Nephron-sparing Surgery Versus Radical Nephroureterectomy in the Treatment of Organ-localized Upper Urinary Tract Urothelial Carcinoma: A Population-based Study of 1581 Patients

Jianhui Qiu,Hua Liu,Zixin Wang,Jingcheng Zhou,Yanqing Gong,Lin Cai,Kan Gong

DOI: https://doi.org/10.21203/rs.3.rs-111383/v1

2020-01-01

Abstract:Abstract Purpose: Nephron-sparing surgeries (NSS) have been widely used for patients with low-risk upper tract urothelial carcinoma (UTUC). We compared the prognosis of patients with organ-localized UTUC after NSS and radical nephroureterectomy (RN) through a population-based study and a propensity score matching (PSM) method. We further developed prognostic models for these patients using the Cox analysis and the competing-risk models.Materials and methods: The follow-up dataset of 1580 patients with organ-localized UTUC were screened out from the Surveillance, Epidemiology, and End Results (SEER) registry. A PSM process with a caliper width of 0.05 was performed to match the patients accepting NSS with those accepting RN. The Cox analysis was used for overall survival (OS) and competing-risks models based on Fine-Gray regression for cancer-specific survival (CSS). A forward stepwise selection process was applied to select independent predictors for OS and CSS. After selecting the predictors, two nomograms were developed based on the Cox analysis and the competing-risk model and internal validations were performed using 1000 bootstrap resamples.Results: Before PSM, RN was associated with better OS compared with NSS in the overall cohort (5-year OS rate: 65.1% vs. 55.6%, p<0.001) and the patients with renal pelvic lesions (5-year OS rate: 68.6% vs. 52.8%, p=0.001). Baseline variables were well balanced using the PSM method. After PSM, there was no significant difference in OS between the RN and NSS groups. Competing-risks models revealed that NSS was associated with poorer CSS compared with RN in the post-PSM cohort (n=766; subdistribution hazard ratio/ SHR=1.39, 95% confidence interval/ CI 1.03-1.89, p=0.033) and in patients with ureteral lesions from the post-PSM cohort. Four independent predictors, which were age at diagnosis, AJCC T stage, histologic grade, and surgery methods, were selected out for OS and CSS through forward stepwise selection methods. Two nomograms for OS and CSS were respectively developed. Conclusions: Patients accepting RN had a comparative or better outcome in each sample group. NSS achieved a similar prognosis for selective patients with organ-localized UTUC. Two nomograms were developed to predict OS and CSS with C-indexes of 0.672 and 0.643.