基于超声指标预测产后压力性尿失禁列线图模型建立与验证

Establishment and validation of a Nomogram model for predicting postpartum stress urinary incontinence based on ultrasound indicators

目的:建立基于超声指标预测产后压力性尿失禁(SUI)列线图模型并进行验证。方法:选择2021年1月至2023年1月在达州市中西医结合医院分娩并于产后3个月行盆底超声检查的产妇413例进行研究,以7∶3比例分为模型组289例与验证组124例;收集患者一般资料,对患者进行二维、三维盆底超声检查及耻骨直肠肌实时剪切波弹性成像检查;根据有无产后SUI将模型组产妇分为2组,比较2组一般资料及超声指标,并以LASSO回归筛选潜在因素后用logistic回归评估出独立性影响因素,采用R语言建立列线图模型并进行验证。结果:模型组289例患者中共有48例(16.61%),SUI组与无SUI组患者孕前体质量指数、产次、分娩方式、会阴裂伤、便秘情况、新生儿出生体质量、Ry、Vy、Rα、Vα、Vβ、BND、URA、膀胱颈漏斗、LHA1、LHA2、E1、E2及△E差异均具有统计学意义(P<0.05);在LASSO回归基础上行多因素logistic回归分析结果显示:孕前体质量指数、Ry、Vy、BND、URA、膀胱颈漏斗、LHA1及△E为产后SUI的独立性影响因素;列线图模型ROC分析结果显示,模型组列线图模型预测产后SUI发生的AUC为0.951(95%CI:0.924~0.977);验证组AUC为0.936(95%CI;0.907~0.964)。采用Bootstrap法重复抽样1000次,并以验证组进行验证,校准曲线结果显示:模型曲线与理想模型基本拟合成对角线。模型组决策曲线分析结果显示:预测概率阈值0.10~0.90时使用本研究模型预测产后SUI的净获益最高。结论:产后SUI的发生受体质量指数、Ry、Vy等因素的影响,该列线图模型有望用于预测产后SUI。

Objective:To establish a Nomogram model for predicting postpartum stress urinary incontinence(SUI)based on ultrasound indicators and validate it.Methods:A total of 413 women who gave birth in Dazhou Integrated TCM&Western Medicine Hospital and underwent pelvic floor ultrasonography 3 months after deliv-ery from January 2021 to January 2023 were selected for the study and divided into model group with 289 cases and validation group with 124 cases in a 7:3 ratio.The general data of the patients were collected,and two-dimen-sional and three-dimensional pelvic floor ultrasound examination and real-time shear wave elastography of the pu-borectalis muscles were performed.According to the presence or absence of postpartum SUI,the model group of parturients was divided into two groups.The general data and ultrasound indicators of the two groups were com-pared,and potential factors were screened using LASSO regression.Independent influencing factors were evalu-ated using logistic regression,and the Nomogram model was established using R language for validation.Re-sults:In the model group of 289 patients,there were a total of 48 cases(16.61%).There were significant differ-ences in pre-pregnancy BMI,parity,delivery mode,perineal laceration,constipation,newborn birth weight,Ry,Vy,Rα,Vα,Vβ,BND,URA,bladder neck funnel,LHA1,LHA2,E1,E2 and△E between the SUI group and the non-SUI group(P<0.05).The results of multiple lo-gistic regression analysis based on LASSO regression showed that pre-pregnancy BMI,Ry,Vy,BND,URA,bladder neck funnel,LHA1 and△E were independent influencing factors of postpartum SUI.The ROC analysis results of the Nomogram model showed that the AUC of the Nomogram model in the model group predicting the occurrence of postpartum SUI was 0.951(95%CI:0.924-0.977).The AUC of the verification group was 0.936(95%CI:0.907-0.964).The Bootstrap method was used for repeated sampling for 1,000 times,and the verification group was used for verification.The calibra-tion curve results showed that the model curve was basically fitted to the diagonal line of the ideal model.The re-sults of decision curve analysis in the model group showed that when the prediction probability threshold was 0.10-0.90,the net benefit of the model in predicting postpartum SUI was the highest.Conclusion:The occur-rence of postpartum SUI is affected by BMI,Ry,Vy and other factors.The Nomogram model is expected to be used to predict postpartum SUI.