结直肠腺瘤危险因素列线图的建立与评价

Establishment and Evaluation of Nomogram for Risk Factors of Colorectal Adenoma

目的对结直肠腺瘤的危险因素进行临床分析,构建结直肠腺瘤发生风险的列线图预测模型并评价预测效能。方法回顾性收集2018年1月~2019年6月于青岛市市立医院行结肠镜检查的1066例患者,其中496例经病理诊断为结直肠腺瘤的患者纳入腺瘤组,另外570例结肠镜检查未见明显异常,或病理诊断为炎性或增生性息肉者纳入非腺瘤组。通过单因素及多因素分析评价结直肠腺瘤发生的独立危险因素;应用R软件建立预测结直肠腺瘤发生风险的列线图模型,用Bootstrap法进行模型内部验证,采用Calibration校准曲线及受试者工作特征曲线评价列线图的预测效能。结果单因素及多因素分析结果显示,高龄(OR=1.068,95%CI:1.049~1.087,P<0.001)、男性(OR=0.593,95%CI:0.396~0.886,P=0.011)、嗜酒(OR=1.847,95%CI:1.085~3.144,P=0.024)、有肿瘤家族史(OR=1.778,95%CI:1.241~2.547,P=0.002)、大便习惯改变(OR=3.508,95%CI:2.496~4.930,P<0.001)、有糖尿病史(OR=1.832,95%CI:1.179~2.846,P=0.007)、高中性粒细胞计数与淋巴细胞计数比值(OR=2.861,95%CI:2.055~3.982,P<0.001)、高癌胚抗原(OR=1.391,95%CI:1.229~1.574,P<0.001)是结直肠腺瘤发生的独立危险因素,列线图预测模型的C指数为0.780,验证模型的C指数为0.774。结论该研究构建的列线图预测模型具有较高的预测准确性,为筛选结直肠腺瘤发病的高危人群提供了依据,为临床医生内镜检查前提供了参考。

Objective To analyze the risk factors of colorectal adenoma,construct a nomogram prediction model for the risk of colorectal adenoma,and evaluate its predictive efficiency.Methods A total of 1066 patients who underwent colonoscopy in Qingdao Municipal Hospital from January 2018 to June 2019 were collected retrospectively.Among them,496 patients who were pathologically diagnosed with colorectal adenoma were included in the adenoma group,the other 570 patients had no obvious abnormality in colonoscopy,or pathological diagnosis of inflammatory or hyperplastic polyps were included in the non-adenoma group.The independent risk factors for colorectal adenoma were analyzed by univariate and multivariate analysis;R software was used to establish a nomogram model for predicting the risk of colorectal adenoma,and the Bootstrap method was used for internal validation of the model.The Calibration curve and receiver operator characteristic curve was used to evaluate the predictive efficiency of the nomogram.Results The results of univariate and multivariate analysis showed that advanced age(OR=1.068,95%CI:1.049-1.087,P<0.001),male(OR=0.593,95%CI:0.396-0.886,P=0.011),alcoholism(OR=1.847,95%CI:1.085-3.144,P=0.024),family history of tumor(OR=1.778,95%CI:1.241-2.547,P=0.002),changes in bowel habit(OR=3.508,95%CI:2.496-4.930,P<0.001),diabetes(OR=1.832,95%CI:1.179-2.846,P=0.007),high neutrophil/lymphocyte ratio(OR=2.861,95%CI:2.055-3.982,P<0.001),high carcinoembryonic antigen(OR=1.391,95%CI:1.229-1.574,P<0.001)were independent risk factor for colorectal adenoma.The C-index of the nomogram prediction model was 0.780,and the C-index of the validation model was 0.774.Conclusion The nomogram prediction model constructed in this study has high prediction accuracy,which can provide a basis for screening high-risk groups of colorectal adenoma and provide a reference for clinicians before endoscopy.