基于常规超声及超声造影定量参数的Logistic回归模型预测移植肾功能延迟恢复的临床价值

Clinical value of Logistic regression model based on quantitative parameters of conventional ultrasound and contrast-enhanced ultrasound in predicting delayed graft function

目的探讨基于常规超声、超声造影定量参数的Logistic回归模型在移植肾功能延迟恢复(DGF)中的早期预测价值。方法选取佛山市第一人民医院泌尿外科接受同种异体肾移植术的患者212例,其中移植肾功能正常恢复157例(NGF组),DGF 55例(DGF组)。比较两组患者临床资料、常规超声及超声造影参数的差异;Logistic回归分析DGF的独立影响因素,并建立诊断模型;绘制受试者工作特征(ROC)曲线分析该模型预测DGF的诊断效能。结果DGF组身高、体质量、血尿素氮(BUN)均高于NGF组,血红蛋白低于NGF组,差异均有统计学意义(均P<0.05);常规超声检查结果显示,DGF组叶间动脉收缩期峰值流速(PSV)及阻力指数(RI)、段动脉PSV及RI、肾门动脉RI、动脉吻合口RI均增高,与NGF组比较差异均有统计学意义(均P<0.05);超声造影检查结果显示,DGF组起始强度(AI)增高,峰值强度(PI)、增强强度(A)、曲线下面积(AUC)均减低,与NGF组比较差异均有统计学意义(均P<0.05)。Logistic回归分析显示,BUN、叶间动脉RI、AUC均为DGF的独立影响因素(OR=1.184、44.413、0.999,均P<0.05),建立诊断模型为:Logit(P)=-3.363+0.169×BUN+3.794×叶间动脉RI-0.001×AUC。ROC曲线分析显示,该模型截断值为0.332时,预测DGF的灵敏度为78.2%,特异度为89.8%,约登指数为68.0%,阳性预测值为72.9%,阴性预测值为92.1%;曲线下面积为0.909(P<0.01),95%可信区间为0.863~0.955。结论基于常规超声及超声造影定量参数的Logistic回归模型对早期预测DGF有较高的诊断效能。

Objective To explore the early prediction value of Logistic regression model based on quantitative parameters of conventional ultrasound and contrast-enhanced ultrasound in delayed graft function(DGF).Methods A total of 212 patients with renal allograft transplantation admitted to the department of urology,the first people’s hospital of foshan were retrospectively analyzed,including 157 cases with normal graft function(NGF group)and 55 cases with DGF(DGF group).The clinical data,parameters of conventional ultrasound and contrast-enhanced ultrasound between the two groups were compared.Logistic regression analysis was used to analyze the independent influencing factors of DGF,and a diagnostic model was established.Receiver operating characteristic(ROC)curve was drawn to analyze the diagnostic efficacy of the model in predicting DGF.Results The height,weight,and blood urea nitrogen(BUN)of the DGF group were higher than those of the NGF group,and the hemoglobin was lower than that of the NGF group,with statistically significant differences(all P<0.05).The results of conventional ultrasound examination showed the peak systolic velocity(PSV)and resistance index(RI)of interlobar artery,segmental artery PSV and RI,renal hilar artery RI and anastomotic artery RI of the DGF group were increased than those of the NGF group,with statistically significant differences(all P<0.05).The results of contrast-enhanced ultrasound examination showed that the initial intensity(AI)of the DGF group was increased than that of the NGF group,the peak intensity(PI),enhancement intensity(A)and area under the curve(AUC)of the DGF group were decreased than those of the NGF group,with statistically significant differences(all P<0.05).Logistic regression analysis showed that BUN,interlobar artery RI and AUC were all independent influencing factors of DGF(OR=1.184,44.413,0.999,all P<0.05).The established model was:Logi(t P)=-3.363+0.169×BUN+3.794×interlobar artery RI-0.001×AUC.ROC curve analysis showed that the diagnostic threshold of the model was 0.332,and the sensitivity,specificity,Youden index,positive predictive value and negative predictive value of DGF were 78.2%,89.8%,68.0%,72.9%and 92.1%,respectively.The area under the curve was 0.909(P<0.01),and the 95%confidence interval was 0.863~0.955.Conclusion The Logistic regression model based on parameters of conventional ultrasound and contrastenhanced ultrasound has high diagnostic efficiency for early prediction of DGF.