绝经后期子宫内膜增厚与子宫内膜癌的风险因素分析及列线图预测模型构建

Construction of risk prediction nomogram for endometrial cancer in postmenopausal patients with endometrial thickening

目的 探索绝经后期子宫内膜增厚患者罹患子宫内膜癌或不典型增生的相关危险因素并建立预测模型。方法 回顾性分析同济大学附属第四人民医院、第十人民医院收治的绝经后期子宫内膜增厚患者397例。记录患者年龄、绝经年限、身高、体重、孕次、产次、有无阴道流血排液等症状、内科合并症(高血压、糖尿病、高脂血症)、影像学表现、血清学指标、术中情况等,最终以石蜡病理结果作为确诊标准。在R软件中进行数据分析,通过倾向性匹配评分对患者合并症进行匹配后,行多因素Logistic回归分析确定独立危险因素,应用确立的危险因素构建Logistic回归预测模型并验证。结果 绝经后期子宫内膜增厚的397例患者,癌变者41例(10.3%),子宫内膜不典型增生2例(0.5%),其余均为良性病变。单因素Logistic回归分析显示,D-二聚体水平、单核细胞数、平均血小板体积、子宫体积、子宫内膜厚度是子宫内膜癌/不典型增生的危险因素。多因素Logistic回归分析显示,子宫体积>94 cm~3、平均血小板体积<11.1 fL以及D-二聚体水平>0.35 mg/L是绝经后期子宫内膜增厚患者罹患子宫内膜癌/不典型增生的独立危险因素(P<0.05)。Logistic预测模型对于子宫内膜病变的预测有良好的表现(AUC=0.917,95%CI:0.858~0.975)。结论 子宫体积增大、D-二聚体水平增高、平均血小板体积变小是绝经后期子宫内膜增厚患者罹患子宫内膜癌或不典型增生的独立危险因素,以此构建的预测模型对绝经后期子宫内膜增厚患者罹患子宫内膜癌或不典型增生具有良好的预测能力。

Objective To investigate the risk factors of endometrial cancer or atypical hyperplasia in postmenopausal patients with endometrial thickening and to establish a prediction nomogram. Methods A retrospective study was performed on 397 postmenopausal patients with endometrial thickening admitted to the Fourth People’s Hospital and Tenth People’s Hospital Affiliated to Tongji University. The following clinical characteristics of patients were recorded: age, age of menopause, height, weight, number of pregnancies, parity, vaginal bleeding and discharge, medical complications(hypertension, diabetes, hyperlipidemia), imaging findings, serology Indicators, intraoperative conditions, etc. Paraffin pathological results were used as the diagnostic criteria. R software was used for data analysis, independent risk factors were determined through binary multivariate Logistic regression analysis, and a nomogram prediction model was established and validated. Results Among the 397 patients with postmenopausal endometrial thickening, there were 41 cases of endometrial cancer(10.3%), 2 cases of atypical endometrial hyperplasia(0.5%), and 354 cases of benign lesions. Univariate regression analysis showed that D-dimer, monocyte count, mean platelet volume, uterus volume, and endometrial thickness were risk factors for endometrial cancer or atypical hyperplasia. Multivariate Logistic regression analysis showed that D-dimer>0.35 mg/L, mean platelet volume(MPV)<11.1 fL, uterus volume >94 cm~3 were independent risk factors for endometrial cancer or atypical hyperplasia in postmenopausal patients with endometrial thickening(P<0.05). The nomogram model has good performance in predicting endometrial lesions(AUC=0.917). Conclusion Increased uterus volume and D-dimer level, decreased mean platelet volume are the risk factors for endometrial cancer or atypical hyperplasia in postmenopausal patients with endometrial thickening. The nomogram model shows a good performance in predicting endometrial cancer or atypical hyperplasia.