预测建模用于在食道缩手术后早期检测耐火食道收缩:来自机器学习研究的洞察力
Chuanping Xie1, Dingding Wang1, An Wang1
1Department of Neonatal Surgery, Beijing Childrens Hospital, Capital Medical University, National Center for Children's Health, No. 56 Nalishi Road, Xicheng District, Beijing, 100045, China.
识别下缩 (LEA) 和解漏等危险因素可以帮助预测经过缩 (EA) 修复后的耐火性食道收缩 (RES). 机器学习模型,特别是随机森林,在预测可再生能源发展方面表现出很高的准确性.
科学领域:
- 儿科手术 儿科手术
- 胃肠病学 胃肠病学
- 医疗信息学 医疗信息学
背景情况:
- 难治性食道收缩 (RES) 是食道缩 (EA) 修复后的一个重大并发症.
- 在临床决策中,早期识别有RES风险的患者至关重要.
- 关于在EA修复后预测RES的研究有限.
研究的目的:
- 确定在EA修复后开发可再生能源的风险因素.
- 使用机器学习 (ML) 构建一个可再生能源的预测模型.
主要方法:
- 对92名接受内镜气球扩张 (EBD) 治疗的患者进行了回顾性审查,以治疗食道狭窄 (ES) 后EA修复.
- 分析基线特征,早期术后并发症和ES特征.
- 开发和验证ML模型 (决策树,随机森林,多层感知器).
主要成果:
- 下部胃肠道内镜 (LGEA),静脉漏,下部收缩水平,异常食道狭窄症和较高的上袋收缩指数 (U-SI) 是 RES 的显著风险因素.
- 随机森林模型显示了最高的预测准确性 (AUC=0.756).
- SHAP分析确定了U-SI,rTEF,狭窄水平和道泄漏作为关键预测特征.
结论:
- LGEA,动脉漏,较低的收缩水平,偏心收缩和较高的U-SI是RES后EA修复的危险因素.
- 基于ML的预测模型可以有效和准确地预测可再生能源的发生.
- 一个基于射频的临床决策支持系统被开发用于动态 RES 预测.
更多相关视频
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
05:57ADSC-sheet Transplantation to Prevent Stricture after Extended Esophageal Endoscopic Submucosal Dissection
Published on: February 10, 2017
相关概念视频
Esophageal Strictures-II: Clinical Features and Management
Healthcare providers should gather a comprehensive medical history and conduct a physical examination for diagnosis. If esophageal stricture is...
Esophageal Strictures-I: Introduction
Etiology
The primary cause of esophageal strictures is long-standing gastroesophageal reflux disease (GERD), accounting for about 70 to 80% of adult cases. Chronic acid reflux can lead to injury and scarring of the esophageal lining, culminating in...
Barrett Esophagus-I: Introduction
This constant acid exposure transforms the esophagus's pink mucosal lining (stratified squamous epithelium) into a type of lining more...
