开发和验证一个可解释的预测模型,用于儿科慢性鼻炎的术后复发
Sijie Jiang1,2,3,4, Bo Qi5, Shaobing Xie1,2,3,4
1Department of Otolaryngology-Head and Neck Surgery, Xiangya Hospital of Central South University, Changsha, People's Republic of China.
概括
这项研究开发了一种可解释的机器学习模型,用于预测儿科慢性鼻炎 (CRS) 的术后复发. 该模型使用关键炎症标志物准确识别有风险的患者,帮助临床决策.
科学领域:
- 耳鼻喉科 耳鼻喉科 耳鼻喉科
- 医疗信息学 医疗信息学
- 机器学习在医学中的应用
背景情况:
- 儿童慢性鼻炎 (CRS) 在手术后复发是一个重大的临床挑战.
- 对术后复发的准确预测对于优化患者管理和结果至关重要.
研究的目的:
- 开发和验证可解释的机器学习 (ML) 模型,用于预测儿科CRS的术后复发.
- 确定与CRS复发相关的关键临床和炎症因素.
主要方法:
- 对148名接受功能性内镜外科手术的儿科CRS患者的回顾性分析.
- 训练和比较三个ML算法,通过AUC评估性能.
- 使用莎普利的附加解释和Ceteris Paribus来解释模型的可解释性.
主要成果:
- 随机森林 (RF) 模型实现了最高的区分能力 (AUC=0.728).
- 一个具有四个特征 (SII,PIV,E%,L%) 的优化射频模型显示出强大的预测性能 (AUC=0.779).
- 模型解释揭示了非线性关系,并突出了L%和E%的贡献.
结论:
- 一个可解释的ML模型在预测儿科CRS复发方面显示出前景.
- 关键的炎症标志物 (SII,PIV,L%,E%) 是重要的预测因素.
- 开发的模型可以通过临床实用性的Web应用程序访问.
相关概念视频
01:30Epistaxis
127
Epistaxis, or nosebleeds, occurs when small, swollen blood vessels in the nasal mucous membrane rupture. Typically, the anterior septum is the primary site of occurrence.
Etiology
Possible causes of this condition include high blood pressure, trauma, low humidity, upper respiratory tract infections, allergies, foreign bodies, nasal inhalation of corticosteroids or illicit drugs, excessive use of decongestant nasal sprays, facial or nasal surgery, anatomic malformation, tumors, or systemic...
Etiology
Possible causes of this condition include high blood pressure, trauma, low humidity, upper respiratory tract infections, allergies, foreign bodies, nasal inhalation of corticosteroids or illicit drugs, excessive use of decongestant nasal sprays, facial or nasal surgery, anatomic malformation, tumors, or systemic...
127
01:29Suctioning the Nasopharyngeal Airway
275
Nasopharyngeal suctioning is a procedure to remove secretions from the upper part of the respiratory tract that the patient cannot clear independently. It helps maintain airway patency and prevents complications such as aspiration pneumonia.
Equipment Required
Equipment Required
275