Immunology Department, Weizmann Institute of Science, Rehovot 7610001, Israel.
Immunology Department, Weizmann Institute of Science, Rehovot 7610001, Israel; Cancer-Microbiome Division Deutsches Krebsforschungszentrum (DKFZ), Neuenheimer Feld 280, Heidelberg 69120, Germany.
Med. 2021 Jun 11;2(6):642-665. doi: 10.1016/j.medj.2021.04.006. Epub 2021 Apr 30.
Machine learning is increasingly integrated into clinical practice, with applications ranging from pre-clinical data processing, bedside diagnosis assistance, patient stratification, treatment decision making, and early warning as part of primary and secondary prevention. However, a multitude of technological, medical, and ethical considerations are critical in machine-learning utilization, including the necessity for careful validation of machine-learning-based technologies in real-life contexts, unbiased evaluation of benefits and risks, and avoidance of technological over-dependence and associated loss of clinical, ethical, and social-related decision-making capacities. Other challenges include the need for careful benchmarking and external validations, dissemination of end-user knowledge from computational experts to field users, and responsible code and data sharing, enabling transparent assessment of pipelines. In this review, we highlight key promises and achievements in integration of machine-learning platforms into clinical medicine while highlighting limitations, pitfalls, and challenges toward enhanced integration of learning systems into the medical realm.
机器学习正越来越多地融入临床实践，其应用范围包括临床前数据处理、床边诊断辅助、患者分层、治疗决策以及作为初级和二级预防一部分的早期预警。然而，在机器学习的应用中，存在许多技术、医学和伦理方面的考虑因素至关重要，包括需要在实际环境中仔细验证基于机器学习的技术，公正评估收益和风险，避免过度依赖技术以及随之而来的丧失临床、伦理和社会相关决策能力。其他挑战包括需要仔细进行基准测试和外部验证、将计算专家的终端用户知识传播到现场用户，以及负责任地共享代码和数据，从而能够对管道进行透明评估。在这篇综述中，我们强调了将机器学习平台集成到临床医学中的关键承诺和成就，同时突出了学习系统融入医疗领域所面临的局限性、陷阱和挑战。
