预测缺偿性肝硬化患者的门口压力梯度:一种非侵入性的深度学习模型
Zi-Wen Liu1, Tao Song2, Zhong-Hua Wang1
1Department of Gastroenterology, Shandong Provincial Hospital Affiliated to Shandong First Medical University, 324, Jing 5 Rd, Ji'nan, Shandong Province, China.
Digestive diseases and sciences
|October 28, 2024
概括
一个新的深度学习模型准确地预测了肝硬化患者的门口压力梯度 (PPG). 这种非侵入性工具可以识别高风险的门性高血压,帮助做出早期干预决策.
科学领域:
- 肝病学 肝病学是一种肝病学.
- 医疗成像医学成像
- 人工智能在医学中的应用
背景情况:
- 高门压梯度 (PPG) 与非补偿性肝硬化症的严重并发症相关.
- 目前的PPG测量是侵入性的,阻碍了常规的临床应用.
- 非侵入性PPG评估对于患者管理至关重要.
研究的目的:
- 开发和验证一个深度学习模型来预测非补偿性肝硬化中的PPG.
- 确定患有高风险门性高血压 (HRPH) 的患者,以便进行潜在的早期干预.
- 促进及时考虑跨内肝移植系统分流 (TIPS).
主要方法:
- 对428名接受TIPS治疗的失补偿性肝硬化患者进行了回顾性分析.
- 使用实验室和成像数据开发人工神经网络模型.
- 通过递归特征消除和外部验证进行特征选择.
主要成果:
- 一个三参数 (3P) 模型 (INR,门静脉直径,WBC) 实现了87.5%的准确性.
- 外部验证在单独的数据集上显示了高准确度 (85.40%和90.80%).
- 该模型有效地识别了HRPH,在外部验证中AUROC为0.842.
结论:
- 开发的3P深度学习模型准确地预测了非补偿性肝硬化中的PPG.
- 该模型可靠地区分高风险门高血压,指导临床决策.
- 这种非侵入性方法为治疗肝硬化患者提供了一个有前途的工具.
相关概念视频
Cirrhosis II: Pathophysiology
Cirrhosis is a progressive chronic liver injury caused by prolonged inflammation, excessive fibrotic remodeling, and impaired regeneration. Over time, repeated hepatic insults disrupt the liver’s architecture and function, leading to reduced blood flow, impaired bile drainage, and diminished metabolic capacity.Pathophysiology of cirrhosisCirrhosis arises from three main responses to chronic liver damage: inflammation, immune activation, and hepatocyte death. These processes lead to structural...
Portal Hypertension
Portal hypertension is an increase in blood pressure within the portal venous system. Normally, this pressure is less than 5 mmHg. It is considered clinically significant when it rises above 10 mmHg. At this threshold, complications from altered blood flow and venous congestion emerge.EtiologyPortal hypertension arises from conditions that impede blood flow through the liver. The most common cause is cirrhosis, in which chronic liver injury leads to fibrotic scarring. This fibrosis narrows or...


