Comparative outcomes and prognosis in patients with ureteral upper tract urothelial carcinoma undergoing segmental ureterectomy versus radical nephroureterectomy: a multicenter cohort study

作者：Ailing Yu, Xianfeng Meng, Zhaonan Hou, Xiaochen Du, Mengxin Chen, Xiaowen Zong, Yi‐Hao Chen, Minjie Wang, Ruihang Dai, Yike Li, Zixuan Li, Pei Dong, Xuesong Li, Bo Fan · 发表于：International Journal of Surgery · 年份：2025 · DOI：10.1097/js9.0000000000003292 · 被引用次数：4 · 研究领域：Bladder and Urothelial Cancer Treatments、Ureteral procedures and complications、Kidney Stones and Urolithiasis Treatments

BACKGROUND: Although radical nephroureterectomy (RNU) with bladder cuff excision remains the gold standard treatment for upper tract urothelial carcinoma (UTUC), segmental ureterectomy (SU) may offer a nephron-sparing alternative. Studies comparing oncologic outcomes and renal functional outcomes between SU and RNU have yielded controversial, often conflicting results. Furthermore, investigations specifically involving ureteral UTUC patients in northern China are scarce. METHODS: We retrospectively enrolled 546 ureteral urothelial carcinoma patients (282 SU and 264 RNU cases) from three hospitals between October 2003 and September 2024 to assess overall survival (OS) and intravesical recurrence-free survival (IV-RFS) as oncologic outcomes, and to evaluate renal functional outcomes (preoperative, postoperative [2-month] and delta change in estimated glomerular filtration rate [eGFR] and serum creatinine [Scr]). The prognostic impact of surgical approach and clinicopathological variables were evaluated using Kaplan-Meier analysis with log-rank tests and Cox regression models. Significant predictors were incorporated into a validated nomogram for personalized 3-, 4-, and 5-year OS and IV-RFS probability estimation. Furthermore, eight machine learning algorithms (Lasso Cox, random survival forest [RSF], CoxBoost, generalized boosted regression modeling [GBM], support vector machine [Survival-SVM], eXtreme Gradient Boosting [XGBoost], supervised principal components [SuperPC], and...