通过前瞻性研究,建立基于临床的列线图,以预测外头型(ECV)的成功率。这是一项单中心前瞻性研究,收集了符合条件的臀位孕妇。152名参与者参加了培训队列,接受由单个操作员执行的ECV程序的人。我们使用训练队列建立回归方程和预测模型。这些变量包括母亲因素(年龄,手术胎龄,孕前BMI(身体质量指数),手术BMI,BMI增加,multipara),超声因素(胎儿体重估计,羊水指数,胎盘位置,臀位的类型,脊柱位置),和麻醉。采用单变量和多变量分析筛选影响ECV成功的因素。基于这些因素建立了列线图评分模型。和C指数,DCA(决策曲线分析)和校准曲线,采用Hosmer-Lemeshow检验对模型的预测效果进行验证。最后,33名参与者被纳入测试队列,他们接受了无限制操作员的ECV。我们使用了C指数,DCA(决策曲线分析),和Hosmer-Lemeshow验证了预测模型的应用价值。绘制训练群组和测试群组两者的校准曲线和ROC曲线以用于模型的内部和外部验证。训练队列的ECV成功率为62.5%。单因素分析显示,与ECV成功率相关的预测因素为年龄,BMI增加值,AFI(羊水指数),臀位类型,胎盘位置,脊柱位置,麻醉,和多段。根据Youden指数计算相应指标的预测阈值。多因素logistic回归分析显示BMI增加≥3.85kg/m2,AFI≥10.6cm,麻醉,multipara,非前胎盘是ECV成功的独立预测因子。通过内部和外部验证,模型具有良好的标定和预测能力。我们的列线图具有很好的预测ECV成功率的能力。
To establish a clinical-based nomogram for predicting the success rate of external cephalic version (ECV) through a prospective study. This was a single-center prospective study that collected eligible breech pregnant women. 152 participants were enrolled in the training cohort, who received ECV procedures performed by a single operator. We used the training cohort to establish regression equations and prediction models. These variables include maternal factors (age, operation gestational age, pre-pregnancy BMI (Body Mass Index), operation BMI, BMI increase, multipara), ultrasound factors (fetal weight estimation, amniotic fluid index, placental location, type of breech presentation, spinal position), and anesthesia. Univariate and multivariable analyses were used to screen the factors affecting the success of ECV. A nomogram scoring model was established based on these factors. And C-index, DCA (Decision Curve Analysis) and calibration curve, Hosmer-Lemeshow test was used to verify the prediction effect of the model. Finally, 33 participants were enrolled in the testing cohort who received ECV with an unrestricted operator. We used C-index, DCA (decision curve analysis), and Hosmer-Lemeshow to verify the application value of the prediction model. The calibration curves and ROC curves of both the training cohort and testing cohort are plotted for internal and external validation of the model. The ECV success rate of the training cohort was 62.5%. Univariate analysis showed that the predictors related to the success rate of ECV were age, BMI increase value, AFI (amniotic fluid index), breech type, placental location, spinal position, anesthesia, and multipara. The prediction thresholds of the corresponding indexes were calculated according to the Youden index. Multivariate logistic regression analysis showed that BMI increase ≥ 3.85 kg/m2, AFI ≥ 10.6 cm, anesthesia, multipara, and non-anterior placenta were independent predictors of ECV success. Through the internal and external validation, it is confirmed that the model has a good calibration and prediction ability. Our nomogram has a good ability to predict the success rate of ECV.