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 Peking University First Hospital https://orcid.org/0000-0002-4393-612X Hua Liu Peking University First Hospital Zixin Wang Peking University First Hospital Jingcheng Zhou Peking University First Hospital Yanqing Gong Peking University First Hospital Lin Cai Peking University First Hospital Kan Gong (  gongkan_pku@126.com ) Department of Urology, Peking University First Hospital, Beijing, China.; Institute of Urology, Peking University, Beijing, China. National Urological Cancer Center, Beijing, China. https://orcid.org/0000-00017195-677X


Background
Upper urinary tract urothelial carcinoma (UTUC) are rare tumors that account for 5%-10% of all urothelial malignancies with an incidence rate of 1.45-2.06 cases per 100,000 person-years [1,2]. Radical nephroureterectomy (RN) with bladder cuff resection remains the gold standard therapy for patients with UTUC [3], but it is associated with some serious complications, such as chronic renal failure and failureinduced other system morbidities. Given that the elderly are more like to get affected by UTUC [2], RN is not suitable for all patients as an invasive procedure with the risk of complications. Besides some patients with medical comorbidities that make invasive procedures risky, patients with solitary kidney may also prefer nephron-sparing surgery (NSS). After carefully weighing the advantages and disadvantages, NSS, which spares the morbidity associated with RN in the avoid of compromising oncological outcomes and kidney function, has been advocated in selected patients with low-risk UTUC [3][4][5].
Due to the rarity of UTUC, many studies about UTUC were limited by sample size and retrospective case series didn't have enough power to get conclusions. Surveillance, Epidemiology, and End Results (SEER) program gathered information on treatment and prognosis, which currently covers more than a fth of the US population. In this study, a population-based study including patients with UTUC registered in SEER was performed to compare prognosis between patients accepting RN or NSS. To avoid bias and derive a more persuasive conclusion, a propensity score matching (PSM) method was performed to adjust baseline characteristics between the RN and NSS groups.
As UTUC is more common in the elderly, comorbidities of patients can't be ignored during follow-up and the process of survival analysis. Deaths from other noncancer causes can be competing events to cancer-speci c deaths. The risk of competing mortality increases and the effect of treatment on survival diminishes with advancing age. In the presence of competing events, conventional Kaplan-Meier and Cox analysis methods may not be the best choices because these two methods regard competing events as independent censored events and overrate the incidence of target events [6,7]. Instead, competing risk models based on the Fine-Gray regression can better discriminate the effects of clinical factors on special events [7]. Thus, competing-risks models were used in this study to construct prognostic models of organlocalized UTUC. Nomograms visualizing the prognostic models were also made to predict the 3-year and 5-year prognosis [8,9].

Study population
The nal diagnosis recorded in the SEER data (https://seer.cancer.gov/data/) was coded by the International Classi cation of Disease for Oncology-3 (ICD-O-3). We identi ed all subjects diagnosed with UTUC (topography codes: C65.9 and C66.9; histological code: 8120/3, 8122/3, 8130/3, 8131/3) between 2004 and 2015 in the SEER registry. Clinicopathological data of subjects, such as age, gender, race, American Joint Committee on Cancer (AJCC) TNM stage, histological grade, SEER registry location, marital status, survival time, clinical outcome, and other demographic characteristics needed in the study were exported. Only T1N0M0 and T2N0M0 patients with complete data were included in the study. Patients with more than one primary malignant tumor, or not receiving NSS or RN were excluded. Local tumor destruction (e.g., photodynamic therapy, cryosurgery, thermal ablation, and electrocautery), local tumor excision (e.g., polypectomy and excisional biopsy), partial nephrectomy for pelvis tumors, and partial ureterectomy for ureter tumors were all categorized as NSS. The SEER registry de nes histologic grade as follows: G1, well-differentiated; G2, moderately differentiated; G3, poorly differentiated; G4, undifferentiated or anaplastic.

Propensity score matching
The PSM method was used to balance the baseline characteristics between the NSS and RN groups.
Multivariate logistic regression analysis was used to get propensity scores for each subject based on gender, age at diagnosis, laterality (left or right), tumor location (pelvis or ureter), marital status, AJCC T stage, and histologic grade of tumors. NSS and RN groups were matched 1: 1 through a caliper width of 0.05 for the propensity score and the nearest neighbor matching.

Statistical analysis
Continuous variables were described using medians and ranges, while categorical variables were reported by proportions. The chi-square tests were adopted to compare the baseline categorical variables between the NSS group and RN group before and after PSM and the Mann-Whitney U tests for continuous variables. The study outcomes were cancer-speci c survival (CSS) and overall survival (OS) according to the record of the SEER database. Kaplan-Meier method as well as log-rank tests were used to compare the OS between groups. Univariate and multivariate Cox regression analysis models were adopted to detect factors in uencing OS. Competing risk models were used to assess the predictive factors of CSS. X-tile version 3.6.1 (Robert L Camp, Yale University; https://medicine.yale.edu/lab/rimm/research/software/) was used to classify continuous variables (age at diagnosis) [10]. All analyses were performed using the Stata/SE version 15.1 (StataCorp, College Station, Texas) , IBM SPSS Statistics version 23.0 (SPSS Inc., Chicago, IL, USA) and R statistical software (version 3.6.3; https://www.r-project.org/). All statistical tests were two-sided with P<0.05 considered to be indicative of statistical signi cance. The "regplot", "ggplot2" and "riskRegression" R packages were used to construct and verify prognostic models. 14,135 patients diagnosed with UTUC between 2004 and 2015 were collected from the SEER database. Finally, 1580 eligible T1-2N0M0 patients with UTUC as the only primary tumor were included for further analysis as shown in Figure 1. 306 failure events and 302 competition events were observed in the overall cohort. Median follow-up period for the overall cohort and after PSM was 44 months and 41 months.

Study population
Baseline characteristics of patients between the RN and NSS groups before and after PSM were shown in Table 1. For the entire cohort, 1187 (75.1%) patients were treated with RN, and 393 (24.9%) were treated with NSS. Patients undergoing NSS were associated with increasing age and renal pelvis tumors compared with those receiving RN (p<0.001). After PSM, a cohort of 766 patients was generated, 383 patients in each group. 10 of 393 patients from the NSS group didn't have the matched individuals from the RN group. The baseline characteristics were well balanced between the two groups in the PSM cohort (Table 1; Figure 2). 167 failure events and 163 competition events were recorded in the post-PSM cohort. G4 was the most common histologic grade before and after PSM, accounting for above 40% in both the overall cohort and the post-PSM cohort.

Differences of prognosis between the RN and NSS groups
As shown in Figure 3, in the overall cohort, patients in the NSS group were associated with poorer OS compared with those in the RN group (5-year OS rate: 55.6% vs. 65.1%, p<0.001; Figure 3. A). After stratifying patients according to locations of lesions, for patients with ureteral tumors, there was no signi cant difference in OS between the RN and NSS groups (453 and 284 patients, respectively; 5-year OS rate: 59.6% vs. 56.9, p=0.232; Figure 3. B). For patients with renal pelvic tumors, RN was associated with better OS compared with NSS (734 and 109 patients, respectively; 5-year OS rate: 68.6% vs. 52.8%, p=0.001; Figure 3. C).
In the post-PSM cohort, NSS seemed to be associated with poorer OS compared with RN, whose differences between them were almost signi cant (5-year OS rate, 62.8% vs. 55.2%, p=0.085; Figure 3. D).
This trend still existed in patients with renal pelvic tumors, for which RN was almost associated with better overall survival compared with NSS (5-year OS rate 67.5% vs. 52.8%: p=0.055; Figure 3. E). For ureteral tumors, OS between the two groups showed no signi cant difference (5-year OS rate, RN vs. NSS: 60.8% vs. 56.3%, p=0.404; Figure 3. F).

Independent prognostic factors of OS and CSS in the overall cohort
Age at diagnosis was divided into two groups as ">78.5 years old" and "<78.5 years old" using X-tile software for further building prognostic models. The results of univariate and multivariate Cox analysis for OS in the overall cohort were listed in Table 2. Univariate Cox analysis demonstrated that some characteristics of patients, which were age groups, tumor locations (renal pelvic tumors or ureteric tumors), AJCC T stage, histologic grade, marital status, surgery methods, were associated with OS in the overall cohort. Further multivariate Cox analysis including statistically signi cant variates in univariate Cox analysis by the forward-stepwise model selection revealed that age at diagnosis, AJCC T stage, and histologic grade were independent prognostic factors for OS in the overall cohort ( Competing-risks models based on the Fine-Gray models were used to assess CSS in the entire cohort. The results of univariate analysis for CSS in the overall cohort were shown in Table 3. Age groups, tumor location, AJCC T stage, histologic grade, and surgery methods were associated with CSS. Gender, race, tumor laterality, and marital status seemed not to be related to CSS. Statistically signi cant variables in the univariate analysis in the competing-risks models were included in multivariate analysis through forward-stepwise selection methods, which validated that older age at diagnosis, advanced AJCC T stage, advanced histologic grade, and NSS were independent predictors of CSS (Table 4). Age at diagnosis, AJCC T stage, and histologic grade hold a similar predictive power for competing risk of CSS.
Compared with the younger group (age at diagnosis < 78.5 years old), the older group (age at diagnosis > 78. 5

Developing nomograms for OS and CSS
As mentioned above, the nal prognostic models for OS and CSS both included four factors, which were age at diagnosis, surgery methods, histologic grade, and AJCC T stage of tumors. The prognostic nomograms for OS ( Figure 5) and CSS ( Figure 6) using the overall cohort that integrated all these four signi cant independent factors were shown in Figure 5 and Figure 6. The nomogram for OS was based on multivariate Cox analysis to predict the 3-year and 5-year OS, while the nomogram for CSS was based on the competing risk model mentioned above to predict the 3-year and 5-year CSS. After adding up scores corresponding to each value and check against the bottom probability axis, the possibility of allcause mortality or cancer-speci c mortality was then got. The score of each value was shown in Table 5.
Then the overall cohort was used for internal validation of prognostic models. The nomogram for OS displayed a C-index of 0.672, as well as the nomogram for CSS, displayed a C-index of 0.643. Using 1000 resampled bootstrap data sets for cross-validation, the calibration plots for the two nomograms in the overall cohort were shown in Figure 7 and Figure 8. The Brier scores of corresponding models were also shown in the gures.

Discussion
This study reported a population-based study of organ-localized UTUC from the SEER registry using the PSM method to compare the prognosis after accepting RN and NSS. Two prognostic models for OS and CSS were developed and validated through multivariate Cox and competing-risks analysis. Nomograms visualizing the two models were shown in Figure 5 and Figure 6.
NSS has cut a gure in the eld of UTUC in recent years. For certain groups of patients with low-risk UTUC, NSS was considered to avoid RN-related comorbidities without compromising oncologic control. Previous studies were often limited by sample capacity due to its rarity [3,11,12]. A previous multicenter retrospective study reported that patients treated with NSS had similar CSS and better OS compared with patients accepting RN in multivariate Cox analysis [12]. A recent systematic review conducted by Seisen et al including 22 studies got the conclusion that similar survival existed between KSS, which were ureteroscopic as well as percutaneous methods, and RN only for low-grade and noninvasive UTUC. Segmental ureterectomy may let selective patients with invasive or high-grade UTUC achieve comparable oncologic outcomes with RN. Thus, Seisen et al advocated that segmental ureterectomy could be one of the optional managements for selected patients with high-risk UTUC [11]. A meta-analysis conducted by Yakoubi et al included eight published retrospective studies reported that endoscopic NSS had similar OS and CSS with RN using the pool data on 1002 patients with localized UTUC. However, the power of the results was limited by the heterogeneity from the included studies and the low level of evidence [13]. There is a lack of high-quality research focusing on whether NSS is suitable for patients with UTUC.
This study simultaneously used the follow-up data from the SEER registry and the PSM method to control selected biases and other potential bias. The SEER registry covers plenty of individual patient data nearly representing the overall population [14]. The PSM method has been widely used in clinical researches to balance the baseline variables between the treatment and control groups [15,16]. A cohort of eligible 1580 patients who were diagnosed with organ-localized UTUC between 2004 to 2015 was screened out according to Figure 1. After PSM, a cohort of 766 patients was created for further analysis. The baseline characteristics of the RN and NSS groups in the post-PSM cohort showed no signi cant difference (Table  1). We chose OS and CSS as the target events of this study. In the overall cohort, RN was associated with better OS compared with NSS (p<0.001). RN was still associated with better OS in patients with renal pelvic UTUC (p=0.001), whereas NSS achieved a similar OS with RN in patients with ureteral UTUC. After reducing bias through PSM, the OS between RN and NSS groups showed no signi cant difference (p=0.085). NSS had a tendency to be associated with poorer OS in comparison to RN in the post-PSM cohort but there was still no signi cant difference (0.055).
Competing risk models have been adopted in many previous studies [17,18]. Patients with advanced age often had comorbidities with an increased risk of competing mortality. The common Cox and Kaplan-Meier methods analyzed a combined endpoint, aggregating one or more deaths from other competing events. Thus, competing risk models based on Fine-Gray regression were used in this study to assess the CSS of patients as deaths from other noncancer causes were regarded as competing events to cancer-speci c deaths. In the post-PSM cohort, RN was associated with better CSS (p=0.033, Figure 4. A). After stratifying by tumor locations, RN was still related to better CSS in patients with ureteral tumors (p=0.044). CSS of patients with renal pelvic tumors was similar between the RN and NSS groups (p=0.471). Taking the above results for OS and CSS into consideration, NSS achieved a similar prognosis and comparative oncologic control with RN for selective patients. Patients accepting RN had comparative or better outcomes in each sample group.
Seldom studies have published competing risk models and nomograms for organ localized UTUC [19,20]. A previous published prognostic model for UTUC con rmed tumor location, tumor grade, and tumor architecture as independent predictors of nonorganic localized disease [21]. In this study, four factors, which were age at diagnosis, histologic grade, AJCC T stage, and surgery methods, were included in the prognostic models. Two nomograms based on the Cox model and the competing risk model respectively were developed and validated. Because of the lack of external follow-up data, only the inner validation of nomograms using 1000 resampled bootstrap methods was performed [22]. Surgeons and patients could simply and quickly get the 3-year or 5-year prognosis using the nomograms.
The power of our prognostic models was the large population-based sample and Fine-Gray regression that can distinguish target events and competing-risks events. There are some shortcomings in our prognostic models. Some characteristics were not available, such as family history, smoking history, treatment detail, chemical exposure, and laboratory indexes [23,24]. Additionally, there was a lack of external datasets to validate the models. In the future, we aimed at the validation of the models using follow-up data of our center.

Conclusions
NSS could achieve a comparative oncologic outcome and overall outcome with RN in selective patients with organ-localized UTUC. As the gold standard treatment for UTUC, RN was appropriate for patients from each subgroup. Age at diagnosis, tumor location, histologic grade, and surgery methods were independent predictors for OS and NSS. Two nomograms were developed on the basis of these four factors. Multicentral prospective randomized controlled clinical trials are urgently needed to assess the oncologic outcomes of UTUC with NSS or RN management.

Declarations Ethics approval and consent to participate
Since all the follow-up information analysed in this article were from the SEER database (https://seer.cancer.gov/data/), informed consents were not needed. As long as data usege agreement was signed, data in the SEER database become available to the public.

Consent of publication
Not applicable.

Availability of data and materials
The datasets for this study can be found at https://seer.cancer.gov/data/.

Competing interests
The authors declare that the research was conducted in the absence of any commercial or nancial relationships that could be construed as a potential con ict of interest.  Tables   Table 1. Baseline characteristics of patients between NSS and RN group before and after PSM. Abbreviations: PSM, propensity-score matching; NSS, nephron-sparing surgery; RN, radical nephroureterectomy; NA, not available; IQR, interquartile range; The word "Unmarried" in marital status means unmarried, divorced, widowed, separated and never married, etc.