基于临床及CT特征的列线图模型评估急性复杂性阑尾炎风险

Nomogram model of clinical and CT features for evaluating the risk of acute complicated appendicitis

目的探索基于临床及CT特征的风险列线图模型对急性复杂性阑尾炎(ACA)的诊断准确率。方法回顾性收集从2020年12月—2022年12月经我院术后病理证实为急性阑尾炎(AA)的478例患者信息,构建预测模型,结合术后病理结果及临床治疗方案,将AA患者分为ACA(包括急性坏疽性阑尾炎、急性穿孔性阑尾炎及阑尾脓肿,需要手术治疗)与急性非复杂性阑尾炎(AUA)(包括急性单纯性阑尾炎与急性化脓性阑尾炎,可选择保守治疗);收集临床特征、实验室检查及术前CT征象(含阑尾管径增粗、阑尾壁增厚、阑尾粪石、腔外积气、阑尾周围渗液);采用单因素分析比较各项参数在ACA组和AUA组之间的差异,使用多因素Logistic回归分析筛选出高危因素,并构建相应列线图模型。另外收集从2023年1月—2023年8月经我院术后病理证实为AA的146例患者信息用于构建验证模型,采用受试者工作特征(ROC)曲线与决策曲线分析(DCA)对模型结果进行内部验证,并与传统的Alvarado评分与阑尾炎炎性反应(AIR)评分法进行比较。结果全部患者中有84例诊断为ACA,占13.5%(84/624)。患者年龄、阑尾炎发病症状持续时间、体温、中性粒细胞百分比以及术前CT特征是ACA的高风险因素。ROC曲线显示基于此构建的列线图模型曲线下面积(AUC)为0.831(95%CI:0.783~0.876),敏感度0.820,特异度0.722,准确度0.771;Alvarado评分法AUC为0.640(95%CI:0.585~0.696),敏感度0.734,特异度0.462,准确度0.598;AIR评分法AUC为0.738(95%CI:0.585~0.890),敏感度0.668,特异度0.727,准确度0.698。相比于Alvarado、AIR评分法,列线图模型呈现出更好的临床获益效果。结论基于临床及CT特征的风险列线图模型在诊断ACA中具有较高的准确性和可靠性。

Objective To explore the diagnostic accuracy of risk nomogram model based on clinical and CT features for acute complicated appendicitis(ACA).Methods 478 patients with pathologically confirmed acute appendicitis(AA)from December 2020 to December 2022 were retrospectively classified as ACA(acute gangrenous appendicitis,acute perforated appendicitis,appendiceal abscess requiring surgery)and acute uncomplicated appendicitis(AUA:acute simple appendicitis,acute suppurative appendicitis).The clinical features,laboratory findings,and preoperative CT appearance between ACA and AUA groups were compared using single factor analysis.The high-risk factors were screened by multivariate logistic regression analysis and the corresponding nomogram model was constructed.Data from a separate group of 146 patients with AA from January 2023 to August 2023 were used to construct the validation model.Receiver operating characteristic(ROC)curve and decision curve analysis were used for internal validation of model results and compared with traditional Alvarado and the appendicitis inflammatory response(AIR)scoring methods.Results Of the 624 patients,13.5%(84/624)had ACA.Regression analysis showed that age,duration of appendicitis symptoms,body temperature,percentage of neutrophil,and preoperative CT features were high risk factors for ACA.The nomogram model based on these factors had better diagnostic performance with area under the ROC curve(AUC)of 0.831(95%CI:0.783-0.876),82.0%sensitivity,72.2%specificity,and 77.1%accuracy compared to the Alvarado scoring method(AUC:0.640,95%CI:0.585-0.696;sensitivity:73.4%;specificity:46.2%;accuracy:59.8%)and AIR scoring method(AUC:0.738,95%CI:0.585-0.890;sensitivity:66.8%;specificity:72.7%;accuracy:69.8%).Conclusion Risk nomogram model based on clinical and CT features has high accuracy and reliability in diagnosing ACA.