• 中国核心期刊数据库收录期刊
  • 中文科技期刊数据库收录期刊
  • 中国期刊全文数据库收录期刊
  • 中国学术期刊综合评价数据库统计源期刊等

快速检索引用检索图表检索高级检索

中国医药导刊 ›› 2025, Vol. 27 ›› Issue (2): 201-206.

• 专栏:药物临床试验 • 上一篇    下一篇

基于机器学习方法的临床研究报告规范和建议

郭颖1, 王点墨2, 邹碧君1, 何莲珠1, 李苏1*   

  1. 1.中山大学肿瘤防治中心,华南肿瘤国家重点实验室,癌症医学协同创新中心,广东 广州 510060;
    2.中山大学中山医学院,广东 广州 510275
  • 收稿日期:2025-02-13 修回日期:2025-03-14 出版日期:2025-02-28 发布日期:2025-02-28
  • 基金资助:
    中山大学教学质量工程建设项目(2023)

Recommendations and Guidelines for Reporting Clinical Research with Machine Learning Methods

  1. 1.Sun Yat-sen University Cancer Center State Key Laboratory of Oncology in South China
    Collaborative Innovation Center for Cancer Medicine Guangdong Guangzhou 510060, China
    2.Zhong Shan School of Medicine Sun Yat-sen UniversityGuangdong Guangzhou 510275, China
  • Received:2025-02-13 Revised:2025-03-14 Online:2025-02-28 Published:2025-02-28

摘要:

确保临床试验数据的科学性和可靠性,是药物临床试验发展的关键,也是推动我国医药产业进步的重要举措。近年来,随着临床研究数据集复杂性的不断提升,传统统计方法的应用受到限制,机器学习(ML)方法被大量应用于观察性研究和预测模型中。机器学习能有效分析和利用海量数据,提供个性化的诊断、治疗和预后预测,从而提高临床研究和决策的准确性和效率。然而,目前由于研究报告撰写规则尚不统一,研究报告质量参差不齐。2020年,学术期刊《Cardiovascular Quality and Outcomes》发表的文章“临床研究中机器学习分析报告的建议”,探讨了机器学习分析结果进行透明和结构化报告的必要性,对采用机器学习技术的临床研究报告提出了系列规范建议。20244月,TRIPOD指南也更新了人工智能(AI)相关的报告规范。本研究以该分析报告为重点,借鉴TRIPOD+AI指南内容,探讨并整理了基于机器学习方法下我国临床研究报告的撰写规范,并提出建议,旨在让临床医生熟悉机器学习的基本原理,进一步提高我国药物临床试验专家学者应用机器学习方法促进临床研究的可重复性,加强临床研究报告撰写的规范性。


关键词: 生物信息学, 机器学习, 预后, 结果可重复性, 研究报告, 临床研究

Abstract:

Ensuring the scientificity and reliability of clinical trial data is the key to the development of drug clinical trials and it is also an important measure to promote the progress of the pharmaceutical industry in China.In recent years with the growing complexity of clinical data sets the application of traditional statistical methods has become limited and the adoption of machine learning ML techniques has become widespread in observational studies and predictive models. Machine learning can effectively analyze and leverage massive amounts of data to provide personalized diagnosis treatment and prognosis predictions thereby enhancing the accuracy and efficiency of clinical research and decision-making. However at present since the rules for writing research reports are not yet uniformthe quality of research reports varies. In 2020 the academic journal Cardiovascular Quality and Outcomes published an article "Recommendations for Reporting Machine Learning Analyses in Clinical Research"which examined the need for transparent and structured reporting of machine learning analysis results and put forward a series of normative recommendations for reporting clinical research using machine learning technology. In April 2024 the TRIPOD guidelines also updated the reporting guidelines related to artificial intelligence AI.Focusing on the analysis report and refer to the content of TRIPOD+AI guideline this study discusses and sorts out the specifications and recommendations for reporting clinical research using machine learning methods so as to familiarize clinicians with the basic principles of machine learning further improve the reproducibility of Chinese drug clinical trial experts and scholars applying machine learning methods to promote clinical research and strengthen the standardization of clinical research report writing.


Key words: Bioinformatics , Machine learning , Prognosis , Reproducibility of results , Research report , Clinical research

中图分类号: