本研究旨在开发和验证一种基于双参数磁共振成像（bpMRI）的模型，用于检测未接受活检的患者中临床意义显著的前列腺癌（csPCa）。这项回顾性研究纳入了324名接受bpMRI和MRI针对性融合活检（MRGB）和/或系统活检的患者，其中217名患者随机分配到训练组，107名患者分配到验证组。我们评估了三种基于bpMRI的评分的敏感性和特异性的诊断性能。随后，我们构建了3个模型（模型1、模型2和模型3），将bpMRI评分与临床变量结合起来，并使用受试者工作特征曲线（ROC曲线）下的面积（AUC）进行比较。我们使用DeLong检验评估了这些模型之间的差异的统计学意义。在训练组中，217名患者中有68名获得了病理学上证实的csPCa。评分1的敏感性和特异性分别为64.7%（95% CI 52.2%-75.9%）和80.5%（95% CI 73.3%-86.6%）；评分2的敏感性和特异性分别为86.8%（95% CI 76.4%-93.8%）和73.2%（95% CI 65.3%-80.1%）；评分3的敏感性和特异性分别为61.8%（95% CI 49.2%-73.3%）和80.5%（95% CI 73.3%-86.6%）。多变量回归分析显示，基于bpMRI、年龄和前列腺特异性抗原密度（PSAD）的评分是csPCa的独立预测因子。三个模型的AUC分别为0.88（95% CI 0.83-0.93）、0.90（95% CI 0.85-0.94）和0.88（95% CI 0.83-0.93）。模型2的性能显著优于模型1（P = 0.03）和模型3（P < 0.01）。所有三种评分均具有较好的诊断准确性。与年龄和PSAD结合使用时，预测能力显著改善，其中基于评分2的模型显示出最高性能。 © 2023年。国际癌症成像学会（ICIS）。

This study aimed to develop and validate a model based on biparametric magnetic resonance imaging (bpMRI) for the detection of clinically significant prostate cancer (csPCa) in biopsy-naïve patients.This retrospective study included 324 patients who underwent bpMRI and MRI targeted fusion biopsy (MRGB) and/or systematic biopsy, of them 217 were randomly assigned to the training group and 107 were assigned to the validation group. We assessed the diagnostic performance of three bpMRI-based scorings in terms of sensitivity and specificity. Subsequently, 3 models (Model 1, Model 2, and Model 3) combining bpMRI scorings with clinical variables were constructed and compared with each other using the area under the receiver operating characteristic (ROC) curves (AUC). The statistical significance of differences among these models was evaluated using DeLong's test.In the training group, 68 of 217 patients had pathologically proven csPCa. The sensitivity and specificity for Scoring 1 were 64.7% (95% CI 52.2%-75.9%) and 80.5% (95% CI 73.3%-86.6%); for Scoring 2 were 86.8% (95% CI 76.4%-93.8%) and 73.2% (95% CI 65.3%-80.1%); and for Scoring 3 were 61.8% (95% CI 49.2%-73.3%) and 80.5% (95% CI 73.3%-86.6%), respectively. Multivariable regression analysis revealed that scorings based on bpMRI, age, and prostate-specific antigen density (PSAD) were independent predictors of csPCa. The AUCs for the 3 models were 0.88 (95% CI 0.83-0.93), 0.90 (95% CI 0.85-0.94), and 0.88 (95% CI 0.83-0.93), respectively. Model 2 showed significantly higher performance than Model 1 (P = 0.03) and Model 3 (P < 0.01).All three scorings had favorite diagnostic accuracy. While in conjunction with age and PSAD the prediction power was significantly improved, and the Model 2 that based on Scoring 2 yielded the highest performance.© 2023. International Cancer Imaging Society (ICIS).