Consistency of location and scoring in the diagnosis of clinically significant prostate cancer based on the Prostate Imaging Reporting and Data System version 2.1
Consistency of location and scoring in the diagnosis of clinically significant prostate cancer based on the Prostate Imaging Reporting and Data System version 2.1
Objective:To evaluate the consistency of radiologists with different years of experi-ence in the diagnosis of clinically significant prostate cancer(csPCa)using multi-parametric magnetic resonance imaging(mpMRI)based on the Prostate Imaging Reporting and Data System version 2.1(PI-RADSv2.1).Methods:A total of 583 patients who underwent prostate mpMRI scans with patho-logical results from January 2018 to February 2021 were retrospectively reviewed.All images were in-dependently analyzed by three radiologists with different levels of experience in prostate MRI:(A=8 years,B=6 years,C=3 years)according to PI-RADSv2.1.Kappa and Fleiss'Kappa tests were used to evaluate the consistency of the three radiologists'diagnoses of lesion location and score.Results:The consistency of the lesion location diagnosis among the three radiologists was moderate,with the con-sistency of the middle and high experience radiologists being higher than that of the low and high ex-perience radiologists.The Kappa values of pairwise comparison were moderate,0.63(A vs.B),0.51(B vs.C)and 0.45(A vs.C)respectively,and 0.52(overall comparison).The consistency of score diagno-sis among the three radiologists was moderate,with the Kappa values of pairwise comparisons being 0.70(A vs.B,moderate),0.76(B vs.C,good),and 0.64(A vs.C,moderate),whereas a lowest Kappa value of 0.43 for the overall comparison.The consistency of the diagnosis of lesion location and score of csPCa was statistically significant(P=0.000).Conclusion:The consistency of different levels of expe-rience radiologists in evaluating the location and score of csPCa using PI-RADS v2.1 is moderate.To improve the diagnostic consistency of lesion location and score,professional training in PI-RADS v2.1 still needs to be strengthened.
关键词
前列腺肿瘤/数据系统/多参数磁共振成像
Key words
Prostatic Neoplasms/Data Systems/Multiparametric Magnetic Resonance Imaging