失修复诊疗指南(2022年版)》等权威修复学资料。
  • 评价维度包括决策准确性、逻辑合理性、专业可读性,采用李克特量表评分。
  • 客观指标为ROUGE-L F1(侧重文本知识掌握程度)和BERTScore F1(侧重思维模式与决策逻辑语义一致性)。

临床怎么用

  • CDSS-RAG-Pro可作为根管治疗后牙体缺损修复的客观决策辅助工具,提升通用大语言模型输出的准确性与可解释性。
  • RAG技术可作为小样本量条件下轻量化大语言模型配置策略,赋能口腔修复临床决策。

以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。

原文

Objective: To develop and evaluate Clinical Decision Surpport System-augmented by retrieval-augmented generation (RAG)-Professional (CDSS-RAG-Pro), a RAG-based clinical decision support system, through multi-level comparison of multiple large language models (LLM), and to provide an objective decision-making aid for the clinical restoration of endodontically treated teeth. Methods: A structured knowledge base was constructed by systematically retrieving and incorporating authoritative prosthodontic sources, including the Guideline for the Restoration of Tooth Defects, Dentition Defects and Edentulism (2022 Edition). A RAG module was developed accordingly and integrated into three LLM (DeepSeek-V3, Qwen 2.5-72B-Instruct, and GPT-3.5-turbo) to form the CDSS-RAG-Pro clinical decision support system for post-endodontic restoration. To evaluate the decision-making capability of this system, an expert panel was established. Thirty typical clinical cases were selected from the Department of Prosthodontics, Hospital of Stomatology, Sun Yat-sen University, from May 2024 to December 2025. The three LLM generated restorative plans under both RAG and non-RAG modes. Using a Likert scale, the expert panel scored the restorative plans across three dimensions: decision accuracy, logical coherence, and professional readability. Meanwhile, the expert panel drafted standard restorative reports. Two objective metrics, F1 score of the recall-oriented understudy for gisting evaluation based on longest common subsequence (ROUGE-L F1) and F1 score of the bidirectional encoder representations from Transformer-based semantic similarity score (BERTScore F1), were used to evaluate the semantic similarity between the generated restorative plans and the standard restorative reports. ROUGE-L F1 focused on assessing the model' s mastery of textual knowledge such as restorative guidelines, while BERTScore F1 focused on semantic consistency in reasoning patterns and decision-making logic. Generalized estimating equations were used to analyze differences in decision performance across LLM and between RAG and non-RAG modes. Results: Compared with the non-RAG mode, all models showed improved performance across all evaluation metrics in the RAG mode, with statistically significant differences (all P0.05). The highest BERTScore F1 in the RAG mode was 0.91 (DeepSeek-V3). Significant differences were observed among the three LLM across all evaluation metrics (all P0.05). Significant interaction effects were found between model type and RAG mode for decision accuracy score (interaction χ2=10.11, P0.001) and ROUGE-L F1 (interaction χ2=8.57, P0.05). Conclusions: The RAG-based CDSS-RAG-Pro improves the accuracy and interpretability of general-purpose LLM in decision-making for post-endodontic tooth restoration. This study demonstrates the unique clinical safety and scalability advantages of RAG as a lightweight, small-sample-oriented LLM configuration strategy to empower prosthodontic clinical decision-making. 目的: 基于检索增强生成(RAG)技术构建融合国际前沿修复理念与专业知识的临床决策支持系统(CDSS-RAG-Pro),并通过真实临床病例验证RAG对多种大语言模型(LLM)专科决策能力的提升效果,以期为临床提供客观的决策辅助工具。 方法: 开发RAG模块并系统检索和纳入《牙体缺损、牙列缺损与缺失修复诊疗指南(2022年版)》等口腔修复专业资料构建结构化知识库,据此对CDSS-RAG-Pro支持的DeepSeek、通义千问、ChatGPT 3种LLM进行优化,形成针对根管治疗后牙体缺损修复的临床决策支持系统。为评估该系统决策能力,成立专家小组,自中山大学附属口腔医院口腔修复科2024年5月至2025年12月的临床诊疗记录中筛选30份典型临床病例资料,分别由3种LLM在有无RAG两种模式下出具修复决策报告,参考李克特量表,由专家小组对上述修复决策报告从决策准确性、逻辑合理性、专业可读性3个维度进行评分。同时,由专家小组撰写标准修复报告,随后采用基于最长公共子序列的召回率驱动辅助评估指标的F1分数(ROUGE-L F1)和基于双向编码器表示变换器的语义相似度评分的F1分数(BERTScore F1),评估LLM修复决策报告与标准修复报告的语义相似度,其中ROUGE-L F1侧重于衡量LLM对修复指南等文本知识的掌握程度;BERTScore F1则关注LLM生成的修复决策报告与标准修复报告在思维模式、决策逻辑等语义层面的一致性。采用广义估计方程分析各LLM间及有无RAG模式下的决策表现差异。 结果: 与无RAG模式相比,RAG模式下所有LLM各项评价指标得分均显著提高(均P0.05)。RAG模式下的BERTScore F1最高达0.91(DeepSeek)。所有评价指标层面各LLM间差异均有统计学意义(均P0.05)。模型与RAG模式在决策准确性评分(交互效应χ2=10.11,P0.001)与ROUGE-L F1值(交互效应χ2=8.57,P0.05)上存在显著交互效应。 结论: 基于RAG技术构建的CDSS-RAG-Pro可提升通用LLM在根管治疗后牙体修复决策中的准确性和可解释性,体现出RAG技术作为面向小样本量的轻量化LLM配置策略在牙体缺损修复智能决策方面具有一定优势。.

来源:Zhonghua Kou Qiang Yi Xue Za Zhi,查看原文。版权归原作者所有。

引用

  1. Zhai JR, Guo JW, Su WY, et al. [A retrieval-augmented generation-based clinical decision support system for the restoration of endodontically treated teeth]. Zhonghua Kou Qiang Yi Xue Za Zhi. 2026. PMID: 42394554. doi:10.3760/cma.j.cn112144-20260409-00225

供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。