基于临床数据的输卵管性不孕症手术效果的风险预测模型

Risk Prediction Model for Surgical Outcomes of Tubal Infertility Based on Clinical Data

目的 探讨基于临床数据的输卵管性不孕症手术治疗效果的风险预测模型。方法 选取2018年1月—2020年7月收治的输卵管性不孕症100例,均行宫腹腔镜联合手术。术后12个月统计宫内妊娠率,分为妊娠组、非妊娠组,收集2组临床资料,采用多因素Logistic回归分析输卵管性不孕症疗效影响因素,建立预测模型,验证其预测价值。结果 输卵管性不孕症患者宫腹腔镜手术时间为(45.30±5.41)min,术中出血量(30.08±4.03)ml,并发症发生率为7.00%,术后12个月随访宫内妊娠率为57.29%。年龄、输卵管阻塞部位、盆腔操作史、盆腔疾病史、不孕症类型、盆腔弥散情况是输卵管性不孕症患者术后宫内妊娠影响因素(P<0.05,P<0.01)。将影响因素经多因素Logistic回归模型拟合成联合预测因子,其预测手术疗效受试者工作特征曲线下面积为0.934(95%CI 0.867,0.976),预测敏感度为89.09%,特异度为87.80%。随机抽取1例,将各自变量代入概率预测方程得到概率值P=0.124,小于临界值,预测准确度为87.50%的条件下,该患者手术效果良好。结论 基于临床数据建立输卵管性不孕症手术疗效风险预测模型具有良好预测能力,可为本病疗效预测提供参考。

Objective To explore a risk prediction model for surgical outcome of tubal infertility based on clinical data. Methods In total, 100 cases of tubal infertility treated from January 2018 to July 2020 were selected, all of whom underwent combined hysteroscopic and laparoscopic surgery. The intrauterine pregnancy rate was calculated at 12 months after operation and they were divided into pregnant group and non-pregnant group. The clinical data of the two groups were collected. The factors influencing the curative effect of tubal infertility were analyzed by multivariate Logistic regression, and the predictive model was established to verify its predictive value. Results The duration of hysteroscopic operation was(45.30±5.41) min, the amount of intraoperative blood loss was(30.08±4.03) ml, the complication rate was 7.00%, and the intrauterine pregnancy rate was 57.29% at 12 months after operation in patients with tubal infertility. Age, site of tubal obstruction, history of pelvic operation, history of pelvic disease, type of infertility and pelvic diffusion were the influencing factors of postoperative intrauterine pregnancy in tubal infertility patients(P<0.05,P<0.01). The combined predictive factors were fitted by multivariate Logistic regression model, and the area under the receiver operating characteristic(ROC) curve(AUC) for predicting surgical efficacy was 0.934(95%CI 0.867, 0.976), the predictive sensitivity was 89.09%, and the specificity was 87.80%. One patient was randomly selected and the probability value P=0.124 was obtained by substituting their variables into the probability prediction equation, which was less than the critical value and the prediction accuracy was 87.50%. The patient had good surgical outcome. Conclusion The risk prediction model of surgical efficacy for tubal infertility based on clinical data has good predictive ability and can provide reference for the prediction of curative effect of tubal infertility.