Automatic disease diagnosis has become increasingly valuable in clinical practice. The advent of large language models (LLMs) has catalyzed a paradigm shift in artificial intelligence, with growing evidence supporting the efficacy of LLMs in diagnostic tasks. Despite the growing attention in this field, many critical research questions remain under-explored. For instance, what diseases and LLM techniques have been investigated for diagnostic tasks? How can suitable LLM techniques and evaluation methods be selected for clinical decision-making? To answer these questions, we performed a comprehensive analysis of LLM-based methods for disease diagnosis. This scoping review examined the types of diseases, associated organ systems, relevant clinical data, LLM techniques, and evaluation methods reported in existing studies. Furthermore, we offered guidelines for data preprocessing and the selection of appropriate LLM techniques and evaluation strategies for diagnostic tasks. We also assessed the limitations of current research and delineated the challenges and future directions in this research field. In summary, our review outlined a blueprint for LLM-based disease diagnosis, helping to streamline and guide future research endeavors.

本研究针对大型语言模型在疾病诊断中的应用进行了全面分析，填补了目前研究中对特定疾病和模型技术的探讨不足。论文提出了数据预处理、模型选择和评估策略的指导方针，并概述了当前研究的局限性与未来方向。研究结果为利用大型语言模型进行疾病诊断提供了蓝图，有助于未来研究的系统化和指导。

大型语言模型在疾病诊断中的应用：范围审查