EUROPEAN UROLOGY 56 (2009)

EUROPEAN UROLOGY 56 (2009) 512 519 available at www.sciencedirect.com journal homepage: www.europeanurology.com Urothelial Cancer The Extent of Lymphadenectomy Seems to Be Associated with Better Survival in Patients with Nonmetastatic Upper-Tract Urothelial Carcinoma: How Many Lymph Nodes Should Be Removed? Marco Roscigno a, *, Shahrokh F. Shariat b,1, Vitaly Margulis e, Pierre Karakiewicz m, Mesut Remzi d, Eiji Kikuchi g, Richard Zigeuner h, Alon Weizer i, Arthur Sagalowsky b, Karim Bensalah l, Jay D. Raman k, Christian Bolenz f, Wassim Kassou f, Theresa M. Koppie j, Christopher G. Wood e, Jeffrey Wheat i, Cord Langner h, Casey K. Ng k, Umberto Capitanio m, Roberto Bertini a, Mario I. Fernández c, Shuji Mikami g, Masaru Isida g, Philipp Ströbel f, Francesco Montorsi a a Vita Salute University San Raffaele, Milan, Italy b University of Texas Southwestern Medical Center, Dallas, Texas, USA c Clinica Alemana de Santiago, Santiago, Chile d University of Vienna, Vienna, Austria e University of Texas M.D. Anderson Cancer Center, Houston, TX, USA f University Medical Center Mannheim, University of Heidelberg, Heidelberg, Germany g Keio University School of Medicine, Tokyo, Japan h Medical University Graz, Graz, Austria i University of Michigan, Ann Arbor, MI, USA j University of California Davis, Sacramento, CA, USA k Cornell University, New York, NY, USA l University of Rennes, Rennes, France m University of Montreal, Montreal, QC, Canada Article info Article history: Accepted June 8, 2009 Published online ahead of print on June 18, 2009 Keywords: Lymph node dissection Prognosis Urinary tract cancer Urothelial carcinoma Nephroureterectomy Survival Metastasis Abstract Background: The role and extent of lymphadenectomy in patients with uppertract urothelial carcinoma (UTUC) is debated. Objective: To establish whether the number of lymph nodes (LNs) removed might be associated with better cause-specific survival in patients with UTUC. Design, setting, and participants: The study included 552 consecutive patients who underwent radical nephroureterectomy (RNU) and lymphadenectomy between 1992 and 2006. Intervention: Patients were treated with RNU and lymphadenectomy. Measurements: Univariable and multivariable Cox proportional hazards regression models addressed the association between the number of LNs removed and cause-specific mortality (CSM). The number of LNs removed was coded as a cubic spline to allow for nonlinear effects. Finally, the most informative cut-off for the number of removed LNs was identified. 1 Currently at Memorial Sloan-Kettering Cancer Center, New York, NY, USA. * Corresponding author. Vita Salute University San Raffaele, Via Olgettina 60, 20132, Milan, Italy. Tel. +39 2 2643 4199; Fax: +39 2 26437298. E-mail address: roscigno.marco@hsr.it (M. Roscigno). 0302-2838/$ see back matter # 2009 European Association of Urology. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.eururo.2009.06.004

EUROPEAN UROLOGY 56 (2009) 512 519 513 Please visit www.eu-acme.org/ europeanurology to read and answer questions on-line. The EU-ACME credits will then be attributed automatically. Results and limitations: In the entire population, the number of LNs removed was not associated with CSM in univariable (hazard ratio [HR]: 0.99; p = 0.16) or in multivariable (HR: 0.97; p = 0.12) analyses. In contrast, in the subgroup of pn0 patients (n = 412), the number of LNs removed achieved the independent predictor status of CSM (HR: 0.93; p = 0.02). Eight LNs removed was the most informative cut-off in predicting CSM (HR: 0.42; p = 0.004). The inclusion of the variable defining dichotomously the number of removed LNs (<8 vs8) in the base model (age, Eastern Cooperative Oncology Group performance status, pathologic stage, grade, architecture, and lymphovascular invasion) significantly increased the accuracy in predicting CSM (+1.7%; p < 0.001). Conclusions: The extension of the lymphadenectomy in pn0 UTUC patients seems to be associated with CSM. Longer survival was observed in patients in whom at least eight LNs had been removed. # 2009 European Association of Urology. Published by Elsevier B.V. All rights reserved. 1. Introduction Upper-tract urothelial carcinoma (UTUC) is a relatively rare neoplasm, accounting for about 5% of all urothelial cancers [1]. Up to 30% of patients with muscle-invasive UTUC have metastasis to regional lymph nodes (LNs) [2,3], which represents a well-established poor prognostic factor [4]. Despite the prognostic and predictive effect of LN metastases, the benefit of lymphadenectomy in patients treated with radical nephroureterectomy (RNU) for UTUC remains debatable, and the procedure is still not widely performed [5,6]. Conversely, in other urologic settings such as bladder cancer, lymphadenectomy is recognized as a fundamental step in surgical management. Moreover, the number of LNs removed has been shown to affect survival after cystectomy in negative-node as well as in positive-node patients [7 12]. Regarding UTUC, several studies have already suggested that an extended lymphadenectomy may provide a survival benefit in UTUC patients treated with RNU [13 15]. Unfortunately, those single-institution reports relied on very limited study populations [13 15]. Therefore, we decided to assess the impact of the number of LNs removed on survival in a large, multicenter cohort. Moreover, we aimed to identify the minimum number of LNs that should be removed for accurate staging and potential therapeutic effect. 2. Patients and methods 2.1. Patients An institutional review board approved the study, with all participating sites providing the necessary institutional data-sharing agreements prior to initiation of the study. A total of 13 centers worldwide provided data (Table 1). A computerized data bank 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. Prior to final analysis, the database was frozen, and the final data set was produced for the current analysis. The multi-institutional data set included 1453 patients. The current study included 649 consecutive patients who underwent RNU and lymphadenectomy between 1992 and 2006. Patients with pta disease (n = 82) and ptis disease (n = 15) were excluded for the negligible risk of LN metastases. These exclusions left 552 (85%) patients for analysis. 2.2. Treatment Of the 552 patients, 464 were treated with open RNU (84%) and 88 with laparoscopic RNU (16%). The extent of lymphadenectomy was determined by the treating surgeon based on clinical presentation and by the location and laterality of the primary tumors. Lymphadenectomy typically included the para-aortic, paracaval, and interaortocaval nodes from the renal hilum to the inferior mesenteric artery in the case of renal pelvis and proximally ureteral tumors. In the case of mid- and lowerureteral tumors, nodes from the renal hilum to the bifurcation of the common iliac artery and ipsilateral pelvic nodes were removed, respectively. The extent of the LN dissection (LND) was evaluated by review of the operative and pathologist reports. Standardization of lymphadenectomy was impossible because of the multicenter and retrospective design of the study. Overall, 131 patients (23.1%) received adjuvant chemotherapy, while 34 (6.1%) received neoadjuvant chemotherapy. Chemotherapy was platinum based in 97% of the patients and consisted of methotrexate, vinblastine, doxorubicin, and cisplatin (M-VAC) in the majority. Table 1 Number of patients from each institution. Center No. of patients University of Texas M.D. Anderson 119 Cancer Center, Houston, TX, USA Centre Hospitalier Universitaire 7 Pontchaillou, Rennes, France University of Texas Southwestern 15 Medical Center, Dallas, TX, USA Keio University School of Medicine, Tokyo, Japan 151 University of California Davis, 11 Sacramento, CA, USA University of Graz, Graz, Austria 46 University Vita-Salute, Milan, Italy 84 University Hospital Mannheim, Mannheim, Germany 27 University of Michigan Hospital, Ann Arbor, MI, USA 14 Weill Medical College of 14 Cornell University, New York, NY, USA Clínica Alemana de Santiago, Chile 5 University of Vienna, Vienna, Austria 36 McGill University, Montreal, QC, Canada 23

514 EUROPEAN UROLOGY 56 (2009) 512 519 2.3. Pathologic findings Only primary urothelial carcinoma (UC) tumors were included in this study. All surgical specimens were processed according to standard pathologic work-up procedures, and all slides were re-reviewed by genitourinary pathologists according to identical, strict criteria. All pathologists were blinded to clinical outcomes. LNs were identified macroscopically and/or by palpation. The total number of identified LNs and the number of LNs involved with tumor were documented. All pathologic slides were re-reviewed by genitourinary pathologists according to strict, identical criteria and blinded to the original pathology slides and clinical outcomes. Pathologic staging was performed according to the 2002 TNM classification of the American Joint Committee on Cancer. Tumor grading was assessed according to the 1998 World Health Organization/ International Society of Urologic Pathology consensus classification [16]. Lymphovascular invasion (LVI) was defined as the presence of tumor cells within an endothelium-lined space without underlying muscular walls. The presence of concomitant carcinoma in situ and the tumor architecture were also assessed in every representative section. To ensure the validity of the pathologic data extraction, two pathologists independently reviewed specimens from 145 patients while blinded to patient clinical parameters and the findings of the other reviewer. Inter-reader reliability measured using the intraclass correlation coefficient was >0.95 for each pathologic characteristic. 2.4. Follow-up Patients were followed every 3 4 mo for the first year after RNU, every 6 mo from the second through the fifth year, and annually thereafter. Follow-up consisted of a history, physical examination, routine blood work and serum chemistry studies, urinary cytology, chest radiography, cystoscopic evaluation of the urinary bladder, and radiographic evaluation of the contralateral upper urinary tract. Elective bone scan, chest computed tomography scan, or magnetic resonance imaging scan was performed when clinically indicated. Cause of death was determined by the treating physicians, by chart review corroborated by death certificates, or by death certificates alone. Information abstracted from each death certificate included the date of death and whether UC was noted in part I or II. To reduce bias in attribution of cause of death, only men who had UC listed in part I of the death certificate were considered to have died of UTUC for this study. All patients who were coded as dead of cancer had previous disease recurrence. Perioperative mortality (any death within 30 d of surgery or before discharge) was censored at time of death for cause-specific mortality (CSM) analyses. 2.5. Statistical analysis Univariable and multivariable Cox proportional hazards regression analyses addressed the association between age at diagnosis, tumor stage (non muscle-invasive vs muscle-invasive), tumor grade (high vs low), nodal status, tumor architecture, LVI, adjuvant chemotherapy, tumor necrosis and institution of origin on recurrence, and CSM. Moreover, the same analyses were repeated in the subgroups of pn0 and pn+ patients. Finally, the most informative cut-off for either recurrence or CSM was identified, using an analysis of variance test for every possible cut-off and choosing the lowest p value. All statistical tests were performed using S-Plus Professional v.1 (MathSoft, Table 2 Patient characteristics and descriptive statistics. Variables All patients (n = 552) Number of LNs removed 8 (n = 145) Number of LNs removed <8 (n = 407) p Age, yr 0.59 * Mean (median) 68 (67.1) 66.5 (65.8) 68 (67.4) Range 27 94 31 94 27 92 Pathologic stage, no. (%) <0.001 y pt1 103 (18.7) 7 (4.8) 96 (23.6) pt2 114 (20.6) 33 (22.8) 81 (19.9) pt3 4 335 (60.7) 105 (72.4) 230 (56.5) Pathologic grade, no. (%) 0.513 y Low 98 (17.7) 23 (15.9) 75 (18.4) High 454 (82.3) 122 (84.1) 333 (81.6) Nodal status, no. (%) 0.111 y pn0 412 (74.6) 100 (69.0) 312 (76.7) pn+ 140 (25.4) 45 (31.0) 95 (23.3) ECOG PS, no. (%) <0.001 y 0 395 (71.6) 80 (55.2) 315 (77.4) 1 135 (24.5) 56 (38.6) 79 (19.4) 2 3 22 (3.9) 9 (6.2) 13 (3.2) Tumor necrosis, no. (%) 0.047 y Yes 219 (39.7) 69 (47.6) 150 (36.9) No 333 (60.3) 76 (52.4) 257 (63.1) CIS, no. (%) 132 (32.4) 0.015 y Yes 199 (36.1) 67 (46.2) 275 (67.6) No 353 (63.9) 78 (53.8) LVI, no. (%) 0.235 y Yes 190 (34.4) 57 (39.3) 133 (32.7) No 362 (65.6) 88 (61.7) 274 (67.3) Architecture, no. (%) 0.031 y Papillary 357 (64.7) 82 (56.5) 275 (67.6) Sessile 195 (35.3) 63 (43.5) 132 (32.4) LN = lymph node; ECOG = Eastern Cooperative Oncology Group; PS = performance status; CIS = carcinoma in situ; LVI = lymphovascular invasion. * Mann-Whithney test. y Pearson x 2 test.

EUROPEAN UROLOGY 56 (2009) 512 519 515 Table 3 Univariable and multivariable Cox proportional hazards regression model analyses for prediction of recurrence and causespecific mortality in 552 patients treated with radical nephroureterectomy with lymphadenectomy. Variables Univariable analysis Multivariable analysis Recurrence CSM Recurrence CSM HR p HR p HR p HR p Tumor grade 3.17 <0.001 2.94 <0.001 1.97 0.01 1.84 0.03 No. of LNs removed 0.99 0.19 0.99 0.17 0.97 0.04 0.97 0.12 (continuous variable) Pathologic stage (muscle-invasive 6.16 <0.001 7.09 <0.001 3.83 <0.001 3.70 <0.001 vs non muscle-invasive) Nodal status (N+ vs N0) 1.88 <0.001 1.93 <0.001 1.48 <0.001 1.50 <0.001 LVI 2.34 <0.001 2.72 <0.001 0.99 0.98 1.15 0.44 Age (continuous variable) 1.01 0.09 1.02 0.002 1.01 0.50 1.02 0.04 Institution 0.97 0.42 0.99 0.38 0.98 0.57 0.96 0.11 Tumor necrosis 2.71 <0.001 3.05 <0.001 1.42 0.02 1.54 0.01 Adjuvant chemotherapy 2.35 <0.001 2.02 <0.001 1.29 0.11 0.98 0.98 Tumor architecture 2.83 <0.001 3.14 <0.001 1.71 <0.001 1.89 <0.001 (sessile vs papillary) ECOG PS 1.53 <0.001 1.47 0.002 1.49 <0.001 1.38 0.02 CSM = cause-specific mortality; HR = hazard ratio; LN = lymph node; LVI = lymphovascular invasion; ECOG = Eastern Cooperative Oncology Group; PS = performance status. Seattle, WA, USA). Moreover, all tests were two-sided, with a significance level set at 0.05. 3.2. Association of the number of lymph nodes removed with clinical outcomes in patients with negative lymph nodes 3. Results 3.1. Association of the number of lymph nodes removed with clinical outcomes in all patients Baseline clinical characteristics of the 552 patients are described in Table 2. The median patient age was 68 yr (range: 27 97). The median number of removed LNs was five (range: 1 41) for the entire population, five (range: 1 41) for open RNU, and four (range: 1 21) for laparoscopic RNU ( p = 0.12). The median number of removed LNs was four (range: 1 38) for the patients with UTUC of the renal pelvis and five (range: 1 41) for patients with UTUC of the ureter ( p < 0.001). Median follow-up for patients alive at last follow-up was 48 mo (range: 1 246). Advanced patient age, sessile architecture, muscleinvasive disease, pn+ status, high grade, high Eastern Cooperative Oncology Group (ECOG) performance status (PS), the presence of tumor necrosis, and LVI were associated with recurrence and CSM in univariable analyses (all p < 0.01). In multivariable analysis, sessile architecture, muscle-invasive disease, pn+ disease, high tumor grade, higher ECOG PS, and the presence of tumor necrosis were also independent predictors of recurrence and CSM (all p 0.03; Table 3). The number of LNs removed was independently associated with recurrence (HR: 0.97; p = 0.04) but not with CSM ( p = 0.1; Table 3). Laparoscopic RNU was generally performed in patients with more favorable pathologic features; nonetheless, after adjustment for all covariates, open and laparoscopic RNU had no residual effect on cancer recurrence and mortality ( p = 0.1 for both). In pn0 patients (n = 412), the median number of LNs removed was six (range: 1 41). In univariable analysis, the number of LNs removed was associated with recurrence (HR: 0.97; p = 0.03) and CSM (HR: 0.96; p = 0.04). When the variable was coded as a cubic spline curve to allow for nonlinear effects, the curve depicting the relationship between the number of LNs removed and CSM (Fig. 1) did not plateau but continued to decrease as the number of LNs removed increased. In multivariable analyses, the number of LNs removed, the sessile architecture, and muscle-invasive disease achieved the independent predictor status for recurrence and CSM (all p 0.02; Table 4). The most informative cut-off for the number of LNs removed was eight for recurrence as well as for CSM. Patients treated with an extended lymphadenectomy (eight LNs or more) showed a lower recurrence rate (HR: 0.49; p < 0.01) and decreased CSM probability (HR: 0.42; p < 0.01) relative to patients treated with limited lymphadenectomy (fewer than eight LNs). Fig. 2 depicts the survival of pn0 UTUC patients according to the dichotomized number of LNs removed (8 vs <8). The 5-yr survival of patients treated with an extended lymphadenectomy (8 LNs) was statistically significantly higher than in the patients treated with a limited lymphadenectomy (<8 LNs) (84% vs 73%; p = 0.038). In multivariable analyses, the dichotomized variable defining the number of LNs removed also achieved independent predictor status (HR: 0.49; p = 0.005 and HR: 0.40; p = 0.003) in predicting recurrence and CSM, respectively (Table 4). Finally, the inclusion of the variable defining dichotomously the number of removed LNs (<8vs8) in the base model (age, ECOG PS, pathologic stage, grade, architecture, and LVI) significantly increased the accuracy in predicting CSM (+1.7%; p < 0.001).

516 EUROPEAN UROLOGY 56 (2009) 512 519 LNs removed was not associated with recurrence ( p =0.48) or CSM ( p = 0.74) in univariable or multivariable Cox regression analysis ( p =0.54andp =0.88,respectively). 4. Discussion Fig. 1 The plots depict the (A) linear and (B) nonlinear effects of the number of lymph nodes removed on cause-specific mortality in 412 patients with pn0 upper-tract urothelial carcinoma. LN = lymph node. 3.3. Association of the number of lymph nodes removed with clinical outcomes in patients with positive lymph nodes In pn+ patients (n =140), the median number of LNs removed was five (range: 1 34). The number of In UTUC patients, treatment with RNU possibly provides the best survival outcomes and the lowest local recurrence rates. Despite the recent improvements in imaging techniques, the preoperative assessment of nodal status remains poor, and a routine lymphadenectomy seems necessary for accurate nodal staging. Unfortunately, to date, only a small number of singlecenter studies have reported the impact of lymphadenectomy on clinical outcomes [5,6,13 15]. In a recent study of 169 patients, Kondo et al suggested that the extent of lymphadenectomy improves survival in patients with advanced-stage UTUC [13]. Moreover, Brausi et al demonstrated that lymphadenectomy improves diseasefree survival (DFS) and cancer-specific survival (CSS) in patients with muscle-invasive UTUC [14]. We have previously observed in 132 patients that the number of LNs removed has a statistically significantly effect on DFS and CSS [15]. Although the extent or absolute limit of lymphadenectomy remains to be defined, this growing body of data supports a more extended lymphadenectomy at the time of RNU in patients who are appropriate surgical candidates. The current study represents the largest UTUC patient cohort ever collected. Our data show that an extended lymphadenectomy results in lower recurrence rates and decreased CSM in pn0 UTUC patients treated with RNU. This suggests that the extent of lymphadenectomy improves disease staging, thus better stratifying node-negative and node-positive patients. Indeed, it can be argued that when a patient is diagnosed with a pn0 UTUC, the number of LNs removed is an index of the accuracy of nodal status assessment: Patients diagnosed with pn0 UTUC by a limited Table 4 Univariable and multivariable Cox proportional hazards regression model analyses for prediction of recurrence and causespecific mortality in 412 pn0 patients treated with radical nephroureterectomy with lymphadenectomy. Variables UVA MVA MVA Recurrence CSM Recurrence CSM Recurrence CSM HR p HR p HR p HR p HR p HR p Tumor grade 2.21 <0.01 1.84 0.03 1.49 0.09 1.27 0.44 1.61 0.10 1.26 0.44 Pathologic stage (muscle-invasive 4.86 <0.001 6.05 <0.001 4.61 <0.001 4.92 <0.001 4.68 <0.001 5.27 <0.001 vs non muscle-invasive) LVI 1.73 <0.01 2.48 <0.001 0.99 0.98 1.55 0.07 1.01 0.95 1.62 0.04 Age (continuous variable) 1.01 0.23 1.02 0.05 1.01 0.78 1.01 0.43 1.01 0.84 1.01 0.42 Tumor necrosis 1.79 <0.01 1.89 <0.01 1.29 0.20 1.36 0.18 1.33 0.15 1.41 0.13 Adjuvant chemotherapy 1.95 0.03 1.84 0.02 1.45 0.12 1.23 0.46 1.41 0.15 1.17 0.57 Institutions 0.97 0.81 0.98 0.70 0.99 0.84 0.99 0.79 0.96 0.82 0.99 0.78 Architecture (sessile vs papillary) 2.76 <0.001 2.99 <0.001 2.14 <0.001 2.27 <0.001 2.07 <0.001 2.17 0.001 ECOG PS 1.38 0.04 1.16 0.45 1.49 0.01 1.17 0.45 1.53 <0.01 1.31 0.19 No. of LNs removed (continuous variable) 0.96 0.03 0.96 0.04 0.94 0.01 0.93 0.02 No. of LNs removed (8 vs <8) 0.66 0.04 0.54 0.03 0.49 <0.01 0.42 <0.01 UVA = univariable analysis; MVA = multivariable analysis; CSM = cause-specific mortality; HR = hazard ratio; LVI = lymphovascular invasion; ECOG = Eastern Cooperative Oncology Group; PS = performance status; LN = lymph node.

EUROPEAN UROLOGY 56 (2009) 512 519 517 Fig. 2 Cancer-specific survival according to the number of lymph nodes removed in pn0 patients (n = 412). CSM = cause-specific mortality; LN = lymph node. lymphadenectomy may have been inaccurately staged, and some positive nodes may have been skipped during the surgery. In this case, the improvement in CSS of pn0 patients could simply be the result of better staging, which eliminates the negative impact on CSS for patients with nodal metastases that have been erroneously classified as pn0. In contrast, we could also speculate that an extended lymphadenectomy may provide therapeutic advantages in pn0 patients with UTUC: A higher number of nodes removed in pn0 patients may be related to the removal of undetected LN micrometastases and the subsequent improvement in survival. Moreover, we could hypothesize that the number of LNs removed may act as a proxy for thoroughness of surgical extirpation, thus offering a further reason why lymphadenectomy was associated with lower risk of recurrence. Unfortunately, we are unable to obtain this information from our data set. Although the cubic spline curve analysis showed that the probability of survival after RNU continued to rise in proportion to the number of LNs removed, an attempt was made to identify the most informative cut-off to define and quantify the best lymphadenectomy extent. It was found that patients in whom eight or more LNs had been removed exhibited a lower risk of cancer recurrence and cancerrelated death, also after adjusting for the effects of all possible confounders. In pn+ patients, no association was found between the number of LNs removed and outcomes, after adjustment for all covariates. It is possible that the absence of a statistically significant effect of lymphadenectomy on survival in pn+ patients might be caused by the small sample size, the low median number of LNs removed, and the effect of adjuvant chemotherapy. That being said, also without a statistically significant result, some important aspects regarding the use of an extended lymphadenectomy in pn+ patients deserve to be mentioned. An extended lymphadenectomy may allow surgeons to more accurately detect all possible nodal metastases and better identify candidates for adjuvant chemotherapy, although an extended lymphadenectomy s effect on survival remains controversial. Additionally, an extended lymphadenectomy might reduce the residualtumorburdenandmighthaveapositiveimpact in a subpopulation of patients with pn+ UTUC, as has been previously described in a transitional cell carcinoma bladder setting [17]. The importance of the lymphadenectomy extent was also recently reported by Bolenz et al [18]. The authors demonstrated that LN density significantly affected CSM in UTUC patients. Because LN density is determined by the proportion of the number of positive LNs and the number of LNs removed, such findings reinforce our preliminary hypothesis regarding the potential therapeutic value of an extended lymphadenectomy. The current study is not devoid of limitations. First, the power of our conclusion may be somewhat limited by the retrospective nature of the study. Second, the extent of lymphadenectomy was decided by the surgeon, who could have been influenced by many uncontrolled parameters, and the low median number of nodes removed reflects the lack of consensus regarding the usefulness of lymphadenectomy. Third, the wide range of LNs removed reflects a lack of standardization of the LND, different sampling techniques and processing of tissue by the pathologist, en bloc removal versus presentation in packets, and the age of the patient. Finally, the population in this study underwent RNU by multiple surgeons at different institutions and had their specimens evaluated by multiple pathologists. However, the institution was included as a confounder in the multivariable analyses to test this possible imbalance,

518 EUROPEAN UROLOGY 56 (2009) 512 519 and no impact was found on recurrence and CSM. Moreover, all specimens were re-examined by dedicated genitourinary pathologists according to identical, strict criteria agreed on by all participating pathologists. Although prognostic factors may perform well in the select group of patients operated on by a single surgeon, it remains to be determined whether these are applicable to a greater population of patients with UTUC. Thus, the multiinstitutional nature of our study and the use of local pathologic interpretation may make our findings more relevant and applicable in both the academic and community settings. 5. Conclusions The extent of the lymphadenectomy in pn0 UTUC patients is associated with recurrence and CSM. Longer survival was observed in patients in whom at least eight LNs had been removed, suggesting that the extent of lymphadenectomy improves disease staging and thus better stratifying nodenegative and node-positive patients. Moreover, in nodenegative patients, lymphadenectomy extent may relate to the removal of undetected LN micrometastases and thus may be associated with an improved survival in patients undergoing RNU. Prospective trials are required to clearly state the impact of lymphadenectomy on survival in patients with UTUC. Author contributions: Marco Roscigno 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: Roscigno, Shariat Acquisition of data: Roscigno, Shariat, Margulis, Karakiewicz, Remzi, Kikuchi, Zigeuner, Weizer, Sagalowsky, Bensalah, Raman, Bolenz, Kassou, Koppie, Wood, Wheat, Langner, Ng, Capitanio, Bertini, Fernández, Mikami, Isida, Ströbel, Montorsi Analysis and interpretation of data: Roscigno, Shariat, Capitanio Drafting of the manuscript: Roscigno, Shariat, Bertini Critical revision of the manuscript for important intellectual content: Langner, Wood, Capitanio, Sagalowsky, Montorsi Statistical analysis: Capitanio, Roscigno Obtaining funding: None. Administrative, technical, or material support: None. Supervision: Roscigno 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: Christopher G. Wood has a financial interest and/or other relationship with Antigenics, Inc.; Pfizer, Inc.; Bayer Pharmaceuticals Corporation/Onyx Pharmaceuticals; the Kidney Cancer Association; and ETHICON, Inc. Francesco Montorsi has a financial interest and/or other relationship with Pfizer, Inc.; Bayer Pharmaceuticals Corporation; Ely Lilly and Company; Laboratoires Pierre Fabre; American Medical Systems, Inc.; and GlaxoSmithKline. Funding/Support and role of the sponsor: None References [1] Jemal A, Siegel R, Ward E, Murray T, Xu J, Thun MJ. Cancer statistics, 2007. CA Cancer J Clin 2007;57:43 66. [2] Corrado F, Ferri C, Mannini D, et al. Transitional cell carcinoma of the upper urinary tract: evaluation of prognostic factors by histopathology and flow cytometric analysis. J Urol 1991;145: 1159 63. [3] Schatteman P, Chatzopoulos C, Assenmacher C, et al. Laparoscopic nephroureterectomy for upper urinary tract transitional cell carcinoma: results of a Belgian retrospective multicentre survey. Eur Urol 2007;51:1633 8. [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] Komatsu H, Tanabe N, Kubodera S, et al. The role of lymphadenectomy in the treatment of transitional cell carcinoma of the upper urinary tract. J Urol 1997;157:1622 4. [6] Miyake H, Hara I, Gohji K, et al. The significance of lymphadenectomy in transitional cell carcinoma of the upper urinary tract. BJU Int 1998;82:494 8. [7] Koppie TM, Vickers AJ, Vora K, et al. Standardization of pelvic lymphadenectomy performed at radical cystectomy: can we establish a minimum number of lymph nodes that should be removed? Cancer 2006;107:2368 74. [8] Leissner J, Hohenfellner R, Thuroff JW, et al. Lymphadenectomy in patients with transitional cell carcinoma of the urinary bladder: significance for staging and prognosis. BJU Int 2000;85:817 23. [9] Herr HW, Bochner BH, Dalbagni G, et al. Impact of the number of lymph nodes retrieved on outcome in patients with muscle invasive bladder cancer. J Urol 2002;167:1295 8. [10] Konety BR, Joslyn SA, O Donnell MA. Extent of pelvic lymphadenectomy and its impact on outcome in patients diagnosed with bladder cancer: analysis of data from the Surveillance, Epidemiology and End Results Program data base. J Urol 2003;169:946 50. [11] Abdel-Latif M, Abol-Eneim H, El-Baz M, et al. Nodal involvement in bladder cancer cases treated with radical cystectomy: incidence and prognosis. J Urol 2004;172:85 9. [12] Liedberg F, Månsson W. Lymph node metastasis in bladder cancer. Eur Urol 2006;49:13 21. [13] Kondo T, Nakazawa H, Ito F, Hashimoto Y, et al. Impact of the extent of regional lymphadenectomy on the survival of patients with urothelial carcinoma of the upper urinary tract. J Urol 2007;178: 1212 7. [14] Brausi MA, Gavioli M, De Luca G, et al. Retroperitoneal lymph node dissection (RPLD) in conjunction with nephroureterectomy in the treatment of the infiltrative transitional cell carcinoma (TCC) of the upper urinary tract: impact on survival. Eur Urol 2007;52: 1414 20. [15] 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. [16] Epstein JI, Amin MB, Reuter VR, Mostofi FK. The World Health Organization/International Society of Urological Pathology consensus classification of urothelial (transitional cell) neoplasms of the urinary bladder. Bladder Consensus Conference Committee. Am J Surg Pathol 1998;22:1435 8. [17] Stein JP, Cai J, Groshen S, et al. Risk factors for patients with pelvic lymph node metastases following radical cystectomy with en bloc pelvic lymphadenectomy: the concept of lymph node density. J Urol 2003;170:35 41. [18] Bolenz C, Shariat SF, Fernandez MI, et al. Risk stratification of patients with nodal involvement in upper tract urothelial carcinoma: value of lymph-node density. BJU Int 2009;103:302 6.

EUROPEAN UROLOGY 56 (2009) 512 519 519 Editorial Comment on: The Extent of Lymphadenectomy Seems to Be Associated with Better Survival in Patients with Nonmetastatic Upper-Tract Urothelial Carcinoma: How Many Lymph Nodes Should Be Removed? Guido Dalbagni Memorial Sloan-Kettering Cancer Center, New York, NY, USA dalbagng@mskcc.org The role of lymph node dissection in the management of patients with upper tract urothelial cancer is undefined and not universally embraced. There are important questions to be addressed. Is there a role for lymphadenectomy in upper tract urothelial carcinoma? Several studies have shown a survival benefit in patients who underwent a lymphadenectomy. Brausi et al indicated that lymph node dissection and the T stage were the only factors affecting survival [1]. Kondo et al reported an improvement in the cancerspecific survival of patients with pt3 disease who underwent a lymphadenectomy [2]. To date, however, there has not been any randomized trial assessing the importance of a lymph node dissection in the management of upper tract urothelial carcinoma. Roscigno et al did not find an association between the number of lymph nodes and cause-specific survival in the entire population, which could be a reflection of the limited dissection in some cases [3]. Indeed, the node count was as low as one lymph node in some patients. They did, however, find an association with survival in the subset of patients with N0 disease, which could be due to stage migration rather than the effect of the lymphadenectomy. What is the extent of lymph node dissection? The lymphatic drainage of the kidney ends in the lateral aortic nodes on each side of the abdominal aorta in front of the medial margin of the psoas major, the crus of the diaphragm, and the sympathetic trunk. The lymphatics from the upper part of the ureter join the renal lymphatic system or drain directly in the lateral aortic nodes near the origin of the testicular vessels. Those from the midureter pass to the common iliac nodes, and those from the pelvic part end in the common, the external, and internal iliac nodes. The pattern of metastasis in upper tract urothelial carcinoma follows the lymphatic drainage, as suggested by Kondo et al [4]. In a small retrospective analysis, Kondo et al reported that tumors of the renal pelvis metastasized to the right renal hilum and paracaval and retrocaval nodes, and tumors of the right upper two-thirds of the ureter spread to the retrocaval and interaortic nodes. Tumors of the left renal pelvis end in the left renal hilum and para-aortic nodes, while tumors in the left ureter metastasize to the para-aortic nodes. Tumors of the lower ureter metastasize inferior to the aortic bifurcation. To date, there is no agreed template for lymph node dissection in the setting of upper tract urothelial cancer. A thorough lymph node dissection is not performed in the majority of the patients [5]. It varies from no dissection to lymph node plucking to dissection of the ipsilateral and interaortocaval lymph nodes. How can lymphadenectomy quality be assessed? The number of lymph nodes has been addressed by Roscigno et al [3] as a surrogate to the extent of dissection, but an established number of lymph nodes is not enough to optimize survival, as shown by Koppie et al [6]: Survival continued to rise as the number of lymph nodes increased in a cohort of patients with bladder cancer who underwent a radical cystetomy. References [1] Brausi MA, Gavioli M, De Luca G, et al. Retroperitoneal lymph node dissection (RPLD) in conjunction with nephroureterectomy in the treatment of infiltrative transitional cell carcinoma (TCC) of the upper urinary tract: impact on survival. Eur Urol 2007;52:1414 8. [2] Kondo T, Nakazawa H, Ito F, et al. Impact of the extent of regional lymphadenectomy on the survival of patients with urothelial carcinoma of the upper urinary tract. J Urol 2007;178:1212 7, discussion 1217. [3] Roscigno M, Shariat SF, Margulis V, et al. The extent of lymphadenectomy seems to be associated with better survival in patients with nonmetastatic upper-tract urothelial carcinoma: how many lymph nodes should be removed? Eur Urol 2009;56:512 9. [4] Kondo T, Nakazawa H, Ito F, et al. Primary site and incidence of lymph node metastases in urothelial carcinoma of upper urinary tract. Urology 2007;69:265 9. [5] Secin FP, Koppie TM, Salamanca JI, et al. Evaluation of regional lymph node dissection in patients with upper urinary tract urothelial cancer. Int J Urol 2007;14:26 32. [6] Koppie TM, Vickers AJ, Vora K, et al. Standardization of pelvic lymphadenectomy performed at radical cystectomy: can we establish a minimum number of lymph nodes that should be removed? Cancer 2006;107:2368 74. DOI: 10.1016/j.eururo.2009.06.005 DOI of original article: 10.1016/j.eururo.2009.06.004