Oncological outcomes after radical nephroureterectomy for upper tract urothelial carcinoma: Comparison over the three decades


Vitaly Margulis M.D., Department of Urology, UT Southwestern Medical Center, 5323 Harry Hines Blvd, Dallas, TX 75390-9110, USA. Email: vitaly.margulis@utsouthwestern.edu


Objective:  To evaluate temporal trends in clinicopathological features and oncological outcomes after radical nephroureterectomy for upper tract urothelial carcinoma.

Methods:  Utilizing a multi-institutional database of patients treated with radical nephroureterectomy between 1983 and 2007, we compared clinicopathological features and survival outcomes over the past three decades using the following cohorts: group 1 comprised of patients treated before the 1990s (n = 106), group 2 from 1990 to1999 (n = 655), and group 3 from 2000 to 2007 (n = 701). Survival rates were compared using Kaplan−Meier survival analysis.

Results:  The study included 1462 patients, 992 men and 470 women, with 36 months median follow up (range 1–250 months) after radical nephroureterectomy. Tumors were organ confined (≤T2/N0) in 88% and high-grade in 64%. Neoadjuvant and adjuvant systemic chemotherapy was administered in 47 (3.2%) and 171 (11.7%) patients, respectively. There was a significant increase in the use of laparoscopic radical nephroureterectomy, endoscopic management of urothelial carcinoma and utilization of perioperative chemotherapy between decades 1 to 3. There were no significant differences in pathological stage distribution. The overall 5-year disease-free survival rates were 66 ± 5%, 68.5 ± 2% and 71 ± 2%, and the 5-year cancer-specific survival rates were 75 ± 5%, 72 ± 2%, and 75 ± 2% for groups 1, 2 and 3, respectively, with no significant statistical differences between the three decades (P > 0.05).

Conclusion:  Outcomes after radical nephroureterectomy have not changed significantly over the past three decades, despite staging and surgical refinements. Utilization of perioperative systemic chemotherapy in urothelial carcinoma management remains low. Further improvements in outcomes of urothelial carcinoma patients necessitate rigorous investigation of multimodal treatment approaches.

Abbreviations & Acronyms
BC =

bladder cancer


carcinoma in situ


cancer-specific survival

CT =

computed tomography


disease-free survival


Eastern Cooperative Oncology Group


lymph node dissection


lymphovascular invasion


radical nephroureterectomy


Surveillance, Epidemiology, and End Results


upper tract urothelial carcinoma


There have been modest albeit significant strides in defining optimal treatment for patients with urothelial carcinoma of the bladder during the past two decades. Refinements in surgical technique, improved risk stratification and use of perioperative chemotherapeutic agents have led to survival improvements.1 However, similar advances have not been achieved in patients with UTUC,2–4 Despite accounting for just 5% of all urothelial malignancies, UTUC continues to portend a poor prognosis, with approximately 28% of patients experiencing a recurrence of disease outside the bladder and 23% of patients dying of the disease within 5 years.5 In addition, outcomes of patients with locally advanced UTUC after RNU continues to be dismal.6 Because of its low incidence, trends in clinicopathological features and survival outcomes for UTUC have been difficult to establish. Although few studies reporting on changes in UTUC treatment outcomes over time have been reported,2–4 most of these endeavors have consisted of small series, single-institutional reports or large population-based series, which lack relevant clinicopathological data. Herein, utilizing a large and well-characterized group of patients treated for UTUC during the past three decades, we evaluate temporal trends in clinical parameters and oncological outcomes.


Patient selection and treatment regimens

After combining the data sets from 13 participating centers, a database containing 1462 patients who underwent RNU with ipsilateral bladder cuff resection for UTUC between 1983 and 2007 was generated. Patients were divided into three groups according to the decade of surgery (group 1, 2 and 3 comprised patients who underwent RNU before 1990, from 1990 to the end of 1999, and from 2000 to 2007, respectively). Data analyzed included clinicopathological features and outcomes. Operative notes were used to determine whether LND was carried out or not. The standardization of surgical technique and indications for LND was impossible because of the multicenter and retrospective design of the present study. Patients were chosen for neoadjuvant and adjuvant chemotherapy at the discretion of the treating physician. Institutional review board approvals were obtained from each participating institution.

Pathological evaluation

Only primary UC tumors were included in the present study. We used a database comprising patients from the UTUC Collaboration, representing those from academic centers in Houston and Dallas, Texas; Rennes, France; Tokyo, Japan; Sacramento, California; Graz and Vienna, Austria; Milan, Italy; Mannheim, Germany; Ann Arbor, Michigan; New York City, New York; Santiago, Chile; and Montreal, Quebec, Canada who were treated with nephroureterectomy for clinically non-metastatic UTUC. All surgical specimens were processed according to standard pathological procedures, and all slides were re-reviewed by genitourinary pathologists at each participating institution according to prospectively defined uniform criteria. All pathologists were blinded to clinical outcomes. Tumors were staged according to the 6th edition of the American Joint Committee on Cancer – Union International Contre le Cancer Tumor–Node–Metastasis classification.7 Tumor grades were assessed according to the 1998 World health Organization/International Society of Urologic Pathology consensus classification.8 In addition, all UTUC were evaluated for tumor location (renal pelvis vs ureter), pattern of tumor growth (papillary vs sessile), presence of LVI and concomitant CIS in the nephroureterectomy specimen. Lymphovascular invasion was defined as the presence of tumor cells within an endothelium-lined space without underlying muscular walls. The presence of concomitant CIS and the tumor architecture were also assessed in every representative section. Tumor necrosis was defined as the presence of microscopic coagulative necrosis, whereas the presence of necrosis on gross examination was not included. The extent of tumor necrosis was assessed in a semi-quantitative manner at low magnification (reduced from ×40). Tumor necrosis was recorded as absent or focally present (0–10% of tumor area) or extensively present (>10% of the tumor area).

In order to ensure the validity of the pathological data extraction, two pathologists independently reviewed specimens from 145 randomly selected patients while blinded to patient clinical parameters and the finding of the other reviewer. Inter-reader reliability measured using the intraclass correlation coefficient was greater than 0.95 for each pathological characteristic.

Surveillance regimen

Patients were generally followed every 3 months for the first year after RNU, and every 6 months from the second year. Follow up consisted of a history, physical examination, serum chemistry studies, urinary cytology, chest radiography, cystoscopic evaluation of the bladder and radiographic evaluation of the contralateral upper urinary tract. Elective bone scans, chest CT or magnetic resonance imaging were carried out when clinically indicated.

Outcome evaluation and statistical analysis

A comparison of overall DFS and CSS was carried out between the three decades using Kaplan–Meier survival analysis. Bladder recurrences were not considered in the analysis of DFS rate. Finally, a comparison of survival rates among the three decades based on the definitive pathological stage (organ-confined T1/T2N0 vs non-confined T3/T4 and/or N+) was carried out. Differences were determined by log–rank test with significant P-values <0.05. All statistical tests were carried out with SPSS, version 19.0 (SPSS, Chicago, IL, USA).


Patient clinicopathological characteristics by decade

A total of 1462 RNU patients were identified, including 106 in decade 1, 655 in decade 2 and 701 in decade 3. Median follow up was 36 months (range 1.2–250 months). Table 1 lists the clinicopathological characteristics stratified according to the decade of treatment. There were no significant differences in patient age, sex, performance status, previous history of bladder cancer or tumor location between the three decades. In the past two decades, the majority of patients were diagnosed when tumors were asymptomatic as opposed to those in decade 1. There was a significant increase in the percentage of patients with high-grade, sessile tumor architecture and tumor necrosis.

Table 1. Clinical and pathological characteristics of 1462 patients who underwent RNU for UTUC categorized by decade
Decade1 (<1990s)2 (1990–1999)3 (2000–2007)Total P-value
Total no. patients (%)106 (7.3)655 (44.8)701 (47.9)1462 (100) 
Clinical characteristics     
 Age (years)    0.7
  Mean ± SD68 ± 1067 ± 1169 ± 1168 ± 11
 Sex (%)    0.7
  Male69 (65)441 (67)482 (68)992 (68)
  Female37 (35)214 (33)219 (32)470 (32)
 ECOG (%)    0.07
  049 (88)365 (68)425 (66)839 (68)
  16 (11)138 (26)176 (27)320 (26)
  21 (1)31 (6)42 (7)74 (6)
  30 (0)2 (<0.5)4 (<0.5)6 (<0.5)
 Tumor location (%)    0.76
  Kidney75 (72)421 (61)456 (65)952 (65)
  Ureter31 (28)234 (39)245 (35)510 (35)
 Symptoms (%)    0.0001
  Present13 (54)109 (39)78 (24)200 (32)
  Absent11 (46)170 (61)248 (76)429 (68)
 Previous diagnosis of bladder UC (%)    0.08
  Absent88 (83)491 (75)511 (73)1090 (75)
  Present18 (17)164 (25)190 (27)372 (25)
Pathological characteristics     
 Pathological stage (%)    0.01
  Ta/Tis25 (24)125 (19)185 (26)335 (23)
  T125 (24)140 (21)149 (21)314 (22)
  T215 (14)149 (23)105 (15)269 (18)
  T339 (37)201 (31)235 (34)475 (32)
  T42 (1)40 (6)27 (4)69 (5)
 Pathological grade (%)    0.0001
  Low58 (55)266 (41)205 (29)529 (36)
  High48 (45)389 (59)496 (71)933 (64)
 Pathological nodal status (%)    0.34
  Negative28 (74)225 (77)204 (76)457 (76)
  Positive10 (26)68 (23)65 (24)143 (24)
 Tumor extent    0.54
  Organ-confined (pTa/Tis-T2 and cN0)94 (89)567 (87)620 (88)1281 (88)
  Non-organ-confined (pT3–4 or pN+)12 (11)88 (12)81 (12)181 (12)
 Distant Metastasis    0.72
  Absent106 (100)652 (99)697 (99)1455 (99)
  Present0 (0)3 (1)4 (1)7 (1)
 Concomitant upper tract CIS (%)    0.017
  Absent80 (75)491 (75)479 (68)1050 (72)
  Present26 (25)164 (25)222 (32)412 (28)
 Tumor architecture (%)    0.013
  Papillary89 (85)482 (74)498 (71)1069 (73)
  Sessile16 (15)171 (26)201 (29)388 (27)
 LVI (%)    0.179
  Absent87 (82)484 (74)532 (76)1103 (75)
  Present19 (18)171 (26)169 (24)359 (25)
 Necrosis (%)    0.011
  Absent90 (85)475 (73)553 (79)1118 (77)
  Present16 (15)180 (27)148 (21)344 (33)
 RNU (%)    0.0001
  Open106 (100)643 (98)383 (54)1132 (77)
  Laparoscopic0 (0)12 (2)318 (46)330 (23)
 Lymphadenectomy (%)    0.570.19
  Not carried out60 (57)379 (58)429 (60)862 (59)
  Carried out46 (33)276 (32)278 (40)600 (39)
 Lymphadenectomy for high grade disease (%)    
  Not carried out19 (40)190 (49)263 (53)472 (51)
  Carried out29 (60)199 (51)233 (47)461 (49)
 Endoscopic management of tumor before extirpative treatment5 (5)60 (9.2)139 (20)204 (23)0.0001
 Neoadjuvant chemotherapy (%)    0.0001
  Not administered106 (100)647 (99)662 (94)1415 (97)
  Administered0 (0)8 (1)39 (6)47 (3)
 Adjuvant chemotherapy (%)    0.0001
  Not administered100 (94)584 (89)606 (87)1290 (88)
  Administered6 (6)71 (11)94 (13)171 (12)

There were no significant differences in the distribution of organ-confined versus non-organ-confined disease, performance of LND during RNU or pathological nodal status among the three decades. However, there was a significant increase in laparoscopic RNU and endoscopic management of UTUC between the second and third study periods. The utilization of perioperative chemotherapy significantly increased between decades 1 and 2; and remained similar between the past two decades. However, multimodal therapy was rarely utilized in any period.

Disease-free and cancer-specific survival

Disease recurrence was identified in 379 (28%) UTUC patients at a median of 10.4 months after RNU. A total of 510 (37%) patients died in the follow-up duration, with 313 (61%) of deaths due to UTUC. The median time to all cause and cancer-specific mortality was 24 months and 18.5 months, respectively. Figure 1 shows the DFS and CSS according to study decade. Actuarial 5-year DFS estimates were 66 ± 5%, 68.5 ± 2% and 71 ± 2%, and 5-year CSS estimates were 75 ± 5%, 72 ± 2% and 75 ± 2% for decades 1, 2 and 3, respectively. Similarly, Figure 2 shows the 5-year DFS and CSS per decade for patients with organ-confined (pT ≤2 and pN0/Nx) UTUC, whereas Figure 3 shows outcomes of patients with non-organ-confined (pT ≥3 or pN+) disease. After adjusting for pathological stage, there were no statistically significant differences in DFS and CSS between the three decades of treatment (all log–rank, P > 0.05).

Figure 1.

(a) The 5-year DFS and (b) 5-year CSS of 1462 patients who underwent RNU for UTUC stratified by the decade of surgical treatment. inline image, Decade 1; inline image, Decade 2; inline image, Decade 3.

Figure 2.

(a) The 5-year DFS and (b) 5-year CSS of 1281 patients with organ-confined UTUC (≤pT2, pN0/Nx) who underwent RNU for UTUC stratified by the decade of surgical treatment. inline image, Decade 1; inline image, Decade 2; inline image, Decade 3.

Figure 3.

(a) The 5-year DFS and (b) 5-year CSS of 181 patients with non-organ confined disease (>pT3, and/or pN+) who underwent RNU for UTUC stratified by the decade of surgical treatment. inline image, Decade 1; inline image, Decade 2; inline image, Decade 3.


We have provided an analysis of clinicopathological trends and patient outcomes of UTUC patients undergoing RNU occurring over three decades. The time periods were chosen meaningfully, as open surgical extirpation was the mainstay of treatment in 1980s (period 1), there was an increase in laparoscopic and endoscopic interventions in the 1990s (period 2),9 and more frequent use of perioperative chemotherapy for urothelial cancer in the 2000s era (period 3).10

Similar to earlier reports and despite advances in diagnostics, surgical techniques and systemic chemotherapeutics, there has been no significant improvement in outcomes after RNU over a 30-year period. Munoz and Ellison reviewed national trends of incidence and survival of UTUC using the SEER database.3 Among 9072 cases of UTUC identified from 1973 to 1996, the overall 5-year CSS was 75%, with a 62.6% survival rate for regional disease. In addition, in a smaller series of patients, Brown et al. evaluated treatment outcomes of patients undergoing RNU for UTUC over a 20-year period, and found no difference in patient outcomes among the three time periods.4 One advantage of this analysis compared with analyses using SEER data is that all patients were treated at academic centers with high volume and expertise resulting in more uniform care. Furthermore, we were able to evaluate treatment over the decades and correlate with clinicopathological parameters and surgical approach. Furthermore, the present study found no improvement in patient survival after RNU over time, along with a relatively constant distribution of both organ and non-organ-confined UTUC over three decades. Surprisingly, there was no stage migration identified despite the fact that more patients were diagnosed when asymptomatic and with the benefit of improvements in imaging and ureteroscopy.

The observed troubling trends in outcomes after RNU are disappointing; however, they mirror the lack of progress in management of BC. There is an absence of screening of patients for urothelial cancer and scarce use of multimodal therapies. There has not been any reduction in locally advanced or metastatic disease, and no improvement in performance of LND, which could improve staging and potentially outcomes. BC patients have benefited from early detection and intervention, adequate lymph node dissection during cystectomy, and use of perioperative chemotherapy,1,11 Assuming similar biological characteristics, the aforementioned factors might improve outcomes for UTUC. However, the present results show that there has been no change in the distribution of pathological stage of UTUC between the three decades, reflecting a lack of earlier detection and intervention in UTUC. One possible reason for the lack of stage migration over time lies in the difficulties in proper staging of UTUC. Currently, this is hindered by the fine endoscopic instruments used for UTUC biopsy.12 Consequently, the majority of the prognostic factors, such as stage, grade and LVI, are clinically difficult to obtain before the surgery. Furthermore, a review of 37 patients with UTUC who underwent preoperative CT staging showed understaging in 16.2% and no detection of disease in 24%, with the conclusion that CT did not alter management in any patient.13 Margulis et al. have recently proposed a preoperative multivariable prognostic nomogram for the prediction of non-organ-confined UTUC using clinical parameters readily available before RNU, namely; tumor location, grade and architecture.14 Clinical application of this nomogram might improve preoperative risk-stratification and thoughtful integration of presurgical systemic therapy.

Lack of systematic integration of regional LND at the time of RNU might provide additional explanation for the minimal improvement in outcomes after RNU. Historically, 20–40% of UTUC patients had lymph node metastasis at the time of RNU depending on the stage of the primary tumor,15–17 and yet LND is used sporadically and without consistent surgical templates. There was no statistical difference in the percentage of LND over three decades in the current series, with <40% of patients undergoing LND and <1% increase in total LND carried out between the second and third decades. Importantly, the number of patients with high-grade disease who underwent LND did not change significantly between the decades. Considering the similarity in biological behavior of UC (whether upper or lower tract UC), then the substantial role of LND as a prognostic and therapeutic component cannot be denied. LND in UTUC was reported to be associated with improved CSS, and at times was curative in a subset of patients with regional nodal disease.5,18 Given the significant rates of understaging, performance of LND has been advocated as an adjunct to all RNU.19

Continuous limited use of systemic chemotherapy is another contributing factor to lack of significant improvement in oncological outcomes. Despite the five- to sixfold increase in the use of perioperative chemotherapy between the first and second periods, utilization remained essentially unchanged between the second and third periods. There was an overall higher utilization of adjuvant compared with neoadjuvant chemotherapy in this cohort. This could be explained by the frequent inclusion of all urothelial carcinoma patients by the medical oncology community in adjuvant chemotherapy trials regardless of the site of origin.10

There is compelling evidence for the use of neoadjuvant platinum-based chemotherapy for high-risk BC patients.20,21 The potential advantages of neoadjuvant chemotherapy in UTUC include potential eradication of subclinical metastases, improved patient tolerability before surgical extirpation and ability to deliver higher chemotherapy doses before loss of global renal function.10 Although RNU provides adequate local control, nearly a quarter of patients with UTUC die of their disease in 4–5 years.5 This is thought to be because of unrecognized systemic microscopic metastases.22 In a study by Matin et al., 43 patients received neoadjuvant chemotherapy before RNU and were compared with 107 patients in the control group, there was a significant down-staging of pT2-3 UTUC along with a 14% complete response for patients in the study group.23 Youssef et al. noted 5-year CSS rates of 44% in patients receiving neoadjuvant chemotherapy with biopsy-proven loco-regional nodal metastasis compared with 36% in patients undergoing RNU without chemotherapy with positive lymph nodes.6 In addition, Igawa et al. cited a 53% response rate to neoadjuvant chemotherapy for patients with UTUC in a small study of 15 patients, noting a positive correlation between pathological response and patient prognosis.24 In contrast, the role of adjuvant chemotherapy is not well-defined. Our UTUC collaboration showed that adjuvant chemotherapy (used in 22% of patients) yielded no improvements in overall or CSS. However, it was administered in high-stage and grade disease.25

The present study had several limitations. First are the limitations inherent to a retrospective multicenter study design with potential selection bias. Also, the paucity of patients in the first decade compared with the other time periods might not reflect the true trends and patterns of treatment for that era. In addition, treatment of UTUC in multiple institutions introduces variability in surgical technique, surveillance and pathological interpretation. However, the differences in practice patterns across the institutions might be reflective of the real world, making the conclusions of the present study more generalizable. Including a huge number of patients from the large academic centers, and implementing rigorous clinical and pathological review of data provides so far the most useful tool to evaluate where we stand regarding the management of UTUC, trying to identify the potential space for improvements necessary for better patient outcomes.

Outcomes after RNU have not changed significantly over the past three decades, despite staging and surgical refinements, and availability of active systemic treatment strategies. Utilization of perioperative systemic chemotherapy in UTUC management remains low. Improvements in treatment outcomes necessitate rigorous investigation, and application of multimodal treatment approaches and enhanced identification of high-risk patients.

Conflict of interest

None declared.