胃息肉合并结直肠息肉风险诺谟图预测模型的构建与验证  

Construction and validation of Nomogram prediction model for the risk ofgastric polyps combination with colorectal polyps

作　　者：赵洋洋 李紫琼[1] 马欣[1] 木克热木·依明尼亚孜 高峰[1] ZHAO Yangyang;LI Ziqiong;MA Xin;Mukeremu Yimingniyazi;GAO Feng(Xinjiang Clinical Research Center for Digestive Diseases,Xinjiang Uygur Autonomous Region People’s Hospital,Xinjiang Uygur Autonomous Region,Urumqi 830000,China)

机构地区：[1]新疆维吾尔自治区人民医院新疆消化系统疾病临床医学研究中心,新疆乌鲁木齐830000

出　　处：《中国医药导报》2024年第1期27-33,共7页China Medical Herald

基　　金：新疆维吾尔自治区科技支疆项目计划(指令性)项目(2022E02041)。

摘　　要：目的构建胃息肉合并结直肠息肉风险的诺莫图预测模型,并进一步验证模型的预测效能。方法回顾性收集2021年9月至2022年11月新疆维吾尔自治区人民医院消化内科行胃镜检查诊断胃息肉并完善结肠镜检查的患者679例为建模组;依据结肠镜检查结果将其分为有结直肠息肉组与无结直肠息肉组。另收集2022年12月至2023年5月新疆维吾尔自治区人民医院消化内科诊断胃息肉并完善结肠镜检查的患者241例作为验证组。分析胃息肉合并结直肠息肉的影响因素,建立诺谟图预测胃息肉合并结直肠息肉风险的模型;采用受试者操作特征(ROC)曲线评估预测模型的区分度;校准曲线评估数据拟合度;决策曲线分析图评估使用模型的获益情况。结果建模组与验证组一般资料比较,差异无统计学意义(P>0.05)。建模组中有268例发现结直肠息肉。有结直肠息肉组和无结直肠息肉组非酒精性脂肪性肝病、幽门螺杆菌(H.pylori)感染、高同型半胱氨酸血症、2型糖尿病、高血压病、高甘油三酯血症、年龄比较,差异有统计学意义(P<0.05)。非酒精性脂肪性肝病(OR=3.317)、H.pylori感染(OR=3.018)、高同型半胱氨酸血症(OR=3.697)、高甘油三酯血症(OR=4.248)、年龄(OR=1.065)是胃息肉合并结直肠息肉的独立危险因素(P<0.05)。建模组诺谟图预测模型ROC曲线的曲线下面积(AUC)为0.742,验证组AUC为0.737,模型具有较好的区分度。校正曲线提示实际发生率和预测发生率拟合度好。决策曲线图显示模型获益阈值广泛,具有实用价值。结论由非酒精性脂肪性肝病、H.pylori感染、高同型半胱氨酸血症、高甘油三酯血症、年龄建立的预测模型具有较好的预测能力,可为医务人员判断胃息肉合并结直肠息肉的风险提供参考。Objective To construct Nomogram prediction model for the risk of gastric polyps complicated with colorectal polyps,and further verify the prediction efficiency of the model.Methods A total of 679 patients with gastric polyps diagnosed by gastroscopy and complete colonoscopy in the Department of Gastroenterology,Xinjiang Uygur Autonomous Region People’s Hospital from September 2021 to November 2022 were retrospectively collected as the modeling group;according to the results of colonoscopy,they were divided into colorectal polyps group,and no colorectal polyps group.Another 241 patients with gastric polyps diagnosed by gastroscopy and complete colonoscopy from December 2022 to May 2023 were collected as the validation group.The influencing factors of gastric polyp complicated with colorectal polyp were analyzed,the Nomogram was established to predict the risk of gastric polyps complicated with colorectal polyps;receiver operating characteristic(ROC)curve was used to evaluate the discrimination of the prediction model;calibration curve was used to evaluate the degree of data fitting;and decision curve analysis diagram was used to evaluate the bene fit of using the model.Results There was no significant difference in general data between the modeling group and the verification group(P>0.05).Colorectal polyps were found in 268 of the modeling group.There were statistically significant differences in nonalcoholic fatty liver disease,Helicobacter Pylori(H.pylori)in-fection,type 2 diabetes mellitus,hypertension,hyperhomocysteinemia,hypertriglyceridemia,and age between colorectal polyps group and no colorectal polyps group(P<0.05).Nonalcoholic fatty liver disease(OR=3.317),H.pylori infection(OR=3.018),hyperhomocysteinemia(OR=3.697),hypertriglyceridemia(OR=4.248),and age(OR=1.065)were independent risk factors for gastric polyps with colorectal polyps(P<0.05).The area under the curve(AUC)of the ROC curve predicted by the Nomogram in the modeling group was 0.742,and the AUC of the verification group was 0.737,indicating t

关 键 词：胃息肉 结直肠息肉 诺谟图 预测模型 

分 类 号：R574.6[医药卫生—消化系统]