利用大语言模型提升影像学报告可读性:一项系统综述
本系统综述纳入截至2025年2月PubMed、Scopus和Google Scholar索引的17项研究,评估大语言模型简化患者影像报告可读性的效果;12项定量评估可读性指标的研究均报告改善,但输入、模型与评价指标异质性大,作者提出未来研究最佳实践指南。
要点
- 大语言模型可能提升影像报告对患者的可读性,但现有研究间异质性大。
- 定量评估可读性指标的研究均报告改善,定性评估结果因评分者是否为医生而异。
- 作者提出最佳实践指南以标准化未来大语言模型研究。
核心结论
- 大语言模型展现出提升影像报告对患者可及性的潜力,但现有文献受限于输入、模型和评价指标的异质性。
- 作者提出一套最佳实践指南以标准化未来大语言模型研究。
研究设计与样本
- 系统综述,遵循PRISMA指南,检索PubMed、Scopus和Google Scholar截至2025年2月发表的文献。
- 共识别2126条引文,17项研究纳入定性分析;纳入标准为使用大语言模型简化诊断或介入影像报告供患者使用并评估可读性的研究。
- 采用混合方法评估工具2018进行偏倚评估。
诊断标准与关键数值
- 71%的研究使用单一大型语言模型,29%的研究使用多个大型语言模型。
- 最常用的大语言模型包括ChatGPT、Google Bard/Gemini、Bing Chat、Claude和Microsoft Copilot。
- 所有评估定量可读性指标的研究(n=12)均报告改善。
- 通过定性方法评估简化报告的结果在医生与非医生评分者之间表现不一。
临床应用与阅片要点
- 原文未据此给出推荐。
局限性
- 现有文献受限于输入、模型和评价指标在不同研究间的异质性。
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)。
英文原文(摘要)· 点击展开
BACKGROUND: Patients increasingly have direct access to their medical record. Radiology reports are complex and difficult for patients to understand and contextualize. One solution is to use large language models (LLMs) to translate reports into patient-accessible language.
OBJECTIVE: This review summarizes the existing literature on using LLMs for the simplification of patient radiology reports. We also propose guidelines for best practices in future studies.
EVIDENCE ACQUISITION: A systematic review was performed following Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. Studies published and indexed using PubMed, Scopus, and Google Scholar up to February 2025 were included. Inclusion criteria comprised studies that used LLMs for simplification of diagnostic or interventional radiology reports for patients and evaluated readability. Exclusion criteria included non-English publications, abstracts, conference presentations, review articles, retracted articles, and studies that did not focus on report simplification. The Mixed Methods Appraisal tool 2018 was used for bias assessment. Given the diversity of results, studies were categorized based on reporting methods, and qualitative and quantitative findings were presented to summarize key insights.
EVIDENCE SYNTHESIS: A total of 2,126 citations were identified and 17 were included in the qualitative analysis. Of these studies, 71% used a single LLM, and 29% of studies used multiple LLMs. The most prevalent LLMs included ChatGPT, Google Bard/Gemini, Bing Chat, Claude, and Microsoft Copilot. All studies that assessed quantitative readability metrics (n = 12) reported improvements. Assessment of simplified reports via qualitative methods demonstrated varied results with physician versus nonphysician raters.
CONCLUSION AND CLINICAL IMPACT: LLMs demonstrate the potential to enhance the accessibility of radiology reports for patients, but the literature is limited by heterogeneity of inputs, models, and evaluation metrics across existing studies. We propose a set of best practice guidelines to standardize future LLM research.
引用
- Patwardhan V, Balchander D, Fussell D, et al. Leveraging Large Language Models to Enhance Radiology Report Readability: A Systematic Review. J Am Coll Radiol. 2025. PMID: 40945554. doi:10.1016/j.jacr.2025.09.004
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。