相关实验视频
Updated: Jul 5, 2026
09:24
Transurethral Induction of Mouse Urinary Tract Infection
Published on: August 5, 2010
36.6K
用可解释的AI提高尿路感染的预测准确性:一套TabNet方法
Hongyang Wang1,2, Junpeng Ding3, Shuochen Wang4
1Department of Urology, Capital Institute of Pediatrics, Beijing, China.
Scientific reports
|January 19, 2025
概括
一个新的整体模型准确地预测了小儿皮埃洛塑性手术后的尿路感染 (UTI),改善了手术结果. 这种机器学习和深度学习方法有助于预防感染和再阻塞,减少医疗保健负担.
科学领域:
- 儿科泌尿外科 儿科泌尿外科
- 机器学习在医学中的应用
- 预测分析是一种预测分析.
背景情况:
- 尿道皮层结口阻塞 (UPJO) 是一种常见的儿科疾病,可以用皮叶整形术治疗.
- 手术后的泌尿道感染 (UTI) 影响了超过30%的儿科热塑患者,增加了发病率和医疗费用.
- 目前的尿路感染预测方法有限,需要先进的多因素模型.
研究的目的:
- 开发和评估一个强大的,多因素的预测模型,用于小儿皮叶造形术后的术后尿路感染.
- 将传统机器学习算法的性能与用于UTI预测的深度学习模型进行比较.
- 引入集体学习模型,整合机器学习和深度学习,以提高预测准确度.
主要方法:
- 对764名接受热塑性手术的儿科患者的回顾性分析.
- 提取和分析25个临床特征,包括人口统计,病史和手术细节.
- 对物流回归,SVM,随机森林,XGBoost,LightGBM和TabNet模型进行比较评估,然后开发一个使用SHAP可视化的集合元学习者模型.
主要成果:
- 结合LightGBM和TabNet的整体模型实现了最高的预测精度 (精度:0.80,AUC:0.80),优于单个模型.
- 在功能工程之前,深度学习模型TabNet在传统机器学习算法上表现出优异的性能.
- SHAP分析确定了eGFR和ALB作为皮叶造形术后尿路感染的显著预测因素.
结论:
- 开发的整体模型是第一个整合机器学习和深度学习的模型,用于预测儿科皮叶塑术后的尿路感染.
- 这种方法减少了对特征工程的依赖,并减轻了深度学习模型的过度拟合,特别是在有限的医疗数据的情况下.
- 该模型支持主动干预,可能减少术后感染,再阻塞率和相关的医疗负担.
相关概念视频
01:26Urinary Tract Infection I: Introduction
Urinary tract infections (UTIs) impact various parts of the urinary system, including the kidneys, ureters, bladder, and urethra. These infections are generally bacterial, with Escherichia coli being the most common causative agent, often originating from the gastrointestinal tract. However, other bacteria, such as Staphylococcus saprophyticus, Klebsiella pneumoniae, and Proteus mirabilis, are also known to cause UTIs. The type, location, and underlying complexity of the UTI guide both...
01:30Urinary Tract Infection III: Diagnostic Studies and Interprofessional Care
A healthcare provider can diagnose a urinary tract infection (UTI) through several methods:Medical History and Symptoms: The provider will take a detailed medical history and ask about symptoms such as frequent urination, burning sensation during urination, and lower abdominal pain.Urinalysis: A clean-catch urine sample is collected in a sterile container and tested for the presence of bacteria, white blood cells (leukocytes), nitrites, blood, and protein. The presence of leukocytes and...
01:27Acute Pyelonephritis I: Introduction
Pyelonephritis is a bacterial infection that primarily affects the renal parenchyma and collecting system, including the renal pelvis, tubules, and interstitial tissue of one or both kidneys. It can be classified as either acute—a sudden, severe infection—or chronic, which refers to long-term or recurrent kidney infections.The primary cause of acute pyelonephritis (APN) is bacterial infection, with Escherichia coli accounting for approximately 70-80% of cases. Other bacteria, such as Proteus,...