BPH合并T2DM患者TUPKEP术后LUTS改善的预测模型构建

Establishment of a Predictive Model for Early Improvement of LUTS after TUPKEP in Patients with BPH Complicated by T2DM

摘要: 【目的】目前临床中尚缺乏专门针对良性前列腺增生(Benign Prostatic Hyperplasia, BPH)合并2型糖尿病(Type 2 Diabetes Mellitus, T2DM)患者行经尿道前列腺等离子剜除术(Transurethral Plasma Kinetic Enucleation of the Prostate, TUPKEP)后下尿路症状(Lower Urinary Tract Symptoms, LUTS)改善的预测工具。本研究旨在筛选影响患者术后LUTS改善的术前相关因素构建一个预测模型，为临床术前精准评估患者手术获益提供依据。【方法】收集2021年1月至2025年12月我院泌尿外科收治的、因BPH合并T2DM行TUPKEP治疗的患者临床资料，最终纳入110例患者作为研究对象。收集患者术前临床指标及尿动力学参数，同步收集并记录患者术前、术后3个月的国际前列腺症状评分数据(International Prostate Symptom Score, IPSS)进行记录。以术后总IPSS评分(Total IPSS Subscore, IPSS-T)与术前IPSS-T比值 ≤ 0.5或术后IPSS-T < 7分为手术有效标准，将患者分为LUTS改善组与LUTS改善不佳组。采用单因素及多因素Logistic回归分析确定术后LUTS改善的独立预测因素；构建列线图预测模型。绘制受试者工作特征曲线(Receiver Operating Characteristic Curve, ROC)计算曲线下面积(Area Under Curve, AUC)、校准曲线及决策曲线分析(Decision Curve Analysis, DCA)综合评价模型。【结果】最终纳入的110例患者中，手术有效组78例，术后疗效不佳组32例。多因素Logistic回归分析表明，糖尿病并发症(P = 0.02)、BCI (P < 0.01)、BC (P = 0.03)、IPSS-V/S (P < 0.01)是BPH合并T2DM患者行TUPKEP术后LUTS改善的独立预测因素。基于上述独立因素构建的模型，内部验证AUC为0.85 (95%CI: 0.77~0.92)；校准曲线显示预测概率与实际观察概率拟合度高，Hosmer-Lemeshow检验结果表明模型校准度良好(P > 0.05)；DCA结果显示，在0%~70%阈值区间内，该模型临床净收益率高于“全干预”和“不干预”方案。【结论】术前糖尿病并发症、BCI、BC及IPSS-V/S是BPH合并T2DM患者行TUPKEP术后LUTS改善的独立预测因素；基于上述因素建立的预测模型展现出良好的区分度、校准度及临床适用性。

Abstract: [Objective] Currently, there is a lack of specific predictive tools in clinical practice for the improvement of lower urinary tract symptoms (LUTS) following transurethral plasma kinetic enucleation of the prostate (TUPKEP) (Transurethral Plasma Kinetic Enucleation of the Prostate, TUPKEP). This study aims to identify preoperative factors influencing postoperative improvement in lower urinary tract symptoms (LUTS) and to construct a predictive model, thereby providing a basis for the precise preoperative assessment of surgical benefit in clinical practice. [Methods] Clinical data were collected from patients admitted to the Department of Urology at our hospital between January 2021 and December 2025 who underwent TUPKEP for BPH complicated by T2DM; a total of 110 patients were ultimately included in the study. Preoperative clinical indicators and urodynamic parameters were collected, and International Prostate Symptom Score (IPSS) data were simultaneously recorded at both preoperative and 3-month postoperative time points. A postoperative Total IPSS score (IPSS-T) to preoperative IPSS-T ratio ≤ 0.5 or a postoperative IPSS-T score < 7 was defined as the criterion for surgical efficacy, and patients were classified into the LUTS improvement group and the poor LUTS improvement group. Univariate and multivariate logistic regression analyses were performed to identify independent predictors of postoperative LUTS improvement; a nomogram prediction model was constructed. A receiver operating characteristic (ROC) curve was plotted to calculate the area under the curve (AUC), and the model was comprehensively evaluated using calibration analysis and decision curve analysis (DCA). [Results] Of the 110 patients ultimately included in the study, 78 were in the group with effective surgical outcomes and 32 were in the group with poor postoperative outcomes. Multivariate logistic regression analysis indicated that diabetic complications (P = 0.02), BCI (P < 0.01), BC (P = 0.03), and IPSS-V/S (P < 0.01) were independent predictors of LUTS improvement following TUPKEP in patients with BPH and T2DM. The model constructed based on the above independent factors had an internal validation AUC of 0.85 (95%CI: 0.77~0.92); the calibration curve showed a high degree of fit between predicted and observed probabilities, and the Hosmer-Lemeshow test indicated good model calibration (P > 0.05); DCA results showed that, within the 0%~70% threshold range, the clinical net benefit of this model was higher than that of the “full intervention” and “no intervention” strategies. [Conclusion] Preoperative diabetic complications, BCI, BC, and IPSS-V/S are independent predictors of improvement in LUTS following TUPKEP in patients with BPH and T2DM; the predictive model developed based on these factors demonstrates good discriminatory power, calibration, and clinical applicability.