摘要目的:探讨前列腺成像报告和数据系统（PI-RADS）评分与前列腺癌（PCa）国际泌尿病理协会（ISUP）分级之间的相关性以及PI-RADS评分在预测临床有意义PCa（csPCa）、切缘阳性、病理升级等病理特征方面的作用。方法:将2013年5月至2019年12月间在苏北人民医院泌尿外科行多参数磁共振（mpMRI）病理为阳性患者纳入研究，根据病理检查方式分为活检组（523例）和根治性前列腺切除术（RP）组（215例），评估两组PI-RADS评分与ISUP分级之间的相关性以及预测csPCa的准确性，对RP组进一步探讨PI-RADS评分与术后病理升级或降级情况和切缘阳性之间的相关性，以及受试者工作特征（ROC）曲线下与坐标轴围成的面积（AUC）。结果:活检组和RP组的年龄分别为（72±8）岁和（68±7）岁，前列腺体积分别为48.3（32~57）cm 3和47.2（32~54）cm 3，前列腺特异抗原（PSA）水平分别为26.3（10.2~34.2）μg/L和21.7（9.24~23.95）μg/L；活检组PI-RADS评分≤3、4、5分别为109例、97例、317例，RP组为61例、55例、99例；活检组和RP组不同PI-RADS评分的ISUP分级组成存在差异，且差异有统计学意义（ P均<0.001），两者之间呈正相关（活检组 r=0.493，RP组 r=0.671， P均<0.001）；利用PI-RADS评分预测csPCa，活检组（AUC=0.764， P<0.001，95 %CI：0.710~0.819）和RP组（AUC=0.807， P<0.001，95 %CI：0.735~0.879）均有一定的准确性，结合PSA，可以提高活检组（AUC=0.795， P<0.001，95 %CI：0.746~0.843）和RP组（AUC=0.852， P<0.001，95 %CI：0.789~0.915）对csPCa预测的准确性。RP组术后病理结果与穿刺病理结果相比，52.6%的患者出现了ISUP分级的变化，在升级和降级的患者中PI-RADS评分组成差异无统计学意义（ P>0.05），但是在活检为ISUP 1级的患者中41.7%（27/65）出现了病理升级，PI-RADS≤3分占33.3%，而PI-RADS>3分占66.7%，两组之间差异有统计学意义（ P<0.05）。术后43.3%的患者切缘阳性，PI-RADS评分≤3、4、5分的患者分别为13例（14%）、24例（25.8%）、56例（60.2%），而切缘阴性患者的PI-RADS评分≤3、4、5分分别为48例（39.3%）、31例（25.4%）、43例（35.2%），两者差异有统计学意义（ P<0.001），PI-RADS评分越高，切缘阳性可能性越大。 结论:PI-RADS评分与PCa的ISUP分级存在相关性，联合PSA可较为准确地预测csPCa；同时PI-RADS评分越高，RP后切缘阳性和穿刺时Gleason评分3+3=6分的患者病理升级的可能性越大。

abstractsObjective:To explore the correlation between prostate imaging report and data system (PI-RADS) score and international society of uological pathology (ISUP) grade of prostate cancer (PCa) and the role of PI-RADS score in predicting the pathological features of clinically significant PCa (csPCa), positive surgical margin and pathological upgrade.Methods:The pathologically positive patients with multi-parameter magnetic resonance image (mpMRI) were included in this study. The patients with prostate specific antigen (PSA)<100 μg/L were divided into two groups: biopsy group ( n=523) and RP group ( n=215). The correlation between PI-RADS score and ISUP grade and the accuracy of predicting csPCa in the two groups were evaluated. In the RP group, the correlation between PI-RADS score and postoperative pathological grade or degradation and positive incisal margin was further discussed. The patients with PSA≥100 μg/L (171cases in biopsy group and 6 cases in RP group) were not included in the statistical analysis, and the results were simply described. Results:The age, prostate volume, and PSA level of biopsy group and RP group was (72±8) years vs (68±7) years, 48.3 (32-57) cm 3 vs 47.2 (32-54) cm 3, and 26.3(10.2-34.2)μg/L vs 21.7 (9.24-23.95)μg/L, respectively. The PI-RADS scores ≤ 3,4, and 5 in the biopsy group were 109,97, and 317 respectively, and those in the RP group were 61,55, and 99 respectively. There were significant differences in the composition of ISUP grades of different PI-RADS scores between the two groups ( P<0.001), and there was a positive correlation between the two groups ( r=0.493 in the biopsy group, r=0.671 in the RP group, both P<0.001). Using PI-RADS score to predict csPCa, biopsy group (AUC=0.764, P<0.001, 95 %CI:0.710-0.819) and RP group (AUC=0.807, P<0.001, 95 %CI:0.735-0.879) had certain accuracy. The PI-RADS score combined with PSA could improve the accuracy of csPCa prediction in the biopsy group (AUC=0.795, P<0.001, 95 % CI:0.746-0.843) and the RP group (AUC=0.852, P<0.001, 95 %CI:0.789-0.915). Compared with the pathological results of biopsy in the RP group, 52.6% of the patients showed upgrade and degrade of ISUP, and there was insignificant difference in the composition of PI-RADS scores between upgraded and degraded patients ( P>0.05). However, 41.7%(27/65) of the patients with ISUP grade 1 biopsies had pathological upgrades that the patients with PI-RADS ≤ 3 accounted for 33.3%, while the patients with PI-RADS>3 accounted for 66.7%, and there was significant difference between the two groups ( P<0.05). After RP, 43.3% of the patients had positive surgical margins, and the patients with PI-RADS score ≤ 3, 4 and 5 were 13 (14%), 24 (25.8%) and 56 (60.2%), respectively, while the PI-RADS scores of patients with negative surgical margin were 48 (39.3%), 31(25.4%) and 43(35.2%), respectively. There was significant difference between the two groups ( P<0.001). The higher the PI-RADS score, the greater the possibility of the positive surgical margin. For the patients with PSA ≥ 100 μg/L, 98.8% (169/171) patients in the biopsy group had a PI-RADS score 5. The pathological results of all patients were csPCa, of which 85.4% (146/171) had ISUP grade ≥ 4. Among them, 6 cases underwent RP, 5 cases had ISUP grade ≥ 4, all surgical margin were positive, 5 cases had seminal vesicle invasion, 3 cases had capsule invasion and 3 cases had positive pelvic lymph nodes. Conclusion:ThePI-RADS score is correlated with the ISUP grade of PCa. Combined with PSA can accurately predict csPCa. At the same time, the higher PI-RADS score, the more likely the patients with positive incisal margin after RP and Gleason score of 3+3=6 at the time of puncture will be upgraded pathologically.