基于二维、三维超声影像学及临床因素构建模型预测不孕症介入再通术后粘连研究

Study on the construction of a model to predict adhesions after interventional recanalisation for infertility based on 2D and 3D ultrasound imaging and clinical factors

目的探讨基于二维、三维超声影像学及临床因素构建模型对不孕症介入再通术后粘连预测价值,从而确定合理诊治方式,有利于提高临床妊娠率。方法选取我院不孕症介入再通术后患者335例,按照7∶3比例分为建模人群(n=235)和验证人群(n=100),统计术后3个月输卵管粘连情况、临床因素、二维及三维超声影像学表现,Logistic回归方程分析不孕症介入再通术后粘连影响因素,构建列线图模型,并进行内外验证。结果粘连组和非粘连组盆腔炎症、输卵管堵塞位置、宫腔操作史、流产史、二维及三维超声影像学表现比较差异有统计学意义(P<0.05);盆腔炎症、输卵管堵塞位置、宫腔操作史、卵巢环状强回声、输卵管走形扭曲、盆腔均匀弥散、卵巢包绕是不孕症再通介入术后粘连影响因素(P<0.05);构建列线图预测模型,该模型在建模人群和验证人群中AUC分别为0.929、0.919,DCA曲线显示在范围0.2~0.9、0.0~0.85内该模型在建模人群和验证人群中净获益值较好。结论基于二维、三维超声影像学及临床因素构建模型可用于不孕症介入再通术后粘连预测中,临床可结合相关因素针对性展开后续治疗,以降低术后粘连风险。

Objective To investigate the predictive value of constructing models based on two-dimensional and three-dimensional ultrasound imaging and clinical factors for adhesions after interventional recanalization for infertility,so as to determine reasonable diagnostic and treatment modalities,thereby improving clinical pregnancy rates.Methods A total of 335 patients after interventional recanalisation for infertility in our hospital were selected and divided into a modeling population(n=235)and a validation population(n=100)according to the ratio of 7∶3.Tubal adhesions,clinical factors,2D and 3D ultrasound imaging performance at 3 months after surgery were recorded,and logistic regression equations were used to analyse the influencing factors of adhesions after interventional recanalisation for infertility,a nomogram model was constructed,and internal and external validation was conducted.Results There were significant differences in the comparison of pelvic inflammation,location of tubal occlusion,history of uterine manipulation,history of miscarriage,and 2D and 3D ultrasound imaging performance between the adhesion and non-adhesion groups(P<0.05).Pelvic inflammation,location of tubal occlusion,history of uterine manipulation,ovarian annular strong echogenicity,tubal alignment distortion,pelvic homogeneous diffusion,and ovarian encapsulation were factors influencing adhesions after recanalization intervention for infertility(P<0.05).A nomogram prediction model was constructed,which had AUCs of 0.929 and 0.919 in the modeling and validation populations,respectively,and DCA curves showed that the model had good net benefit values ranging from 0.2 to 0.9 and from 0.0 to 0.85 in the modeling and validation populations.Conclusion A model based on 2D and 3D ultrasound imaging and clinical factors can be used to predict adhesions after interventional recanalization for infertility,and clinical follow-up treatment combined with relevant factors can be tailored to reduce the risk of postoperative adhesions.