定量CT腹部体脂参数联合临床-影像学综合模型预测复发性急性胰腺炎

目的 观察定量CT(QCT)腹部体脂参数联合临床-影像学综合模型预测复发性急性胰腺炎(RAP)的价值。方法 回顾性分析468例急性胰腺炎(AP),包括207例RAP(RAP组)与261例非RAP(非RAP组),记录首诊临床资料、常规CT表现、腹部皮下脂肪面积(SFA)、内脏脂肪面积(VFA)、腹部总脂肪面积(TFA),以及肝脏和胰腺脂肪含量等QCT参数;比较组间临床特征、CT表现及QCT参数值,以多因素logistic回归分析筛选RAP的独立预测因素,并构建临床-影像学模型及结合QCT参数的综合模型,以受试者工作特征(ROC)曲线评估其预测RAP的效能。结果 相比非RAP组,RAP组患者年龄较小,饮酒、胆道结石及高脂血症占比、首诊CT胰腺边缘清晰占比,以及VFA、TFA、肝脏脂肪含量及胰腺脂肪含量均较高(P均<0.05)。饮酒、高脂血症、胆道结石、胰腺边缘及胰腺脂肪含量均为RAP的独立预测因素,以之构建的综合模型预测RAP的曲线下面积为0.860,高于基于前4者的临床-影像学模型(0.701)及胰腺脂肪含量(0.770)(P均<0.001)。结论 胰腺脂肪含量QCT参数联合临床-影像学综合模型可有效预测RAP风险。

Quantitative CT abdominal fat parameters combined with clinical-imaging comprehensive model for predicting recurrent acute pancreatitis

Objective To observe the value of quantitative CT (QCT) abdominal body fat parameters combined with clinical-imaging model for predicting recurrent acute pancreatitis (RAP). Methods Data of 468 patients with acute pancreatitis (AP), including 207 cases of RAP (RAP group) and 261 cases of non-RAP (non-RAP group) were retrospectively analyzed. Clinical information, conventional CT manifestations and QCT parameters such as abdominal subcutaneous fat area (SFA), visceral fat area (VFA), total abdominal fat area (TFA), hepatic fat content and pancreatic fat content at the first visit were recorded or measured. Clinical characteristics, CT manifestations and QCT parameter values were compared between groups, and the independent factors for predicting RAP were selected with multivariate logistic regression analysis. Then a clinical-imaging model and a comprehensive model combining with QCT parameters were constructed, and their efficacies for predicting RAP were evaluated with receiver operating characteristic (ROC) curves, and the area under the curves (AUC) were calculated. Results Compared with non-RAP group, patients in RAP group were younger, had higher percentage of alcohol consumption, biliary stones and hyperlipidemia, as well as of distinct pancreatic margins on CT images, also higher VFA, TFA, liver fat content and pancreatic fat content (all P<0.05). Alcohol consumption, hyperlipidemia, biliary stones, pancreatic margins and pancreatic fat content were all independent predictors of RAP, and the comprehensive model constructed based on these five had higher AUC (0.860) than clinical-imaging model constructed based on the previous 4 factors (0.701) and pancreatic fat content alone (0.770) (both P<0.001). Conclusion QCT abdominal body fat parameters combined with clinical-imaging comprehensive model was effective for predicting the risk of RAP.