Comparison of the Oncological and Renal Function Outcomes of Partial Ureterectomy and Radical Nephroureterectomy in Upper Tract Urothelial Carcinoma: a Meta-analysis

Background To perform a meta-analysis of comparative studies reporting oncological and renal function outcomes of of partial ureterectomy and radical nephroureterectomy in upper tract urothelial carcinoma (UTCC). A literature search of PubMed, Embase, and the Cochrane library was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines, and a meta-analysis was performed to assess cancer-specific survival (CSS), overall survival (OS), recurrence-free survival (RFS), and perioperative variations in the estimated glomerular filtration rate (eGFR). Nineteen studies involving 4940 patients were included in our meta-analysis. No significant differences were found in the 5-year OS (HR=1.20, p=0.40), 5-year RFS (HR=1.21, p=0.37) and CSS (HR=0.89, p=0.20). A better preservation of renal function of PU compared with RNU changes in (eGFR) (WMD=-9.75, p=0.0006) between the patients undergoing the two types of surgery.


Background
To perform a meta-analysis of comparative studies reporting oncological and renal function outcomes of of partial ureterectomy and radical nephroureterectomy in upper tract urothelial carcinoma (UTCC).

Methods
A literature search of PubMed, Embase, and the Cochrane library was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines, and a meta-analysis was performed to assess cancer-specific survival (CSS), overall survival (OS), recurrence-free survival (RFS), and perioperative variations in the estimated glomerular filtration rate (eGFR).

Results
Nineteen studies involving 4940 patients were included in our meta-analysis. No Background Upper tract urothelial carcinomas (UTUCs) are rare and contribute for 5-10% of all urothelial malignancies [1]. Radical nephroureterectomy (RNU) with resection of the bladder cuff is considered a standard curative treatment for UTUC [2]. However, this treatment can increase cardiovascular morbidity as a result of renal unit. Partial ureterectomy, including endoscopic or segmental resection, could be considered a method to treat patients with UTUC. According to the European Association of Urology (EAU) guidelines, segmental resection of the distal ureter can be used even in patients with high-grade invasive distal ureteral urothelial carcinoma (UC) [2]. Partial ureterectomy (PU) consists of resecting the portion of the ureter compromised with UTUC, and endoscopic management can be performed through either antegrade or retrograde access to the upper urinary tract with percutaneous (PC) or ureteroscopic (URS) instruments, respectively. At present, segmental ureterectomy (SU) and progress in endoscopic surgery can be considered nephron-sparing procedures (NSP) and can be used to spare the renal unit in some cases such as those with low-stage tumor, renal insufficiency, or a solitary kidney [3][4][5][6]. Recently, with equal oncological control of these cases, expanded indications were used for patients with bilateral functioning kidneys.
However, the rarity of upper urinary tumors prevents the oncological outcomes after endoscopic surgery or PU from being shown. The endoscopic treatment of UTUC can increase the rate of bladder cancer recurrence. Radical surgical methods are associated with a lower bladder cancer recurrence rate [2,3]. However, in a series by Fang et al and Yakoubi et al, the type of approach was unlikely to impact subsequent bladder cancer, and equal oncological outcomes could be achieved [7,8]. Nevertheless, due to the low incidence of UTUC, evidence-based data that evaluate the oncological efficacy of RNU with a conservative approach that includes three validated treatment modalities are rare. The exclusion criteria were as follows: (1) case reports, reviews, editorial comments, meeting abstracts and articles without applicable data; (2) studies with insufficient data, such as lacking hazard ratios (HRs) or information to calculate the HR and number; (3) studies that were not comparative; and 4 studies comparing PU, endoscopic studies without surgery and studies without comparison groups or single-arm studies. The process of identifying relevant studies is summarized in Figure 1.
Data extraction and quality assessment These two authors extracted data, such as the OS, RFS, and CSS rates. The following data were recorded: (1) baseline comparative data; (2) clinical outcomes; and (3) postoperative complications.
We used the New-Ottawa Scale (NOS) to evaluate the included studies. The NOS scores were evaluated using a 9-point system. An NOS score of 7 or above was considered higher quality, and an NOS score of 3 or below was considered lower quality. Two reviewers (YLJ and LJQ) assessed the quality of the included studies.

Discussion
This meta-analysis shows that PU could achieve equal oncological outcomes and better preservation of renal function to those of RNU. We pooled 19 studies to compare 5-year OS rates, CSS, RFS and renal function. Our meta-analysis suggested that no significant differences were found in the 5-year OS, CSS and RFS between the RNU and PU groups.
Data regarding OS were available in 9 studies. The present study demonstrated that OS exhibited no statistically significant differences between the groups (Figure 2) [24,25]. However, Lucas et al found that among 108 patients who received NU or RNU, 5-year OS was significantly different between the RNU and SU groups (P>0.05) [18]. In this study, they did not perform a subgroup analysis. We also pooled the data of both the adjusted and unadjusted HRs for OS and other oncological outcomes.
Despite a trend towards worse CSS in the PU groups in a multivariate analysis, no significant difference between PU and RNU was reported [10,16]. However, we also found that data were poorly reported in general and that there was a significant selection bias in favour of PU, which was not controlled in survival analyses according to the risk of bias assessment. Although match-paired analyses or multivariate Cox regression models were commonly used to balance or adjust for covariates, the comparison of the clinical outcomes of PU and RNU suffered from a selection bias related to the presence of pathologic and other adverse clinical features in the PU and RNU groups, respectively.
Moreover, Jeldres et al performed a large population-based study that indicated that the operative modality (PU versus RNU) did not influence cancer-specific mortality (CSM) for advanced stage tumors, and they expanded the indication for PU [12]. Bin et al conducted a retrospective study that reviewed 60 patients with isolated primary ureteral tumors.
They found no significant differences in the probability of CSS with regard to either tumor location or surgical approach (P = 0.523 and P = 0.904, respectively)[13]. However, Isbarn et al performed a large population-based study to clarify the prognostic significance of tumor location in patients with UTUC and showed that tumor location did not affect oncological outcomes in a multivariate analysis [30].
In addition, Lughezzani et al studied 2,299 patients with UTUC treated with PU or nephroureterectomy (NU) and did not find that surgery type affected CSM, which agreed with the findings of our study [31]. Roupret et al found that the surgical approach (SU or RNU) had no influence on CSS or RFS (P = 0.94 and P = 0.42, respectively) in multivariate analyses, and pT stage and pN stage were independent predictors for CSS (p<0.05) in a univariate analysis [17]. When compared to RNU, neither total ureterectomy (TU) nor SU significantly affected OS (p>0.05), RFS (p>0.05) or CSS (p>0.05). These results were similar to our results. The presence of urothelial recurrence also did not correlate with CSM (p = 0.73) or overall mortality (p = 0.39).
Fukushima et al included 1,329 patients with UTUC for clinical evaluation. They used a propensity scoring method (PSM) to balance potential baseline differences between groups. They also performed a subgroup analysis of patients with pTa-1 and pT2-4 disease, which indicated that the RFS and CSS of the DUx group were comparable to those of the NUx group (P = 0.50 and P = 0.99, P = 0.64 and P = 0.93, respectively) [25].
In our study, the patients in the PU group had better renal function than those in the RNU group. The use of perioperative chemotherapy was poorly described, but Bagrodia et al and Singla et al reported that patients treated with SU or RNU were more likely to receive adjuvant or neoadjuvant chemotherapy, respectively [16,28].
Only the risk of positive surgical margins was equivalent between the SU and RNU groups, but this finding could simply result from better tumor characteristics in the conservative treatment group. Since the adjusted CSS was equivalent between the two arms, SU could be considered in selected cases of high-risk UTUC [7].
Silberstein et al performed a study including 367 patients with UTUC. They found that the preservation of the renal unit resulted in significantly smaller decreases in renal function than RNU [15]. Furthermore, the patients who underwent PU were more likely to receive adjuvant chemotherapy (for the preservation of the eGFR renal function), which may prolong OS, than those who underwent RNU.
Regarding the comparative renal function outcomes of PU, few robust datasets are available in the current literature. Nevertheless, the main objectives of sparing the renal unit are to prevent the development of postoperative chronic kidney disease and to preserve quality of life. However, our results should be interpreted cautiously, as comparative data may be missing due to the selection criteria that did not systematically capture all of the studies reporting renal function outcomes. Similarly, Seisen et al also performed a systematic review that suggested that all PU patients experienced better postoperative renal function than RNU patients and thus improved quality of life.
We acknowledge that our meta-analysis is limited by the retrospective design of the included studies. As a consequence, different data on the tumor and patient features were missing in the adjusted analyses. Tumor multiplicity should be taken into account in patients with UTUC to predict tumor existence in other locations of the ipsilateral upper urinary tract. A general limitation inherent to meta-analyses of aggregated data extracted from published data is the limited number of potentially influential covariates that can be accounted for. We were unable to perform subgroup meta-analyses or meta-regression that incorporated influential factors such as lesion size, age, and pathological features.
The selection biases of the original studies still exist in our study.

Conclusions
Our meta-analysis showed that PU and RNU provide equal cancer control in patients with UTUC. In addition, PU ensures better preservation of renal function than RNU. Therefore, PU is safe and provides good tumor control for UTUC treatment. Nonetheless, the evidence that supports these findings remains poor and potentially biased. Prospective clinical trials are necessary to adequately compare the oncological and renal function outcomes of PU and RNUin treating UTUC.

Availability of data and materials
All data generated or analyzed during this study are included in this published article.

Competing interests
The authors declare that they have no competing interests.     Forest plot for the5-year OS between the RNU and PU for UTUC  Forest plot for the 5-year RFS between the RNU and PU for UTUC