EUROPEAN UROLOGY 57 (2010)

Transcription

1 EUROPEAN UROLOGY 57 (2010) available at journal homepage: Platinum Priority Urothelial Cancer Editorial by Alexandre R. Zlotta on pp of this issue Should Bladder Cuff Excision Remain the Standard of Care at Nephroureterectomy in Patients with Urothelial Carcinoma of the Renal Pelvis? A Population-based Study Giovanni Lughezzani a,b,1, Maxine Sun a,1, Paul Perrotte c, Shahrokh F. Shariat a, Claudio Jeldres a, Lars Budaus a,d, Ahmed Alasker a, Alain Duclos c, Hugues Widmer c, Mathieu Latour e, Giorgio Guazzoni b, Francesco Montorsi b, Pierre I. Karakiewicz a,c, * a Cancer Prognosis and Health Outcomes Unit, University of Montreal Health Center, Montreal, Québec, Canada b Department of Urology, Vita-Salute San Raffaele University, Milan, Italy c Department of Urology, University of Montreal, Montreal, Québec, Canada d Martini-Clinic, Prostate Cancer Center Hamburg-Eppendorf, Hamburg, Germany e Department of Pathology, University of Montreal, Montreal, Québec, Canada Article info Article history: Accepted December 1, 2009 Published online ahead of print on December 10, 2009 Keywords: Urothelial carcinoma Upper urinary tract Nephroureterectomy Survival Abstract Background: A large, multi-institutional, tertiary care center study suggested no benefit from bladder cuff excision (BCE) at nephroureterectomy in patients with upper tract urothelial carcinoma (UC). Objective: We tested and quantified the prognostic impact of BCE at nephroureterectomy on cancer-specific mortality (CSM) in a large population-based cohort of patients with UC of the renal pelvis. Design, setting, and participants: A cohort of 4210 patients with UC of the renal pelvis were treated with nephroureterectomy with (NUC) or without (NU) a BCE between 1988 and 2006 within 17 Surveillance, Epidemiology, and End Results registries. Measurements: Cumulative incidence plots and competing risks regression models compared CSM after either NUC or NU. Covariates consisted of pathologic T and N stages, grade, age, year of surgery, gender, and race. Results and limitations: Respectively, 2492 (59.2%) and 1718 (40.8%) patients underwent a nephroureterectomy with or without BCE. In univariable and multivariable analyses, BCE omission increased CSM rates in patients with pt3n0/x, pt4n0/x, and pt(any)n1-3 UC of the renal pelvis. For example, in patients with pt3n0/x disease, holding all other variables constant, BCE omission increased CSM in a 1.25-fold fashion ( p = 0.04). Similarly, in patients with pt4n0/x disease, BCE omission resulted in a 1.45-fold increase ( p = 0.02). The main limitation of our study is the lack of data on disease recurrence. Conclusions: Nephroureterectomy with BCE remains the standard of care in the treatment of UC of the renal pelvis and should invariably be performed in patients with locally advanced disease. Conversely, patients with pt1 and pt2 disease could be considered for NU without compromising CSM. However, recurrence data are needed to fully confirm the validity of this option. Crown Copyright # 2009 Published by Elsevier B.V. on behalf of European Association of Urology. All rights reserved. 1 These authors contributed equally to the manuscript. * Corresponding author. Cancer Prognostics and Health Outcomes Unit, University of Montreal Health Center (CHUM), 1058, rue St-Denis, Montréal, Québec, Canada, H2X 3J4. Tel ; Fax: address: pierre.karakiewicz@umontreal.ca (P.I. Karakiewicz) /$ see back matter Crown Copyright # 2009 Published by Elsevier B.V. on behalf of European Association of Urology. All rights reserved. doi: /j.eururo

2 EUROPEAN UROLOGY 57 (2010) Introduction Nephroureterectomy with a bladder cuff excision (BCE) represents the surgical standard of care for upper tract urothelial cancers (UTUCs) [1,2]. However, not all nephroureterectomies are performed with a BCE. Only 43.1% of 1249 nephroureterectomy patients also underwent a BCE, according to a recent multi-institutional tertiary care center study [3]. The practice of nephroureterectomy without a BCE may undermine cancer-control outcomes. Surprisingly, in the tertiary care center study, BCE omission did not affect recurrence or cancer-specific mortality (CSM) rates [3]. Lack of effect might have been due to the population characteristics. Unfortunately, the rate of BCE according to urothelial carcinoma (UC) risk factors (pathologic T and N stages, grade, tumor location) were not provided in that study [3]. Consequently, it cannot be determined whether BCE indications or the lack of benefit of BCE accounted for the absence of protective effect of BCE on recurrence or CSM rates [3]. Because the selection criteria for BCE may be different in the community than in tertiary care centers, we performed a large-scale population-based analysis of the effect of BCE at nephroureterectomy. Our goal was to quantify the effect of nephroureterectomy with or without BCE on CSM. We hypothesized that a protective effect of BCE would exist in this large cohort of patients with UC of the renal pelvis treated with nephroureterectomy with or without BCE. 2. Patients and methods 2.1. Study population We identified patients diagnosed with stages T1-4N0-3/NxM0 renal pelvis UC (ICD-O-2 C65.9 code). All patients underwent either a nephroureterectomy with (NUC; 2-digit surgery codes: 20-30; sitespecific surgery code: 40) or without (NU; 2-digit surgery codes: 40-50; site-specific surgery code: 50) a BCE, between 1988 and 2006 within 17 registries of the US National Cancer Institute Surveillance, Epidemiology Table 1 Descriptive characteristics of 4210 patients treated for urothelial carcinoma of the renal pelvis stratified according to the type of surgery (nephroureterectomy with [NUC] or without [NU] bladder cuff excision) No. patients Overall population (%) NUC (%) NU (%) p value 4210 (100.0) 2492 (59.2) 1718 (40.8) Age, yr 0.01 Mean (median) 70.9 (72.0) 70.5 (72.0) 71.4 (73.0) Range Gender 0.04 Male 2453 (58.3) 1419 (56.9) 1034 (60.2) Female 1757 (41.7) 1073 (43.1) 684 (39.8) Race 0.8 Caucasian 3764 (89.4) 2226 (89.3) 1538 (89.5) Non-Caucasian 446 (10.6) 266 (10.7) 180 (10.5) pt stage pt (29.7) 832 (33.4) 420 (24.4) <0.001 pt2 536 (12.7) 351 (14.1) 185 (10.8) pt (41.5) 978 (39.2) 768 (44.7) pt4 676 (16.1) 331 (13.3) 345 (20.1) pn stage <0.001 pn (73.5) 1905 (76.4) 1191 (69.3) pn (10.5) 148 (5.9) 293 (17.1) pnx 673 (16.0) 439 (17.6) 234 (13.6) Tumor grade <0.001 I 192 (4.6) 133 (5.3) 59 (3.4) II 1249 (29.7) 843 (33.8) 406 (23.6) III 1812 (43.0) 996 (40.0) 816 (47.5) IV 957 (22.7) 520 (20.9) 437 (25.4) Year of surgery (14.1) 382 (15.3) 210 (12.2) (19.7) 494 (19.8) 336 (19.6) (30.2) 706 (28.3) 564 (32.8) (36.1) 910 (36.5) 608 (35.4) Follow-up in censored patients Mean (median) 56.5 (40.0) 58.0 (41.0) 53.8 (39.0) Range No. events Cancer specific 836 (19.9) 393 (15.8) 443 (25.8) Other cause 1396 (33.2) 834 (32.7) 562 (33.5)

3 958 EUROPEAN UROLOGY 57 (2010) and End Results (SEER) program. These registries included the states of Connecticut, Hawaii, Iowa, New Mexico, Alaska, Kentucky, New Jersey, Louisiana, and Utah as well as the following regional areas: Atlanta, Georgia; Detroit, Michigan; San Francisco-Oakland, California; Seattle- Puget Sound, Washington; San Jose-Monterey, California; Los Angeles, California; rural Georgia; and greater California (excluding San Francisco, Los Angeles, and San Jose). The 17 registries represent a 26% sample of the US population. They are considered representative of US demographic composition as well as of cancer incidence and mortality [4]. The quality of the SEER database is highly controlled and continuously improved [5]. Patients with unknown metastatic stage (n = 73), unknown tumor stage (n = 88), and unknown tumor grade (n = 228) (total number of patients with missing variables: n = 304 [7.2%]) were excluded from the current analysis. None of the missing variables appeared to be directly associated with outcomes in univariate survival analysis. The cause of death was defined according to the SEER-specific cause of death code (code 29020). For this analysis, deaths from UC of the renal pelvis were coded as cancer-specific events (ie, CSM). All other deaths were considered to be other-cause mortality (OCM) Statistical analysis Patients were divided in two groups according to surgical procedure: NUC versus NU. The student t test and the x 2 test were used for comparison of means and proportions between the two surgical procedures, respectively. Because a substantial proportion of patients with UTUC die as a result of competing causes of mortality [6], we used competing risks regression analyses to test the predictors of CSM. The use of competing risks regression models, as described by Fine and Gray, account for the effect OCM and provide the most unbiased estimates of CSM [7]. Cumulative incidence plots were used to graphically depict the effect of CSM after accounting for OCM in the overall population. Additionally, stage-stratified cumulative incidence plots examined the effect of NUC versus NU on CSM in patients with pt1n0/x versus pt2n0/x versus pt3n0/x versus pt4n0/x versus pt(any)n1-3 UC of the renal pelvis. The Gray test was used for comparisons of CSM rates between different groups [8]. Univariate and multivariate competing risks regression models were used to test the effect of surgery type (NUC vs NU) on CSM rates. Covariates consisted of pt and pn stages, tumor grade, continuously coded age, race (Caucasian vs non-caucasian), gender, and year of surgery. To quantify the magnitude of the effect related to the type of surgery (NUC vs NU), we repeated all multivariate competing risks regression models after stratifying according to pt and pn stages. Before fitting the competing risks models, the proportional cause-specific hazards assumption, the proportionality of the hazards of the cumulative incidence function assumption, and the linearity assumption were tested [9]. Subsequently, the same analyses were repeated within a dataset with imputed missing values using the multiple imputation method implemented in the S-PLUS aregimpute function. All reported p values were two-sided, and statistical significance was set at <0.05. Statistical analyses were performed with S-Plus Professional software (MathSoft Inc., Seattle, WA, USA). (89.4%) and male (58.3%). Patients who underwent a NUC were significantly younger than patients who underwent a NU (median age: 72.0 vs 73.0 yr; p = 0.01). The proportion of locally advanced (pt3-4) disease was significantly lower in NUC patients relative to NU patients (52.5% vs 64.8%; p < 0.01). Similarly, NUC patients showed a lower rate of grades III and IV UC of the renal pelvis than NU patients (60.9% vs 72.9%; p < 0.001). Overall, 17.6% of NUC patients and 13.6% of NU patients were not staged with a lymph node dissection (pnx; p < 0.001). Moreover, lymph node metastases (pn1-3) were observed in 5.9% versus 17.1% of NUC and NU patients ( p < 0.001), respectively. Cumulative incidence plots illustrate the CSM rates for the entire population (n = 4210; Fig. 1A) and after stratification according to surgery type (NUC vs NU; Fig. 1B), after accounting for OCM. For the whole cohort, the 2- and 5-yr CSM rates were, respectively, 16.4% and 22.0% (Fig. 1A). The CSM rates after NUC versus NU were 12.2% versus 17.7% and 22.4% versus 28.1% for the same time points ( p < 0.001) (Fig. 1B). In stage-stratified cumulative incidence plots, the 2- and 5-yr CSM rates in NUC versus NU patients were, respectively, 4.7% versus 5.6% and 7.1% versus 9.8% for pt1n0/x patients (Fig. 2A); 7.7% versus 10.3% and 10.0 versus 13.8% for pt2n0/x patients (Fig. 2B); 15.1% versus 21.1% and Results The descriptive characteristics of the entire cohort (n = 4210) are shown in Table 1. NUC was performed in 2492 (59.2%) patients. Conversely, 1718 (40.8%) patients underwent a NU. The study cohort was mostly Caucasian Fig. 1 Cumulative incidence plots depicting cancer-specific mortality (A) after accounting for other-cause mortality in the overall population of patients with urothelial carcinoma of the renal pelvis and (B) after stratification according to surgery type (nephroureterectomy with [NUC] or without [NU] bladder cuff excision). CSM = cancer-specific mortality; CI = confidence interval.

4 EUROPEAN UROLOGY 57 (2010) Fig. 2 Cumulative incidence plots depicting cancer-specific mortality after accounting for other-cause mortality in patients with (A) pt1n0/x, (B) pt2n0/x, (C) pt3n0/x, (D) pt4n0/x, and (E) pt(any)n1-3 urothelial carcinoma of the renal pelvis, stratified according to surgery type (nephroureterectomy with [NUC] or without [NU] bladder cuff excision). CSM = cancer-specific mortality; CI = confidence interval. versus 26.7% for pt3n0/x patients (Fig. 2C); and 19.9% versus 33.9% and 26.6% versus 39.7% for pt4n0/x patients (Fig. 2D). Finally, in patients with lymph node metastases (pt[any]n1-3), the 2- and 5-yr CSM rates were 32.2% versus 46.5% and 45.4% versus 56.5% for NUC versus NU patients, respectively (Fig. 2E). A statistically significant absolute CSM survival benefit favoring NUC ranged from 6.0% to 14.0% at 2 yr and from 4.0% to 13.1% at 5 yr for patients with pt3n0/x and pt4n0/x tumors ( p 0.03). Conversely, no statistically significant survival benefit was seen after NUC versus NU for stages pt1n0/x and pt2n0/x disease. Table 2 shows the overall univariate and multivariate competing risks regression analyses, adjusting for pt and pn stages, tumor grade, age, gender, and year of surgery. A 1.31-fold higher rate of CSM was recorded in patients treated with NU versus NUC ( p < 0.001) (Table 2). Subsequently, we repeated the multivariate competing risks regression models in patient subgroups that were defined according to individual pt and pn stages: pt1n0/x, pt2n0/ x, pt3n0/x, pt4n0/x, and pt(any)n1-3 (Table 3). In these analyses, pt3n0/x, pt4n0/x, and pt(any)n1-3 patients treated with NU were at a fold higher risk of CSM than their counterparts treated with NUC (all p values 0.04). As shown previously, patients with pt1n0/x and pt2n0/x disease derived no benefit from a BCE. Taken together, the multivariate models findings indicate that the

5 960 EUROPEAN UROLOGY 57 (2010) Table 2 Univariate and multivariate competing-risks regression models addressing cancer-specific mortality after accounting for othercause mortality, in the overall population of patients treated for urothelial carcinoma of the renal pelvis with a nephroureterectomy with (NUC; n = 2492) versus without (NU; n = 1718) a bladder cuff excision Univariate analysis Multivariate analysis HR (95% CI) p value HR (95% CI) p value Surgery type NU vs NUC 1.75 ( ) < ( ) <0.01 pt stage pt2 vs pt ( ) ( ) 0.07 pt3 vs pt ( ) < ( <0.01 pt4 vs pt ( ) < ( ) <0.001 pn stage pn1-3 vs pn ( ) < ( ) <0.01 pnx vs pn ( ) ( ) 0.81 Grade III vs I and II 3.49 ( ) < ( ) <0.01 IV vs I and II 3.47 ( ) < ( ) <0.01 Age 1.02 ( ) < ( ) <0.01 Year of surgery 0.98 ( ) ( ) <0.01 Gender Female vs male 1.28 ( ) < ( ) 0.03 Race Non-Caucasian vs Caucasian 0.89 ( ) 0.30 HR = hazard ratio; CI = confidence interval. benefit of a BCE on CSM is only applicable to patients with locally advanced (pt3-4n0/x, pt[any]n1-3) UC of the renal pelvis and not to patients with localized disease (pt1-2n0/x). Finally, the use of multiple data imputation within the population that included patients with missing tumor grade and missing tumor and metastatic stages resulted in virtually the same findings. Specifically, failure to perform a BCE resulted in an increase in the hazard ratio (HR) for CSM death (HR: 1.29; 95% confidence interval [CI], ; p < 0.001; in the complete-case analyses, HR: 1.31; 95% CI, ; p < 0.001). Table 3 Multivariable competing-risks regression models addressing cancer-specific mortality after accounting for other-cause mortality in patients treated for urothelial carcinoma of the renal pelvis with a nephroureterectomy with (NUC; n = 2492) versus without (NU; n = 1718) a bladder cuff excision, stratified according to pathologic tumor and node stages: pt1n0/x, pt2n0/x, pt3n0/x, pt4n0/x, and pt(any)n1-3 pt1n0/x pt2n0/x pt3n0/x pt4n0/x pt(any)n1-3 (n = 1232) (n = 520) (n = 1496) (n = 521) (n = 441) HR p HR p HR p HR p HR p Surgery type NU vs NUC pt stage pt2 vs pt pt3 vs pt pt4 vs pt pn stage pnx vs pn Grade III vs I and II 2.46 < < < IV vs I and II < < Age Year of surgery Gender Female vs male HR = hazard ratio.

6 EUROPEAN UROLOGY 57 (2010) Discussion European and North American guidelines recommend BCE at nephroureterectomy for patients with UC of the renal pelvis [1,10]. To date, however, the benefit of a BCE has not been formally quantified [3]. Based on lack of data proving or disproving the need for a BCE at nephroureterectomy, many patients may undergo a nephroureterectomy without a BCE rather than a nephroureterectomy with a formal BCE [11 17]. In a recent multi-institutional tertiary care center analysis, 41.2% and 50.0% of open and laparoscopic nephroureterectomies, respectively, included a BCE [3]. BCE had no effect on recurrence or CSM rates [3]. This phenomenon may be due to lack of benefit of BCE. Alternatively, it may also be due to patient selection. To address the effect of BCE on CSM rates, we performed a population-based analysis in patients with UC of the renal pelvis who were treated with nephroureterectomy. Our population differed from the previous one, as it is represented by individuals identified within 17 SEER registries instead of highly select patients from tertiary care centers. Moreover, our population differed with respect to the rate of BCE, which was 59.2% versus 43.1% in the previous analysis [3]. Our findings showed two important observations. First, in patients with pt3-4n0/x and pt(any)n1-3, the rate of CSM was between 1.25-fold and 1.45-fold higher when BCE was not performed. Based on these results, BCE should be mandatory in these patients. Second, we confirmed the lack of survival benefit for BCE in patients with pt1-2n0/x UC of the renal pelvis. According to these results, in patients with localized disease, BCE omission does not undermine survival. In patients with locally advanced disease, a wider resection, such as occurs if the distal ureter and bladder cuff are excised, may provide better cancer control and may lower the risk of local recurrence and local progression of the disease. Unfortunately, our database does not contain information that can be used to confirm or refute this hypothesis. It is also noteworthy that although BCE improves survival, it may add to morbidity. For example, at open or laparoscopic nephroureterectomy, BCE may require a second incision. Moreover, this procedure requires more time and wider dissection and increases the complexity of surgery. However, to the best of our knowledge, the added morbidity has not been formally quantified. Several limitations apply to our study. First, lack of data on ureteral stump recurrences prevents us from stating that a nephroureterectomy without a BCE should represent a standard of care for all patients with pt1-2n0 disease. BCE omission can only be suggested if our findings are confirmed with recurrence data showing that BCE omission would not result in a higher rate of disease recurrence. Second, lack of central pathology review represents another limitation that was shared with the previous analyses [3,11 14,17,18]. Third, no information was available regarding chemotherapy or radiotherapy delivery. A recent report indicated that adjuvant chemotherapy was rarely used in patients with upper tract UC [19]. Moreover, the effect of a chemotherapy did not translate into an overall or CSM risk reduction [19]. Fourth, the retrospective nature of the study represents a limitation that is shared with all previous studies addressing upper tract UC cancer-control outcomes [3,11 14,17,18]. Fifth, the SEER database does not provide any information regarding other potential UTUC prognostic factors, such as the number of lymph nodes removed, tumor size, tumor multifocality, surgical margin status, and history of bladder cancer. Finally, surgery type (NU vs NUC) represents a field where inaccurate coding may apply, such as cause of death, presence of metastatic disease, tumor grade, and tumor stage. This said, several investigators who examined the type, extent, and specificity of various cancer procedures have previously confirmed their accuracy and validity in the SEER database [20 22]. These studies found high levels of concordance between SEER registry and medical records or claims for surgeries [20 22]. The reliability of the cause of death ascertainment was also examined by comparing the SEER database cause of death item with the clinicianassigned or clinical history-based cause of death in prostate cancer [23] and breast cancer patients [24]. These studies confirmed the accuracy of the SEER cause of death item [23,24]. Taken together, these findings indicate that various SEER codes are highly reliable. 5. Conclusions Despite its limitations, our study demonstrated that CSM may be increased significantly if a BCE is not performed in patients with locally advanced (pt3-4n0/x and pt[any]n1-3) UC of the renal pelvis. Therefore, BCE should represent a standard of care at nephroureterectomy in these patients. Conversely, no increase in CSM was recorded in patients with localized (pt1-2n0/x) UC of the renal pelvis when BCE was omitted. Although BCE omission does not undermine survival, confirmatory recurrence data are required before a nephroureterectomy without BCE may be considered as an option to nephroureterectomy with formal BCE for this patient category. Author contributions: Pierre I. Karakiewicz had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Lughezzani, Sun, Perrotte, Shariat, Karakiewicz. Acquisition of data: Lughezzani, Sun, Jeldres, Alasker, Karakiewicz. Analysis and interpretation of data: Lughezzani, Sun, Jeldres, Latour, Karakiewicz. Drafting of the manuscript: Lughezzani, Sun, Budaus, Karakiewicz. Critical revision of the manuscript for important intellectual content: Duclos, Widmer, Latour, Guazzoni, Montorsi, Karakiewicz. Statistical analysis: Lughezzani, Sun, Jeldres, Budaus, Alasker, Karakiewicz. Obtaining funding: Karakiewicz. Administrative, technical, or material support: Perrotte, Karakiewicz. Supervision: Montorsi, Karakiewicz. Other (specify): None.

7 962 EUROPEAN UROLOGY 57 (2010) 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: Pierre I. Karakiewicz is partially supported by the University of Montreal Health Center Urology Associations, Fonds de la Recherche en Sante du Quebec, the University of Montreal Department of Surgery, and the University of Montreal Health Center (CHUM) Foundation. Funding/Support and role of the sponsor: None. References [1] Oosterlinck W, Solsona E, van der Meijden APM, et al. EAU guidelines on diagnosis and treatment of upper urinary tract transitional cell carcinoma. Eur Urol 2004;46: [2] Zigeuner R, Pummer K. Urothelial carcinoma of the upper urinary tract: surgical approach and prognostic factors. Eur Urol 2008;53: [3] Capitanio U, Shariat SF, Isbarn H, et al. Comparison of oncologic outcomes for open and laparoscopic nephroureterectomy: a multiinstitutional analysis of 1249 cases. Eur Urol 2009;56:1 9. [4] SEER Cancer Statistics Review, US National Cancer Institute Web site. Accessed January [5] Surveillance, Epidemiology and End Results (SEER), quality control activities. NIH Publication Bethesda, MD: US National Institutes of Health; [6] Inman BA, Tran VT, Fradet Y, Lacombe L. Carcinoma of the upper urinary tract: predictors of survival and competing causes of mortality. Cancer 2009;115: [7] Fine JP, Gray RJ. A proportional hazards model for the subdistribution of a competing risk. J Am Stat Assoc 1999;94: [8] Gray RJ. A class of K-sample tests for comparing the cumulative incidence of a competing risk. Ann Stat 1988;16: [9] Pintilie M. Competing risks: a practical perspective. Hoboken, NJ: Wiley; [10] NCCN Clinical Practice Guidelines in Oncology. Bladder cancer. National Comprehensive Cancer Network Web site. org/professionals/physician_gls/f_guidelines.asp. Updated [11] Laguna MP, de la Rosette JJ. The endoscopic approach to the distal ureter in nephroureterectomy for upper urinary tract tumor. J Urol 2001;166: [12] Ubrig B, Boenig M, Waldner M, Roth S. Transurethral approach to the distal ureter in nephroureterectomy: transurethral extraction vs. pluck technique with long-term follow-up. Eur Urol 2004;46: [13] Matin SF, Gill IS. Recurrence and survival following laparoscopic radical nephroureterectomy with various forms of bladder cuff control. J Urol 2005;173: [14] Kurzer E, Leveillee RJ, Bird VG. Combining hand assisted laparoscopic nephroureterectomy with cystoscopic circumferential excision of the distal ureter without primary closure of the bladder cuff is it safe? J Urol 2006;175:63 7, discussion [15] Raman JD, Scherr DS. Management of patients with upper urinary tract transitional cell carcinoma. Nat Clin Pract Urol 2007;4: [16] Macejko AM, Pazona JF, Loeb S, Kimm S, Nadler RB. Management of distal ureter in laparoscopic nephroureterectomy a comprehensive review of techniques. Urology 2008;72: [17] Walton TJ, Sherwood BT, Parkinson RJ, et al. Comparative outcomes following endoscopic ureteral detachment and formal bladder cuff excision in open nephroureterectomy for upper urinary tract transitional cell carcinoma. J Urol 2009;181: [18] Margulis V, Shariat SF, Matin SF, et al. Outcomes of radical nephroureterectomy: a series from the Upper Tract Urothelial Carcinoma Collaboration. Cancer 2009;115: [19] Hellenthal NJ, Shariat SF, Margulis V, et al. Adjuvant chemotherapy for high risk upper tract urothelial carcinoma: results from the upper tract urothelial carcinoma collaboration. J Urol 2009;182: [20] Cooper GS, Virnig B, Klabunde CN, et al. Use of SEER-Medicare data for measuring cancer surgery. Med Care 2002;40(Suppl 8):43 8. [21] Cooper GS, Yuan Z, Stange KC, et al. Agreement of Medicare claims and tumor registry data for assessment of cancer-related treatment. Med Care 2000;38: [22] Warren JL, Harlan LC. Can cancer registry data be used to study cancer treatment? Med Care 2003;41: [23] Penson DF, Albertsen PC, Nelson PS, Barry M, Stanford JL. Determining cause of death in prostate cancer: are death certificates valid? J Natl Cancer Inst 2001;93: [24] Dignam JJ, Huang L, Ries L, et al. Estimating breast cancer-specific and other-cause mortality in clinical trial and population-based cancer registry cohorts. Cancer 2009;115: