Departments of Ultrasound, The Second Affiliated Hospital of Soochow University, Suzhou, Jiangsu, China.
Departments of Urology, The Second Affiliated Hospital of Soochow University, Suzhou, Jiangsu, China.
BMC Urol. 2022 Apr 19;22(1):64. doi: 10.1186/s12894-022-01013-8.
To evaluate the diagnostic performance of multiparametric transrectal ultrasound (TRUS) and to design diagnostic scoring systems based on four modes of TRUS to predict peripheral zone prostate cancer (PCa) and clinically significant prostate cancer (csPCa).
A development cohort involved 124 nodules from 116 patients, and a validation cohort involved 72 nodules from 67 patients. Predictors for PCa and csPCa were extracted to construct PCa and csPCa models based on regression analysis of the development cohort. An external validation was performed to assess the performance of models using area under the curve (AUC). Then, PCa and csPCa diagnostic scoring systems were established to predict PCa and csPCa. The diagnostic accuracy was compared between PCa and csPCa scores and PI-RADS V2, using receiver operating characteristics (ROC) and decision curve analysis (DCA).
Regression models were established as follows: PCa = - 8.284 + 4.674 × Margin + 1.707 × Adler grade + 3.072 × Enhancement patterns + 2.544 × SR; csPCa = - 7.201 + 2.680 × Margin + 2.583 × Enhancement patterns + 2.194 × SR. The PCa score ranged from 0 to 6 points, and the csPCa score ranged from 0 to 3 points. A PCa score of 5 or higher and a csPCa score of 3 had the greatest diagnostic performance. In the validation cohort, the AUC for the PCa score and PI-RADS V2 in diagnosing PCa were 0.879 (95% confidence interval [CI] 0.790-0.967) and 0.873 (95%CI 0.778-0.969). For the diagnosis of csPCa, the AUC for the csPCa score and PI-RADS V2 were 0.806 (95%CI 0.700-0.912) and 0.829 (95%CI 0.727-0.931).
The multiparametric TRUS diagnostic scoring systems permitted better identifications of peripheral zone PCa and csPCa, and their performances were comparable to that of PI-RADS V2.
评估多参数经直肠超声（TRUS）的诊断性能，并设计基于 TRUS 四种模式的诊断评分系统，以预测外周区前列腺癌（PCa）和临床显著前列腺癌（csPCa）。
建立了一个包含 116 例患者 124 个结节的开发队列，以及一个包含 67 例患者 72 个结节的验证队列。从开发队列的回归分析中提取预测 PCa 和 csPCa 的指标，建立基于 PCa 和 csPCa 模型的 PCa 和 csPCa 模型。使用曲线下面积（AUC）进行外部验证，以评估模型的性能。然后，建立 PCa 和 csPCa 诊断评分系统，以预测 PCa 和 csPCa。使用接受者操作特征（ROC）和决策曲线分析（DCA）比较 PCa 和 csPCa 评分与 PI-RADS V2 的诊断准确性。
建立了以下回归模型：PCa=-8.284+4.674×边缘+1.707×Adler 分级+3.072×增强模式+2.544×SR；csPCa=-7.201+2.680×边缘+2.583×增强模式+2.194×SR。PCa 评分范围为 0 至 6 分，csPCa 评分范围为 0 至 3 分。PCa 评分≥5 分和 csPCa 评分≥3 分的诊断性能最佳。在验证队列中，PCa 评分和 PI-RADS V2 诊断 PCa 的 AUC 分别为 0.879（95%置信区间 [CI]0.790-0.967）和 0.873（95%CI0.778-0.969）。对于 csPCa 的诊断，csPCa 评分和 PI-RADS V2 的 AUC 分别为 0.806（95%CI0.700-0.912）和 0.829（95%CI0.727-0.931）。
多参数 TRUS 诊断评分系统可更好地识别外周区 PCa 和 csPCa，其性能与 PI-RADS V2 相当。
