宫腔镜下子宫黏膜下肌瘤切除术后宫腔粘连发生风险的列线图预测模型构建与评估

Construction and evaluation of a nomogram prediction model for the intrauterine adhesions risk factor after hysteroscopic surgery for subumcosal myoma of uterus

目的分析宫腔镜下子宫黏膜下肌瘤切除术后发生宫腔粘连的危险因素,并构建列线图预测模型。方法收集2017年1月至2020年12月在东莞市妇幼保健院接受宫腔镜下子宫黏膜下肌瘤切除术的322例患者的临床资料,采用单因素分析和多因素Logistic回归分析术后发生宫腔粘连的危险因素,根据分析结果使用R软件构建列线图预测模型,并以H-L拟合度曲线及曲线下面积评估该模型的有效性及区分度。结果322例患者术后进行为期1年的随访,47例患者出现宫腔粘连(粘连组),占14.60%;其余275例未出现宫腔粘连(非粘连组)。单因素分析结果显示,盆腔炎、孕次、刮宫史、合并子宫肌瘤、血清转移生长因子(TGF)-β1水平是宫腔镜下黏膜下肌瘤切除术后发生宫腔粘连的危险因素(P<0.05)。Logistic回归分析结果显示,血清TGF-β1水平、盆腔炎、刮宫史、合并子宫肌瘤是宫腔镜下子宫黏膜下肌瘤切除术后发生宫腔粘连的独立危险因素(P<0.05)。以多因素Logistic回归分析结果建立列线图预测模型,并对列线图模型的区分度进行评估,结果显示,曲线下面积为0.854,灵敏度、特异度分别为91.50%、70.50%;对模型的有效性(H-L拟合度曲线)进行评估,结果显示,χ^(2)=7.12,P=0.413。结论血清TGF-β1水平、盆腔炎、刮宫史、合并子宫肌瘤是宫腔镜下子宫黏膜下肌瘤切除术后发生宫腔粘连的独立危险因素,所构建的列线图预测模型具有较好的有效性及区分度,可作为临床早期干预的有效预测工具。

Objective To analyze the risk factors of intrauterine adhesions in patients after hysteroscopic surgery for subumcosal myoma of uterus,and to construct and evaluate a nomogram prediction model.Methods The clinical data of 322 patients underwent hysteroscopic surgery for subumcosal myoma of uterus in Dongguan Maternal and Child Health Hospital from January 2017 to December 2020 were collected.The univariate analysis and multivariate Logistic regression were used to analyze the factors affecting the occurrence of intrauterine adhesions,according to the analysis results,R software was used to construct a nomogram prediction model that affected the occurrence of intrauterine adhesions after hysteroscopic surgery for subumcosal myoma of uterus,and the H-L fit curve and the area under the curve were used to evaluate the effectiveness and discrimination of the model.Results Through a 1-year follow-up,it was found that 47 patients had intrauterine adhesions(adhesions group),accounting for 14.60%;another 275 patients was enrolled in non-adhesions group.The results of univariate analysis showed that combined pelvic inflammatory disease,pregnancy times,history of curettage,combined uterine fibroids,and serumtransforming growth factor(TGF)-β1 level were risk factors that affected the occurrence of intrauterine adhesions(P<0.05).The results of multivariate Logistic regression analysis showed that serum TGF-β1 level,pelvic inflammatory disease,history of curettage and uterine fibroids were independent risk factors that affected the occurrence of intrauterine adhesions(P<0.05).The nomogram prediction model was established with the results of multivariate Logistic regression analysis,and the discrimination of the nomogram model was evaluated,the results showed that the area under the curve was 0.854,and the sensitivity and specificity were 91.50%and 70.50%,respectively.The validity of the model(H-L fit curve)was evaluated and the results showed thatχ^(2)=7.12,P=0.413.Conclusions Serum TGF-β1 level,combined with pelvic inflammatory disease,history of curettage,combined with uterine fibroids are independent risk factors that affect the occurrence of uterine adhesions after hysteroscopic surgery for subumcosal myoma of uterus.The constructed nomogram prediction model has relatively good effectiveness and discrimination.It can be used as an effective predictive tool for early clinical intervention.