Mazzone Elio, Gandaglia Giorgio, Ploussard Guillame, Marra Giancarlo, Valerio Massimo, Campi Riccardo, Mari Andrea, Minervini Andrea, Serni Sergio, Moschini Marco, Marquis Alessandro, Beauval Jean Baptiste, van den Bergh Roderick, Rahota Razvan-George, Soeterik Timo, Roumiguiè Mathieu, Afferi Luca, Zhuang Junlong, Guo Hongqian, Mattei Agostino, Gontero Paolo, Cucchiara Vito, Stabile Armando, Fossati Nicola, Montorsi Francesco, Briganti Alberto
Division of Oncology/Unit of Urology, URI, IRCCS Ospedale San Raffaele, Milan, Italy.
Division of Oncology/Unit of Urology, URI, IRCCS Ospedale San Raffaele, Milan, Italy.
Eur Urol. 2022 Feb;81(2):193-203. doi: 10.1016/j.eururo.2021.07.027. Epub 2021 Aug 13.
Despite the key importance of magnetic resonance imaging (MRI) parameters, risk classification systems for biochemical recurrence (BCR) in prostate cancer (PCa) patients treated with radical prostatectomy (RP) are still based on clinical variables alone.
We aimed at developing and validating a novel classification integrating clinical and radiological parameters.
DESIGN, SETTING, AND PARTICIPANTS: A retrospective multicenter cohort study was conducted between 2014 and 2020 across seven academic international referral centers. A total of 2565 patients treated with RP for PCa were identified.
Early BCR was defined as two prostate-specific antigen (PSA) values of ≥0.2 ng/ml within 3 yr after RP. Kaplan-Meier and Cox regressions tested time and predictors of BCR. Development and validation cohorts were generated from the overall patient sample. A model predicting early BCR based on Cox-derived coefficients represented the basis for a nomogram that was validated externally. Predictors consisted of PSA, biopsy grade group, MRI stage, and the maximum diameter of lesion at MRI. Novel risk categories were then identified. The Harrel's concordance index (c-index) compared the accuracy of our risk stratification with the European Association of Urology (EAU), Cancer of the Prostate Risk Assessment (CAPRA), and International Staging Collaboration for Cancer of the Prostate (STAR-CAP) risk groups in predicting early BCR.
Overall, 200 (8%), 1834 (71%), and 531 (21%) had low-, intermediate-, and high-risk disease according to the EAU risk groups. The 3-yr overall BCR-free survival rate was 84%. No differences were observed in the 3-yr BCR-free survival between EAU low- and intermediate-risk groups (88% vs 87%; p = 0.1). The novel nomogram depicted optimal discrimination at external validation (c-index 78%). Four new risk categories were identified based on the predictors included in the Cox-based nomogram. This new risk classification had higher accuracy in predicting early BCR (c-index 70%) than the EAU, CAPRA, and STAR-CAP risk classifications (c-index 64%, 63%, and 67%, respectively).
We developed and externally validated four novel categories based on clinical and radiological parameters to predict early BCR. This novel classification exhibited higher accuracy than the available tools.
Our novel and straightforward risk classification outperformed currently available preoperative risk tools and should, therefore, assist physicians in preoperative counseling of men candidate to radical treatment for prostate cancer.
尽管磁共振成像（MRI）参数至关重要，但接受根治性前列腺切除术（RP）的前列腺癌（PCa）患者生化复发（BCR）的风险分类系统仍仅基于临床变量。
我们旨在开发并验证一种整合临床和放射学参数的新型分类方法。
设计、设置和参与者：2014年至2020年期间在七个国际学术转诊中心进行了一项回顾性多中心队列研究。共确定了2565例接受RP治疗的PCa患者。
早期BCR定义为RP后3年内两次前列腺特异性抗原（PSA）值≥0.2 ng/ml。采用Kaplan-Meier法和Cox回归分析BCR的时间和预测因素。从总体患者样本中生成开发队列和验证队列。基于Cox衍生系数预测早期BCR的模型构成了外部验证的列线图的基础。预测因素包括PSA、活检分级组、MRI分期以及MRI上病变的最大直径。然后确定新的风险类别。采用Harrel一致性指数（c指数）比较我们的风险分层与欧洲泌尿外科学会（EAU）、前列腺癌风险评估（CAPRA）和国际前列腺癌分期协作组（STAR-CAP）风险组在预测早期BCR方面的准确性。
总体而言，根据EAU风险组，200例（8%）、1834例（71%）和531例（21%）分别患有低、中高风险疾病。3年总体无BCR生存率为84%。EAU低风险组和中风险组之间的3年无BCR生存率无差异（88%对87%；p = 0.1）。新型列线图在外部验证中表现出最佳的区分能力（c指数78%）。根据基于Cox的列线图中的预测因素确定了四个新的风险类别。这种新的风险分类在预测早期BCR方面的准确性（c指数70%）高于EAU、CAPRA和STAR-CAP风险分类（c指数分别为64%、63%和67%）。
我们基于临床和放射学参数开发并外部验证了四个预测早期BCR的新类别。这种新的分类方法比现有工具具有更高的准确性。
我们新颖且简单的风险分类优于目前可用的术前风险工具，因此应有助于医生对前列腺癌根治性治疗候选男性进行术前咨询。
