心脏门手术后ARDS的机器学习预测:在中国西北地区的开发和验证
Xuhua Li1, Hao Chen1, Aoxiang Chen2
1The First Clinical Medical College of Lanzhou University, Lanzhou, China.
Frontiers in cardiovascular medicine
|February 6, 2026
概括
使用XGBoost的AI模型可以预测心脏门置换手术后的急性呼吸困扰综合征 (ARDS). 这种工具有助于早期识别心肺旁路术后患有ARDS高风险的患者.
科学领域:
- 心脏病学 心脏病学
- 人工智能的人工智能
- 临界护理医学 临界护理医学
背景情况:
- 急性呼吸困扰综合征 (ARDS) 是心肺旁路 (CPB) 辅助心脏门置换 (HVR) 后的一个显著并发症.
- 早期识别高风险ARDS患者对于及时干预和改善结果至关重要.
研究的目的:
- 开发和验证基于人工智能的ARDS后CPB辅助HVR预测模型.
- 为了能够早期识别和风险分层的患者易受ARDS.
主要方法:
- 对400名接受CPB辅助HVR的患者进行了回顾性分析.
- 机器学习模型的开发和优化,XGBoost被选为最佳算法.
- 数据集分为培训 (n=280) 和测试 (n=120) 数据集用于模型评估.
主要成果:
- 56名 (14%) 患者在术后出现了ARDS.
- XGBoost模型表现出高预测性能,AUC为0.853和良好的校准.
- 发现的关键预测因素包括年龄,单细胞计数,右心房直径,血液损失,血小板计数和主要肺动脉直径.
结论:
- 开发的XGBoost模型准确地预测了CPB辅助HVR后的ARDS风险.
- 该模型利用了六个临床相关的预测因子来实现有效的风险分层.
- 这种人工智能工具为高风险患者的早期干预提供了有价值的方法.
更多相关视频
07:13Comparison of Predictive Performance of Three Lymph Node Staging Systems in Colorectal Signet Ring Cell Carcinoma Based on Machine Learning Model
Published on: April 18, 2025
744
04:09Predicting Treatment Response to Image-Guided Therapies Using Machine Learning: An Example for Trans-Arterial Treatment of Hepatocellular Carcinoma
Published on: October 10, 2018
8.8K
相关概念视频
Heart Valves
12.2K
The human heart is a complex organ with an intricate system of valves that regulate blood flow. There are two main types of valves: atrioventricular (AV) valves and semilunar valves.
The AV valves prevent the backflow of blood from the ventricles to the atria during ventricular contraction. These valves function with the assistance of the chordae tendineae and papillary muscles. When the ventricles are relaxed, the chordae tendineae are slack, allowing blood to flow from the atria into the...
The AV valves prevent the backflow of blood from the ventricles to the atria during ventricular contraction. These valves function with the assistance of the chordae tendineae and papillary muscles. When the ventricles are relaxed, the chordae tendineae are slack, allowing blood to flow from the atria into the...
12.2K
Development of the Heart
2.5K
The development of the human heart, a crucial organ, commences from the mesoderm on the 18th or 19th day after fertilization. This process initiates in the cardiogenic area, a group of mesodermal cells at the embryo's head end, which evolves into elongated strands known as cardiogenic cords. These cords undergo a transformation to form hollow-centered endocardial tubes.
As the embryo undergoes lateral folding, these paired tubes approach each other, merging into a single primitive heart...
As the embryo undergoes lateral folding, these paired tubes approach each other, merging into a single primitive heart...
2.5K
Reliability and Validity
14.1K
Reliability and validity are two important considerations that must be made with any type of data collection. Reliability refers to the ability to consistently produce a given result. In the context of psychological research, this would mean that any instruments or tools used to collect data do so in consistent, reproducible ways.
14.1K
Anatomy of the Heart
120.0K
The human heart is made up of three layers of tissue that are surrounded by the pericardium, a membrane that protects and confines the heart. The outermost layer, closest to the pericardium, is the epicardium. The pericardial cavity separates the pericardium from the epicardium. Beneath the epicardium is the myocardium, the middle layer, and the endocardium, the innermost layer. There are four chambers of the heart: the right atrium, the right ventricle, the left atrium, and the left ventricle.
120.0K
In Vitro Drug Release Testing: Overview, Development and Validation
357
In vitro dissolution and drug release tests assess how quickly and how much of a drug is released from its dosage form into an aqueous medium under standardized laboratory conditions. These tests are essential tools in pharmaceutical development and quality assurance, offering insight into the drug's performance before clinical use.During formulation development, dissolution testing identifies incomplete or inconsistent drug release issues. It also supports decisions on selecting the optimal...
357
Predicting Molecular Geometry
46.0K
VSEPR Theory for Determination of Electron Pair Geometries
46.0K
