国际妇产科学杂志 ›› 2024, Vol. 51 ›› Issue (2): 189-193.doi: 10.12280/gjfckx.20230951

• 产科生理及产科疾病:论著 • 上一篇    下一篇

胎儿胸腔积液影响因素的决策树分析

董越, 张慧芳, 郭晓珮, 李珠银, 姬若楠, 罗晓华()   

  1. 450052 郑州大学第三附属医院产科
  • 收稿日期:2023-11-25 出版日期:2024-04-15 发布日期:2024-04-19
  • 通讯作者: 罗晓华,E-mail:luoxiaohua620@163.com
  • 基金资助:
    河南省医学科技公关计划联合共建项目(LHGJ20210466)

Decision Tree Analysis of Influencing Factors of Fetal Pleural Effusion

DONG Yue, ZHANG Hui-fang, GUO Xiao-pei, LI Zhu-yin, JI Ruo-nan, LUO Xiao-hua()   

  1. Department of Obstetrics, The Third Affiliated Hospital of Zhengzhou University, Zhengzhou 450052, China
  • Received:2023-11-25 Published:2024-04-15 Online:2024-04-19
  • Contact: LUO Xiao-hua, E-mail: luoxiaohua620@163.com

摘要:

目的: 采用决策树模型分析胎儿胸腔积液预后的影响因素。方法: 选取郑州大学第三附属医院2016年1月—2023年1月收治的胎儿胸腔积液患者148例为研究对象,按胎儿有无水肿分为胎儿水肿组(83例)和胎儿无水肿组(65例)2组,采用χ2检验或t检验进行单因素分析,采用二分类Logistic回归分析胎儿胸腔积液病例发生胎儿水肿的影响因素,并进行决策树算法(classification regression tree,CRT)分析,生成胎儿胸腔积液的临床诊断决策树,对胎儿胸腔积液预后的影响因素进行探讨。结果: 单因素分析结果显示,胎儿水肿组与胎儿无水肿组的首诊孕周(χ2=2.044,P=0.043)、胸腔积液位置(χ2=5.628,P=0.018)、积液严重程度(χ2=12.955,P=0.001)、胸腔积液转归(χ2=11.704,P=0.001)比较,差异均有统计学意义。Logistic回归分析显示,首诊孕周越晚(OR=0.909,95%CI:0.844~0.978,P=0.011)为胎儿水肿的保护因素,重度胸腔积液(OR=6.045,95%CI:1.980~18.451,P=0.002)为胎儿水肿的危险因素。决策树分析结果显示,首诊孕周、胸腔积液的严重程度和胸腔积液位置是胎儿胸腔积液导致水肿的影响因素,首诊孕周与胎儿胸腔积液导致水肿的关联性最强。结论: 胎儿胸腔积液常同时合并有胎儿水肿,当首诊孕周越早、胎儿胸腔积液越严重、胎儿双侧胸腔积液,就越容易导致胎儿发生水肿,胎儿的预后就越差。

关键词: 胎儿, 胸腔积液, 决策树, 胎儿水肿, 胎儿预后

Abstract:

Objective: To analyze the influencing factors of fetal pleural effusion complicated with edema by using decision tree model. Methods: The patients with fetal pleural effusion admitted to the Third Affiliated Hospital of Zhengzhou University from January 2016 to January 2023 were selected for the study. The 148 patients were divided into the fetal hydrops group (83 cases) and the fetal no hydrops group (65 cases) according to whether the fetus was edema or not. The χ2 test or t test was used for single factor test, and the factors with statistical difference between the two groups were screened out and incorporated into the decision tree model (classification regression tree, CRT). The factors influencing the prognosis of fetal pleural effusion were explored. Results: The results of univariate analysis showed that in the single-factor analysis, these factors for the first diagnosis of pregnancy week ( χ2=2.044, P=0.043), the location of fetal pleural effusion ( χ2=5.628, P=0.018), the severity of fetal pleural effusion ( χ2=12.955, P=0.001), the return of fetal pleural effusion during pregnancy ( χ2=11.704, P=0.001) between the two groups had statistically difference. The results of Logistic regression analysis suggested that late gestational week at first diagnosis (OR=0.909, 95%CI: 0.844-0.978, P=0.011) was a protective factor for hydrops fetalis, and severe pleural effusion (OR=6.045, 95%CI: 1.980-18.451, P=0.002) was a risk factor for hydrops fetalis. Decision tree analysis showed that the earlier the first gestational age, the severity of pleural effusion, and bilateral pleural effusion were the influencing factors for edema caused by fetal pleural effusion, and the strongest association was between the first gestational age and fetal pleural effusion leading to edema. Conclusions: Fetal pleural effusion is often accompanied with fetal edema. The earlier the first gestation week, the more serious the pleural effusion, and the bilateral pleural effusion of the fetus, the more likely it is to cause fetal edema and the worse the prognosis of the fetus.

Key words: Fetus, Pleural effusion, Decision trees, Hydrops fetalis, Fetal prognosis