完全植入式静脉输液港植入后感染风险的列线图预测模型构建与验证:一项回顾性研究

Development and validation of a Nomogram model for predicting infection risk after totally implantable venous access ports implantation:a retrospective study

目的:构建并验证列线图预测模型,用于预测完全植入式静脉输液港(TIVAP)植入后的感染风险。方法:对2017年1月至2022年8月,在广州市第一人民医院接受TIVAP植入手术的794例患者进行回顾性分析。通过随机抽样将研究对象按7∶3分为训练集(555例)和验证集(239例)。首先应用LASSO回归分析筛选出最佳预测因素,并将其纳入多因素Logistic回归分析以构建列线图模型。利用受试者工作特征(ROC)曲线、校准曲线和临床决策曲线(DCA)评估预测模型的区分能力、准确性和临床有效性。结果:多因素Logistic回归分析确认年龄、手术时间、血液系统肿瘤和BMI为TIVAP植入后感染的独立影响因素。基于这些因素构建列线图预测模型,ROC曲线分析显示训练集和验证集的AUC值分别为0.948(95%CI:0.914~0.982)和0.935(95%CI:0.876~0.994),表明模型具有良好的区分能力和预测准确性。结论:本研究开发的TIVAP植入后感染风险预测模型在临床上具有实用价值,可以帮助临床评估患者的感染风险,从而制定更加个性化的治疗计划。

Objective:To develop and validate a Nomogram predictive model for estimating the risk of infection following the implantation of a totally implantable venous access ports(TIVAP).Methods:A retrospective analysis was conducted on 794 patients who underwent TIVAP implantation surgery at Guangzhou First People's Hospital from January 2017 to August 2022.Subjects were randomly divided into a training set(555 cases) and a validation set(239 cases) in a 7∶3 ratio.LASSO regression analysis was initially applied to identify the optimal predictive factors,which were then incorporated into a multifactorial Logistic regression to develop the Nomogram model.The predictive model's discriminative ability,accuracy,and clinical utility were assessed using the receiver operating characteristic(ROC) curve,calibration curve,and decision curve analysis(DCA).Results:Multifactorial Logistic regression analysis identified age,operation duration,hematological malignancies,and BMI as independent factors affecting infection after TIVAP implantation.Based on these factors,a Nomogram predictive model was constructed.ROC curve analysis revealed AUC values of 0.948(95%CI:0.914-0.982) and 0.935(95%CI:0.876-0.994) for the training and validation sets,respectively,indicating good discriminative ability and predictive accuracy of the model.Conclusion:The TIVAP post-implantation infection risk predictive model developed in this study has practical clinical value.It can assist clinicians in assessing patient infection risks,thus facilitating the formulation of more personalized treatment plans.