摘要目的:探讨 18F-前列腺特异性膜抗原(PSMA)PET/CT PRIMARY评分联合多参数磁共振(mpMRI)前列腺影像报告和数据系统(PI-RADS)评分对有临床意义前列腺癌(CsPCa)的诊断价值。 方法:回顾性分析2019年1月至2023年12月北京医院收治的63例行根治性前列腺切除术的前列腺癌患者的病例资料。中位年龄70(64，75)岁，前列腺特异性抗原(PSA)8.46(5.40，14.80)ng/ml。所有患者术前均行 18F-PSMA PET/CT和mpMRI检查，术后前列腺标本制作病理大切片。2名影像科医生和1名病理科医生分别对前列腺病灶进行诊断和定位。根据病理大切片结果，Gleason评分≥3＋4分诊断为CsPCa，Gleason评分＝6分或阴性为非CsPCa。PSMA PET/CT图像采用PRIMARY评分(5分法)：1分为无以下对应的模式，2分为移行带或中央带非局灶性高摄取，3分为移行带局灶性高摄取，4分为外周带局灶性高摄取，5分为SUV max≥12.0。以前列腺病灶SUV max/邻近SUV max比值(LBR)半定量评价其对显像剂摄取的程度。mpMRI图像采用PI-RADS Version2.1进行评分。以患者作为研究对象比较CsPCa组和非CsPCa组的年龄、PSA、病灶数量的差异；以病灶作为研究对象比较CsPCa组和非CsPCa组的PRIMARY评分、LBR、PI-RADS评分、病灶分布构成的差异。采用多因素logistic回归分析CsPCa的独立预测因素。采用受试者工作特征(ROC)曲线确定各独立预测因素的最佳诊断阈值，分别构建PRIMARY评分、PI-RADS评分、二者联合应用的预测模型，比较各模型对CsPCa的诊断效能。 结果:本研究63例，CsPCa组54例(85.7%)，非CsPCa组9例(14.3%)，PSA分别为9.64(6.1，15.3)ng/ml和5.6(4.6，7.6)ng/ml，差异有统计学意义( P<0.05)，年龄[71(64，75)岁与65(63，69)岁]、病灶数量[2(1，2)个与2(1，3)个]差异均无统计学意义( P>0.05)。CsPCa组81个病灶，Gleason评分3＋4分49个、4＋3分16个、8分14个、9～10分2个。非CsPCa组28个病灶，良性前列腺病变14个，Gleason评分3＋3分14个。CsPCa组和非CsPCa组的PRIMARY评分分别为4(3，5)分和2(1，4)分、LBR分别为2.69(2.08，4.48)和1.89(1.45，2.48)，PI-RADS评分分别为4(3，5)分和2(2，3)分，差异均有统计学意义( P<0.05)。CsPCa组和非CsPCa组病灶位于移行带分别为15个(18.5%)和8个(28.6%)，位于外周带分别为66个(81.5%)和20个(71.4%)，差异无统计学意义( P>0.05)。多因素logistic回归分析结果显示，PRIMARY评分( OR＝2.134，95% CI 1.429～3.187， P<0.01)和PI-RADS评分( OR＝2.689，95% CI 1.618～4.469， P<0.01)是诊断CsPCa的独立预测因素。ROC曲线显示，PRIMARY评分和PI-RADS评分诊断CsPCa的最佳阈值均为3分。PRIMARY评分、PI-RADS评分、二者联合诊断CsPCa的准确性分别为72%、67%、83%，敏感性分别为72%、63%、91%，特异性分别为75%、79%、57%，阳性预测值分别为89%、89%、86%，阴性预测值分别为48%、42%、70%。PRIMARY评分、PI-RADS评分、二者联合预测CsPCa的ROC曲线的曲线下面积分别为0.733(95% CI 0.624～0.842)、0.708(95% CI 0.599～0.817)、0.743(95% CI 0.623～0.862)，二者联合的诊断效能高于单独应用PRIMARY评分或PI-RADS评分(均 P<0.01)。 结论:18F-PSMA PET/CT PRIMARY评分和mpMRI PI-RADS评分对CsPCa均具有良好的诊断价值，二者联合应用可提高诊断的准确性、敏感性和阴性预测值，诊断效能更高。

abstractsObjective:To explore the diagnostic value of 18F-prostate specific membrane antigen (PSMA) PET/CT PRIMAY score combined with multiparameter MRI (mpMRI) PI-RADS score for clinically significant prostate cancer (CsPCa). Methods:The data of 63 patients with prostate cancer who underwent radical prostatectomy at Beijing Hospital from January 2019 to December 2023 were retrospectively analyzed. The median age was 70 (64, 75) years old with prostate-specific antigen (PSA) level of 8.46 (5.40, 14.80) ng/ml. All patients underwent 18F-PSMA PET/CT and mpMRI examination before surgery, and pathological large sections of prostate specimens were made after surgery. The prostate lesions were diagnosed and located by two radiologists and one pathologist respectively. Lesions with Gleason scores (GS)≥3+ 4 from the surgical pathology were diagnosed with CsPCa, and lesions with negative or GS=6 were diagnosed with non-CsPCa. The PSMA PET/CT images were evaluated using the PRIMARY study criteria (5-level PRlMARY score): no pattern (score of 1), diffuse transition zone or central zone(not focal) (score of 2), focal transition zone(score of 3), focal peripheral zone(score of 4), or an SUV max of at least 12 (score of 5). The degree of uptake of imaging agent in prostate lesions was semi-quantitatively evaluated using lesion-to-background ratios (LBR) of SUV max. MpMRI was evaluated according to the Prostate Imaging Reporting and Data System (PI-RADS) version 2.1. The patients were divided into CsPCa group and non-CsPCa group based on patients and lesions. Mann-Whitney U test and chi-square test were used to compare the differences between groups. Multivariate logistic regression analysis was performed to determine the independent predictive factors of CsPCa. Receiver operator characteristic (ROC) curve was used to determine the optimal diagnostic threshold for each independent predictor. Predictive models were constructed for PRIMARY score, PI-RADS score, and their combined application, and the diagnostic performance of each model for CsPCa was compared. Results:Of all 63 patients, there were 54 cases in CsPCa group (85.7%) and 9 cases in non-CsPCa group (14.3%).There was significant difference between CsPCa group and non-CsPCa group in the serum PSA level [9.64 (6.1, 15.3) ng/ml vs. 5.6 (4.6, 7.6) ng/ml]( P<0.05). There was no statistically significant difference in age [71 (64, 75) years vs. 65 (63, 69) years], and number of lesions [2 (1, 2) vs. 2 (1, 3)] (all P>0.05). Of all 109 lesions, there were 81 lesions in CsPCa group(including 49 lesions with Gleason score = 3+ 4, 16 lesions with Gleason score=4+ 3, 14 lesions with Gleason score = 8, and 2 lesions with Gleason score>8) and 28 lesions in non-CsPCa group(including 14 lesions with Gleason score = 3+ 3 and 14 with benign prostate lesions). There was significant difference between CsPCa group and non-CsPCa group in PRIMARY score [4 (3, 5) vs. 2 (1, 4)], LBR [2.69 (2.08, 4.48) vs. 1.89 (1.45, 2.48)], PI-RADS score [4 (3, 5) vs. 2 (2, 3)] (all P<0.05). There was no statistically significant difference in the lesion distribution including the number of lesions located in the transition zone [15(18.5%) vs. 8(28.6%)] and in the peripheral zone[66(81.5%) vs. 20(71.4%)]( P>0.05). Multivariate logistic regression analysis indicated that PRIMARY score ( OR=2.134, 95% CI 1.429-3.187) and PI-RADS score ( OR=2.689, 95% CI 1.618-4.469) were independent predictors of CsPCa (both P<0.01). ROC curves analysis revealed that the cut-off value for diagnosing CsPCa was both 3 for PRIMARY score and PI-RADS score. The accuracy for PRIMARY score, PI-RADS score, and their combined complication in diagnosing CsPCa was 72%, 67%, and 83%, respectively. The sensitivity was 72%, 63%, and 91%, and the specificity was 75%, 79%, and 57%, respectively. The positive predictive value was 89%, 89%, and 86%, and the negative predictive value was 48%, 42%, and 70%, respectively. The area under the curve of the PRIMARY score, PI-RADS score, and their combined complication of the ROC curve for CsPCa were 0.733 (95% CI 0.624-0.842), 0.708 (95% CI 0.599-0.817), and 0.743 (95% CI 0.623-0.862), respectively. The diagnostic efficacy of their combined complication was higher than PRIMARY score or PI-RADS score alone (both P<0.01). Conclusions:Both the 18F-PSMA PET/CT PRIMAY score and the mpMRI PI-RADS score have good diagnostic value for CsPCa. The combined application of the two imaging parameters can improve the accuracy, sensitivity, and negative predictive value, which have a higher diagnostic efficiency of CsPCa.