中国血液净化

中国血液净化 ›› 2026, Vol. 25 ›› Issue (05): 415-419.doi: 0.3969/j.issn.1671-4091.2026.05.011

• 综述 • 上一篇    下一篇

人工智能在急性肾损伤预测模型中的研究进展

陈鹏伟   李月红   

  1. 102218 北京，1清华大学北京清华长庚医院肾内科
  • 收稿日期:2025-09-22 修回日期:2025-11-21 出版日期:2026-05-12 发布日期:2026-05-12
  • 通讯作者: 李月红 E-mail:lyha01051@btch.edu.cn

Research progress in artificial intelligence in acute kidney injury prediction models

CHEN Peng-wei, LI Yue-hong   

  1. Department of Nephrology, Beijing Tsinghua Changgung Hospital, School of Clinical Medicine, Tsinghua Medicine, Tsinghua University, Beijing 102218, China
  • Received:2025-09-22 Revised:2025-11-21 Online:2026-05-12 Published:2026-05-12
  • Contact: 102218 北京，1清华大学北京清华长庚医院肾内科 E-mail:lyha01051@btch.edu.cn

摘要： 急性肾损伤(acute kidney injury，AKI)是临床常见的急危重症，病因复杂多样、进展迅速，显著增加患者短期及长期不良结局，包括死亡、慢性肾脏病及心血管事件等。随着电子健康记录和大规模临床数据库的发展，人工智能(artificial intelligence，AI)在AKI领域展现出重要潜力，能够整合人口学、实验室指标和生命体征等多维度信息，实现AKI预警及早期识别。本文综述了AI在AKI预测模型的研究进展，包括不同算法类型、关键预测因子、模型验证及临床应用挑战，旨在为未来AKI风险预测研究及临床转化提供参考。

关键词: 急性肾损伤, 人工智能, 风险预测模型

Abstract: Acute kidney injury (AKI) is a common and severe syndrome with poor prognosis. Artificial intelligence (AI) is a promising tool for early AKI prediction. This review summarizes recent advances in AI-based AKI prediction models including different modeling approaches, key predictors, validation strategies and clinical challenges, aiming to provide references for AKI risk prediction and translational medicine.

Key words: Acute kidney injury, Artificial intelligence, Risk prediction model

中图分类号: