探究膀胱影像报告与数据系统对肌层浸润性膀胱癌的诊断效能及观察者间一致性的研究

Investigating the Diagnostic Efficacy and Inter-Observer Agreement with the Vesical Imaging-Reporting and Data System(VI-RADS)for Muscle-Invasive Bladder Cancer

目的探讨VI-RADS对膀胱癌浸润肌层的诊断效能和评估观察者间一致性。方法共纳入122例接受膀胱多参数MRI检查的患者。两位影像医师按照膀胱影像报告与数据系统(VI-RADS)规定的评分方法和标准进行阅片。VI-RADS评分观察者之间的一致性通过组内相关系数(ICCs)来评估,采用受试者工作特征曲线(ROC)对VI-RADS评分诊断膀胱癌浸润肌层的诊断效能进行分析,获得ROC曲线下面积(AUC),并计算不同阈值条件下的诊断敏感度、特异度、阳性预测值(PPV)、阴性预测值(NPV)及准确度。以P<0.05为差异有统计学意义。结果共纳入122例患者,男96例,女26例。平均年龄为66岁,膀胱肿瘤多发19例,单发103例。病理诊断为肌层浸润性膀胱癌(MIBC)33例,非肌层浸润性膀胱癌(NMIBC)89例。两名观察者一致性良好,ICC值为0.94,95%置信区间(0.91,0.96)。<5.4%MIBC出现在VI-RADS 1分和2分中;对于VI-RADS 3、4、5分的病变,MIBC比例分别为28.6%(2/7),47.1%(8/17)和80%(20/25)。当阈值VI-RADS≥3定为MIBC时,观察者1和观察者2的敏感度、特异度、PPV、NPV及准确度分别为91%、80%、63%、96%、83%和91%、76%、59%、96%、80%;如果将定义MIBC的分界点移至VI-RADS≥4,则观察者1和观察者2的敏感度、特异度、PPV、NPV及准确度分别为85%、85%、68%、94%、85%和91%、81%、64%、96%、83%。观察者1、2的AUC分别为0.89和0.91。结论VI-RADS评分能够准确地鉴别MIBC和NMIBC,观察者间一致性总体较好。

Objective The aim of this study was to investigate the diagnostic efficacy and inter-observer agreement with the Vesical Imaging-Reporting and Data System(VI-RADS)for muscle-invasive bladder cancer.Methods A total of 122 patients who underwent multiparametric MRI of the bladder were included.Two radiologists read and scored all mpMRI data according to the scoring method and criteria specified by VI-RADS.Inter-observer agreement was assessed by intraclass correlation coefficients(ICCs).The diagnostic efficacy of the VI-RADS scores of muscle-invasive bladder cancer was analyzed using the ROC curve to obtain the area under the ROC curve,and calculate the diagnostic sensitivity,specificity,positive predictive value,negative predictive value and the accuracy.A difference of P<0.05 is considered statistically significant.Results A total of 122 patients,96 males and 26 females,were included in this study.The mean age was 66 years,and there were 19 cases of multiple bladder tumors and 103 cases of solitary tumors.The pathological diagnosis was MIBC in 33 cases and NMIBC in 89 cases.There was good agreement between the two observers with ICC value of 0.94 and 95%confidence interval(0.91,0.96).<5.4%MIBC was present in VI-RADS 1 and 2 scores.For VI-RADS scores 3,4,and 5,the proportion of MIBC was 28.6%(2/7),47.1%(8/17)and 80%(20/25)in VI-RADS 3,4,and 5 lesions,respectively.When the threshold VI-RADS≥3 was set to MIBC,thesensitivity,specificity,PPV,NPV and accuracy of observer 1 and 2 were 91%,80%,63%,96%,83%,and 91%,76%,59%,96%,80%,respectively.If the cutoff point defining MIBC is shifted to VI-RADS≥4,the sensitivity,specificity,PPV,NPV and accuracy of observer 1 and 2 were 85%,85%,68%,94%,85%and 91%,81%,64%,96%and 83%,respectively.The area under the ROC curve for observers 1 and 2 was 0.89 and 0.91,respectively.Conclusion VI-RADS scores with overall good interobserver agreement are able to accurately identify MIBC and NMIBC.