Objective:To determine the interobserver reproducibility of the Prostate Imaging Reporting and Data System(PI-RADS)version 1 and version 2.To compare the diagnostic accuracy of PI-RADS v1 and PI-RADS v2 in the detection of clinically significant prostate cancer(csPCa).Materials and Methods:Between January 2015 and April 2018,a total of 165 consecutive patients underwent mpMRI before biopsy in our institution.Multiparametric MRI including T2-weighted imaging,diffusion-weighted imaging(DWI)and dynamic contrast-enhanced imaging(DCE)was acquired according to ESUR guidelines on a 3 Tesla system using phased-array pelvic coil.The examinations were performed before puncture,operation,frozen,endocrine and radiation related treatment.Two independent readers evaluated the likelihood of csPCa according to PI-RADS v1 and PI-RADS v2.The reader was blinded to biopsy results.The csPCa was defined as Gleason score ≥ 7.The lesions were categorized into transitional zone(tz)and peripheral zone(pz)lesions.The agreement between the inter-observer agreement for each version was determined using Cohen’s kappa statistics.The agreement was defined excellent(κ>0.81),good(κ=0.61-0.80),moderate(κ=0.41-0.60),fair(κ=0.21-0.40)and poor(κ≤0.20).If the diagnosis between the two physicians is highly consistent,Receiver operating characteristic(ROC)curve analysis for PI-RADS v1 and v2 was performed for reader one as well as for the tz and pz,using histopathological results as the gold standard.Youden-selected thresholds were determined.Finally,Sensitivity,specificity,negative predictive value(NPV),positive predictive value(PPV)and accuracy at the threshold were recorded.The level of significance was set at 0.05.Receiver operating characteristic(ROC)curve analysis for PI-RADS v1 and v2 was performed separately for each reader as well as for the tz and pz,when there is poor consistency between the two readers.For all statistical analyses SPSS for Windows version 22.0(SPSS,Chicago,Ill)and MedCalc V16.2 was used.Results:Concordance between the examiners for PI-RADS v1 and for V2 showed substantial agreement(version 1:weighted kappa 0.618;version 2:weighted kappa 0.817).A total of 169 patients were enrolled in the study(92 lesions in peripheral zone(pz)and 77 lesions in transitional zone(tz)),including 94 cases Pca(including Gleason≥ 3+4,clinically significant cancer,and Gleason=3+3),of which 63 cases of peripheral zone PCa and 31 cases of transitional zone PCa;There were 76 cases of significant clinical carcinoma,including 53 cases of peripheral zone significant clinical PCa and 23 cases of transitional zone significant clinical PCa.For peripheral zone lesions,PI-RADS v2 showed a signifcantly larger discrimination ability of clinically significant prostate cancer:V1 AUC 0.881(95%CI:0.797-0.939)and V2 AUC 0.939(95%CI:0.869-0.978),(Z=2.691,P=0.0071).For transition zone lesions,PI-RADS v1 showed more discrimination:V1 AUC 0.935(95%CI:0.855-0.978)and V20.911(95%CI:0.824-0.964),but the difference was not signifcant(Z=0.747,P=0.4448).With PI-RADS v2 scores best diagnostic accuracy in the pz was reached with a cut-off>3 at a Youden-Index of 0.8219(sensitivity 92.45%,specificity 89.74%,positive prediction92.5%,negative prediction values89.7%),while in the tz a cut-off>3 revealed diagnostic accuracy(sensitivity 86.96%,specificity 83.33%,positive prediction69.0%,negative prediction values93.7%,Youden-Index 0.7029).With PI-RADS v1 scores best diagnostic accuracy in the pz was reached with a cut-off>3 at a Youden-Index of 0.7842(sensitivity 88.68%,specificity 89.74%,positive prediction92.2%,negative prediction values85.4%),while in the tz a cut-off>3 revealed highest diagnostic accuracy(sensitivity 95.65%,specificity 77.78%,positive prediction 64.7%,negative prediction values 97.7%,Youden-Index 0.7343).Conclusion:The inter-reader agreement for PI-RADS v1 was good,PI-RADS v2 achieved excellent reproducibility.PI-RADS v2 demonstrated better accuracy in detecting PZ csPCa,both versions revealed high diagnostic accuracy for detection of CZ csPCa.PI-RADS v2 could be more practicable for clinical routine. |