开发和验证一种可解释的机器学习模型,用于预测柔性尿路透镜光三后的败血症风险
Ruichen Li1, Biao Zhang2, Liying Zeng2
1Graduate Collaborative Training Base of Yiyang, Hengyang Medical School, University of South China, Hengyang Hunan, 421001, China.
Urolithiasis
|September 18, 2025
概括
一个新的机器学习模型准确地预测了柔性尿路透镜光三 (fURL) 后的败血症风险. 这种工具有助于临床医生识别高风险患者,改善结石治疗的结果.
科学领域:
- 泌尿器科 泌尿器科 泌尿器科 泌尿器科
- 腎臟病學 (nephrology) 是一種醫學專業.
- 医疗信息学 医疗信息学
背景情况:
- 败血症是一个严重的风险后,灵活的尿路透镜石 (fURL),一个常见的结石治疗.
- 精确预测败血症风险对于及时干预和改善患者结果至关重要.
研究的目的:
- 开发和验证一种机器学习 (ML) 模型,用于预测fURL后的败血症风险.
- 为了提高模型的可解释性,使用Shapley增量解释 (SHAP).
主要方法:
- 使用来自1386名 (衍生) 和604名 (外部验证) 患者的数据,对ML模型的回顾性开发和验证.
- 败血症诊断基于败血症-3.0指导方针.
- 评估了15个ML算法,通过SHAP分析选择了额外树木 (ET) 模型,以其性能和可解释性.
主要成果:
- 额外树木 (ET) 模型在培训组中实现了0.90的AUC,在内部验证中为0.87,在外部验证中为0.81.
- 该ET模型确定了八个关键特征,这些特征有助于败血症预测.
- SHAP分析为增强模型解释性提供了重要的特征.
结论:
- 成功开发了一种可解释的额外树木 (ET) ML模型,以预测fURL后的败血症风险.
- 该模型显示出高准确性和作为患者风险分层的临床工具的潜力.
- 作为Web应用程序的部署可以促进临床采用,以改善患者护理.
相关概念视频
01:25Urinary Tract Calculi VI: Surgical Management
384
Procedures for Kidney StonesMedical intervention is necessary when kidney stones or renal calculi are too large to pass spontaneously (typically greater than 5 millimeters) when stones are accompanied by symptomatic infection (such as fever or pyelonephritis), when they impair kidney function, or when they cause persistent symptoms like severe pain, nausea, or urinary retention. Additionally, patients with only one kidney or those who cannot be treated with medical management also require...
384
01:30Urinary Tract Calculi III: Medical Management
196
The diagnosis of renal calculi involves several imaging techniques, including non-contrast CT scans and ultrasound. These methods help visualize kidney stones, assess their size and location, and detect possible obstructions. Additionally, Measuring urine pH is useful for diagnosing specific stone types, such as struvite (alkaline pH) and uric acid stones (acidic pH). Cystine stones are primarily linked to cystinuria, a genetic condition. A urinalysis helps detect blood in the urine (hematuria)...
196