肝癌肝切除术后感染风险预测模型的建立与评价

Establishment and validation of a nomogram risk prediction model for infection complications in patients after hepatectomy for liver cancer

目的探讨肝癌肝切除术后发生感染的危险因素,建立并验证风险预测模型。方法收集2020年1月—2022年4月于武汉大学人民医院行肝切除术的167例原发性肝癌患者的临床资料。根据术后早期是否发生感染,将所有患者分为术后感染组(n=28)和非感染组(n=139)。计量资料两组间比较采用t检验或Mann-Whitney U检验,计数资料两组间比较采用χ^(2)检验。采用单因素分析与Logistic回归分析筛选肝癌肝切除术后感染的影响因素,并建立术后发生感染的列线图风险预测模型,将所有患者按7∶3随机分为训练集(n=119)和验证集(n=48),采用Bootstrap法对模型进行内部验证,应用模型校准曲线和受试者工作特征曲线(ROC曲线)来评价列线图模型的校准度和区分度。结果167例患者中28例(16.8%)发生术后感染。Logistc多因素分析结果显示,糖尿病、CONUT评分≥4分、术前NLR、手术时间、术中失血量、引流管放置时间>7 d是肝癌肝切除术后发生感染的独立危险因素(P值均<0.05)。基于上述6个危险因素构建的列线图,训练集和验证集的ROC曲线下面积分别为0.848、0.853。列线图模型校准曲线显示预测值与实际观测值基本一致,表明列线图模型预测的准确度较好。结论基于糖尿病、CONUT评分≥4分、术前NLR、手术时间、术中失血量、引流管放置时间>7 d建立的个体化列线图风险预测模型预测效能良好,对高风险患者具有较高的预测价值。

Objective To investigate the risk factors of infection after hepatectomy for liver cancer,and to establish and validate a risk prediction model.Methods The clinical data of 167 patients with primary liver cancer who underwent hepatectomy in People’s Hospital of Wuhan University from January 2020 to March 2022 were retrospectively collected.All patients were divided into postoperative infection group(n=28)and non-infection group(n=139)according to whether postoperative infection complications occurred.The t-test or Mann-Whitney U test was used for comparison of continuous data between two groups and the chi-square test was used for comparison of categorical data between two groups.Univariate analysis and logistic regression analysis were used to screen the risk factors of infection after hepatectomy for hepatocellular carcinoma,and a nomogram risk prediction model for postoperative infection was established.All patients were randomly divided into training cohort(n=119)and the validation cohort(n=48)according to the ratio of 7∶3,the Bootstrap method was used for internal validation of the model,and the model calibration curve and ROC curve were used to evaluate the calibration and discrimination of the nomogram model.Results Postoperative infection occurred in 28 of 167 patients(16.8%).Logistic regression analysis showed that diabetes,CONUT score≥4 points,preoperative NLR,operation time,intraoperative blood loss,and drainage tube placement time>7 d were independent risk factors for infection after hepatectomy for liver cancer(all P<0.05).Based on the nomogram constructed from the above six risk factors,the area under the ROC curve of the training cohort and the validation cohort was 0.848,and 0.853,respectively.The calibration curve of the nomogram model shows that the predicted value is basically consistent with the actual observed value,indicating that the accuracy of the nomogram model prediction is better.Conclusion The individualized nomogram risk prediction model based on diabetes,CONUT score≥4 points,preoperative NLR,operation time,intraoperative blood loss,and drainage tube placement time>7 d has good predictive performance and has high predictive value for high-risk patients.