结肠癌患者术后感染的影响因素分析及列线图模型构建

Analysis of influencing factors of postoperative infection in patients with colon cancer and construction of nomogram model

目的分析影响结肠癌患者术后感染的危险因素并构建列线图模型。方法回顾性分析安徽省肿瘤医院2019年5月至2022年6月220例结肠癌患者的临床资料。其中,发生术后感染55例(感染组),未发生术后感染165例(非感染组)。采用受试者工作特征(ROC)曲线分析各指标预测结肠癌患者发生术后感染的效能;采用多因素Logistic回归分析影响结肠癌患者发生术后感染的独立危险因素。采用R语言3.5.2软件构建预测结肠癌患者发生术后感染的列线图模型并验证和评估。结果两组性别构成、体质量指数、肿瘤分期、术中输血、高血压、吸烟史、饮酒史、肿瘤直径和血红蛋白比较差异无统计学意义(P>0.05);感染组年龄、糖尿病比例、手术时间和排气时间明显大于非感染组[(49.60±4.40)岁比(47.20±4.12)岁、63.64%(35/55)比30.30%(50/165)、(197.80±12.55)min比(192.23±12.05)min和(3.42±1.18)d比(2.60±0.80)d],白蛋白明显低于非感染组[(28.29±3.02)g/L比(32.80±3.21)g/L],差异有统计学意义(P<0.01)。ROC曲线分析结果显示,年龄、手术时间、排气时间和白蛋白预测结肠癌患者发生术后感染的曲线下面积分别为0.672、0.610、0.706和0.846,最佳截断值分别为49岁、184 min、3 d和30 g/L。多因素Logistic回归分析结果显示,年龄(>49岁)、糖尿病、手术时间(>184 min)、排气时间(>3 d)、白蛋白(≤30 g/L)是影响结肠癌患者发生术后感染的独立危险因素(OR=2.131、1.758、1.449、1.841和2.325,95%CI 1.269~2.696、1.354~3.059、1.201~1.965、1.018~2.365和1.582~3.051,P<0.01)。以年龄、糖尿病、手术时间、排气时间、白蛋白作为预测因子构建了列线图模型。此列线图模型预测结肠癌患者发生术后感染的校正曲线趋近于理想曲线(C-index为0.764,95%CI 0.657~0.834);决策曲线分析结果显示,风险阈值>0.07时,此列线图模型提供临床净收益;且模型的临床净收益高于年龄、糖尿病、手术时间、排气时间和白蛋白。结论年龄(>49岁)、糖尿病、手术时间(>184 min)、排气时间(>3 d)和白蛋白(≤30 g/L)是影响结肠癌患者发生术后感染的独立危险因素,且基于以上变量构建的列线图模型可以对患者术后感染进行较好的预测。

Objective To analyze the risk factors of postoperative infection in patients with colon cancer,and construct a nomogram model.Methods The clinical data of 220 patients with colon cancer in Anhui Cancer Hospital from May 2019 to June 2022 were retrospectively analyzed.Among them,55 patients developed postoperative infection(infection group),and 165 patients did not develop postoperative infection(non-infection group).The receiver operating characteristic(ROC)curve was used to analyze the efficacy of each index in predicting postoperative infection in patients with colon cancer.Multivariate Logistic regression analysis was used to analyze the independent risk factors of postoperative infection in patients with colon cancer.R language 3.5.2 software was used to construct a nomogram model for predicting postoperative infection in patients with colon cancer,and it was verified and evaluated.Results There were no significant differences in gender composition,body mass index,tumor stage,intraoperative blood transfusion,hypertension,smoking history,alcohol consumption history,tumor diameter and hemoglobin between the two groups(P>0.05);the age,diabetes mellitus ratio,operation time and exhaust time in the infection group were significantly higher than those in the non-infection group:(49.60±4.40)years old vs.(47.20±4.12)years old,63.64%(35/55)vs.30.30%(50/165),(197.80±12.55)min vs.(192.23±12.05)min and(3.42±1.18)d vs.(2.60±0.80)d,the albumin was significantly lower than that in the non-infected group:(28.29±3.02)g/L vs.(32.80±3.21)g/L,and there were statistical differences(P<0.01).ROC curve analysis result showed that the area under the curve of age,operation time,exhaust time and albumin for predicting postoperative infection in patients with colon cancer were 0.672,0.610,0.706 and 0.846,and the optimal cut-off values were 49 years old,184 min,3 d and 30 g/L,respectively.Multivariate Logistic regression analysis result showed that age(>49 years old),diabetes mellitus,operation time(>184 min),exhaust time(>3 d)and albumin(≤30 g/L)were independent risk factors of postoperative infection in patients with colon cancer(OR=2.131,1.758,1.449,1.841 and 2.325;95%CI 1.269 to 2.696,1.354 to 3.059,1.201 to 1.965,1.018 to 2.365 and 1.582 to 3.051;P<0.01).A nomogram model was constructed with age,diabetes mellitus,operation time,exhaust time,and albumin as predictors for predicting postoperative infection in patients with colon cancer.The correction curve of the nomogram model for predicting postoperative infection in patients with colon cancer was close to the ideal curve(C-index=0.764,95%CI 0.657 to 0.834);decision curve analysis result showed that the nomogram model provided clinical net benefit when the risk threshold was>0.07;and the clinical net benefit of the model was higher than that of age,diabetes mellitus,operation time,exhaust time and albumin.Conclusions The age(>49 years old),diabetes mellitus,operation time(>184 min),exhaust time(>3 d)and albumin(≤30 g/L)are the independent risk factors of postoperative infection in patients with colon cancer,and the nomogram model based on the above variables could predict postoperative infection.