基于客观因素的儿童坏疽穿孔性阑尾炎预测模型的建立与验证  

Development and validation of a scoring model based upon objective factors for predicting gangrenous/perforated appendicitis in children

作　　者：黄卓 徐诺 毛连春 马丹丹 金炜东 Huang Zhuo;Xu Nuo;Mao Lianchun;Ma Dandan;Jin Weidong(School of Medicine,Wuhan University of Science&Technology,Hubei Wuhan 430065,China;Department of General Surgery,General Hospital of Chinese People´s Liberation Army Central Theater,Hubei Wuhan 430070,China)

机构地区：[1]武汉科技大学医学部医学院,湖北武汉430065 [2]中国人民解放军中部战区总医院普通外科,湖北武汉430070

出　　处：《腹部外科》2024年第6期406-412,共7页Journal of Abdominal Surgery

基　　金：湖北省自然科学基金面上项目(2021CFB500)。

摘　　要：目的建立并验证基于客观因素的儿童急性阑尾炎(acute appendicitis,AA)发生坏疽穿孔的列线图预测模型。方法回顾性收集2010年1月至2024年9月就诊于中国人民解放军中部战区总医院普通外科的350例AA患儿的临床资料,选择2010年1月至2019年12月的患儿为研究对象(n=225),选择2020年1月至2024年9月的患儿为验证对象(n=125)。所有患儿在诊断明确后行阑尾切除手术,依据术后送检阑尾病理结果,分为无坏疽性/穿孔性阑尾炎(non-gangrenous/perforated appendicitis,N-GPA)组和坏疽性/穿孔性阑尾炎(gangrenous/perforated appendicitis,GPA)组。进行两组间比较,对有意义的指标通过LASSO回归、多因素logistic回归分析筛选出患儿发生坏疽穿孔的独立危险因素,并对其建立列线图预测模型。同时基于列线图模型对验证对象进行内部验证,分别采用受试者操作特征(receiver operating characteristic,ROC)曲线的曲线下面积(area under the curve,AUC)和Hosmer-Lemeshow(H-L)拟合优度检验评估列线图模型的区分度和校准度。结果术前CT提示阑尾管腔内是否有粪石、白细胞计数、症状持续时间、体温和术前CT检查阑尾直径为AA患儿发生坏疽穿孔的独立危险因素(P<0.05)。列线图预测研究对象和验证对象中AA患儿发生坏疽穿孔的AUC分别为0.896(95%CI:0.848~0.943)和0.899(95%CI:0.825~0.972),结果提示模型具有良好区分度;研究对象和验证对象的H‐L拟合优度检验P值分别为0.459(χ^(2)=6.72)和0.272(χ^(2)=9.91),表示所建立的列线图模型有较好的预测精准能力。结论研究所建立的列线图预测模型具有较高的准确性,可方便地用于AA患儿发坏疽穿孔的早期识别和风险预测。Objective To establish and validate a nomogram prediction model based upon objective factors for predicting gangrenous/perforated appendicitis(GPA)in children with acute appendicitis(AA).Methods Between January 2010 and September 2024,the relevant clinical data were retrospectively reviewed for 350 AA children.And patients admitted between January 2010 and December 2019 were allocated into study group(n=225)while those admitted between January 2020 and September 2024 were selected as validation group(n=125).Appendectomy was performed after a definite diagnosis.They were assigned into two groups of non-gangrenous/perforated appendicitis(N-GPA)and gangrenous/perforated appendicitis(GPA)according to the postoperative results of pathological examination.A comparison was made between two groups.And independent risk factors for GPA were identified through LASSO regression and multivariate Logistic regression analysis.Then these factors were utilized for constructing a nomogram prediction model.Internal validation of nomogram model was conducted with validation group.The area under the receiver operating characteristic curve(AUC)and Hosmer-Lemeshow(H-L)goodness-of-fit test were applied for assessing the model’s discrimination and calibration.Results Preoperative computed tomography(CT)findings of fecalith within appendiceal lumen,leukocyte count,duration of symptoms,body temperature and appendiceal diameter as measured by preoperative CT were identified as independent risk factors for GPA in AA children(P<0.05).AUC for nomogram prediction model was 0.896(95%CI:0.848-0.943)in study group and 0.899(95%CI:0.825-0.972)in validation group,indicating an excellent discriminatory power.H-L test results were P=0.459(χ^(2)=6.72)for study group and P=0.272(χ^(2)=9.91)for validation group,indicating that nomogram model had a decent predictive accuracy.Conclusion The nomogram prediction model developed here demonstrates high accuracy and it may be readily applied for early identification and risk prediction of GPA in AA children

关 键 词：急性阑尾炎 坏疽穿孔 预测模型 儿童 

分 类 号：R656.8[医药卫生—外科学]