Diamand Romain, Oderda Marco, Albisinni Simone, Fourcade Alexandre, Fournier Georges, Benamran Daniel, Iselin Christophe, Fiard Gaelle, Descotes Jean-Luc, Assenmacher Grégoire, Svistakov Ilyas, Peltier Alexandre, Simone Giuseppe, Di Cosmo Giacomo, Roche Jean-Baptiste, Bonnal Jean-Louis, Van Damme Julien, Rossi Maxime, Mandron Eric, Gontero Paolo, Roumeguère Thierry
Urology Department, Hôpital Erasme, University Clinics of Brussels, Université Libre de Bruxelles, Brussels, Belgium.
Urology Department, Città della Salute e della Scienza di Torino, University of Turin, Turin, Italy.
Urol Oncol. 2020 Nov;38(11):847.e9-847.e16. doi: 10.1016/j.urolonc.2020.04.011. Epub 2020 May 26.
To validate a nomogram predicting lymph node invasion (LNI) in prostate cancer patients undergoing radical prostatectomy taking into consideration multiparametric-magnetic resonance imaging (mp-MRI) parameters and targeted biopsies in a western European cohort.
A total of 473 men diagnosed by targeted biopsies, using software-based MRI-ultrasound image fusion system, and operated by radical prostatectomy with extended pelvic lymph node dissection across 11 Europeans centers between 2012 and 2019 were identified. Area under the curve of the receiver operator characteristic curve, calibration plot and decision curve analysis were used to evaluated the performance of the model.
Overall, 56 (11.8%) patients had LNI on final pathologic examination with a median (IQR) of 13 (9-18) resected nodes. Significant differences (all P < 0.05) were found between patients with and without LNI in terms of preoperative PSA, clinical stage at DRE and mp-MRI, maximum diameter of the index lesion, PI-RADS score, Grade Group on systematic and targeted biopsies, total number of dissected lymph nodes, final pathologic staging and Grade Group. External validation of the prediction model showed a good accuracy with an area under the curve calculated as 0.8 (CI 95% 0.75-0.86). Graphic analysis of calibration plot and decision curve analysis showed a slight underestimation for predictive probability for LNI between 3% and 22% and a high net benefit. A cut-off at 7% was associated with a risk of missing LNI in 2.6%, avoiding unnecessary surgeries in 55.9%.
We report an external validation of the nomogram predicting LNI in patients treated with extended pelvic lymph node dissection in a western European cohort and a cut-off at 7% seems appropriate.
在一个西欧队列中，考虑多参数磁共振成像（mp-MRI）参数和靶向活检，验证用于预测接受根治性前列腺切除术的前列腺癌患者淋巴结侵犯（LNI）的列线图。
通过基于软件的MRI-超声图像融合系统经靶向活检确诊，并于2012年至2019年间在11个欧洲中心接受根治性前列腺切除术及扩大盆腔淋巴结清扫术的473名男性被纳入研究。采用受试者操作特征曲线下面积、校准图和决策曲线分析来评估模型的性能。
总体而言，56名（11.8%）患者在最终病理检查中发现有LNI，切除淋巴结的中位数（IQR）为13个（9 - 18个）。有LNI和无LNI的患者在术前前列腺特异性抗原、直肠指诊和mp-MRI时的临床分期、索引病灶最大直径、PI-RADS评分、系统活检和靶向活检的分级组、清扫淋巴结总数、最终病理分期和分级组方面存在显著差异（均P < 0.05）。预测模型的外部验证显示准确性良好，曲线下面积计算为0.8（95%CI 0.75 - 0.86）。校准图的图形分析和决策曲线分析显示，LNI预测概率在3%至22%之间略有低估，净效益较高。7%的截断值与2.6%的漏诊LNI风险相关，可避免55.9%的不必要手术。
我们报告了在一个西欧队列中对接受扩大盆腔淋巴结清扫术患者预测LNI的列线图的外部验证，7%的截断值似乎合适。
