Global Robotics Institute, Advent Health Celebration Health, Celebration, Florida.
Data Methods and Systems Statistical Laboratory, University of Brescia, Brescia, Italy.
J Urol. 2021 Oct;206(4):942-951. doi: 10.1097/JU.0000000000001895. Epub 2021 May 25.
Prediction of potency recovery following robot-assisted radical prostatectomy (RARP) is useful for better patient counseling and postoperative treatment strategies. In this study we propose a preoperative and postoperative nomogram to predict postoperative potency recovery following RARP.
Patients from development set (6,502) were selected to develop the nomograms, and patients in validation set (2,706) were used for validation. Cox regression models were fitted on the development cohort to predict potency recovery after RARP using as prognostic factors the covariates selected. Two nomograms were drawn using the regression coefficients of the preoperative and postoperative Cox models.
The discrimination ability of the preoperative model was evaluated on the development cohort using the receiver operator curves estimated at 3, 6, 12 and 24 months. The AUC at these time points was 0.726, 0.734, 0.754, and 0.778, respectively. The areas under the curve of the postoperative model at 3, 6, 12 and 24 months were 0.746, 0.756 and 0.777, and 0.801, respectively. Preoperative and postoperative predictive models were validated using a separate set of 2,706 patients. The AUCs of the preoperative model at 3, 6, 12 and 24 months were 0.789, 0.772, 0.768, and 0.778, respectively. The ROC curves of the postoperative model at 3, 6, 12 and 24 months with AUCs of 0.807, 0.797, 0.793 and 0.798, respectively. Along with age and preoperative sexual function, nerve-sparing technique determines the potency outcomes justifying better AUC for postoperative model vs the preoperative model.
The above nomograms help us to predict with good accuracy the probability of potency recovery within 3, 6, 12 and 24 months following surgery taking into consideration preoperative and postoperative factors. This is a novel tool for the care giver to predict realistic expectation of potency outcomes to the patients, while preoperative and immediate postoperative counseling.
预测机器人辅助前列腺根治性切除术（RARP）后的勃起功能恢复情况有助于更好地为患者提供咨询和制定术后治疗策略。本研究旨在建立术前和术后列线图，以预测 RARP 后的术后勃起功能恢复情况。
选择开发集（6502 例）中的患者建立列线图，验证集（2706 例）中的患者用于验证。使用 Cox 回归模型，将选择的预后因素作为协变量，对开发队列进行分析，预测 RARP 后的勃起功能恢复情况。使用术前和术后 Cox 模型的回归系数绘制两个列线图。
通过在开发队列中估计 3、6、12 和 24 个月的接收者操作曲线，评估术前模型的区分能力。在这些时间点，AUC 值分别为 0.726、0.734、0.754 和 0.778。术后模型在 3、6、12 和 24 个月时的 AUC 值分别为 0.746、0.756 和 0.777、0.801。使用 2706 例独立患者对术前和术后预测模型进行验证。在 3、6、12 和 24 个月时，术前模型的 AUC 值分别为 0.789、0.772、0.768 和 0.778。术后模型在 3、6、12 和 24 个月时的 AUC 值分别为 0.807、0.797、0.793 和 0.798。年龄和术前性功能与神经保留技术共同决定了勃起功能的结局，这使得术后模型的 AUC 优于术前模型。
上述列线图有助于考虑术前和术后因素，准确预测术后 3、6、12 和 24 个月内勃起功能恢复的概率。这是一种新的工具，可帮助护理人员在术前和术后咨询时，向患者预测现实的勃起功能结局预期。
