- Email: [email protected]
Impact of Tumor Location on Prognosis for Patients with Upper Tract Urothelial Carcinoma Managed by Radical Nephroureterectomy
EUROPEAN UROLOGY 57 (2010) 1072–1079
available at www.sciencedirect.com journal homepage: www.europeanurology.com
Urothelial Cancer
Impact of Tumor Location on Prognosis for Patients with Upper Tract Urothelial Carcinoma Managed by Radical Nephroureterectomy Jay D. Raman a,b,*, Casey K. Ng a, Douglas S. Scherr a, Vitaly Margulis c, Yair Lotan d, Karim Bensalah e, Jean-Jacques Patard e, Eiji Kikuchi f, Francesco Montorsi g, Richard Zigeuner h, Alon Weizer i, Christian Bolenz j, Theresa M. Koppie k, Hendrik Isbarn l, Claudio Jeldres l, Wareef Kabbani d, Mesut Remzi m, Mathias Waldert m, Christopher G. Wood c, Marco Roscigno g, Mototsuga Oya f, Cord Langner h, J. Stuart Wolf i, Philipp Stro¨bel j, Mario Ferna´ndez n, Pierre Karakiewcz l, Shahrokh F. Shariat c,o a
New York-Presbyterian Hospital, Weill Cornell Medical Center, New York, NY, USA
b
Penn State Milton S. Hershey Medical Center, Hershey, PA, USA
c
University of Texas, MD Anderson Cancer Center, Houston, TX, USA
d
University of Texas, Southwestern Medical Center, Dallas, TX, USA
e
University of Rennes, Rennes, France
f
Keio University School of Medicine, Tokyo, Japan
g
Vita-Salute University, Milan, Italy
h
Medical University Graz, Graz, Austria
i
University of Michigan, Ann Arbor, MI, USA
j
University Hospital Mannheim, Mannheim, Germany
k l
University of California Davis, Sacramento, CA, USA
University of Montreal, Montreal, Quebec, Canada
m
Medical University of Vienna, Vienna, Austria
n
Clinica Alemana de Santiago, Santiago de Chile, Chile
o
Memorial Sloan-Kettering Cancer Center, New York, NY, USA
Article info
Abstract
Article history: Accepted July 3, 2009 Published online ahead of print on July 15, 2009
Background: There is a lack of consensus regarding the prognostic significance of ureteral versus renal pelvic upper tract urothelial carcinoma (UTUC). Objective: To investigate the association of tumor location on outcomes for UTUC in an international cohort of patients managed by radical nephroureterectomy (RNU). Design, setting, and participants: A retrospective review of institutional databases from 10 institutions worldwide identified patients with UTUC. Intervention: The 1249 patients in the study underwent RNU with ipsilateral bladder cuff resection between 1987 and 2007. Measurements: Data accrued included age, gender, race, surgical approach (open vs laparoscopic), tumor pathology (stage, grade, lymph node status), tumor location, use of perioperative chemotherapy, prior endoscopic therapy, urothelial
Keywords: Urothelial carcinoma Renal pelvis Ureter Recurrence Survival
* Corresponding author. Penn State Milton S. Hershey Medical Center, Division of Urology, 500 University Drive, C4830B, Hershey, PA 17033, USA. Tel. +1 717 531 6979; Fax: +1 717 531 4475. E-mail address: [email protected] (J.D. Raman). 0302-2838/$ – see back matter # European Association of Urology. Published by Elsevier B.V. All rights reserved.
doi:10.1016/j.eururo.2009.07.002
EUROPEAN UROLOGY 57 (2010) 1072–1079
1073
carcinoma recurrence, and mortality from urothelial carcinoma. Tumor location was divided into two groups (renal pelvis and ureter) based on the location of the dominant tumor. Results and limitations: The 5-yr recurrence-free and cancer-specific survival estimates for this cohort were 75% and 78%, respectively. On multivariate analysis, only pathologic tumor (pT) classification ( p < 0.001), grade ( p < 0.02), and lymph node status ( p < 0.001) were associated with disease recurrence and cancer-specific survival. When adjusting for these variables, there was no difference in the probability of disease recurrence (hazard ratio [HR]: 1.22; p = 0.133) or cancer death (HR: 1.23; p = 0.25) between ureteral and renal pelvic tumors. Adding tumor location to a base prognostic model for disease recurrence and cancer death that included pT stage, tumor grade, and lymph node status only improved the predictive accuracy of this model by 0.1%. This study is limited by biases associated with its retrospective design. Conclusions: There is no difference in outcomes between patients with renal pelvic tumors and with ureteral tumors following nephroureterectomy. These data support the current TNM staging system, whereby renal pelvic and ureteral carcinomas are classified as one integral group of tumors. # European Association of Urology. Published by Elsevier B.V. All rights reserved.
1.
Introduction
Upper tract urothelial carcinoma (UTUC) specifically comprises cancers extending from the renal calyces to the ipsilateral ureteral orifice. UTUC accounts for approximately 5% of urothelial malignancies and <10% of renal tumors [1]. Presently, radical nephroureterectomy (RNU) with removal of an ipsilateral bladder cuff constitutes the primary therapeutic modality for individuals with bulky, invasive, and/or highgrade UTUC and a normal contralateral renal unit. Primary tumor classification, pathologic grade, lymph node status, and extent of surgery have been implicated as significant prognostic factors in patients with UTUC [2–6]. Another potential prognostic variable is the location of the tumor within the upper urinary tract. Although tumors within the renal pelvis are three to four times more common than ureteral lesions [5,7], studies have suggested that ureteral disease often confers a worse prognosis [8,9]. These reports, however, are predominantly single-institution series that are potentially limited by small patient numbers and case selection bias. Moreover, the homogeneity of the study population in these series (as related to diagnosis, patient selection, staging, pathologic evaluation, and treatment) raises concerns regarding the ability to generalize these findings. Data compiled from multiple institutions will be less likely to suffer from biases inherent in single-center series. Therefore, we constructed a database incorporating comprehensive RNU data and outcomes from 13 centers worldwide using strict accrual criteria and pathologic rereview of all slides. In the current study, we assess the impact of tumor location on clinical outcomes in >1200 patients managed by RNU for UTUC. 2.
Patients and methods
2.1.
Patient selection
agreements before initiation of the study. A total of 13 centers worldwide provided data: New York-Presbyterian Hospital, Weill Cornell Medical Center, New York, NY, USA; University of Texas, MD Anderson Cancer Center, Houston, TX, USA; University of Texas, Southwestern Medical Center, Dallas, TX, USA; University of Rennes, Rennes, France; Keio University School of Medicine, Tokyo, Japan; VitaSalute University, Milan, Italy; Medical University Graz, Graz, Austria; University of Michigan, Ann Arbor, MI, USA; University Hospital Mannheim, Mannheim, Germany; University of California Davis, Sacramento, CA, USA; University of Montreal, Montreal, Quebec, Canada; Medical University of Vienna, Vienna, Austria; Clinica Alemana de Santiago, Santiago de Chile, Chile. Data included age, gender, race, surgical approach (open vs laparoscopic), tumor pathology (stage, grade, lymph node status), tumor location, use of perioperative chemotherapy, prior endoscopic therapy, urothelial carcinoma recurrence, and mortality from urothelial carcinoma. A computerized databank was generated for data transfer. After combining the data sets, reports were generated for each variable to identify data inconsistencies and other data integrity problems. Through regular communication with all sites, resolution of all identified anomalies was achieved before analysis. Before final analysis, the database was frozen and the final data set was produced for the current analysis. This study comprised 1249 patients treated at participating institutions with comprehensive clinical and pathologic data available for review who underwent RNU with ipsilateral bladder cuff resection between 1987 and 2007. Tumor location was divided into two groups (renal pelvis and ureter) based on the location of the dominant tumor as identified in the final pathologic specimen. The dominant lesion was defined as that with the highest pathologic tumor (pT) stage. For multifocal tumors with the same stage, tumor size was then used to define the index lesion for location classification.
2.2.
Pathologic evaluation
All surgical specimens were processed according to standard pathologic procedures, and all slides were re-reviewed by institutional genitourinary pathologists according to identical strict criteria. Pathologists were blinded to clinical outcomes. All specimens were histologically confirmed to be urothelial carcinoma. Tumor classification was assessed according to the 2002 American Joint Committee on Cancer/Union Internationale Contre le Cancer TNM classification. Tumor grading was
This was an institutional review board–approved study with all
assessed according to the 1998 World Health Organization/International
participating sites providing the necessary institutional data use
Society of Urologic Pathology consensus classification.
1074
EUROPEAN UROLOGY 57 (2010) 1072–1079
Table 1 – Descriptive characteristics of 1249 patients with upper tract urothelial carcinoma managed by radical nephroureterectomy stratified by tumor location Variable
All patients (n = 1249)
Mean (median) age, yr; range
Tumor location Renal pelvis (n = 823)
Ureter (n = 426)
p value
68 (70); 27–97
68 (69); 27–97
69 (70); 28–97
0.109
Gender, No. (%) Male Female
846 (67.7) 403 (32.3)
543 (66.0) 280 (34.0)
303 (71.1) 123 (29.9)
0.065
pT stage, No. (%) pTa pTis pT1 pT2 pT3 pT4
283 24 298 237 365 42
187 12 183 127 275 39
96 12 115 110 90 3
pN stage, No. (%) pNX pN0 pN+
768 (61.5) 405 (32.4) 76 (6.1)
520 (63.2) 248 (30.1) 55 (6.7)
248 (58.2) 157 (36.9) 21 (4.9)
Tumor grade, No. (%) Low High
498 (39.9) 751 (60.1)
338 (41.1) 485 (58.9)
160 (37.6) 266 62.4)
*
(22.7) (1.9) (23.9) (19) (29.2) (3.4)
(22.7) (1.5) (22.2) (15.4) (33.4 (4.7)
(22.5) (2.8) (27.0) (25.8) (21.1) (0.70)
<0.001
0.404*
0.230
pN+ vs pN0/pNx.
2.3.
expressed as proportions and compared with the Mantel-Haenszel
Surveillance regimen
test. All reported p values are two-sided, and statistical significance Follow-up for each patient was determined according to individual
was set at 0.05. No adjustments were made for the number of p
institution protocols. In general, patients were followed every 3–4 mo
values calculated. Statistical analysis was performed with S-Plus
for the first year following RNU, every 6 mo from the second through the
Professional (MathSoft Inc, Seattle, WA, USA).
fifth year, and annually thereafter. Follow-up consisted of a history, physical examination, routine blood work and serum chemistry
3.
Results
upper urinary tract. Elective bone scans, chest computed tomography,
3.1.
Association of tumor location with outcomes in all patients
or magnetic resonance imaging were performed when clinically
(n = 1249)
studies, urinary cytology, chest radiography, cystoscopic evaluation of the urinary bladder, and radiographic evaluation of the contralateral
indicated. Disease recurrence was defined as local failure in the tumor bed, regional lymph nodes, or distant metastasis. Bladder recurrences were not considered in the analysis of recurrence-free survival rate. Cause of death was determined by the treating physicians, by chart review corroborated by death certificates, or by death certificates alone. Perioperative mortality (any death within 30 d of surgery or before discharge) was censored at time of death for urothelial disease-specific survival analyses.
2.4.
Statistical analysis
The x2 test was used to evaluate the association between categorical variables, and the Mann Whitney test assessed for differences in variables with a continuous distribution across dichotomous categories. Univariate recurrence and survival probabilities after RNU were determined using the Kaplan-Meier method with differences estimated using the log-rank test. Univariate and multivariate Cox proportional hazards regression models addressed time to recurrence and disease-specific mortality after RNU. In all models, proportional hazards
assumptions
were
systematically
verified
using
the
Grambsch-Therneau residual-based test. The change in predictive accuracy resulting from the addition of a variable of interest to standard predictor variables was quantified with the Harrell concordance index [10,11]. Internal validation was performed using 200 bootstrap resamples [10,11]. Predictive accuracy estimates were
3.1.1.
Clinicopathologic characteristics
Table 1 summarizes the clinical and pathologic characteristics of patients as stratified by tumor location. There were 846 men (68%) and 403 women (32%) included with a median age of 70 yr. Overall, 66% of tumors were classified as renal pelvic and 34% were ureteral. The stage distribution of UTUC tumors in this cohort was 23% pTa, 2% pTis, 24% pT1, 19% pT2, 29% pT3, and 3% pT4. Sixty percent of patients had high-grade urothelial tumors, and 6% had positive lymph nodes. Regional lymphadenectomy was generally performed in patients with enlarged lymph nodes on preoperative axial imaging or with adenopathy detected during intraoperative examination. As such, 61% of patients in this cohort did not undergo a lymphadenectomy (pNx). When comparing renal pelvic tumors and ureteral tumors, there were no differences in age, gender distribution, nodal status (pN+ vs pN0/pNx), tumor grade, or followup interval (Table 1). Interestingly, for pT classification, renal pelvic tumors had a more advanced presentation than ureteral disease ( p < 0.001). Specifically, 38% of renal pelvis versus 22% of ureteral tumors presented with locally advanced (T3/T4) tumors.
EUROPEAN UROLOGY 57 (2010) 1072–1079
3.1.2.
Urothelial cancer recurrence
Median follow-up for this patient cohort was 49 mo (mean: 62.7; range: 0.1–250). Cancer recurrence occurred in 277 patients (22%) with a 5-yr recurrence-free survival estimate of 75% (Fig. 1a). On univariate analyses, pT classification ( p < 0.001), tumor grade ( p < 0.001), and lymph node status ( p < 0.001) were all associated with disease recurrence. No difference was noted in the probability of urothelial cancer recurrence between ureteral and renal pelvic tumors (HR: 1.0; log rank p value: 0.7) (Fig. 1a). On multivariate analysis, only pT classification ( p < 0.001), grade ( p < 0.001), and lymph node status ( p < 0.001) were associated with disease recurrence. When adjusting for the effects of these variables, there was no difference in the probability of recurrence-free survival between renal pelvis and ureteral tumors (HR: 1.220; p = 0.133; Table 2). The predictive accuracy of a base prognostic model for disease recurrence that included T classification, tumor
1075
grade, and lymph node status was 79.6%. Addition of tumor location did not improve the predictive accuracy of this base model (area under the curve [AUC]: 79.7%; 0.1% gain). 3.1.3.
Cancer-specific survival
Urothelial cancer death occurred in 238 patients (19%), and 436 patients (34.9%) died of causes unrelated to urothelial cancer. The 5-yr cancer-specific survival estimate was 78% (Fig. 1b). On univariate analyses, pT classification ( p < 0.001), grade ( p < 0.001), and lymph node status ( p < 0.001) were associated with cancerspecific survival. However, there was no difference in the probability of cancer-specific survival between ureteral and renal pelvic tumors (HR: 1.03; log rank p value: 0.80) (Fig. 1b). On multivariate analysis, pT classification ( p < 0.001), grade ( p = 0.017), and lymph node status ( p < 0.001) were associated with worse
Fig. 1 – (a) Kaplan-Meier estimates for recurrence-free survival; (b) cancer-specific survival in 1249 patients managed by radical nephroureterectomy for upper tract urothelial carcinoma. HR = hazard ratio.
1076
EUROPEAN UROLOGY 57 (2010) 1072–1079
Table 2 – Multivariate Cox regression models predicting disease recurrence and cancer-specific survival in 1249 patients following radical nephroureterectomy for urothelial carcinoma of the upper urinary tract UTUC recurrence
pT classification pT1 vs pTa/pTis pT2 vs pTa/pTis pT3 vs pTa/pTis pT4 vs pTa/pTis Tumor grade High vs low
UTUC death
HR
95% CI
p value
HR
95% CI
2.388 5.130 11.733 34.307
1.152–4.952 2.585–10.179 6.082–22.636 16.532–71.192
<0.001 0.019 <0.001 <0.001 <0.001
2.071 5.098 9.827 25.588
0.848–5.058 2.212–11.753 4.384–22.029 9.854–66.445
2.310
1.627–3.281
<0.001
1.697
pN classification pN0 vs pNx pN+ vs pNx
0.501 1.321
0.378–0.663 0.974–1.792
<0.001 <0.001 0.073
Tumor location Ureter vs RP
1.220
0.942–1.580
0.133
p value <0.001 0.110 <0.001 <0.001 <0.001
1.100–2.617
0.017
0.418 1.371
0.279–0.624 0.890–2.112
<0.001 <0.001 0.152
1.230
0.864–1.750
0.250
CI = confidence interval; HR = hazard ratio; RP = renal pelvis; UTIC = upper tract urothelial carcinoma.
cancer-specific survival. There was no difference in cancer-specific survival between renal pelvic and ureteral tumors (HR: 1.230; p = 0.25; Table 2). The predictive accuracy of a base model for cancerspecific survival was 80.4%. Addition of tumor location improved the predictive accuracy of this base model by only 0.1% (AUC: 80.5%). 3.2.
Association of tumor location with outcomes in patients
3.2.1.
Clinicopathologic characteristics
Table 3 highlights the clinical and pathologic characteristics of patients with organ-confined disease stratified by tumor location. In this subgroup, when comparing renal pelvic to ureteral tumors, there were no differences in age, gender distribution, or follow-up interval. Ureteral tumors did have a higher distribution of pT2 lesions (31.9% vs 24.3%; p = 0.028) and high-grade tumors (53.4% vs 44.2%; p = 0.010).
with organ-confined (pT2 or lower, pN0/pNx) disease (n = 824) 3.2.2.
We hypothesized that patients with advanced (pT3/T4) local disease or node-positive disease would be more likely to receive neoadjuvant or adjuvant chemotherapy. To limit this as a potential confounding variable when studying the impact of tumor location, we performed a subgroup analysis of patients with organ-confined disease who did not receive perioperative chemotherapy.
Urothelial cancer recurrence
Median follow-up for this patient cohort was 52.8 mo (mean: 66.2; range: 0.1–250). Cancer recurrence occurred in 86 patients (10%) with a 5-yr recurrence-free survival estimate of 88% (Fig. 2a). On univariate analyses, pT classification ( p < 0.001), tumor grade ( p < 0.001), and tumor location ( p = 0.02) were all associated with disease recurrence. Specifically, patients with ureteral tumors
Table 3 – Descriptive characteristics of 824 patients with organ-confined [pT2 or lower] upper tract urothelial carcinoma managed by radical nephroureterectomy Variable
Mean (median) age, yr; range
All patients (n = 824)
Tumor location Renal pelvis (n = 498)
Ureter (n = 326)
p value
68 (69); 30–96
68 (68); 30–96
69 (70); 30–91
0.195
Gender, No. (%) Male Female
571 (69.3) 253 (31.7)
335 (67.3) 163 (32.7)
236 (72.4) 90 (27.6)
0.119
pT stage, No. (%) pTa pTis pT1 pT2
283 24 292 225
187 12 178 121
96 12 114 104
0.028
pN stage, No. (%) pNX pN0
563 (68.3) 261 (31.7)
356 (71.5) 142(28.5)
207 (63.5) 119 (36.5)
Tumor grade, No. (%) Low High
430 (52.2) 394 (47.8)
278 (55.8) 220 (44.2)
152 (46.6) 174 (53.4)
(34.3) (2.9) (35.4) (27.3)
(37.6) (2.4) (35.7) (24.3)
(29.5) (3.7) (35.0) (31.9)
0.016
0.010
1077
EUROPEAN UROLOGY 57 (2010) 1072–1079
Fig. 2 – (a) Kaplan-Meier estimates for recurrence-free survival; (b) cancer-specific survival in 824 patients with organ-confined upper tract urothelial carcinoma managed by radical nephroureterectomy. HR = hazard ratio.
had a lower 3-yr (87% vs 92%), 5-yr (86% vs 91%), and 7-yr (84% vs 89%) recurrence-free survival probability compared with renal pelvic tumors (HR: 1.6; log rank p value: 0.02; Fig. 2a). However, on multivariate analysis, only pT
classification ( p < 0.001) and grade ( p = 0.002) were associated with disease recurrence. When adjusting for the effects of these two variables, there was no difference in the probability of recurrence-free survival between
Table 4 – Multivariate Cox regression models predicting disease recurrence and cancer-specific survival in 824 patients with organ-confined upper tract urothelial carcinoma (UTUC; pT2 or lower, pN0/pNx) following radical nephroureterectomy UTUC recurrence
UTUC death
HR
95% CI
p value
HR
95% CI
pT classification pT1 vs pTa/pTis pT2 vs pTa/pTis
2.507 5.023
1.320–4.761 2.718–9.282
<0.001 0.005 <0.001
2.605 5.764
1.253–5.416 2.884–11.520
Tumor grade High vs low
1.951
1.280–2.972
0.002
1.604
1.022–2.518
0.040
Tumor location Ureter vs RP
1.228
0.844–1.788
0.283
1.333
0.877–2.025
0.178
CI = confidence interval; HR = hazard ratio; RP = renal pelvis.
p value <0.001 0.010 <0.001
1078
EUROPEAN UROLOGY 57 (2010) 1072–1079
renal pelvis and ureteral tumors (HR: 1.228; p = 0.283; Table 4). The predictive accuracy of a base prognostic model for disease recurrence for patients with organ-confined disease that included tumor classification and grade was 73.0%. Addition of tumor location only marginally improved the predictive accuracy of this base model (AUC: 73.5%; 0.5% gain). 3.2.3.
Cancer-specific survival
In this subgroup, urothelial cancer death occurred in 74 patients (9%), and 220 patients (26.7%) died of nonurothelial causes. The 5-yr cancer-specific survival estimate was 90%. On univariate analyses, pT classification ( p < 0.001), tumor grade ( p < 0.001), and tumor location ( p = 0.02) were all associated with disease recurrence. Similar to disease recurrence, ureteral tumors had lower 3-yr (90% vs 93%), 5-yr (87% vs 92%), and 7-yr (84% vs 90%) cancer-specific survival probability compared with renal pelvic tumors (HR: 1.8; log rank p value: 0.02; Fig. 2b). On multivariate analysis, pT classification ( p < 0.001) and grade ( p = 0.04) were associated with worse cancer-specific survival. There was no statistical difference in cancer-specific survival, however, between renal pelvic and ureteral tumors (HR: 1.333; p = 0.178; Table 4). The predictive accuracy of a base model for cancerspecific survival for patients with organ-confined UTUC using pT stage and grade was 73.0%. Addition of tumor location improved the predictive accuracy of this base model by only 0.5% (AUC: 73.5%). 4.
Discussion
Nephroureterectomy with ipsilateral bladder cuff excision is considered the gold-standard therapy for the management of invasive, high-grade, or bulky UTUC [2,12]. Despite adequate surgical therapy, a significant percentage of patients experience either local or distant failure [13]. Identification of prognostic variables for relapse and survival is essential to provide more accurate patient counseling, guidance for operative strategies, as well as recommendations for postoperative surveillance and adjuvant therapy regimens. Obtaining critical information for UTUC, however, is difficult owing to the rarity of the disease and the absence of randomized, prospective studies. Indeed, only a few reports of patients with UTUC have systematically examined potential prognostic factors for disease relapse. In this series, we presented data from an international cohort of >1200 patients treated by RNU for UTUC. The overall 5-yr recurrence-free and cancer-specific survival was 75% and 78%, respectively. These oncologic outcomes are comparable with other contemporary series, underscoring that RNU can achieve durable local control even for locally advanced disease. Similar to previous studies, we further confirmed that pT stage, grade, and lymph node status were associated with disease recurrence and cancerspecific survival (Table 2) [2–5]. We noted several interesting differences in pathologic characteristics depending on tumor location. Somewhat
surprisingly, renal pelvic tumors presented at a more advanced tumor stage than ureteral disease when considering locally advanced (T3/T4) cancers (Table 1). Conversely, when only considering organ-confined disease, ureteral tumors were more likely to be muscle invasive (pT2) and high grade (Table 3). Other variables including age, gender, and nodal status, however, were similar between renal pelvis and ureteral tumors. Overall, we failed to detect a difference in urothelial cancer recurrence or cancer-specific survival when stratifying patients by ureteral and renal pelvic tumor location (Fig. 1a and 1b; Table 2). The literature evaluating the impact of tumor location on UTUC outcomes is conflicting. In a single-institution study of 86 UTUC patients treated with RNU, Park and colleagues found that after adjusting for pathologic features, ureteral tumor location was significantly associated with disease recurrence and mortality [9]. These findings were supported by another single-institution study of 72 patients [8]. The hypothesis supporting the worse prognosis of ureteral compared with renal pelvic tumors was attributed to the thin layer of surrounding ureteral adventitia containing an extensive plexus of blood vessels and lymphatic channels making tumor invasion and metastasis easier [8,9,14]. In addition, it was postulated that the renal parenchyma and perihilar adipose tissue surrounding the renal pelvis may act as a barrier against early spread. Our observations from this study contradict this hypothesis. In fact, we found that ureteral tumors were less likely to present with non-organ-confined (pT3/4) disease compared with renal pelvic tumors. What is the potential explanation for our findings of higher tumor classification in renal pelvic disease compared with ureteral tumors? One likely theory is that renal pelvic tumors remain asymptomatic for a longer duration of time, whereas ureteral carcinomas are diagnosed earlier (hence lower stage at diagnosis) due to obstruction that can occur in its narrow lumen. This occurrence can potentially balance out the fact that there is a thinner adventitial wall for the ureteral tumors to invade and may explain the absence of survival differences in this study. Indeed, our classification of tumor location simple reflects a snapshot in time of the upper tract neoplasm, and thus it fails in part to capture the chronicity and biology of disease. In an attempt to minimize confounding effects of neoadjuvant and adjuvant chemotherapy regimens, we further evaluated the subgroup of patients with organconfined disease. In this group, on univariate analysis, patients with ureteral tumors did demonstrate a higher likelihood for disease recurrence (HR: 1.6) and death (HR: 1.8). However, when adjusting for tumor stage and grade, this difference was not significant (Table 4). Ultimately, the value in identifying a prognostic variable is the ability to incorporate it into a predictive model that can allow more accurate risk stratification to tailor follow-up therapy. In this regard, tumor location provided only a minimal incremental predictive gain of 0.1–0.5% when used in conjunction with more established variables including stage, grade, and lymph node status.
EUROPEAN UROLOGY 57 (2010) 1072–1079
Our findings are in concordance with the work from Hall et al [2], who reported similar clinical outcomes between ureteral and renal pelvis tumors in a cohort of 250 consecutive patients with UTUC. This study and ours are the two largest studies examining the association of tumor location with clinical outcomes after RNU. Moreover, our study is the only multicenter study, thereby controlling for biases associated with single-center studies such as specificities related to genetics, diagnosis, patient selection, staging, pathologic evaluation, and management. There are several limitations to the current study. First, it is inherently limited by biases associated with its retrospective design. Data were collected by chart review at participating institutions, thus introducing variability in the interpretation of study variables. Second, variability in practice patterns, surveillance, and patient populations introduces a degree of bias with regard to surgical intervention, patient follow-up, disease severity, access to care, determination of recurrences, and assessment of outcomes. We believe, however, that such variability may allow for these data to be more readily generalizable to a greater population of patients with UTUC. Finally, pathology specimens were not subjected to a centralized review. Rather, all specimens were reexamined by dedicated genitourinary pathologists according to strict criteria agreed on by all participating pathologists. Nonetheless, this may introduce some bias in interpretation of pathologic data.
1079
Waldert, Wood, Roscigno, Oya, Langner, Wolf, Stro¨bel, Ferna´ndez, Karakiewcz, Shariat. Statistical analysis: Raman, Shariat. Obtaining funding: None. Administrative, technical, or material support: None. Supervision: None. Other (specify): None. Financial disclosures: I certify that all conflicts of interest, including specific financial interests and relationships and affiliations relevant to the subject matter or materials discussed in the manuscript (eg, employment/affiliation, grants or funding, consultancies, honoraria, stock ownership or options, expert testimony, royalties, or patents filed, received, or pending), are the following: None. Funding/Support and role of the sponsor: None.
References [1] Jemal A, Siegel R, Ward E, et al. Cancer statistics, 2006. CA Cancer J Clin 2006;56:106–30. [2] Hall MC, Womack S, Sagalowsky AI, Carmody T, Erickstad MD, Roehrborn CG. Prognostic factors, recurrence, and survival in transitional cell carcinoma of the upper urinary tract: a 30-year experience in 252 patients. Urology 1998;52:594–601. [3] Roscigno M, Cozzarini C, Bertini R, et al. Prognostic value of lymph node dissection in patients with muscle-invasive transitional cell carcinoma of the upper urinary tract. Eur Urol 2008;53:794–802. [4] Ozsahin M, Zouhair A, Villa S, et al. Prognostic factors in urothelial renal pelvis and ureter tumours: a multicentre Rare Cancer Network study. Eur J Cancer 1999;35:738–43. [5] Huben RP, Mounzer AM, Murphy GP. Tumor grade and stage as
5.
Conclusions
prognostic variables in upper tract urothelial tumors. Cancer 1988;62:2016–20.
In this large international series of patients treated by RNU, there was no difference in outcomes between patients with renal pelvic and ureteral tumors following nephroureterectomy. The results of our study favor the current concept of tumor staging, in which renal pelvic and ureteral carcinomas are classified together.
[6] Margulis V, Shariat SF, Matin SF, et al. Outcomes of radical
Author contributions: Jay D. Raman had full access to all the data in the
[9] Park S, Hong B, Kim CS, Ahn H. The impact of tumor location on
study and takes responsibility for the integrity of the data and the
prognosis of transitional cell carcinoma of the upper urinary tract.
nephroureterectomy: a series from the Upper Tract Urothelial Carcinoma Collaboration. Cancer 2009;115:1224–33. [7] Murphy DM, Zincke H, Furlow WL. Primary grade 1 transitional cell carcinoma of the renal pelvis and ureter. J Urol 1980;123:629–31. [8] Akdogan B, Dogan HS, Eskicorapci SY, Sahin A, Erkan I, Ozen H. Prognostic significance of bladder tumor history and tumor location in upper tract transitional cell carcinoma. J Urol 2006;176:48–52.
accuracy of the data analysis.
J Urol 2004;171:621–5. [10] Harrell Jr FE, Califf RM, Pryor DB, Lee KL, Rosati RA. Evaluating the
Study concept and design: Raman, Margulis, Shariat.
yield of medical tests. JAMA 1982;247:2543–6.
Acquisition of data: Raman, Ng, Scherr, Margulis, Lotan, Bensalah, Patard,
[11] Harrell Jr FE, Lee KL, Mark DB. Multivariable prognostic models:
Kikuchi, Montorsi, Zigeuner, Weizer, Bolenz, Koppie, Isbarn, Jeldres,
issues in developing models, evaluating assumptions and ade-
Kabbani, Remzi, Waldert, Wood, Roscigno, Oya, Langner, Wolf, Stro¨bel, Ferna´ndez, Karakiewcz, Shariat.
quacy, and measuring and reducing errors. Stat Med 1996;15:
Analysis and interpretation of data: Raman, Shariat. Drafting of the manuscript: Raman, Ng, Scherr, Margulis, Lotan, Bensalah, Patard, Kikuchi, Montorsi, Zigeuner, Weizer, Bolenz, Koppie, Isbarn,
361–87. [12] Krogh J, Kvist E, Rye B. Transitional cell carcinoma of the upper urinary tract: prognostic variables and post-operative recurrences. Br J Urol 1991;67:32–6.
Jeldres, Kabbani, Remzi, Waldert, Wood, Roscigno, Oya, Langner, Wolf, Stro¨bel, Ferna´ndez, Karakiewcz, Shariat.
[13] Heney NM, Nocks BN, Daly JJ, Blitzer PH, Parkhurst EC. Prognostic
Critical revision of the manuscript for important intellectual content:
[14] Miyake H, Hara I, Gohji K, Arakawa S, Kamidono S. The significance
Raman, Ng, Scherr, Margulis, Lotan, Bensalah, Patard, Kikuchi, Montorsi,
of lymphadenectomy in transitional cell carcinoma of the upper
Zigeuner, Weizer, Bolenz, Koppie, Isbarn, Jeldres, Kabbani, Remzi,
urinary tract. Br J Urol 1998;82:494–8.
factors in carcinoma of the ureter. J Urol 1981;125:632–6.
available at www.sciencedirect.com journal homepage: www.europeanurology.com
Urothelial Cancer
Impact of Tumor Location on Prognosis for Patients with Upper Tract Urothelial Carcinoma Managed by Radical Nephroureterectomy Jay D. Raman a,b,*, Casey K. Ng a, Douglas S. Scherr a, Vitaly Margulis c, Yair Lotan d, Karim Bensalah e, Jean-Jacques Patard e, Eiji Kikuchi f, Francesco Montorsi g, Richard Zigeuner h, Alon Weizer i, Christian Bolenz j, Theresa M. Koppie k, Hendrik Isbarn l, Claudio Jeldres l, Wareef Kabbani d, Mesut Remzi m, Mathias Waldert m, Christopher G. Wood c, Marco Roscigno g, Mototsuga Oya f, Cord Langner h, J. Stuart Wolf i, Philipp Stro¨bel j, Mario Ferna´ndez n, Pierre Karakiewcz l, Shahrokh F. Shariat c,o a
New York-Presbyterian Hospital, Weill Cornell Medical Center, New York, NY, USA
b
Penn State Milton S. Hershey Medical Center, Hershey, PA, USA
c
University of Texas, MD Anderson Cancer Center, Houston, TX, USA
d
University of Texas, Southwestern Medical Center, Dallas, TX, USA
e
University of Rennes, Rennes, France
f
Keio University School of Medicine, Tokyo, Japan
g
Vita-Salute University, Milan, Italy
h
Medical University Graz, Graz, Austria
i
University of Michigan, Ann Arbor, MI, USA
j
University Hospital Mannheim, Mannheim, Germany
k l
University of California Davis, Sacramento, CA, USA
University of Montreal, Montreal, Quebec, Canada
m
Medical University of Vienna, Vienna, Austria
n
Clinica Alemana de Santiago, Santiago de Chile, Chile
o
Memorial Sloan-Kettering Cancer Center, New York, NY, USA
Article info
Abstract
Article history: Accepted July 3, 2009 Published online ahead of print on July 15, 2009
Background: There is a lack of consensus regarding the prognostic significance of ureteral versus renal pelvic upper tract urothelial carcinoma (UTUC). Objective: To investigate the association of tumor location on outcomes for UTUC in an international cohort of patients managed by radical nephroureterectomy (RNU). Design, setting, and participants: A retrospective review of institutional databases from 10 institutions worldwide identified patients with UTUC. Intervention: The 1249 patients in the study underwent RNU with ipsilateral bladder cuff resection between 1987 and 2007. Measurements: Data accrued included age, gender, race, surgical approach (open vs laparoscopic), tumor pathology (stage, grade, lymph node status), tumor location, use of perioperative chemotherapy, prior endoscopic therapy, urothelial
Keywords: Urothelial carcinoma Renal pelvis Ureter Recurrence Survival
* Corresponding author. Penn State Milton S. Hershey Medical Center, Division of Urology, 500 University Drive, C4830B, Hershey, PA 17033, USA. Tel. +1 717 531 6979; Fax: +1 717 531 4475. E-mail address: [email protected] (J.D. Raman). 0302-2838/$ – see back matter # European Association of Urology. Published by Elsevier B.V. All rights reserved.
doi:10.1016/j.eururo.2009.07.002
EUROPEAN UROLOGY 57 (2010) 1072–1079
1073
carcinoma recurrence, and mortality from urothelial carcinoma. Tumor location was divided into two groups (renal pelvis and ureter) based on the location of the dominant tumor. Results and limitations: The 5-yr recurrence-free and cancer-specific survival estimates for this cohort were 75% and 78%, respectively. On multivariate analysis, only pathologic tumor (pT) classification ( p < 0.001), grade ( p < 0.02), and lymph node status ( p < 0.001) were associated with disease recurrence and cancer-specific survival. When adjusting for these variables, there was no difference in the probability of disease recurrence (hazard ratio [HR]: 1.22; p = 0.133) or cancer death (HR: 1.23; p = 0.25) between ureteral and renal pelvic tumors. Adding tumor location to a base prognostic model for disease recurrence and cancer death that included pT stage, tumor grade, and lymph node status only improved the predictive accuracy of this model by 0.1%. This study is limited by biases associated with its retrospective design. Conclusions: There is no difference in outcomes between patients with renal pelvic tumors and with ureteral tumors following nephroureterectomy. These data support the current TNM staging system, whereby renal pelvic and ureteral carcinomas are classified as one integral group of tumors. # European Association of Urology. Published by Elsevier B.V. All rights reserved.
1.
Introduction
Upper tract urothelial carcinoma (UTUC) specifically comprises cancers extending from the renal calyces to the ipsilateral ureteral orifice. UTUC accounts for approximately 5% of urothelial malignancies and <10% of renal tumors [1]. Presently, radical nephroureterectomy (RNU) with removal of an ipsilateral bladder cuff constitutes the primary therapeutic modality for individuals with bulky, invasive, and/or highgrade UTUC and a normal contralateral renal unit. Primary tumor classification, pathologic grade, lymph node status, and extent of surgery have been implicated as significant prognostic factors in patients with UTUC [2–6]. Another potential prognostic variable is the location of the tumor within the upper urinary tract. Although tumors within the renal pelvis are three to four times more common than ureteral lesions [5,7], studies have suggested that ureteral disease often confers a worse prognosis [8,9]. These reports, however, are predominantly single-institution series that are potentially limited by small patient numbers and case selection bias. Moreover, the homogeneity of the study population in these series (as related to diagnosis, patient selection, staging, pathologic evaluation, and treatment) raises concerns regarding the ability to generalize these findings. Data compiled from multiple institutions will be less likely to suffer from biases inherent in single-center series. Therefore, we constructed a database incorporating comprehensive RNU data and outcomes from 13 centers worldwide using strict accrual criteria and pathologic rereview of all slides. In the current study, we assess the impact of tumor location on clinical outcomes in >1200 patients managed by RNU for UTUC. 2.
Patients and methods
2.1.
Patient selection
agreements before initiation of the study. A total of 13 centers worldwide provided data: New York-Presbyterian Hospital, Weill Cornell Medical Center, New York, NY, USA; University of Texas, MD Anderson Cancer Center, Houston, TX, USA; University of Texas, Southwestern Medical Center, Dallas, TX, USA; University of Rennes, Rennes, France; Keio University School of Medicine, Tokyo, Japan; VitaSalute University, Milan, Italy; Medical University Graz, Graz, Austria; University of Michigan, Ann Arbor, MI, USA; University Hospital Mannheim, Mannheim, Germany; University of California Davis, Sacramento, CA, USA; University of Montreal, Montreal, Quebec, Canada; Medical University of Vienna, Vienna, Austria; Clinica Alemana de Santiago, Santiago de Chile, Chile. Data included age, gender, race, surgical approach (open vs laparoscopic), tumor pathology (stage, grade, lymph node status), tumor location, use of perioperative chemotherapy, prior endoscopic therapy, urothelial carcinoma recurrence, and mortality from urothelial carcinoma. A computerized databank was generated for data transfer. After combining the data sets, reports were generated for each variable to identify data inconsistencies and other data integrity problems. Through regular communication with all sites, resolution of all identified anomalies was achieved before analysis. Before final analysis, the database was frozen and the final data set was produced for the current analysis. This study comprised 1249 patients treated at participating institutions with comprehensive clinical and pathologic data available for review who underwent RNU with ipsilateral bladder cuff resection between 1987 and 2007. Tumor location was divided into two groups (renal pelvis and ureter) based on the location of the dominant tumor as identified in the final pathologic specimen. The dominant lesion was defined as that with the highest pathologic tumor (pT) stage. For multifocal tumors with the same stage, tumor size was then used to define the index lesion for location classification.
2.2.
Pathologic evaluation
All surgical specimens were processed according to standard pathologic procedures, and all slides were re-reviewed by institutional genitourinary pathologists according to identical strict criteria. Pathologists were blinded to clinical outcomes. All specimens were histologically confirmed to be urothelial carcinoma. Tumor classification was assessed according to the 2002 American Joint Committee on Cancer/Union Internationale Contre le Cancer TNM classification. Tumor grading was
This was an institutional review board–approved study with all
assessed according to the 1998 World Health Organization/International
participating sites providing the necessary institutional data use
Society of Urologic Pathology consensus classification.
1074
EUROPEAN UROLOGY 57 (2010) 1072–1079
Table 1 – Descriptive characteristics of 1249 patients with upper tract urothelial carcinoma managed by radical nephroureterectomy stratified by tumor location Variable
All patients (n = 1249)
Mean (median) age, yr; range
Tumor location Renal pelvis (n = 823)
Ureter (n = 426)
p value
68 (70); 27–97
68 (69); 27–97
69 (70); 28–97
0.109
Gender, No. (%) Male Female
846 (67.7) 403 (32.3)
543 (66.0) 280 (34.0)
303 (71.1) 123 (29.9)
0.065
pT stage, No. (%) pTa pTis pT1 pT2 pT3 pT4
283 24 298 237 365 42
187 12 183 127 275 39
96 12 115 110 90 3
pN stage, No. (%) pNX pN0 pN+
768 (61.5) 405 (32.4) 76 (6.1)
520 (63.2) 248 (30.1) 55 (6.7)
248 (58.2) 157 (36.9) 21 (4.9)
Tumor grade, No. (%) Low High
498 (39.9) 751 (60.1)
338 (41.1) 485 (58.9)
160 (37.6) 266 62.4)
*
(22.7) (1.9) (23.9) (19) (29.2) (3.4)
(22.7) (1.5) (22.2) (15.4) (33.4 (4.7)
(22.5) (2.8) (27.0) (25.8) (21.1) (0.70)
<0.001
0.404*
0.230
pN+ vs pN0/pNx.
2.3.
expressed as proportions and compared with the Mantel-Haenszel
Surveillance regimen
test. All reported p values are two-sided, and statistical significance Follow-up for each patient was determined according to individual
was set at 0.05. No adjustments were made for the number of p
institution protocols. In general, patients were followed every 3–4 mo
values calculated. Statistical analysis was performed with S-Plus
for the first year following RNU, every 6 mo from the second through the
Professional (MathSoft Inc, Seattle, WA, USA).
fifth year, and annually thereafter. Follow-up consisted of a history, physical examination, routine blood work and serum chemistry
3.
Results
upper urinary tract. Elective bone scans, chest computed tomography,
3.1.
Association of tumor location with outcomes in all patients
or magnetic resonance imaging were performed when clinically
(n = 1249)
studies, urinary cytology, chest radiography, cystoscopic evaluation of the urinary bladder, and radiographic evaluation of the contralateral
indicated. Disease recurrence was defined as local failure in the tumor bed, regional lymph nodes, or distant metastasis. Bladder recurrences were not considered in the analysis of recurrence-free survival rate. Cause of death was determined by the treating physicians, by chart review corroborated by death certificates, or by death certificates alone. Perioperative mortality (any death within 30 d of surgery or before discharge) was censored at time of death for urothelial disease-specific survival analyses.
2.4.
Statistical analysis
The x2 test was used to evaluate the association between categorical variables, and the Mann Whitney test assessed for differences in variables with a continuous distribution across dichotomous categories. Univariate recurrence and survival probabilities after RNU were determined using the Kaplan-Meier method with differences estimated using the log-rank test. Univariate and multivariate Cox proportional hazards regression models addressed time to recurrence and disease-specific mortality after RNU. In all models, proportional hazards
assumptions
were
systematically
verified
using
the
Grambsch-Therneau residual-based test. The change in predictive accuracy resulting from the addition of a variable of interest to standard predictor variables was quantified with the Harrell concordance index [10,11]. Internal validation was performed using 200 bootstrap resamples [10,11]. Predictive accuracy estimates were
3.1.1.
Clinicopathologic characteristics
Table 1 summarizes the clinical and pathologic characteristics of patients as stratified by tumor location. There were 846 men (68%) and 403 women (32%) included with a median age of 70 yr. Overall, 66% of tumors were classified as renal pelvic and 34% were ureteral. The stage distribution of UTUC tumors in this cohort was 23% pTa, 2% pTis, 24% pT1, 19% pT2, 29% pT3, and 3% pT4. Sixty percent of patients had high-grade urothelial tumors, and 6% had positive lymph nodes. Regional lymphadenectomy was generally performed in patients with enlarged lymph nodes on preoperative axial imaging or with adenopathy detected during intraoperative examination. As such, 61% of patients in this cohort did not undergo a lymphadenectomy (pNx). When comparing renal pelvic tumors and ureteral tumors, there were no differences in age, gender distribution, nodal status (pN+ vs pN0/pNx), tumor grade, or followup interval (Table 1). Interestingly, for pT classification, renal pelvic tumors had a more advanced presentation than ureteral disease ( p < 0.001). Specifically, 38% of renal pelvis versus 22% of ureteral tumors presented with locally advanced (T3/T4) tumors.
EUROPEAN UROLOGY 57 (2010) 1072–1079
3.1.2.
Urothelial cancer recurrence
Median follow-up for this patient cohort was 49 mo (mean: 62.7; range: 0.1–250). Cancer recurrence occurred in 277 patients (22%) with a 5-yr recurrence-free survival estimate of 75% (Fig. 1a). On univariate analyses, pT classification ( p < 0.001), tumor grade ( p < 0.001), and lymph node status ( p < 0.001) were all associated with disease recurrence. No difference was noted in the probability of urothelial cancer recurrence between ureteral and renal pelvic tumors (HR: 1.0; log rank p value: 0.7) (Fig. 1a). On multivariate analysis, only pT classification ( p < 0.001), grade ( p < 0.001), and lymph node status ( p < 0.001) were associated with disease recurrence. When adjusting for the effects of these variables, there was no difference in the probability of recurrence-free survival between renal pelvis and ureteral tumors (HR: 1.220; p = 0.133; Table 2). The predictive accuracy of a base prognostic model for disease recurrence that included T classification, tumor
1075
grade, and lymph node status was 79.6%. Addition of tumor location did not improve the predictive accuracy of this base model (area under the curve [AUC]: 79.7%; 0.1% gain). 3.1.3.
Cancer-specific survival
Urothelial cancer death occurred in 238 patients (19%), and 436 patients (34.9%) died of causes unrelated to urothelial cancer. The 5-yr cancer-specific survival estimate was 78% (Fig. 1b). On univariate analyses, pT classification ( p < 0.001), grade ( p < 0.001), and lymph node status ( p < 0.001) were associated with cancerspecific survival. However, there was no difference in the probability of cancer-specific survival between ureteral and renal pelvic tumors (HR: 1.03; log rank p value: 0.80) (Fig. 1b). On multivariate analysis, pT classification ( p < 0.001), grade ( p = 0.017), and lymph node status ( p < 0.001) were associated with worse
Fig. 1 – (a) Kaplan-Meier estimates for recurrence-free survival; (b) cancer-specific survival in 1249 patients managed by radical nephroureterectomy for upper tract urothelial carcinoma. HR = hazard ratio.
1076
EUROPEAN UROLOGY 57 (2010) 1072–1079
Table 2 – Multivariate Cox regression models predicting disease recurrence and cancer-specific survival in 1249 patients following radical nephroureterectomy for urothelial carcinoma of the upper urinary tract UTUC recurrence
pT classification pT1 vs pTa/pTis pT2 vs pTa/pTis pT3 vs pTa/pTis pT4 vs pTa/pTis Tumor grade High vs low
UTUC death
HR
95% CI
p value
HR
95% CI
2.388 5.130 11.733 34.307
1.152–4.952 2.585–10.179 6.082–22.636 16.532–71.192
<0.001 0.019 <0.001 <0.001 <0.001
2.071 5.098 9.827 25.588
0.848–5.058 2.212–11.753 4.384–22.029 9.854–66.445
2.310
1.627–3.281
<0.001
1.697
pN classification pN0 vs pNx pN+ vs pNx
0.501 1.321
0.378–0.663 0.974–1.792
<0.001 <0.001 0.073
Tumor location Ureter vs RP
1.220
0.942–1.580
0.133
p value <0.001 0.110 <0.001 <0.001 <0.001
1.100–2.617
0.017
0.418 1.371
0.279–0.624 0.890–2.112
<0.001 <0.001 0.152
1.230
0.864–1.750
0.250
CI = confidence interval; HR = hazard ratio; RP = renal pelvis; UTIC = upper tract urothelial carcinoma.
cancer-specific survival. There was no difference in cancer-specific survival between renal pelvic and ureteral tumors (HR: 1.230; p = 0.25; Table 2). The predictive accuracy of a base model for cancerspecific survival was 80.4%. Addition of tumor location improved the predictive accuracy of this base model by only 0.1% (AUC: 80.5%). 3.2.
Association of tumor location with outcomes in patients
3.2.1.
Clinicopathologic characteristics
Table 3 highlights the clinical and pathologic characteristics of patients with organ-confined disease stratified by tumor location. In this subgroup, when comparing renal pelvic to ureteral tumors, there were no differences in age, gender distribution, or follow-up interval. Ureteral tumors did have a higher distribution of pT2 lesions (31.9% vs 24.3%; p = 0.028) and high-grade tumors (53.4% vs 44.2%; p = 0.010).
with organ-confined (pT2 or lower, pN0/pNx) disease (n = 824) 3.2.2.
We hypothesized that patients with advanced (pT3/T4) local disease or node-positive disease would be more likely to receive neoadjuvant or adjuvant chemotherapy. To limit this as a potential confounding variable when studying the impact of tumor location, we performed a subgroup analysis of patients with organ-confined disease who did not receive perioperative chemotherapy.
Urothelial cancer recurrence
Median follow-up for this patient cohort was 52.8 mo (mean: 66.2; range: 0.1–250). Cancer recurrence occurred in 86 patients (10%) with a 5-yr recurrence-free survival estimate of 88% (Fig. 2a). On univariate analyses, pT classification ( p < 0.001), tumor grade ( p < 0.001), and tumor location ( p = 0.02) were all associated with disease recurrence. Specifically, patients with ureteral tumors
Table 3 – Descriptive characteristics of 824 patients with organ-confined [pT2 or lower] upper tract urothelial carcinoma managed by radical nephroureterectomy Variable
Mean (median) age, yr; range
All patients (n = 824)
Tumor location Renal pelvis (n = 498)
Ureter (n = 326)
p value
68 (69); 30–96
68 (68); 30–96
69 (70); 30–91
0.195
Gender, No. (%) Male Female
571 (69.3) 253 (31.7)
335 (67.3) 163 (32.7)
236 (72.4) 90 (27.6)
0.119
pT stage, No. (%) pTa pTis pT1 pT2
283 24 292 225
187 12 178 121
96 12 114 104
0.028
pN stage, No. (%) pNX pN0
563 (68.3) 261 (31.7)
356 (71.5) 142(28.5)
207 (63.5) 119 (36.5)
Tumor grade, No. (%) Low High
430 (52.2) 394 (47.8)
278 (55.8) 220 (44.2)
152 (46.6) 174 (53.4)
(34.3) (2.9) (35.4) (27.3)
(37.6) (2.4) (35.7) (24.3)
(29.5) (3.7) (35.0) (31.9)
0.016
0.010
1077
EUROPEAN UROLOGY 57 (2010) 1072–1079
Fig. 2 – (a) Kaplan-Meier estimates for recurrence-free survival; (b) cancer-specific survival in 824 patients with organ-confined upper tract urothelial carcinoma managed by radical nephroureterectomy. HR = hazard ratio.
had a lower 3-yr (87% vs 92%), 5-yr (86% vs 91%), and 7-yr (84% vs 89%) recurrence-free survival probability compared with renal pelvic tumors (HR: 1.6; log rank p value: 0.02; Fig. 2a). However, on multivariate analysis, only pT
classification ( p < 0.001) and grade ( p = 0.002) were associated with disease recurrence. When adjusting for the effects of these two variables, there was no difference in the probability of recurrence-free survival between
Table 4 – Multivariate Cox regression models predicting disease recurrence and cancer-specific survival in 824 patients with organ-confined upper tract urothelial carcinoma (UTUC; pT2 or lower, pN0/pNx) following radical nephroureterectomy UTUC recurrence
UTUC death
HR
95% CI
p value
HR
95% CI
pT classification pT1 vs pTa/pTis pT2 vs pTa/pTis
2.507 5.023
1.320–4.761 2.718–9.282
<0.001 0.005 <0.001
2.605 5.764
1.253–5.416 2.884–11.520
Tumor grade High vs low
1.951
1.280–2.972
0.002
1.604
1.022–2.518
0.040
Tumor location Ureter vs RP
1.228
0.844–1.788
0.283
1.333
0.877–2.025
0.178
CI = confidence interval; HR = hazard ratio; RP = renal pelvis.
p value <0.001 0.010 <0.001
1078
EUROPEAN UROLOGY 57 (2010) 1072–1079
renal pelvis and ureteral tumors (HR: 1.228; p = 0.283; Table 4). The predictive accuracy of a base prognostic model for disease recurrence for patients with organ-confined disease that included tumor classification and grade was 73.0%. Addition of tumor location only marginally improved the predictive accuracy of this base model (AUC: 73.5%; 0.5% gain). 3.2.3.
Cancer-specific survival
In this subgroup, urothelial cancer death occurred in 74 patients (9%), and 220 patients (26.7%) died of nonurothelial causes. The 5-yr cancer-specific survival estimate was 90%. On univariate analyses, pT classification ( p < 0.001), tumor grade ( p < 0.001), and tumor location ( p = 0.02) were all associated with disease recurrence. Similar to disease recurrence, ureteral tumors had lower 3-yr (90% vs 93%), 5-yr (87% vs 92%), and 7-yr (84% vs 90%) cancer-specific survival probability compared with renal pelvic tumors (HR: 1.8; log rank p value: 0.02; Fig. 2b). On multivariate analysis, pT classification ( p < 0.001) and grade ( p = 0.04) were associated with worse cancer-specific survival. There was no statistical difference in cancer-specific survival, however, between renal pelvic and ureteral tumors (HR: 1.333; p = 0.178; Table 4). The predictive accuracy of a base model for cancerspecific survival for patients with organ-confined UTUC using pT stage and grade was 73.0%. Addition of tumor location improved the predictive accuracy of this base model by only 0.5% (AUC: 73.5%). 4.
Discussion
Nephroureterectomy with ipsilateral bladder cuff excision is considered the gold-standard therapy for the management of invasive, high-grade, or bulky UTUC [2,12]. Despite adequate surgical therapy, a significant percentage of patients experience either local or distant failure [13]. Identification of prognostic variables for relapse and survival is essential to provide more accurate patient counseling, guidance for operative strategies, as well as recommendations for postoperative surveillance and adjuvant therapy regimens. Obtaining critical information for UTUC, however, is difficult owing to the rarity of the disease and the absence of randomized, prospective studies. Indeed, only a few reports of patients with UTUC have systematically examined potential prognostic factors for disease relapse. In this series, we presented data from an international cohort of >1200 patients treated by RNU for UTUC. The overall 5-yr recurrence-free and cancer-specific survival was 75% and 78%, respectively. These oncologic outcomes are comparable with other contemporary series, underscoring that RNU can achieve durable local control even for locally advanced disease. Similar to previous studies, we further confirmed that pT stage, grade, and lymph node status were associated with disease recurrence and cancerspecific survival (Table 2) [2–5]. We noted several interesting differences in pathologic characteristics depending on tumor location. Somewhat
surprisingly, renal pelvic tumors presented at a more advanced tumor stage than ureteral disease when considering locally advanced (T3/T4) cancers (Table 1). Conversely, when only considering organ-confined disease, ureteral tumors were more likely to be muscle invasive (pT2) and high grade (Table 3). Other variables including age, gender, and nodal status, however, were similar between renal pelvis and ureteral tumors. Overall, we failed to detect a difference in urothelial cancer recurrence or cancer-specific survival when stratifying patients by ureteral and renal pelvic tumor location (Fig. 1a and 1b; Table 2). The literature evaluating the impact of tumor location on UTUC outcomes is conflicting. In a single-institution study of 86 UTUC patients treated with RNU, Park and colleagues found that after adjusting for pathologic features, ureteral tumor location was significantly associated with disease recurrence and mortality [9]. These findings were supported by another single-institution study of 72 patients [8]. The hypothesis supporting the worse prognosis of ureteral compared with renal pelvic tumors was attributed to the thin layer of surrounding ureteral adventitia containing an extensive plexus of blood vessels and lymphatic channels making tumor invasion and metastasis easier [8,9,14]. In addition, it was postulated that the renal parenchyma and perihilar adipose tissue surrounding the renal pelvis may act as a barrier against early spread. Our observations from this study contradict this hypothesis. In fact, we found that ureteral tumors were less likely to present with non-organ-confined (pT3/4) disease compared with renal pelvic tumors. What is the potential explanation for our findings of higher tumor classification in renal pelvic disease compared with ureteral tumors? One likely theory is that renal pelvic tumors remain asymptomatic for a longer duration of time, whereas ureteral carcinomas are diagnosed earlier (hence lower stage at diagnosis) due to obstruction that can occur in its narrow lumen. This occurrence can potentially balance out the fact that there is a thinner adventitial wall for the ureteral tumors to invade and may explain the absence of survival differences in this study. Indeed, our classification of tumor location simple reflects a snapshot in time of the upper tract neoplasm, and thus it fails in part to capture the chronicity and biology of disease. In an attempt to minimize confounding effects of neoadjuvant and adjuvant chemotherapy regimens, we further evaluated the subgroup of patients with organconfined disease. In this group, on univariate analysis, patients with ureteral tumors did demonstrate a higher likelihood for disease recurrence (HR: 1.6) and death (HR: 1.8). However, when adjusting for tumor stage and grade, this difference was not significant (Table 4). Ultimately, the value in identifying a prognostic variable is the ability to incorporate it into a predictive model that can allow more accurate risk stratification to tailor follow-up therapy. In this regard, tumor location provided only a minimal incremental predictive gain of 0.1–0.5% when used in conjunction with more established variables including stage, grade, and lymph node status.
EUROPEAN UROLOGY 57 (2010) 1072–1079
Our findings are in concordance with the work from Hall et al [2], who reported similar clinical outcomes between ureteral and renal pelvis tumors in a cohort of 250 consecutive patients with UTUC. This study and ours are the two largest studies examining the association of tumor location with clinical outcomes after RNU. Moreover, our study is the only multicenter study, thereby controlling for biases associated with single-center studies such as specificities related to genetics, diagnosis, patient selection, staging, pathologic evaluation, and management. There are several limitations to the current study. First, it is inherently limited by biases associated with its retrospective design. Data were collected by chart review at participating institutions, thus introducing variability in the interpretation of study variables. Second, variability in practice patterns, surveillance, and patient populations introduces a degree of bias with regard to surgical intervention, patient follow-up, disease severity, access to care, determination of recurrences, and assessment of outcomes. We believe, however, that such variability may allow for these data to be more readily generalizable to a greater population of patients with UTUC. Finally, pathology specimens were not subjected to a centralized review. Rather, all specimens were reexamined by dedicated genitourinary pathologists according to strict criteria agreed on by all participating pathologists. Nonetheless, this may introduce some bias in interpretation of pathologic data.
1079
Waldert, Wood, Roscigno, Oya, Langner, Wolf, Stro¨bel, Ferna´ndez, Karakiewcz, Shariat. Statistical analysis: Raman, Shariat. Obtaining funding: None. Administrative, technical, or material support: None. Supervision: None. Other (specify): None. Financial disclosures: I certify that all conflicts of interest, including specific financial interests and relationships and affiliations relevant to the subject matter or materials discussed in the manuscript (eg, employment/affiliation, grants or funding, consultancies, honoraria, stock ownership or options, expert testimony, royalties, or patents filed, received, or pending), are the following: None. Funding/Support and role of the sponsor: None.
References [1] Jemal A, Siegel R, Ward E, et al. Cancer statistics, 2006. CA Cancer J Clin 2006;56:106–30. [2] Hall MC, Womack S, Sagalowsky AI, Carmody T, Erickstad MD, Roehrborn CG. Prognostic factors, recurrence, and survival in transitional cell carcinoma of the upper urinary tract: a 30-year experience in 252 patients. Urology 1998;52:594–601. [3] Roscigno M, Cozzarini C, Bertini R, et al. Prognostic value of lymph node dissection in patients with muscle-invasive transitional cell carcinoma of the upper urinary tract. Eur Urol 2008;53:794–802. [4] Ozsahin M, Zouhair A, Villa S, et al. Prognostic factors in urothelial renal pelvis and ureter tumours: a multicentre Rare Cancer Network study. Eur J Cancer 1999;35:738–43. [5] Huben RP, Mounzer AM, Murphy GP. Tumor grade and stage as
5.
Conclusions
prognostic variables in upper tract urothelial tumors. Cancer 1988;62:2016–20.
In this large international series of patients treated by RNU, there was no difference in outcomes between patients with renal pelvic and ureteral tumors following nephroureterectomy. The results of our study favor the current concept of tumor staging, in which renal pelvic and ureteral carcinomas are classified together.
[6] Margulis V, Shariat SF, Matin SF, et al. Outcomes of radical
Author contributions: Jay D. Raman had full access to all the data in the
[9] Park S, Hong B, Kim CS, Ahn H. The impact of tumor location on
study and takes responsibility for the integrity of the data and the
prognosis of transitional cell carcinoma of the upper urinary tract.
nephroureterectomy: a series from the Upper Tract Urothelial Carcinoma Collaboration. Cancer 2009;115:1224–33. [7] Murphy DM, Zincke H, Furlow WL. Primary grade 1 transitional cell carcinoma of the renal pelvis and ureter. J Urol 1980;123:629–31. [8] Akdogan B, Dogan HS, Eskicorapci SY, Sahin A, Erkan I, Ozen H. Prognostic significance of bladder tumor history and tumor location in upper tract transitional cell carcinoma. J Urol 2006;176:48–52.
accuracy of the data analysis.
J Urol 2004;171:621–5. [10] Harrell Jr FE, Califf RM, Pryor DB, Lee KL, Rosati RA. Evaluating the
Study concept and design: Raman, Margulis, Shariat.
yield of medical tests. JAMA 1982;247:2543–6.
Acquisition of data: Raman, Ng, Scherr, Margulis, Lotan, Bensalah, Patard,
[11] Harrell Jr FE, Lee KL, Mark DB. Multivariable prognostic models:
Kikuchi, Montorsi, Zigeuner, Weizer, Bolenz, Koppie, Isbarn, Jeldres,
issues in developing models, evaluating assumptions and ade-
Kabbani, Remzi, Waldert, Wood, Roscigno, Oya, Langner, Wolf, Stro¨bel, Ferna´ndez, Karakiewcz, Shariat.
quacy, and measuring and reducing errors. Stat Med 1996;15:
Analysis and interpretation of data: Raman, Shariat. Drafting of the manuscript: Raman, Ng, Scherr, Margulis, Lotan, Bensalah, Patard, Kikuchi, Montorsi, Zigeuner, Weizer, Bolenz, Koppie, Isbarn,
361–87. [12] Krogh J, Kvist E, Rye B. Transitional cell carcinoma of the upper urinary tract: prognostic variables and post-operative recurrences. Br J Urol 1991;67:32–6.
Jeldres, Kabbani, Remzi, Waldert, Wood, Roscigno, Oya, Langner, Wolf, Stro¨bel, Ferna´ndez, Karakiewcz, Shariat.
[13] Heney NM, Nocks BN, Daly JJ, Blitzer PH, Parkhurst EC. Prognostic
Critical revision of the manuscript for important intellectual content:
[14] Miyake H, Hara I, Gohji K, Arakawa S, Kamidono S. The significance
Raman, Ng, Scherr, Margulis, Lotan, Bensalah, Patard, Kikuchi, Montorsi,
of lymphadenectomy in transitional cell carcinoma of the upper
Zigeuner, Weizer, Bolenz, Koppie, Isbarn, Jeldres, Kabbani, Remzi,
urinary tract. Br J Urol 1998;82:494–8.
factors in carcinoma of the ureter. J Urol 1981;125:632–6.