系统审查:神经放射学中的代理人工智能:技术承诺与有限的临床证据
Sara Salehi1, Varekan Keishing2, Yashbir Singh3
1Radiology Informatics Lab, Department of Radiology, Mayo Clinic, Rochester, MN, 55905, USA. salehi.sara@mayo.edu.
Journal of imaging informatics in medicine
|February 2, 2026
概括
神经放射学中的代理人工智能 (AI) 显示出技术上的希望,但缺乏临床证据. 目前的证据不足以部署,需要进行多中心试验,以确保患者的安全性和结果.
科学领域:
- 人工智能的人工智能
- 神经辐射学神经辐射学
- 医疗成像医学成像
背景情况:
- 建议使用代理人工智能 (AI) 系统,结合代推理和自主工具使用,以克服神经放射学中大语言模型 (LLM) 的局限性.
- 这些在神经辐射学中的代理AI系统的临床实施和验证在很大程度上尚未评估.
研究的目的:
- 系统地审查和评估在神经辐射学中代理AI的实施和临床验证.
- 评估现场代理人工智能应用的证据稀缺性,方法严谨性和临床实用性.
主要方法:
- 在PubMed,Web of Science和Scopus (2022年1月至2025年8月) 进行了系统的文献搜索.
- 如果他们实施了代理人工智能,定义为需要代推理加上自主工具使用或多代理合作,就包括了这些研究.
- 六位独立审查员使用调整的 QUADAS-AI 标准评估了研究质量,重点关注实施,验证和结果.
主要成果:
- 在230个记录中,只有9个 (3.90%) 符合纳入标准,这表明严重的证据稀缺.
- 很大一部分研究 (30%) 错误地将他们的AI描述为代理,缺乏真正的自主性或多代理合作.
- 唯一的随机对照试验显示了高的技术性能,但没有可测量的临床益处,突出了技术准确性和临床实用性之间的差距. 安全评估普遍没有.
结论:
- 神经放射学中的代理人工智能在技术上是有前途的,但在临床上尚未得到证实,仍处于早期研究阶段.
- 目前的证据不足以支持临床部署,原因是方法上的局限性和缺乏已证明的临床实用性.
- 在考虑负责任的临床实施之前,严格的多中心前性试验,专注于以患者为中心和安全结果是必不可少的.
更多相关视频
相关概念视频
02:55The Evidence for Evolution
48.2K
Genetic variations accumulating within populations over generations give rise to biological evolution. Evolutionary changes can result in the formation of novel varieties and entire new species. These changes are responsible for the diverse forms of life inhabiting the planet. The evidence for evolution suggests that all living organisms descended from common ancestors.
48.2K
01:10Review and Preview
8.4K
In statistics, several tools are used to interpret the data. Measures of central tendency represent the characteristics of the data, such as mean, median, and mode. Additionally, measures of variance like standard deviation and range are used to find the spread of data from the mean. Relative standing measures the distance between data locations. Commonly used measures of relative standings are percentile, z score, and quartiles.
Percentiles are a type of fractile that partition data into...
Percentiles are a type of fractile that partition data into...
8.4K
01:13Review and Preview
11.3K
Data are individual items of information obtained from a population or sample. Data may be classified as qualitative (categorical), quantitative continuous, or quantitative discrete. Because it is not practical to measure the entire population in a study, researchers use samples to represent the population. A random sample is a representative group from the population chosen by using a method that gives each individual in the population an equal chance of being included in the sample. Random...
11.3K
02:27Limiting Reactant
70.1K
The relative amounts of reactants and products represented in a balanced chemical equation are often referred to as stoichiometric amounts. However, in reality, the reactants are not always present in the stoichiometric amounts indicated by the balanced equation.
70.1K
01:29The Number e as a Limit
87
The number e is a fundamental constant in calculus, playing a central role in describing continuous change, particularly exponential growth. It is most naturally defined through its relationship with the natural logarithm, which is the inverse of the exponential function with base e. This relationship allows e to be characterized using basic principles of differentiation rather than as an arbitrary numerical constant.A key property of the natural logarithm function, ln x, is that its derivative...
87
01:20Random and Systematic Errors
14.9K
Scientists always try their best to record measurements with the utmost accuracy and precision. However, sometimes errors do occur. These errors can be random or systematic. Random errors are observed due to the inconsistency or fluctuation in the measurement process, or variations in the quantity itself that is being measured. Such errors fluctuate from being greater than or less than the true value in repeated measurements. Consider a scientist measuring the length of an earthworm using a...
14.9K