Tumor Location Based Segmentation in Upper-Tract Urothelial Carcinoma Impacts on the Urothelial Recurrence-Free Survival: A Multi-Institutional Database Study

Urothelial carcinoma arising in sites from the pelvis to the ureter has not been studied in combination with each other, and in combination with cancers arising in the bladder. Also, comparison of the proximal to distal portions of the ureter is very rare. Increasing evidence indicates that risk factors and survival rates of disease may differ across these sites [1,2,3,4]. For example, in patients with ureteral tumors, lymphovascular invasion (LVI) is the only significant prognosticator for both cancer-specific survival and metastasis-free survival in multivariate analysis and the prognostic value of LVI was not observed in pyelocaliceal tumors [5]. Raman et al. [6] adjusting for pathologic T stage, grade, and lymph node status, found no difference in the probability of disease recurrence or cancer death between ureteral and renal pelvic tumors. In non-muscle-invasive tumor stages, urothelial bladder cancer patients were more likely to experience disease recurrence and mortality compared to renal pelvic tumor patients but not ureteral tumors [7]. In pT2 and pT3 tumors, there was no difference in survival outcomes between the three tumor locations (pelvis, ureter, and bladder), while in pT4 tumors, patients with ureteral and pelvicalyceal tumors were more likely to experience disease recurrence and mortality compared to bladder cancer patients [7]. These studies suggest that differences in biological characteristics and risk factors across cancer site within the upper urinary tract and bladder may translate into differences in survival. Favaretto et al. [8] reported that tumor location was not an independent predictor for recurrence, and there was no difference in the probability of disease recurrence between ureteral and renal pelvic tumors. This finding is contrary to recent meta-analysis which indicated that ureteral and multifocal tumors are independent prognosticators of disease progression and cancer-specific survival in patients with upper-tract urothelial carcinoma (UTUC) treated with radical nephroureterectomy [4]. One recent study first suggested that finer distinctions in tumor localization (i.e., subsites within the upper, middle, or lower ureter) are informative of metastatic behavior, although this study did not describe the contribution of the tumor location to survival differences [9]. Using data from the Nishinihon Uro-Oncology Collaborative Group Registry, we defined the segments of the ureter and evaluated differences in urothelial recurrence-free survival after surgery according to tumor site.

The study population included men and women diagnosed with/without invasive primary urothelial carcinoma in the upper urinary tract who were registered in the Nishinihon Uro-Oncology Collaborative Group (NUOCG). The NUOCG is a resource representing a collaboration between multiple study centers in Japan which was approved by the review board of the NUOCG (approval no. H30-048). The above institutional review board clarified that the present types of research are eligible for waiver of consent prior to enrollment in the study. Enrollment protocols and eligibility criteria have been described in detail elsewhere. The present analysis was restricted to cases enrolled through 4 sites (Yamaguchi University, Ube, Yamaguchi, Japan; Osaka Medical College, Takatsuki, Osaka, Japan; Nara Medical University, Kashihara, Nara, Japan; and Shimane University School of Medicine, Izumo, Shimane, Japan). Eligible cases were diagnosed between May 1994 and September 2009, with ages at diagnosis ranging from 29 to 93 years (n = 1,349). A total of 462 cases were treated with chemotherapy or had unspecified ureteral tumors and were excluded from the present analysis (Table 1, 2). Vital status and date of death were determined via passive follow-up. As shown in Table 3, the tumor subsite was determined based on a review of pathological records. Proximal ureteral cancers were defined as cancers arising in the ureter between the ureteropelvic junction and the upper limit of the pelvic bone (U1), cancers in the ureter overlying the pelvic bone were classified as middle ureteral cancers (U2), and cancers in the ureter between the lower border of the pelvic bone and the ureteral orifice were grouped as distal ureteral cancers (U3). We used the Cox proportional hazards regression to evaluate the association between tumor location and survival after cancer diagnosis, where the time axis was months since diagnosis and urothelial recurrence was the outcome of interest. In separate analysis, we evaluated associations with grouped tumor sites and with individual tumor subsites (i.e., pelvis, proximal ureter, middle ureter, and distal ureter). Proportional hazards assumptions were verified by testing for a non-zero slope of the scaled Schoenfeld residuals on ranked failure times. Due to a violation of proportional hazards when evaluating associations across the full duration of the study follow-up, the follow-up was truncated at 5 years post-diagnosis; and thus, all cases still alive at 5 years post-diagnosis were censored at that time. Regression models were adjusted for age at diagnosis (continuous), gender, tumor location, presence of hydronephrosis, urine cytology result, performance status, LVI, pT, and pN.

Previously, open radical nephroureterectomy with open excision of the distal ureter with a bladder cuff was performed. From June 2003 to the present, the approach we use is a conventional 4-trocar nephrectomy with a transperitoneal or retroperitoneal approach. The kidney is freed by a laparoscopic maneuver with the entire length of the ureter and adjacent segment of the bladder cuff, and is then enclosed in an EndoCatch® bag and removed through a widened incision.

After a median follow-up of 34 months, 636 (74.5%) of enrolled patients were no longer alive, and of those who died, 620 (97.5%) died in the first 5 years after diagnosis. Compared to cases who were alive at the end of follow-up or at 5 years post-diagnosis (whichever came first), patients who died within 5 years of diagnosis were more likely to be female (25.1 vs. 29.5%), more likely to be older (median 69 vs. 75 years), more likely to have hydronephrosis (38.2 vs. 50.6%), and more likely to have advanced tumors (T2/T3/T4 stage: 16.9/33.9/3.1 vs. 16.1/22.5/0.4%).

Cases with pelvic tumors accounted for 475 cases (53.6%) of the study population; U1, U2, and U3 segments accounted for 96 (10.8%), 87 (9.8%), and 176 (19.8%) of cases, respectively (Table 3). There were 52 patients who had multifocal tumors (5.9%) as follows: 8 (0.9%) in the pelvis and ureter, 11 (1.2%) in U1 + U2, 1 (0.1%) in U1 + U3, 27 (3.0 %) in U2 + U3, and 6 (0.7%) in U1 + U2 + U3 (Table 3). The distribution of tumor location was similar for long-term survivors (patients who were alive at the end of follow-up or at 5 years post-diagnosis) and short-term survivors (patients who died within 5 years of diagnosis), with the exception that long-term survivors had a smaller proportion of multiple tumors (for pelvis + ureter portion/Ul+U2 portion/U1 + U3 portion/ U2 + U3 portion/U1 + U2 + U3 portion: 0.7/0.4/0/3.4/0 vs. 1.0/1.6/0.2/2.9/1.0%, respectively). Because the presence of concomitant bladder cancer predicted urothelial recurrence-free survival, we tried to evaluate the predictive value of the tumor location for concomitant bladder cancer (Fig. 1). Logistic regression analysis of gender, age, hydronephrosis, cytology, performance status, grade, LVI, T stage, N stage, and tumor focality, found that tumor location was associated with the presence of concomitant bladder cancer (p = 0.004, HR = 1.265, Table 4). When the tumor location was stratified into multiple segments, including multifocal tumors, only the U3 segment remained as predictive for the presence of concomitant bladder cancer (p = 0.002, HR = 2.872, Table 5). When analysis was restricted to unifocal disease, lower ureter tumors (combination of U2 and U3) had a worse prognosis for urothelial recurrence than pelvic tumors or upper ureteral tumors (U1) (p < 0.001 for lower ureteral tumors vs. pelvic tumors, p = 0.322 for upper ureteral tumor vs. pelvic tumor by log rank, Fig. 2). Multivariate analysis showed that the lower ureter location remained as a prognostic factor for urothelial recurrence after adjusting for gender, age, hydronephrosis, urine cytology, LVI, pT, and pN (p < 0.001, HR = 1.469, Table 6). Because the presence of bladder cancer at diagnosis affects urothelial recurrence, we conducted COX regression analysis restricted in patients without concomitant bladder cancer at diagnosis (Table 6). We found that lower ureter location (U2 and U3) still remained as a prognostic factor for urothelial recurrence after adjusting for gender, age, hydronephrosis, urine cytology, LVI, pT stage, and pN stage (p = 0.003, HR = 1.446, Table 6). Lower ureteral tumors had a higher prevalence of deaths (HR = 2.227) with reference to upper ureter tumors (Table 7).

In this cohort of individuals with localized urothelial carcinoma in the upper urinary tract managed by surgery, we found that those with tumors in the distal ureter experienced higher all-cause mortality through follow-up periods and in the 5 years after diagnosis relative to individuals with proximal ureteral carcinoma. Comparisons of the hazard ratio for concomitant bladder cancer by tumor site showed that cancer in the ureter between the lower border of the pelvic bone and the ureteral orifice was associated with the highest hazard ratio for the presence of bladder cancer, whereas cancer arising in the ureter between the ureteropelvic junction and the upper limit of the pelvic bone was associated with the least relevance with bladder cancer. Certain limitations should be considered in interpreting these findings. In particular, inclusion in the study population was contingent on surviving long enough after cancer diagnosis. Selection bias is possible if otherwise eligible cases who died before they could be enrolled differed from included cases with respect to tumor location or covariates. Additionally, we were unable to evaluate treatment variables as confounders or effect modifiers since we did not have this information. However, it is conceivable that treatment modality differed substantially by tumor location. We had a large study population for UTUC, and some case groups became small after stratification by tumor location. Although 32% of cases were followed beyond 5 years post-diagnosis, we truncated follow-up at 5 years in this analysis to avoid violations of proportional hazards assumptions. As a result, patients who died more than 5 years post-diagnosis were censored before they experienced the study outcome. In most studies, tumor stage and grade were considered the most important prognostic factors for UTUC. Other variables, including age, tumor location, lymph node involvement, multifocality, and tumor architecture, have also been reported as significantly relevant factors [10,11]. Prior study suggested that ureteral cancers of a high stage are associated with greater mortality [12]. Tai et al. [12] suggested that ureteral tumors are an independent risk factor that indicates a poor prognosis for cancer-specific and recurrence-free survival in patients with pathologic pT3 UTUC. Consistent with our findings, there was a study suggesting that ureteral and multifocal tumors had a worse prognosis than renal pelvic tumors [13]. Hoang et al. [14] assessed the incidence and clinical significance of 'skip lesions' that are present in proximal but not in distal ureteric sections, and which are occasionally found during the pathological examination of ureteric margins during radical cystectomy. On multivariate analysis, skip lesions correlated with lower median overall survival (OS, inestimable vs. 8.2 years, p = 0.014) in patients with pT0 or pTa disease and a trend towards lower OS (2.7 vs. 8.8 years, p = 0.066) in pTis disease [14]. Conversely, renal pelvic tumors that invade the renal parenchyma have been associated with a better prognosis than tumors invading peripelvic or periureteral fat [15]. No study except for the present study evaluated the impact of the ureteral segments on the prognosis and recurrence.

In summary, mortality after diagnosis of UTUC differed significantly for patients with tumors located in the distal ureter relative to patients with proximal ureteral carcinoma. For individuals with distal ureteral carcinoma, mortality risk with urothelial recurrence significantly differed according to the pT status. Conversely, proximal ureteral carcinoma was associated with the most favorable outcomes. These findings are consistent with a more frequent urothelial recurrence rate in distal ureteral carcinoma.

This study was supported, in part, by a Grant-in-Aid for Scientific Research from the Japan Society for the Promotion of Science. The authors thank the members of the NUOCG for their contributions to this study.