Lymph Node Density Affects Cancer-Specific Survival in Patients with Lymph Node Positive Urothelial Bladder Cancer Following Radical Cystectomy

Transcription

1 EUROPEAN UROLOGY 59 (2011) available at journal homepage: Platinum Priority Bladder Cancer Editorial by James W.F. Catto on pp of this issue Lymph Node Density Affects Cancer-Specific Survival in Patients with Lymph Node Positive Urothelial Bladder Cancer Following Radical Cystectomy Matthias May a,1, Edwin Herrmann b,1, Christian Bolenz c, Arne Tiemann b, Sabine Brookman-May d, Hans-Martin Fritsche d, Maximilian Burger d, Alexander Buchner f, Christian Gratzke f, Christian Wülfing b, Lutz Trojan c,jörg Ellinger e, Derya Tilki f, Christian Gilfrich a, Thomas Höfner g, Jan Roigas i, Mario Zacharias j, Sven Gunia k, Wolf F. Wieland d, Markus Hohenfellner g, Maurice S. Michel c, Axel Haferkamp g,h, Stefan C. Müller e, Christian G. Stief f, Patrick J. Bastian f, * a Klinik für Urologie, St. Elisabeth Klinikum Straubing, Straubing, Germany b Urologische Universitätsklinik, Universität Münster, Münster, Germany c Universitätsmedizin Mannheim, Mannheim, Germany d Klinik für Urologie der Universität Regensburg, Caritas Krankenhaus St. Josef, Regensburg, Germany e Klinik und Poliklinik für Urologie und Kinderurologie, Universitätsklinikum Bonn, Bonn, Germany f Urologische Universitätsklinik, LMU München, München, Germany g Urologische Universitätsklinik, Universität Heidelberg, Heidelberg, Germany h Urologische Universitätsklinik, Universität Frankfurt/Main, Frankfurt/Main, Germany i Urologische Klinik, Vivantes Kliniken Am Urban und Im Friedrichshain Berlin, Berlin, Germany j Urologische Klinik, Vivantes Klinikum AVK Berlin, Berlin, Germany k Pathologisches Institut, Helios-Klinikum Bad Saarow, Bad Saarow, Germany Article info Abstract Article history: Accepted January 14, 2011 Published online ahead of print on January 28, 2011 Keywords: Bladder cancer Urothelial cancer Radical cystectomy Lymphadenectomy Lymph node metastases Number of lymph nodes Lymph node density Prognosis Background: The prognosis for patients with lymph node (LN) positive bladder cancer (BCa) is likely affected by the extent of lymphadenectomy in radical cystectomy (RC) cases. Specifically, the prognostic significance of the LN density (ratio of positive LNs to the total number removed) has been demonstrated. Objective: To evaluate the prognostic signature of lymphadenectomy variables, including the LN density, for a large, multicentre cohort of RC patients with LNpositive BCa. Design, setting, and participants: The clinical and histopathologic data from 477 patients with LN-positive urothelial BCa (pn1 2) were analysed. The median follow-up period for all living patients was 28 mo. Measurements: Multivariable Cox regression analysis was used to test the effect of various pelvic lymph node dissection (PLND) variables on cancer-specific survival (CSS) based on colinearity in various models. 1 These authors contributed equally to this study. * Corresponding author. Department of Urology, Klinikum der Universität München Campus Großhadern, Ludwig-Maximilians-Universität München, Marchioninistr. 15, Munich, Germany. Tel ; Fax: address: Patrick.bastian@med.uni-muenchen.de (P.J. Bastian) /$ see back matter # 2011 European Association of Urology. Published by Elsevier B.V. All rights reserved. doi: /j.eururo

2 EUROPEAN UROLOGY 59 (2011) Results and limitations: The median number of LNs removed was 12 (range: 1 66), and the median number of positive LNs was 2 (range: 1 25). Two hundred ninety (60.8%) of the patients presented with stage pn2 disease. The median and mean LN density was 17.6% and 29% (range: ), respectively, where 268 (56.2%) and 209 (43.8%) patients exhibited am LN density of 20% and >20%, respectively. In separate multivariable Cox regression models adjusted for age, sex, ptn stage, grade, associated Tis, and adjuvant chemotherapy, the interval-scaled LN density (hazard ratio [HR]: 1.01; p = 0.002) and the LN density, ordinal-scaled by 20% (HR: 1.65; p < 0.001) exhibit independent effects on CSS. In addition, an independent contribution appears from the pt but not the pn stage. Limitations include surgeon selection bias when determining the extent of lymphadenectomy. Conclusions: Our results support the prognostic relevance of LN density in patients with LN-positive BCa, where a threshold value of 20% stratifies the population into two prognostically distinct groups. Before LN density is integrated into the clinical decision-making process, these results should be validated by prospective studies with defined LN templates and standardised histopathologic methods. # 2011 European Association of Urology. Published by Elsevier B.V. All rights reserved. 1. Introduction The goal of pelvic lymph node dissection (PLND) during radical cystectomy (RC) for invasive bladder cancer (BCa) is to remove the regional lymph nodes (LN), particularly the typically affected LN sites [1,2]. The histopathologic results and the extent of LN removal have significance as prognostic criteria and thus as indicators for adjuvant therapy [3 5]. Moreover, there is an increasing perception that extensive PLND is an important therapeutic measure associated with improvement in cancer-specific survival (CSS) in both LN-negative and LN-positive patients [3 5]. This viewpoint is supported by the observed histories of long-term survival in patients with LN-positive cancers [6 14]. The available data on the therapeutic effect of PLND with LN-positive BCa are inconsistent, so that no evidence-based recommendations are presently specified [6 14]. Furthermore, the lowest number of LNs to be removed and the necessity for an extensive PLND is debatable, made more confusing by the variety of definitions for standard versus extensive PLND [3,4]. Among the results are mostly retrospective patient studies, with only a small number of cases involving the stratification of prognostic criteria in LN-positive BCa. The prognosis criteria described include the pn stage, the number of positive LNs and the number of removed LNs, their derived ratio (LN density), and the initiation of adjuvant chemotherapy [3 14]. The significance of LN density was first pointed out in two separate studies in 2003 by Stein et al and by Herr et al, where a prognosis-relevant threshold value of 20% was established [6,7,12 14]. The goal of our multicentre data analysis study was to evaluate prognosis criteria using a large cohort of RC patients with LN-positive urothelial BCa and to define factors associated with long-term survival. 2. Patients and methods Clinical and pathologic information was gathered from 2556 patients with urothelial BCa who had undergone RC in one of eight German urology clinics (six university clinics, two maximum care clinics) over a 20-yr period ( ); this information was merged into a multiinstitutional database. The indication for RC in all participating clinics was consistent with current guideline recommendations. Among the patients in this database, 25.5% (n = 653) exhibited LN-positive cancer. The exclusion criteria were having received neoadjuvant chemotherapy (n = 31), morphological evidence of organ metastasis on imaging (n = 13), and data incongruent with the study criteria (n = 128). Four patients with an LN status of pn3 (one or more LN metastases >5 cm) were likewise excluded. The study group in the present analysis thus included 477 patients with LN-positive tumours (pn1 2, M0) for whom data were available concerning the following study criteria: pt, pn stage, grading, associated Tis, adjuvant chemotherapy or radiation therapy (RT), number of LNs removed, and number of positive LNs. LN density was calculated using the following formula: number of positive LNs times 100% divided by the total number of removed LNs. In addition to the evaluation of LN density as an interval-scaled continuous variable, an additional analysis was conducted using an ordinal-scaled dichotomised variable with a threshold value of 20% (20% vs >20%), a figure previously selected in many publications [6,7,12 14]. The mean observation period for patients who were still alive at the end of the study was 41.3 mo (range: 2 240; median: 28 mo). The extent of the preoperative staging examination, the histopathologic workup of the RC specimens, as well as the type of follow-up testing were defined through standardised though individual in-house protocols for the respective clinics. All histopathology reports were based on the TNM classification (6th edition) [15]; prior to 2002, TNM classification was made by the study pathologist (S.G.). If problems occurred with an assignment, the original histopathology report was reviewed. Total survival and CSS were calculated by the Kaplan-Meier method, and group differences were indicated by the log-rank test. Multivariable Cox regression models were developed with the goal of identifying those factors that would enable an independent prognosis of CSS (each model with stepwise backward elimination). Becauseofthe colinear association ofthe LNdensity withthe number of positive LNs (R = 0.67) and the total number of removed LNs (R = 0.46), separate regression models were developed. Model 1 included the number of positive LNs and the total number of removed LNs; Model 2 contained the LNdensity (model 2awas interval scaled; model 2bwas ordinal scaled). Moreover, an additional multivariable Cox regression analysis checked for the independent effect ofthe ordinal-scaled LNdensity dichotomised atthe 20% level on the CSS above and below a total number of removed LNs of n = 12 (models 2b-1 and 2b-2). The following parameters were the basis

3 714 EUROPEAN UROLOGY 59 (2011) determinants of the multivariable Cox regression models: age, sex, pt stage, pn stage, grading, associated Tis, year of treatment, and initiation of adjuvant chemotherapy or RT. A potential colinearity of the variables included in the final regression model was tested beforehand, with bivariatecorrelation analysesof all thevariables with one another, wherea positive interfactor correlation with an R < 0.5 was accepted. Finally, the analysis of the best threshold value for the LN density with respect to the dichotomised boundary between two prognostically different groups was carried out using the maximal x 2 statistic proposed by Boulesteix [16]. Achievement of a significance level of p 0.05 was adopted for all statistical tests. 3. Results The clinical and histopathologic criteria for the 477 patients with LN-positive urothelial BCa are shown in Table 1. Table 1 Clinical and histopathologic criteria for 477 patients with lymph node positive urothelial bladder cancer following radical cystectomy Variable No. (%) or variable unit Patient age, yr Median 66.3 Mean (range) 65.6 (33 86) Sex, No. (%) Female 101 (21.2) Male 376 (78.8) RC time frame, No. (%) (10.5) (37.5) (52.0) Follow-up, mo Median for all patients 16 Mean for all patients, range 26.8 Median for all patients still alive 28 by study end Mean for all patients still alive by 41.3 (2 240) study end (range) Median of deceased patients 12 Mean for deceased patients (range) 18 (0 142) pt stage, No. (%) <pt2 24 (5.0) pt2 103 (21.6) >pt2 350 (73.4) pn stage, No. (%) pn1 187 (39.2) pn2 290 (60.8) Grade, No. (%) <G3 79 (16.6) G3 398 (83.4) Associated Tis, No. (%) 134 (28.1) Adjuvant chemotherapy, No. (%) 159 (33.3) Total number of removed LNs: Median 12 Mean (range) 14 (1 66) Total number of removed LNs: Median 12 Mean (range) 14 (1 66) Total number of positive LNs Median 2 Mean (range) 3.2 (1 25) LN density, % Median 17.6 Mean (range) 29 ( ) LN density, dichotomised, No. (%) 20% 268 (56.2) >20% 209 (43.8) A tumour stage that extends beyond the muscles (pt3 or higher at 73%) and an LN status of pn2 (n = 290; 61%) appear to be dominant. The median number of removed LNs was 12 (range: 1 66), and the median number of positive LNs was 2 (range: 1 25). The median and mean LN density were 17.6% and 29% (range: ), respectively, where 268 (56.2%) and 209 (43.8%) patients exhibited an LN density 20% and >20%, respectively. For the entire study population after 1, 3, and 5 yr, the CSS was 73%, 49%, and 39%, respectively. Patients in tumour stages pt2 or lower and pt3 or higher exhibited a CSS of 49% and 35% ( p = 0.003), respectively, after 5 yr (Fig. 1). The corresponding results for patients with an LN density 20% and >20% were 46% and 31%, respectively ( p < 0.001; Fig. 2). Table 2 shows the results of the univariable and multivariable Cox regression analyses relating to CSS. The pn stage lost its independent effect once the LN density parameter (either interval scaled or ordinal scaled) was included. In contrast, the pt stage (three-fold ordinal scaling) exhibited an independent contribution in each of the three Cox regression models. The total number of removed LNs had an independent effect (interval scaled; hazard ratio [HR]: 0.97; p = 0.002), while the total number of positive LNs did not. The decisive variable in the Cox model with respect to CSS was the LN density, where an HR of 1.65 (95% confidence interval [CI], ; p < 0.001) resulted from inclusion of the parameter dichotomised at the 20% level (Table 2). In two additional Cox regression models for CSS, an LN density 20% versus >20% made an independent contribution regardless of whether the total number of removed LNs was < 12 or 12 (Table 3). [()TD$FIG] Fig. 1 Kaplan-Meier curve for cancer-specific survival following radical cystectomy in 477 patients with lymph node positive bladder cancer stratified according to the pathologic tumour stage lower than pt3 vs pt3 or higher.

4 EUROPEAN UROLOGY 59 (2011) Table 2 Univariable and multivariable Cox regression analyses for cancer-specific survival following radical cystectomy Variable Univariate analysis HR (95% CI), p value analysis HR (95% CI), p value Model 1 analysis HR (95% CI), p value Model 2a * analysis HR (95% CI), p value Model 2b ** Patient age (continuous) 1.01 ( ), p = Sex: female vs male 1.07 ( ), p = RC time frame: >2002 vs vs < ( ), p = pt stage: >pt2 vs pt2 vs <pt ( ), p = ( ), p = ( ), p = ( ), p = pn stage: pn2 vs pn ( ), p = ( ), p = Grade: G3 vs <G ( ), p = Associated Tis: yes vs no 0.97 ( ), p = Adjuvant chemotherapy: yes vs no 0.89 (0, ), p = Total no. LNs (continuous) 0.97 ( ), p = ( ), p = Total no. positive LNs (continuous) 1.02 ( ), p = LN density (continuous) 1.01 ( ), p = ( ), p = LN density: >20% vs 20% 1.70 ( ), p < ( ), p < HR = hazard ratio; CI = confidence interval; = not stated; RC = radical cystectomy; LN = lymph node. * Model 2a: Exclusion of total number of removed LNs and total number of positive LNs as well as integration of the interval-scaled LN density. ** Model 2b-1: Exclusion of LN density. Model 2b: Exclusion of total number of removed LNs and total number of positive LNs as well as integration of the ordinal-scaled LN density. In a multivariable Cox regression model (with exclusion of patients undergoing adjuvant chemotherapy), only pt stage (HR: 1.39; 95% CI, ; p = 0.033) and LN density with a cut-off value of 20% (HR: 1.68; 95% CI, ; p = 0.002) revealed an independent influence on CSS. By limiting the analysis to patients undergoing adjuvant treatment, only for pn stage was a significant influence on CSS assessed (HR: 1.75; 95% CI, ; p = 0.024). The independent influence of pt stage and LN density of 20% in this model was not significant. The inclusion of LN density with a threshold value of 20 led to an approximately 58.7% improvement ( p < 0.001) in the [()TD$FIG] Table 3 Multivariable Cox regression analyses for cancer-specific survival following radical cystectomy Variable analysis HR (95% CI), p value Model 2b-1 * (n = 214) analysis HR (95% CI), p value Model 2b-2 ** (n = 263) Patient age (continuous) 1.03 ( ), p = Sex: female vs male RC time frame: >2002 vs vs <1997 pt stage: >pt2 vs pt2 vs<pt ( ), p = pn stage: pn2 vs pn1 Grade: G3 vs <G3 Associated Tis: yes vs no Adjuvant chemotherapy: yes vs no LN density: >20% vs 20% 1.83 ( ), p = ( ), p = HR = hazard ratio; CI = confidence interval; = not stated; RC = radical cystectomy; LN = lymph node. * Model 2b-1: <12 LN. ** Model 2b-2: 12 LN. predictive value of Cox model 2b with respect to CSS. Alternatively, raising the threshold value for LN density to 25% resulted in an improvement in the predictive value of only 36% ( p < 0.001). By integration of different LN density cut-off values, the predictive ability of the multivariable Cox model 2b was influenced as follows: an LN density of 4% showed an improvement of 7% ( p = 0.056), and an LN density of 11% showed an improvement of 25.6% ( p =0.022). 4. Discussion Fig. 2 Kaplan-Meier curve for cancer-specific survival following radical cystectomy in 477 patients with lymph node (LN) positive bladder cancer stratified according to the LN density =20% vs >20%. The present RC cohort represents the largest evaluation of patients with LN-positive urothelial BCa with the exception

5 716 EUROPEAN UROLOGY 59 (2011) Table 4 Literature survey of the existing studies on the prognostic value of lymph node density Study, yr No. Total no. of LN, median (range) Total no. of positive LN, median (range) Threshold value of the LN density Results Stein et al, 2003[6] (1 96) 2 (1 63) 20 vs >20 Significantly poorer RFS and OS above the threshold Herr, 2003 [7] (2 32) 3 (1 14) 20 vs >20 Significantly poorer CSS above the threshold Konety et al, 2003 [17] 361 N/A N/A <25, 26 50, 51 75, >75 No significantly poorer CSS above the multiple threshold values (MRA) Abdel-Latif et al, 2004 [18] (mean) 4.1 (mean) <10, 10 20, >20 No significantly poorer RFS above the multiple threshold values (MRA) Fleischmann et al, 2005 [14] (10 43) N/A <20 vs 20 No significantly poorer RFS above the threshold Kassouf et al, 2006 [19] (1 58) 2 (1 10) 25 vs >25 Significantly poorer RFS and OS above the threshold Steven and Poulsen, 2007 [13] (11 49) N/A 20 vs >20 Significantly poorer RFS and OS above the threshold value (URA) Kassouf et al, 2008 [12] (2 58) 2 (1 14) 20 vs >20 Significantly poorer CSS above the threshold Wright et al, 2008 [8] (1 48) 2 (1 18) <12.6, , 25,1 50, >50 Significantly poorer CSS and OS above the multiple threshold values (MRA) Wiesner et al, 2009 [10] (15 77) 3 (1 28) 11 vs >11 Significantly poorer CSS above the threshold Osawa et al, 2009 [11] (1 80) 2 (1 12) 25 vs >25 Significantly poorer OS above the threshold Bruins et al, 2009 [9] 181 * N/A 1 (1 2) 4 vs>4 Significantly poorer RFS and OS above the threshold Present study (1 66) 2 (1 25) 20 vs >20 Significantly poorer CSS above the threshold LN = lymph node; RFS = relapse-free survival; OS = overall survival; MRA = multiple regression analyses conducted; CSS = cancer-specific survival; N/A = not available; URA = univariable regression analyses only. * Exclusively evaluations of patients with 1 2 positive LNs. of the Surveillance Epidemiology and End Results population [8]. Consideration of our results and those of the previous studies (Table 4) points to LN density with a threshold value of 20% as a criterion for stratifying the rather heterogeneous prognoses of patients in stage pn+. Our study is the first to document that LN density (threshold value = 20%) has a significant effect on CSS, both for limited PLND (<12 LNs) as well as for standard/extensive PLND (12 LNs; Table 3). Before applying these findings to clinical practice, however, one must address whether LN density is superior to pn stage in terms of its prognostic value. The standard TNM classification for LN metastases (N category) is based on an evaluation of the number and size of positive LNs, dependent on the removal of positive LNs by the surgeon. Understaging can occur as a result of a failure to remove a sufficient number of LNs. Various studies have compared the prognostic value of LN density and pn stage. Herr investigated the 1997/ 2002 N staging system (valid through 2010), LN density (threshold value of 20%), and the number of positive LNs (threshold value of 4) with respect to their effect on CSS [7]. OnlyLNdensity(p = 0.002) exhibited an independent effect in a multivariable analysis (pn stage with p = 0.21; number of positive LNs with p =0.6) [7]. Comparable results were presented by Kassouf et al, who evaluated the databases of both the MD Anderson Cancer Centre and the Memorial Sloan-Kettering Cancer Centre, likewise with the 1997/2002 N staging system [12]. In a multivariable Cox regression model adjusted inter alia for adjuvant chemotherapy having been carried out, only a higher LN density (threshold value of 20%) correlated with a significant worsening of CSS (HR: 2.75; p < 0.01) [12]. In a univariate analysis conducted by Fleischmann and colleagues at Bern, significant differences were found between patients with different LN densities (threshold value of 20%) in relapse-free survival ( p =0.0034)andin overallsurvival(os;p =0.002) [14]. Stratification of the patients with respect to pn stage (pn1 vs pn2, likewise according to the 1997/2002 staging system) yielded no significant difference in the univariate analyses. In a multivariable Cox model that included LN density, ptn stage, number of positive LNs, and extracapsular spread of LN metastases, only the last-named parameter made an independent contribution [14]. In our analysis, LN density exhibited a stronger effect on CSS compared to pn stage according to the 1997/2002 staging system. Whether interval scaled (continuous) or ordinal scaled (threshold value of 20%), LN density crowded the pn stage out of the Cox model (Table 2). The parameters found in our study to have the strongest prognostic signature (LN density and pt stage) can be used to delineate groups characterised by different mortality risks. Patients at tumour stage pt2 or lower with an LN density 20% exhibit a CSS of 56% after 5 yr, while patients at tumour stage pt3 or higher with an LN density >20% exhibit a 5-yr CSS of 28% ( p < 0.001). To further address the clinical application of using LN density as a prognostic criterion requires an evaluation of the unresolved issues that remain regarding LN density. Previous studies relating to the prognostic value of LN density are summarised in Table 4. The various authors defined different threshold values for LN density, with the most frequently used value being 20%. This threshold value was confirmed through our investigation and was found to be superior when compared to a threshold value of 25% using multivariable Cox analysis.

6 EUROPEAN UROLOGY 59 (2011) It is unclear whether the prognostic significance of LN density extends to patients who have undergone adjuvant chemotherapy, while the number of such patients is increasing at the present time. Furthermore, it must be defined what sort of additional procedures will be offered to patients with an LN density above a certain threshold value (eg, 20%). Would they be suitable candidates for adjuvant chemotherapy, while patients below the threshold value would not? Would the prognostic discriminatory power of LN density be maintained if adjuvant chemotherapy were indicated for all patients? All of the Cox regression models developed in our investigation were adjusted for adjuvant chemotherapy, although this was not the case in most of the other previous studies [7,8,10,14,17 19]. A limitation of this study was the retrospective design, with a long timeframe for the investigation, which covered two decades and thus spanned modifications in the staging modalities and necessarily also the further development of surgical techniques, including PLND techniques. This also includes the removed LN range from 1 to 66, suggesting that some patients had an incomplete LN dissection or an inadequate pathologic review. Considering this, one must keep in mind that LN density in a group of patients with a complete LN dissection reflects the volume of metastatic disease, which results in a poor outcome. However, LN density in patients who underwent LN plucking does not reflect volume of disease, and the poor survival might other be the result of an inadequate LN dissection or higher metastatic load. Other limitations include the absence of centralised pathologic reviews and the different, clinicspecific diagnostic protocols, therapies, and follow-up care provided to the patients included in the study. Moreover, no information was available to us regarding patient comorbidities, which could also indirectly influence the surgeon s decision on the extent of the PLND; similarly, no information was available on the extracapsular spread of LN metastases. Finally, consideration must also be given to the fact that 19.6% of the LN-positive patients (128 of 653) were excluded from the total database because of incongruent or missing data. 5. Conclusions Despite its limitations, the results of this large RC cohort of LN-positive patients are important with respect to the heterogeneous prognoses in this patient population. LN density with a threshold value of 20% is presented as an instrument for stratifying these patients on the basis of their distinct oncologic progression. Before LN density can be integrated into the pn classification and the clinical decisionmaking process, however, the present results should be validated by prospective studies with defined LN dissection areas and standardised histopathologic examination methods. Overall, evidence has emerged from our investigation that patients with LN-positive urothelial BCa will benefit from an extended PLND in the course of an RC [20,21]. Author contributions: Patrick J. Bastian 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: May, Herrmann, Bolenz, Tiemann, Fritsche, Burger, Buchner, Bastian. Acquisition of data: May, Herrmann, Bolenz, Tiemann, Brookman-May, Fritsche, Burger, Buchner, Gratzke, Wülfing, Trojan, Ellinger, Tilki, Gilfrich, Höfner, Roigas, Zacharias, Gunia, Wieland, Hohenfellner, Michel, Haferkamp, Müller, Stief, Bastian. Analysis and interpretation of data: May, Herrmann, Bolenz, Tiemann, Fritsche, Burger, Buchner, Bastian. Drafting of the manuscript: May, Herrmann, Bolenz, Tiemann, Fritsche, Burger, Buchner, Gratzke, Trojan, Bastian. Critical revision of the manuscript for important intellectual content: May, Herrmann, Bolenz, Tiemann, Brookman-May, Fritsche, Burger, Buchner, Gratzke, Wülfing, Trojan, Ellinger, Tilki, Gilfrich, Höfner, Roigas, Zacharias, Gunia, Wieland, Hohenfellner, Michel, Haferkamp, Müller, Stief, Bastian. Statistical analysis: May, Herrmann, Bolenz, Tiemann, Fritsche, Burger, Buchner, Bastian. Obtaining funding: May, Herrmann, Bolenz, Tiemann, Brookman-May, Fritsche, Burger, Buchner, Gratzke, Wülfing, Trojan, Ellinger, Tilki, Gilfrich, Höfner, Roigas, Zacharias, Gunia, Wieland, Hohenfellner, Michel, Haferkamp, Müller, Stief, Bastian. Administrative, technical, or material support: May, Herrmann, Bolenz, Tiemann, Brookman-May, Fritsche, Burger, Buchner, Gratzke, Wülfing, Trojan, Ellinger, Tilki, Gilfrich, Höfner, Roigas, Zacharias, Gunia, Wieland, Hohenfellner, Michel, Haferkamp, Müller, Stief, Bastian. Supervision: May, Herrmann, Bolenz, Tiemann, Brookman-May, Fritsche, Burger, Buchner, Gratzke, Wülfing, Trojan, Ellinger, Tilki, Gilfrich, Höfner, Roigas, Zacharias, Gunia, Wieland, Hohenfellner, Michel, Haferkamp, Müller, Stief, Bastian. 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] Fang AC, Ahmad AE, Whitson JM, Ferrell LD, Carroll PR, Konety BR. Effect of a minimum lymph node policy in radical cystectomy and pelvic lymphadenectomy on lymph node yields, lymph node positivity rates, lymph node density, and survivorship in patients with bladder cancer. Cancer 2010;116: [2] Quek ML, Flanigan RC. The role of lymph node density in bladder cancer prognostication. World J Urol 2009;27: [3] Sharir S, Fleshner NE. Lymph node assessment and lymphadenectomy in bladder cancer. J Surg Oncol 2009;99: [4] Buscarini M, Josephson DY, Stein JP. Lymphadenectomy in bladder cancer: a review. Urol Int 2007;79: [5] Leissner J. Lymphadenectomy for bladder cancer. Diagnostic and prognostic significance as well as therapeutic benefit. Urologe A 2005;44: [6] Stein JP, Cai J, Groshen S, Skinner DG. Risk factors for patients with pelvic lymph node metastases following radical cystectomy with en bloc pelvic lymphadenectomy: concept of lymph node density. J Urol 2003;170: [7] Herr HW. Superiority of ratio based lymph node staging for bladder cancer. J Urol 2003;69: [8] Wright JL, Lin DW, Porter MP. The association between extent of lymphadenectomy and survival among patients with lymph node metastasesundergoing radical cystectomy. Cancer 2008;112:

7 718 EUROPEAN UROLOGY 59 (2011) [9] Bruins HM, Huang GJ, Cai J, Skinner DG, Stein JP, Penson DF. Clinical outcomes and recurrence predictors of lymph node positive urothelial cancer after cystectomy. J Urol 2009;182: [10] Wiesner C, Salzer A, Thomas C, et al. Cancer-specific survival after radical cystectomy and standardized extended lymphadenectomy for node-positive bladder cancer: prediction by lymph node positivity and density. BJU Int 2009;104: [11] Osawa T, Abe T, Shinohara N, et al. Role of lymph node density in predicting survival of patients with lymph node metastases after radical cystectomy: a multi-institutional study. Int J Urol 2009;16: [12] Kassouf W, Agarwal PK, Herr HW, et al. Lymph node density is superior to TNM nodal status in predicting disease-specific survival after radical cystectomy for bladder cancer: analysis of pooled data from MDACC and MSKCC. J Clin Oncol 2008;26: [13] Steven K, Poulsen AL. Radical cystectomy and extended pelvic lymphadenectomy: survival of patients with lymph node metastasis above the bifurcation of the common iliac vessels treated with surgery only. J Urol 2007;178: [14] Fleischmann A, Thalmann GN, Markwalder R, Studer UE. Extracapsular extension of pelvic lymph node metastases from urothelial carcinoma of the bladder is an independent prognostic factor. J Clin Oncol 2005;23: [15] Wittekind C, Tischoff I. Tumor classifications. Pathologe 2004;25: [16] Boulesteix AL. Maximally selected chi-square statistics and binary splits of nominal variables. Biom J 2006;48: [17] 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: [18] Abdel-Latif M, Abol-Enein H, El-Baz M, Ghoneim MA. Nodal involvement in bladder cancer cases treated with radical cystectomy: incidence and prognosis. J Urol 2004;172:85 9. [19] Kassouf W, Leibovici D, Munsell MF, Dinney CP, Grossman HB, Kamat AM. Evaluation of the relevance of lymph node density in a contemporary series of patients undergoing radical cystectomy. J Urol 2006;176:53 7. [20] Herr HW. Extent of pelvic lymph node dissection during radical cystectomy: where and why! Eur Urol 2010;57: [21] Karl A, Carroll PR, Gschwend JE, et al. The impact of lymphadenectomy and lymph node metastasis on the outcomes of radical cystectomy for bladder cancer. Eur Urol 2009;55: