Wittenborn Julia, Kupec Tomas, Iborra Severine, Stickeler Elmar, Najjari Laila, Kennes Lieven N
Department of Obstetrics and Gynecology, University Hospital Aachen, Aachen, Germany.
Department of Economics and Business Administration, University of Applied Sciences Stralsund, Stralsund, Germany.
Geburtshilfe Frauenheilkd. 2022 Aug 16;82(12):1387-1396. doi: 10.1055/a-1857-6470. eCollection 2022 Dec.
This study aimed to identify predictors for the presence of cervical dysplasia in diagnostic LEEPs (Loop Electrical Excision Procedure) of the cervix. The study was designed as a retrospective single-institution cohort analysis of all patients who underwent LEEP without prior proof of high-grade intraepithelial lesion (diagnostic LEEP) between 2015 and 2020 in the Department of Obstetrics and Gynecology of University Hospital Aachen. In order to identify the most meaningful predictive variables for CIN status (CIN2+ or non-CIN2+), multivariate logistic regression was performed and a machine-learning method was used. A total of 849 patients with an indication for loop excision of the cervix were assessed for eligibility. Finally, 125 patients without prior proof of CIN2+ were included into the study. Based on the final multivariate logistic regression model, multiple high-risk HPV infections (p = 0.001), the presence of a T2 transformation zone (p = 0.003) and major lesion changes (p = 0.015) as a result of the colposcopy examination were found to be statistically significant for CIN status based on the diagnostic LEEP. Subsequent ROC analysis showed a high predictive value for the model of 88.35% (AUC). The machine-learning technique (recursive partitioning) identified similar variables as important for CIN status with an accuracy of 75%. For clinical decision-making, the result of the colposcopy examination (T2, major change) as well as the results of HPV testing (multiple high-risk HPV infections) are stronger indicators for clinicians to perform diagnostic excisional procedures of the cervix than the presence of high-grade cytological abnormalities.
本研究旨在确定宫颈环形电切术（LEEP）诊断中宫颈发育异常存在的预测因素。该研究设计为对2015年至2020年间在亚琛大学医院妇产科接受LEEP（无高级别上皮内病变先前证据的诊断性LEEP）的所有患者进行回顾性单机构队列分析。为了确定CIN状态（CIN2+或非CIN2+）最有意义的预测变量，进行了多因素逻辑回归分析并使用了机器学习方法。共有849例有宫颈环形切除术指征的患者接受了资格评估。最终，125例无CIN2+先前证据的患者被纳入研究。基于最终的多因素逻辑回归模型，发现多重高危型人乳头瘤病毒感染（p = 0.001）、T2转化区的存在（p = 0.003）以及阴道镜检查导致的主要病变变化（p = 0.015）对于基于诊断性LEEP的CIN状态具有统计学意义。随后的ROC分析显示该模型具有88.35%（AUC）的高预测价值。机器学习技术（递归划分）识别出对CIN状态同样重要的变量，准确率为75%。对于临床决策，阴道镜检查结果（T2，主要变化）以及HPV检测结果（多重高危型HPV感染）比高级别细胞学异常的存在更能有力地指示临床医生进行宫颈诊断性切除手术。
