Identification of Risk Factors and Development of a Prediction Model for Postoperative Urinary Retention Following Radical Hysterectomy for Cervical Cancer Based on Preoperative Urodynamic Profiling.
TL;DR
This model demonstrated good discrimination, calibration, and clinical utility for predicting the risk of POUR following radical hysterectomy for cervical cancer and may facilitate the early identification of high-risk patients and support personalized management strategies.
Abstract
Objective
This study aimed to develop and internally validate a risk prediction model for postoperative urinary retention (POUR) after radical hysterectomy for cervical cancer by integrating preoperative urodynamic parameters and clinical indicators.
Methods
Clinical data of 295 patients with cervical cancer undergoing radical hysterectomy at our institution (July 2023-June 2025) were retrospectively analysed. Patients were assigned to a training cohort (n = 207) and an internal validation cohort (n = 88) at a ratio of 7:3. All underwent preoperative urodynamic examinations documenting first sensation of filling (FSF), maximum cystometric capacity (MCC), maximum flow rate (Qmax), detrusor pressure at Qmax (PdetQmax), and bladder compliance (BC). In the training cohort, patients were divided into urinary retention and non-retention groups based on postoperative urinary retention status. Multivariate logistic regression identified influencing factors and developed a nomogram prediction model, whose performance was evaluated by receiver operating characteristic (ROC) curve, decision curve analysis (DCA), and calibration curve.
Results
The POUR group had a higher prevalence of diabetes and higher postvoid residual volume (PVR), but lower Qmax, FSF, MCC, BC, and PdetQmax than the non-POUR group (all p < 0.05). Multivariate logistic regression identified diabetes history, Qmax, FSF, MCC, BC, PdetQmax, and PVR as independent factors for POUR (p < 0.05). The nomogram achieved under the curves (AUCs) (95% confidence interval (CI)) of 0.901 (0.859-0.942) in the training cohort and 0.883 (0.804-0.931) in the validation cohort. DCA showed a positive net benefit across all thresholds, exceeding both default strategies. Internal validation revealed close alignment of the calibration curve with the ideal curve (Hosmer-Lemeshow χ2 = 5.841, p = 0.557).
Conclusions
A nomogram incorporating preoperative urodynamic parameters and clinical features was successfully developed and validated. This model demonstrated good discrimination, calibration, and clinical utility for predicting the risk of POUR following radical hysterectomy for cervical cancer. It may facilitate the early identification of high-risk patients and support personalized management strategies.