Effects of Bony Pelvic and Prostate Dimensions on Surgical Difficulty of Robot-Assisted Radical Prostatectomy: An Original Study and Meta-analysis.

Abstract

Due to the deep location of the prostate within the pelvic cavity, procedures of robot-assisted radical prostatectomy (RARP) might be challenged by the prostate size and the limited pelvic cavity space. This study aimed to investigate the roles of bony pelvic and prostate dimensions in RARP procedures by an original study coupled with a meta-analysis. In the original study, patients undergoing multiport RARP between 2021 and 2022 were consecutively assessed. The associations of anatomic features with operative time (OT), estimated blood loss (EBL), and positive surgical margin (PSM) were evaluated using linear and logistic regression analyses as well as restricted cubic spline (RCS) analysis. Based on machine-learning algorithms, this study established predictive models for surgical difficulty and interpreted the model using SHapley Additive exPlanation (SHAP). In the meta-analysis, three databases were searched for eligible studies. Quantitative syntheses were subsequently performed. Overall, 219 patients were enrolled in the original study. Prostate volume (PV) and the prostate volume-to-pelvic cavity index (PCI) ratio (PV-to-PCI ratio) were significantly associated with longer OT (P < 0.05). In the RCS models, U-shaped associations were observed between the prostate anteroposterior diameter (PAD) and OT, and between the prostate height (PH) and EBL, and an L-shaped association was observed between the anteroposterior diameter of the pelvic inlet (API) and EBL. The XGBoost model was superior to the logistic regression model in predicting prolonged OT. The meta-analysis demonstrated that greater PV was significantly associated with longer OT (β = 0.20; 95% confidence interval [CI] 0.12-0.27; odds ratio [OR] = 1.05; 95% CI 1.00-1.11), and a smaller PV could increase the risk of PSM (OR = 0.82; 95% CI 0.77-0.88). A large prostate within a narrow and deep pelvis might suggest increased surgical difficulty of RARP. The size of the pelvic inlet also had a great impact on RARP. For PAD and PH, there seemed to be an optimal range with the lowest surgical difficulty. Machine-learning models based on the XGBoost algorithm could be successfully applied to predict the surgical difficulty of RARP.