摘要目的:探讨双参数磁共振成像（bpMRI）定位解剖分区预测前列腺癌危险度的价值。方法:回顾性收集2017年1月至2021年12月在空军军医大学第一附属医院经根治性前列腺切除术后病理证实前列腺癌患者92例。所有患者均行bpMRI扫描，扫描序列包括平扫和扩散加权成像（DWI）。按照2014年国际泌尿系统病理学学会（ISUP）分级对患者进行分组：将≤2 级定义为低危组，共纳入26例，年龄71（64.0，75.2）岁；将≥3级定义为高危组，共纳入66例，年龄70.5（63.0，74.0）岁。采用组内相关系数（ICC）评价测量表现扩散系数（ADC）值的观察者间一致性。比较两组总前列腺特异性抗原（tPSA）差异，比较移行区和外周区前列腺癌危险度差异。以前列腺癌高低危险度作为因变量，纳入解剖分区、tPSA、ADC平均值（ADC mean值）、ADC最小值（ADC min值）和年龄等因素，采用多因素logistic回归分析前列腺癌高低危险度发生的相关因素。绘制受试者工作特征（ROC）曲线评估解剖分区、tPSA及解剖分区+tPSA的联合模型诊断前列腺癌危险度的效能。 结果:测量ADC mean值和ADC min值的观察者间ICC分别为0.906和0.885，一致性良好。低危组tPSA低于高危组tPSA［19.64（10.29，35.18）ng/ml比72.42（24.79，187.98）ng/ml， P<0.001］；外周区前列腺癌危险度高于移行区前列腺癌危险度（ P<0.01）。多因素logistic回归分析显示解剖分区（ OR=0.120，95% CI：0.029~0.501）和tPSA（ OR=1.059，95% CI：1.022~1.099）是前列腺癌高低危险度发生的相关因素（均 P<0.05）。联合模型（AUC=0.895，95% CI：0.831~0.958）的诊断效能均优于解剖分区（AUC=0.717，95% CI：0.597~0.837）和tPSA（AUC=0.801，95% CI：0.714~0.887）的单一模型预测效能（ Z=3.91，2.47，均 P<0.05）。 结论:外周区前列腺癌恶性程度高于移行区者。bpMRI定位解剖分区联合tPSA预测前列腺癌危险度效能最高，有望为患者制定个性化治疗策略提供支持。

abstractsObjective:To investigate the anatomic zone localization based on biparametric magnetic resonance imaging (bpMRI) for the prediction of the risk degree in patients with prostate cancer.Methods:A total of 92 patients with prostate cancer confirmed by radical surgery in First Affiliated Hospital, Air Force Medical University, from January 2017 to December 2021 were collected. All patients underwent bpMRI (non-enhanced scan and DWI). According to ISUP grade, those patients were divided into low-risk group [≤grade 2, n=26, aged 71 (64.0, 5.2) years] and high-risk group[≥grade 3, n=66, aged 70.5 (63.0, 74.0) years]. The interobserver consistency test for ADC values was evaluated using the intraclass correlation coefficients (ICC). The differences in total prostate specific antigen (tPSA) between the two groups were compared and the χ 2 test was used to compare the differences in the risk of prostate cancer in the transitional and peripheral zone. Independent correlation factors for prostate cancer risk were analyzed by logistic regression using high and low risk of prostate cancer as dependent variables, including factors such as anatomical zone, tPSA, apparent diffusion coefficient mean (ADC mean), apparent diffusion coefficient minimum (ADC min) and age. Receiver operating characteristic (ROC) curves were plotted to assess the efficacy of the combined models of anatomical zone, tPSA, and anatomical partitioning+tPSA for diagnosing prostate cancer risk. Results:The ICC values of the ADC mean and ADC min between the observers were 0.906 and 0.885, respectively, with good agreement. The tPSA in the low-risk group was lower than that in the high-risk group [19.64 (10.29, 35.18) ng/ml vs 72.42 (24.79, 187.98) ng/ml; P<0.001]; the risk of prostate cancer in the peripheral zone was higher than that in the transitional zone, and the difference was statistically significant ( P<0.01). Multifactorial regression showed that anatomical zones ( OR=0.120, 95% CI:0.029-0.501, P=0.004) and tPSA ( OR=1.059, 95% CI:1.022-1.099, P=0.002) were risk factors for prostate cancer risk. The diagnostic efficacy of the combined model (AUC=0.895, 95% CI: 0.831-0.958) was better than the predictive efficacy of the single model for both anatomical partitioning (AUC=0.717, 95% CI:0.597-0.837) and tPSA (AUC=0.801, 95% CI: 0.714-0.887) ( Z=3.91, 2.47; all P<0.05). Conclusions:The malignant degree of prostate cancer in peripheral zone was higher than that in transitional zone. Combination of anatomic zone located by bpMRI and tPSA can be used to predict the risk of prostate cancer before surgery, expected to provide support for patients to develop personalized treatment strategies.