基于CT的输尿管结石自发排出预测的可行性分析

Feasibility analysis for predicting spontaneous passage of ureteral calculi based on CT

目的:分析预测输尿管结石自发排出的CT特征因素。方法:回顾性分析2019年12月—2021年12月在浙江医院检查和治疗的115例输尿管结石病例。所有患者初诊均行腹部CT检查并行三维重建,收集输尿管结石位置、结石体积、结石CT值、结石部位输尿管壁厚度和肾积水分级。按照治疗结果不同,将患者分为自发排出(spontaneous passage,SP)组和非自发排出(no-spontaneous passage,No-SP)组。计数资料比较采用χ^(2)检验。计量资料比较采用Mann-Whitney U检验。单因素分析差异有统计学意义的上述CT特征进行logistic回归分析,筛选出输尿管结石自发排出的预测因素。通过ROC曲线检验模型的预测效果。结果:CT共检出输尿管结石115例。其中SP组73例,男59例,女14例,中位年龄37.00岁;No-SP组42例,男31例,女11例,中位年龄43.50岁;2组性别和年龄比较差异无统计学意义(P>0.05)。2组结石位置、肾积水分级比较差异有统计学意义(χ^(2)=8.34、32.74;P<0.05)。输尿管盆段-膀胱壁内段结石(77%)比腹段结石(51%)自发排出率高。0~2级肾积水结石(83%)比3~4级肾结石(30%)自发排出率明显增高。SP组结石体积、结石CT值、结石部位输尿管壁厚度中位数分别为0.07 cm^(3)、562 HU、0.14 cm;No-SP组结石体积、结石CT值、结石部位输尿管壁厚度中位数分别为0.39 cm^(3)、947 HU、0.21 cm。2组比较差异有统计学意义(Z=-7.10、-5.38、-3.39;P<0.05)。结石体积、结石CT值、结石部位输尿管壁厚度的AUC分别为0.90、0.80、0.69,最佳诊断临界值分别为0.08 cm3、802.5 HU、0.11 cm。Logistic回归分析显示结石体积(OR=0.02,95%CI:0.00~0.47)和肾积水分级(OR=0.22,95%CI:0.08~0.62)是结石自发排出的独立预测因素(P<0.05)。该模型的准确度为80.90%,灵敏度为91.80%,特异度为61.90%,阳性预测值为80.72%,阴性预测值为81.25%。ROC曲线显示该模型预测效果好(AUC=0.91)。结论:输尿管结石位置、结石体积、结石CT值、结石部位输尿管壁厚度及肾积水分级是结石自发排出的影响因素。结石体积和肾积水分级为独立预测因素。这为输尿管结石的治疗提供有价值的影像学信息。

Objective To analyze the CT characteristic factors for predicting spontaneous passage of ureteral calculi.Methods A retrospective analysis was performed on 115 cases of ureteral calculi examined and treated in Zhejiang Hospital from December 2019 to December 2021.All patients underwent abdominal CT examination and 3D reconstruction at initial diagnosis.The location,volume,CT value of ureteral calculi,ureteral wall thickness at the calculi site and hydronephrosis grade were measured in the post-processing workstation.According to different therapeutic results,patients were divided into SP and No-SP groups.The chi-square test was used to compare the count data.Measurement data were compared by Mann-Whitney U test.Logistic regression analysis was performed to screen out the predictive factors of spontaneous passage of ureteral calculi.The prediction effect of the model was tested by ROC curve.Results A total of 115 cases of ureteral calculi were detected by CT.Seventy-three cases in SP group(59 males,14 females;Median age:37 years)and 42 cases in No-SP group(31 males,11 females;Median age 43.50 years)were discovered,and there were no significant differences in gender or age between the two groups(P>0.05).There were significant differences in location of of ureteral calculi and hydronephrosis grade between the two groups(χ^(2)=8.34,32.74,P<0.05).The spontaneous passage rate of pelvic and intramural segment of bladder calculi(77%)was higher than that of abdominal segment calculi(51%).The spontaneous passage rate of grade 0-2 hydronephrosis stones(83%)was significantly higher than that of grade 3-4 hydronephrosis stones(30%).In SP group,the median calculi volume,calculi CT value and ureteral wall thickness were 0.07 cm^(3),562 HU and 0.14 cm,respectively.In No-SP group,the median calculi volume,calculi CT value and ureteral wall thickness at calculi site were 0.39 cm^(3),947 HU and 0.21 cm,respectively.There were significant differences between the two groups(Z=-7.10,-5.38,-3.39,P<0.05).The AUC were 0.90,0.80 and 0.69,and the cutoff value were 0.08 cm^(3),802.5 HU and 0.11 cm,respectively.Logistic regression analysis showed that volume of ureteral calculi(OR=0.02,95%CI:0.00-0.47)and hydronephrosis grade(OR=0.22,95%CI:0.08-0.62)were independent predictors of spontaneous passage of stone(P<0.05).The accuracy,sensitivity,specificity,PPV and NPV were 80.90%,91.80%,61.90%,80.72%and 81.25%,respectively.ROC curve showed that the model had a good predictive effect(AUC=0.91).Conclusion Location,volume,CT value of calculi,ureteral wall thickness at stone site and hydronephrosis grade based on CT were the influencing factors of spontaneous passage of ureteral calculi.Volume of ureteral calculi and hydronephrosis grade were independent predictors.These provide valuable imaging information for the treatment of ureteral calculi.