农村地区大肠癌发病因素分析及其列线图预测模型的建立  

Analysis of the incidence factors of colorectal cancer in rural area and establishment of a nomogram prediction model

作　　者：彭斌[1] 伊纪瑛 杜励[1] 斯雪娇 陈刚[1] PENG Bin;YI Jiying;DU Li;SI Xuejiao;CHEN Gang(Department of Community Medical,Second Clinical Medical College of North Sichuan Medical College,Nanchong Central Hospital,Nanchong,Sichuan 637000,China;Department of Oncology,Second Clinical Medical College of North Sichuan Medical College,Nanchong Central Hospital,Nanchong,Sichuan 637000,China)

机构地区：[1]川北医学院第二临床医学院、南充市中心医院社区医疗科,四川南充637000 [2]川北医学院第二临床医学院、南充市中心医院肿瘤科,四川南充637000

出　　处：《安徽医药》2024年第2期335-339,共5页Anhui Medical and Pharmaceutical Journal

基　　金：四川省基层卫生事业发展研究中心项目(SWFZ21-Y-25)。

摘　　要：目的 分析农村地区大肠癌发病因素,并构建列线图预测模型。方法 选择2019年1月至2022年3月南充市8个乡镇的186例大肠癌新发病例为病例组,以性别、年龄±5岁进行1∶1配对选择同期行大肠癌筛查的同一乡镇健康村民186例为对照组。采取面对面问卷调查收集可能引起大肠癌发病的因素,比较两组上述各因素,并采用多因素logistic回归筛选出独立性影响因素。根据多因素分析结果构建列线图模型,并进行内部验证。结果 单因素分析基础上行逐步多因素二元logistic回归分析结果显示,有吸烟史、有饮酒史、有肠道息肉史、便潜血阳性、红肉摄入≥100 g/d为农村地区大肠癌发病的危险因素,豆类摄入情况≥250克/周为保护性因素(P<0.05)。根据多因素分析结果构建列线图模型,以ROC分析法评估列线图模型的预测价值,曲线下面积0.96,95%CI为(0.94,0.98)(P<0.05);再以Bootstrap法对列线图进行内部验证,以原始数据重复抽样1 000次,结果显示,平均绝对误差为0.004,模型表现与理想模型基本拟合。结论 农村地区大肠癌的发生主要受吸烟史、饮酒史、便潜血结果等因素的影响,本研究构建的列线图模型对于农村地区大肠癌的发病预测准确度与区分度均较高。Objective To analyze the incidence factors of colorectal cancer in rural area,and to construct a nomogram prediction model.Methods From January 2019 to March 2022,186 new cases of colorectal cancer in 8 townships in Nanchong city were select⁃ed as the case group,and 186 healthy villagers in the same township who underwent colorectal cancer screening at the same time were selected as the control group by 1∶1 matching based on gender and age±5 years old.A face-to-face questionnaire survey was used to collect the factors that may cause the incidence of colorectal cancer,the above factors were compared between the two groups,and the independent influencing factors were screened out by multivariate logistic regression.A nomogram model was constructed based on the multivariate analysis results and internally validated.Results On the basis of univariate analysis,the results of stepwise multivariate binary logistic regression analysis showed that smoking history,drinking history,intestinal polyps history,positive fecal occult blood and red meat intake≥100 g/d were risk factors for colorectal cancer in rural area,and the consumption of beans≥250 g/week were pro⁃tective factors(P<0.05).A nomogram model was constructed according to the results of multivariate analysis,and the predictive value of the nomogram model was evaluated by ROC analysis.Internal verification was carried out,and the original data was repeatedly sam⁃pled 1000 times.The results showed that the average absolute error was 0.004,and the performance of the model basically fitted the ideal model.Conclusions The incidence of colorectal cancer in rural areas is mainly affected by factors such as smoking history,drinking history,and fecal occult blood results.The nomogram model constructed in this study has a high degree of accuracy and dis⁃crimination in predicting the incidence of colorectal cancer in rural areas.

关 键 词：结直肠肿瘤 农村 危险因素 多因素分析 列线图 

分 类 号：R735.34[医药卫生—肿瘤]