实用医学杂志 ›› 2025, Vol. 41 ›› Issue (14): 2224-2230.doi: 10.3969/j.issn.1006-5725.2025.14.015

• 临床研究 • 上一篇    

慢性非细菌性前列腺炎患者临床特征及风险预测模型的构建和验证

乔玉海,杜春花,赵新鸿,孟晓东,张剑飞()   

  1. 中国人民解放军联勤保障部队第九八〇医院泌尿外科 (河北 石家庄 050000 )
  • 收稿日期:2025-03-27 出版日期:2025-07-25 发布日期:2025-07-29
  • 通讯作者: 张剑飞 E-mail:jianfeixiake@163.com
  • 基金资助:
    河北省医学科学研究项目(20231342)

Construction and validation of a risk prediction model for clinical characteristics of patients with chronic non⁃bacterial prostatitis

Yuhai QIAO,Chunhua DU,Xinhong ZHAO,Xiaodong MENG,Jianfei. ZHANG()   

  1. Department of Urology,the 980th Hospital of the PLA Joint Logistics Support Force,Shijiazhuang 050000,Hebei,China
  • Received:2025-03-27 Online:2025-07-25 Published:2025-07-29
  • Contact: Jianfei. ZHANG E-mail:jianfeixiake@163.com

摘要:

目的 分析慢性非细菌性前列腺炎(CAP)患者临床特点,分析CAP相关因素,构建和验证CAP的风险预测模型。 方法 收集2022年6月至2024年12月医院收治的252例疑似CAP患者临床资料,按7∶3比例将患者分为建模集(n = 177)和验证集(n = 75)。基于建模集数据,采用Lasso筛选CAP相关预测因子,采用logistic多因素模型分析CAP的独立影响因素,并构建风险预测模型。采用验证集患者数据绘制ROC与DCA,对预测模型进行验证。 结果 建模集中86例CAP,占48.59%;验证集中32例CAP,占42.67%。logistic多因素回归分析,结果显示BMI、腰臀比、前列腺液IL-8、COX-2及PGE2异常升高是CAP的独立影响因素(P < 0.05),基于此构建Nomogram列线图。ROC分析显示,该模型判断建模集与验证集患者发生CAP的敏感度分别为0.814和0.802,特异度分别为0.673和0.703。DCA分析显示,列线图用于建模集与验证集的净效益阈值分别为0.1 ~ 0.9和0.2 ~ 1.0。 结论 CAP的发生与患者BMI、腰臀比、前列腺液IL-8、COX-2及PGE2水平相关,据此建立的预测模型准确性高,有助于CAP筛查。

关键词: 慢性非细菌性前列腺炎, 影响因素, 风险预测模型

Abstract:

Objective To investigate the clinical characteristics of patients with chronic abacterial prostafitis (CAP),the CAP related factors were analyzed, and a risk prediction model for CAP were constructed and validated. Methods The clinical dataes of 252 suspected CAP patients admitted to the hospital from June 2022 to December 2024 were collected, the patients were divided into modeling set (n = 177) and validation set (n = 75) by 7∶3 ratio. Based on the modeling set dataes, the Lasso was used to screen CAP related predictive factors, a logistic multiple factor model was used to analyze the independent influence factors of CAP and a risk prediction model was constructed. The validation set patient dataes were used to plot ROC and DCA and validate the prediction model. Results There were 86 cases of CAP in the modeling set, accounting for 48.59%; 32 cases of CAP in the validation set, accounting for 42.67%. The Logistic multiple regression analysis showed that BMI, waist to hip ratio, abnormal elevation of IL-8, COX-2, and PGE2 in prostate fluid were independent influence factors of CAP (P < 0.05), a Nomogram column chart based on this was established.The ROC analysis showed that the sensitivity of the model for detecting CAP in the modeling and validation sets were 0.814 and 0.802, respectively, and the specificity were 0.673 and 0.703, respectively. The DCA analysis showed that the net benefit thresholds for modeling and validation sets by column charts are 0.1 ~ 0.9 and 0.2 ~ 1.0, respectively. Conclusions The occurrence of CAP is related to the patient's BMI, waist to hip ratio, the levels of IL-8, COX-2, and PGE2 in prostate fluid. The predictive model established based on this is highly accurate and it can help for CAP screening.

Key words: chronic abacterial prostafitis, influence factor, risk prediction model

中图分类号: