Development and validation a predictive model for distinguishing malignant pleural effusion
10.13602/j.cnki.jcls.2025.09.12
- VernacularTitle:构建与验证用于区分恶性胸腔积液的预测模型
- Author:
Jinling JI
1
;
Qiong WANG
1
;
Ting SHI
1
;
Yuzhang JIANG
1
;
Chang LI
1
Author Information
1. 南京医科大学附属淮安第一医院检验科,江苏淮安 223300
- Publication Type:Journal Article
- Keywords:
pleural effusion;
predictive model;
diagnosis;
metastasis;
parameter
- From:
Chinese Journal of Clinical Laboratory Science
2025;43(9):702-709
- CountryChina
- Language:Chinese
-
Abstract:
Objective To development and validate a predictive model for distinguishing between malignant pleural effusion(MPE)and benign pleural effusion(BPE).Methods A total of 428 patients diagnosed with pleural effusion(PE)and hospitalized at the First Hospital of Huai'an Affiliated to Nanjing Medical University from July 2020 to May 2022 were selected.The patients were divided into BPE group(211 cases)and MPE group(217 cases)according to diagnostic criteria.The basic information and clinical data of these patients were collected.Boruta method was used for univariate screening,followed by multivariate Logistic regression to construct a basic nomogram model.Bootstrap method was used for internal validation to evaluate the performance of the nomogram,including dis-crimination,accuracy,and clinical applicability.Results The model included 8 key variables:dyspnea,chest pain,general symp-toms,X-ray/CT with malignant tumor features,serum carcinoembryonic antigen,serum neuron-specific enolase,pleural lactate dehy-drogenase,and pleural carcinoembryonic antigen.Internal validation showed that the area under the receiver operating characteristic curve(AUCROC)of the model was 0.933(95%confidence interval:0.912-0.954),with good accuracy(P＞0.05).Decision curve a-nalysis(DCA)indicated that this predictive model for predicting MPE risk had a significant net benefit when the probability threshold exceeded 1%.Conclusion The constructed prediction model could effectively distinguish between MPE and BPE.
