Department of Urology, Jules Bordet Institute-Erasme Hospital, Hôpital Universitaire de Bruxelles, Université Libre de Bruxelles, Brussels, Belgium.
Department of Urology, Vivantes Klinikum Am Urban, Berlin, Deutschland.
World J Urol. 2024 May 15;42(1):322. doi: 10.1007/s00345-024-04962-x.
Utility of prostate-specific antigen density (PSAd) for risk-stratification to avoid unnecessary biopsy remains unclear due to the lack of standardization of prostate volume estimation. We evaluated the impact of ellipsoidal formula using multiparametric magnetic resonance (MRI) and semi-automated segmentation using tridimensional ultrasound (3D-US) on prostate volume and PSAd estimations as well as the distribution of patients in a risk-adapted table of clinically significant prostate cancer (csPCa).
In a prospectively maintained database of 4841 patients who underwent MRI-targeted and systematic biopsies, 971 met inclusions criteria. Correlation of volume estimation was assessed by Kendall's correlation coefficient and graphically represented by scatter and Bland-Altman plots. Distribution of csPCa was presented using the Schoots risk-adapted table based on PSAd and PI-RADS score. The model was evaluated using discrimination, calibration plots and decision curve analysis (DCA).
Median prostate volume estimation using 3D-US was higher compared to MRI (49cc[IQR 37-68] vs 47cc[IQR 35-66], p < 0.001). Significant correlation between imaging modalities was observed (τ = 0.73[CI 0.7-0.75], p < 0.001). Bland-Altman plot emphasizes the differences in prostate volume estimation. Using the Schoots risk-adapted table, a high risk of csPCa was observed in PI-RADS 2 combined with high PSAd, and in all PI-RADS 4-5. The risk of csPCa was proportional to the PSAd for PI-RADS 3 patients. Good accuracy (AUC of 0.69 and 0.68 using 3D-US and MRI, respectively), adequate calibration and a higher net benefit when using 3D-US for probability thresholds above 25% on DCA.
Prostate volume estimation with semi-automated segmentation using 3D-US should be preferred to the ellipsoidal formula (MRI) when evaluating PSAd and the risk of csPCa.
由于前列腺体积估计缺乏标准化，前列腺特异性抗原密度（PSAd）在风险分层以避免不必要的活检中的效用仍不清楚。我们评估了使用多参数磁共振（MRI）的椭球公式和使用三维超声（3D-US）的半自动分割对前列腺体积和 PSAd 估计以及在临床显著前列腺癌（csPCa）的适应风险表中患者分布的影响。
在一项前瞻性维护的 4841 例接受 MRI 靶向和系统活检的患者数据库中，有 971 例符合纳入标准。通过 Kendall 相关系数评估体积估计的相关性，并通过散点图和 Bland-Altman 图直观表示。使用基于 PSAd 和 PI-RADS 评分的 Schoots 适应风险表呈现 csPCa 的分布。使用鉴别力、校准图和决策曲线分析（DCA）评估模型。
与 MRI 相比，使用 3D-US 估计的前列腺体积中位数更高（49cc[IQR 37-68] 比 47cc[IQR 35-66]，p<0.001）。观察到两种成像方式之间存在显著相关性（τ=0.73[CI 0.7-0.75]，p<0.001）。Bland-Altman 图强调了前列腺体积估计的差异。使用 Schoots 适应风险表，PI-RADS 2 合并高 PSAd 和所有 PI-RADS 4-5 均观察到 csPCa 的高风险。PI-RADS 3 患者的 csPCa 风险与 PSAd 成正比。在 DCA 中，当使用概率阈值大于 25%时，3D-US 具有良好的准确性（使用 3D-US 和 MRI 的 AUC 分别为 0.69 和 0.68）、适当的校准和更高的净收益。
在评估 PSAd 和 csPCa 的风险时，应优先使用 3D-US 半自动分割的前列腺体积估计，而不是椭球公式（MRI）。
