| 注册
首页|期刊导航|全科护理|基于Logistic回归和决策树构建妇科恶性肿瘤病人术后并发盆腔淋巴囊肿风险预测模型

基于Logistic回归和决策树构建妇科恶性肿瘤病人术后并发盆腔淋巴囊肿风险预测模型

王欢欢 丁传英

全科护理2025,Vol.23Issue(19):3627-3632,6.
全科护理2025,Vol.23Issue(19):3627-3632,6.DOI:10.12104/j.issn.1674-4748.2025.19.005

基于Logistic回归和决策树构建妇科恶性肿瘤病人术后并发盆腔淋巴囊肿风险预测模型

Construction of a risk prediction model for pelvic lymphocele in patients with gynecological malignancies postoperatively based on Logistic regression and decision trees

王欢欢 1丁传英1

作者信息

  • 1. 266035,山东大学齐鲁医院(青岛)
  • 折叠

摘要

Abstract

Objective:To construct a risk prediction model for postoperative pelvic lymphocele(PL)in patients with gynecological malignancies based on logistic regression and decision trees.Methods:A total of 240 patients with gynecological malignancies admitted to our hospital between March 1,2022 and March 31,2024 were enrolled in the study.The patients were divided into a cyst group(n=73)and a control group(n=167)based on the occurrence of postoperative PL.Clinical data were collected from the patients.A decision tree model and a Logistic regression model were constructed for the prediction of postoperative PL in patients with gynecological malignancies,and their predictive performance was compared.Results:A total of 73 patients with gynecological malignancies developed PL among 240 patients,with an incidence of 30.42%.Multivariate analysis showed that tumor stage Ⅲ-Ⅳ[OR=5.752,95%CI(2.527,13.093)],number of lymph nodes≥20[OR=3.812,95%CI(1.804,8.053)],para-aortic lymph node clearance[OR=5.401,95%CI(2.271,12.848)],postoperative chemoradiotherapy[OR=7.053,95%CI(3.269,15.219)],and postoperative serum albumin<34 g/L[OR=3.510,95%CI(1.632,7.548)]were risk factors for postoperative PL in patients with gynecological malignancies(P<0.05).A decision tree model was constructed based on these risk factors,with 4 in model depth,13 in the number of nodes,and 7 in the number of terminal nodes.Postoperative chemoradiotherapy,tumor stage,para-aortic lymphadenectomy,number of lymph nodes cleared,and postoperative serum albumin were selected as explanatory variables.Postoperative chemoradiotherapy was the most important predictor of postoperative PL in patients with gynecological malignancies.The area under curve(AUC)of the Logistic regression model was 0.884,95%CI[0.837,0.930],and the AUC of the decision tree model was 0.846[95%CI(0.792,0.900)].There was no statistically significant difference between the two models(Z=1.369,P=0.091).Conclusion:Tumor stage Ⅲ-Ⅳ,≥20 lymph node dissections,para-aortic lymphadenectomy,postoperative chemoradiotherapy,and postoperative serum albumin<34 g/L are risk factors for postoperative PL in patients with gynecological malignancies.The Logistic regression and decision tree models constructed in this study have high predictive value.

关键词

盆腔淋巴囊肿/妇科恶性肿瘤/决策树/危险因素/淋巴结清扫术

Key words

pelvic lymphocele/gynecological malignancies/decision tree/risk factors/lymph node dissection

引用本文复制引用

王欢欢,丁传英..基于Logistic回归和决策树构建妇科恶性肿瘤病人术后并发盆腔淋巴囊肿风险预测模型[J].全科护理,2025,23(19):3627-3632,6.

全科护理

1674-4748

访问量0
|
下载量0
段落导航相关论文