Radical Nephrectomy with or without Lymph Node Dissection for High Risk Nonmetastatic Renal Cell Carcinoma: A Multi-Institutional Analysis

Adult Urology Oncology: Adrenal/Renal/Upper Tract/Bladder Radical Nephrectomy with or without Lymph Node Dissection for High Risk Nonmetastatic Renal Cell Carcinoma: A Multi-Institutional Analysis Boris Gershman,* R. Houston Thompson, Stephen A. Boorjian, Alessandro Larcher, Umberto Capitanio, Francesco Montorsi, Cristina Carenzi, Roberto Bertini, Alberto Briganti, Christine M. Lohse, John C. Cheville and Bradley C. Leibovich From the Division of Urology, Rhode Island Hospital and The Miriam Hospital, Providence (BG), Rhode Island, Departments of Urology (RHT, SAB, BCL), Health Sciences Research (CML) and Pathology (JCC), Mayo Clinic, Rochester, Minnesota, and Department of Urology, San Raffaele Scientific Institute (AL, UC, FM, CC, RB, AB), Milan, Italy

Purpose: Lymph node dissection may benefit patients at increased risk for lymph node metastases from renal cell carcinoma. Therefore, we evaluated the association of lymph node dissection with survival in patients at high risk undergoing radical nephrectomy for renal cell carcinoma. Materials and Methods: We identified 2,722 patients with M0 renal cell carcinoma who underwent radical nephrectomy with or without lymph node dissection at 2 international centers from 1990 to 2010. The associations of lymph node dissection with the development of distant metastases, and cancer specific and all cause mortality were evaluated using propensity score techniques and traditional multivariable Cox regression. Subset analyses were done to examine patients at increased risk of lymph node metastases. Results: Overall 171 patients (6.3%) had pN1 disease. Median followup was 9.6 years. Clinicopathological features were well balanced after propensity score adjustment. Lymph node dissection was not significantly associated with a reduced risk of distant metastases, or cancer specific or all cause mortality in the overall cohort, among patients with preoperative radiographic lymphadenopathy (cN1), or across an increasing probability of pN1 disease from 0.10 or greater to 0.50 or greater. Neither extended lymph node dissection nor the extent of lymph node dissection was associated with improved oncologic outcomes. Conclusions: The current analysis of a large, international cohort indicates that lymph node dissection is not associated with improved oncologic outcomes among patients at high risk who undergo radical nephrectomy for M0 renal cell carcinoma. This includes patients with radiographic lymphadenopathy (cN1) and across increasing probability thresholds of pN1 disease. Key Words: lymph node excision; nephrectomy; carcinoma, renal cell; survival; propensity score

THE role of LND in the management of RCC has been controversial with conflicting data on its oncologic benefit. While older retrospective studies provided initial support for an oncologic

benefit,1e4 more recent investigations have not demonstrated improved survival.5e7 Still, a small subset of patients with isolated LN involvement has achieved long-term survival

0022-5347/18/1995-1143/0 THE JOURNAL OF UROLOGY® Ó 2018 by AMERICAN UROLOGICAL ASSOCIATION EDUCATION AND RESEARCH, INC.

https://doi.org/10.1016/j.juro.2017.11.114 Vol. 199, 1143-1148, May 2018 Printed in U.S.A.

Abbreviations and Acronyms ACM ¼ all cause mortality CSM ¼ cancer specific mortality CT/MRI ¼ computerized tomography/magnetic resonance imaging ECOG ¼ Eastern Cooperative Oncology Group IPW ¼ inverse probability weight LN ¼ lymph node LND ¼ lymph node dissection PS ¼ propensity score RCC ¼ renal cell carcinoma RN ¼ radical nephrectomy Accepted for publication November 30, 2017. No direct or indirect commercial incentive associated with publishing this article. The corresponding author certifies that, when applicable, a statement(s) has been included in the manuscript documenting institutional review board, ethics committee or ethical review board study approval; principles of Helsinki Declaration were followed in lieu of formal ethics committee approval; institutional animal care and use committee approval; all human subjects provided written informed consent with guarantees of confidentiality; IRB approved protocol number; animal approved project number. * Correspondence: Rhode Island Hospital and The Miriam Hospital, Warren Alpert Medical School of Brown University, 195 Collyer St., Suite 201, Providence, Rhode Island 02904 (telephone: 401-272-7799; FAX: 401-453-9078; e-mail: Boris. [email protected]).

www.jurology.com

j

1143

1144

NEPHRECTOMY WITH OR WITHOUT LYMPH NODE DISSECTION FOR RENAL CELL CANCER

following resection of LN metastases.8e10 Proponents of LND have suggested that it may benefit patients at increased risk for LN metastases.11,12 Given the rarity of LN metastases, prior studies may have been underpowered to detect a potential benefit in such a small subset of patients at high risk. In a recent study of patients with nonmetastatic RCC we found no association of LND with oncologic outcomes regardless of the risk of nodal involvement.7 However, that study was limited by a relatively small number of patients at high risk and single institution experience. In the current study we used a large multi-institutional cohort to evaluate the association of LND with survival in patients undergoing RN for RCC with a focus on patients at high risk.

PATIENTS AND METHODS Patient Population After obtaining institutional review board approval at each study site we identified a total of 2,722 patients who underwent RN for sporadic, unilateral, M0 RCC between 1990 and 2010 at Mayo Clinic and San Raffaele Scientific Institute. Of these patients 1,215 (45%) underwent concomitant LND. The decision to perform LND was based on surgeon discretion. A standardized template was not used.

Clinicopathological Features Recorded clinicopathological features included study site, surgery year, patient age at surgery, gender, symptoms at presentation, constitutional symptoms at presentation, ECOG performance status, Charlson comorbidity score, surgical approach, pathological tumor size, RCC histological subtype, TNM stage according to the 2010 AJCC (American Joint Committee on Cancer) classification, tumor grade according to the WHO/ISUP (International Society of Urological Pathology) classification and the presence of coagulative tumor necrosis or sarcomatoid differentiation. Radiographic features were abstracted from the medical records, including radiographic lymphadenopathy (cN1), renal vein involvement on CT/MRI, inferior vena cava tumor thrombus on CT/ MRI and radiographic evidence of extrarenal extension, adrenal involvement, or cystic or indeterminate cysts.

Statistical Methods Clinicopathological and radiographic features were summarized with the median and IQR or the frequency count and percent. They were separately compared by study site and receipt of LND using the Wilcoxon rank sum and chisquare tests. A PS for LND was obtained using a logistic regression model with LND as the outcome and clinicopathological features (supplementary table 1, http:// jurology.com/).13 Of the 2,722 patients in the combined cohort 2,657 had nonmissing data on all covariates studied, of whom 1,183 (45%) underwent LND. We excluded 150 patients from PS analyses because they had a PS that did not fall within the common range. The final PS

analysis cohort comprised 2,507 patients, of whom 1,063 (42%) underwent LND. Of these patients 103 with a missing cause of death were excluded from analyses of CSM and 104 with missing information on postoperative distant metastases were excluded from analyses of distant metastases. Associations of LND with time to distant metastases, CSM and ACM were evaluated by Cox proportional hazards regression models and summarized as the HR and 95% CI. Two PS techniques were used in these models, including adjustment for PS quintiles and reweighting by stabilized IPWs.13,14 Stabilized IPWs were truncated by setting weights below the first percentile to the value of the first percentile and weights above the 99th percentile to the value of the 99th percentile. As a sensitivity analysis we evaluated the associations of LND with oncologic outcomes using traditional multivariable Cox models including all covariates studied for the 2,657 patients with nonmissing data. To evaluate the associations of LND with oncologic outcomes among patients at increased risk for pN1 disease we constructed Cox models including an indicator for LND, an indicator for risk of pN1 disease and the interaction term between the 2 indicators. We specifically examined patients with preoperative radiographic lymphadenopathy (cN1) as well as patients with a predicted 0.10 to 0.50 probability of pN1 disease. The predicted probability of pN1 disease was estimated using a multivariable logistic regression model (supplementary methods, http://jurology. com/). Due to the small subset size there was residual imbalance in several characteristics among patients with a 0.50 or greater predicted probability of pN1 disease. Therefore, we further adjusted these analyses for those covariates with residual imbalance. As a secondary analysis we examined the association of extended LND, defined as the removal of 13 or more lymph nodes,15 with oncologic outcomes using multivariable regression adjusted for all study covariates. We also evaluated the association of LND extent (ie the number of LNs removed as a continuous variable) among patients who underwent LND with oncologic outcomes using multivariable regression adjusted for all study covariates. Statistical analyses were performed with SASÒ, version 9.4 and R, version 3.1.1 (https://www.r-project.org/). All tests were 2-sided with p <0.05 considered statistically significant.

RESULTS Of the 2,722 patients included in the overall cohort 1,215 (45%) underwent LND, 237 (9%) had cN1 disease and 171 (6%) had pN1 disease. A median of 6 LNs (IQR 3e11) were removed. Supplementary table 1 (http://jurology.com/) summarizes clinicopathological and radiographic features stratified by study site. There were statistically significant differences in most patient characteristics according to study site, reflecting differences in patient cohorts and clinical practice. Supplementary table 2 (http://jurology.com/) shows clinicopathological and radiographic

NEPHRECTOMY WITH OR WITHOUT LYMPH NODE DISSECTION FOR RENAL CELL CANCER

1145

Table 1. Clinicopathological features stratified by lymph node dissection in 2,437 patients in propensity score cohort after stabilized IPW reweighting No LND No. pts No. study site (%): Milan Rochester No. surgery yr (%): 1990e1995 1996e2000 2001e2005 2006e2010 Median age at surgery (IQR) 18e54 55e64 65e72 73 or Greater No. female (%): No. male (%): Median Charlson score (IQR) No. symptom (%) No. constitutional symptom (%) No. ECOG performance status (%): 0 1 2 3 4 No. cN1 (%) No. involvement on CT/MRI (%): Renal vein Inferior vena cava No. radiographic evidence (%): Extrarenal extension Adrenal involvement Cystic or indeterminate cysts No. radical nephrectomy type (%): Open Laparoscopic No. RCC histological subtype (%): Clear cell Papillary Chromophobe Collecting duct Not otherwise specified Mucinous tubular and spindle cell Translocation-associated Clear cell papillary Tumor size No. 2010 pT stage (%): pT1a pT1b pT2a pT2b pT3a pT3b pT3c pT4 No. grade (%): 1 2 3 4 No. Coagulative tumor necrosis (%) No. Sarcomatoid differentiation (%)

LND

1,398

p Value

1,039 0.37

456 942

(33) (67)

357 682

(34) (66)

404 331 376 287

(29) (24) (27) (21)

282 243 288 226

(27) (23) (28) (22)

349 355 372 322 460 938 1 721 231

(25) (25) (27) (23) (33) (67) (0e2) (52) (17)

267 257 278 237 349 690 1 544 177

(26) (25) (27) (23) (34) (66) (0e2) (52) (17)

944 333 102 17 2 47

(68) (24) (7) (1) (less than 1) (3)

688 261 72 18 0 47

(66) (25) (7) (2) (4)

0.15

146 97

(10) (7)

127 80

(12) (8)

0.17 0.48

49 5 179

(3) (less than 1) (13)

41 6 125

(4) (1) (12)

0.55 0.57 0.57 0.65

1,241 157

(89) (11)

929 110

(89) (11)

0.27

0.81

0.73 0.60 0.69 0.75 0.55

0.54 1,130 (81) 162 (12) 83 (6) 3 (less than 1) 15 (1) 1 (less than 1)

845 (81) 112 (11) 66 (6) 3 (less than 1) 7 (1) 0

1 (less than 1) 3 (less than 1) 6.3 (4.3e9.0)

0 6 (1) 6.5 (4.5e9.0)

299 401 182 96 321 64 23 12

(21) (29) (13) (7) (23) (5) (2) (1)

195 301 136 71 253 54 18 11

(19) (29) (13) (7) (24) (5) (2) (1)

104 668 531 95 473 34

(7) (48) (38) (7) (34) (2)

70 471 413 85 372 34

(7) (45) (40) (8) (36) (3)

0.15 0.12

0.09

features stratified by the performance of LND. Although patients who underwent LND had more advanced pT stage and more aggressive tumor characteristics, clinicopathological and

0.31 0.26

radiographic features were well balanced after PS adjustment. There were no statistically significant differences after reweighting the combined cohort by IPWs (table 1).

1146

NEPHRECTOMY WITH OR WITHOUT LYMPH NODE DISSECTION FOR RENAL CELL CANCER

Median followup among survivors was 9.6 years (IQR 5.6e14.3). During this time distant metastases developed in 787 patients and 1,397 died, including 622 of RCC. We first examined the associations of LND with oncologic outcomes in the overall cohort (table 2 and supplementary material, http://jurology.com/). LND was not significantly associated with CSM or ACM using PS techniques or traditional multivariable regression. LND was associated with a modestly increased risk of distant metastases using IPWs (HR 1.17, 95% CI 1.00e1.36, p ¼ 0.046) and traditional regression (HR 1.20, 95% CI 1.01e1.42, p ¼ 0.040) which just reached statistical significance. Next we examined the associations of LND with oncologic outcomes among patients at increased risk for pN1 disease (table 3). LND was not associated with the development of distant metastases, CSM or ACM in patients with preoperative radiographic lymphadenopathy (cN1). No consistent association of LND with improved oncologic outcomes was noted across increasing probability thresholds of pN1 disease from 0.10 or greater to 0.50 or greater. Specifically among patients with a threshold predicted probability of pN1 disease ranging from 0.10 or greater to 0.40 or greater LND was not significantly associated with improved oncologic outcomes using PS techniques or traditional multivariable regression. Among patients with a 0.50 or greater predicted probability of pN1 disease there was evidence of improved oncologic outcomes with LND using PS techniques. However, given the small size of this subset, there was residual imbalance in several covariates. After further adjustment for residually imbalanced covariates there was no consistent association of LND with oncologic outcomes. Similarly using traditional multivariable regression LND was not significantly associated with improved oncologic outcomes.

Table 2. LND associations with oncologic outcomes in overall cohort adjusted by propensity score quintiles, stabilized IPW reweighting and multivariable adjustment for all covariates PS Technique Distant metastasis: PS quintiles IPWs Multivariable Ca specific mortality: PS quintiles IPWs Multivariable All cause mortality: PS quintiles IPWs Multivariable

LND vs no LND HR (95% CI)

p Value

1.09 (0.91e1.29) 1.17 (1.00e1.36) 1.20 (1.01e1.42)

0.35 0.046 0.040

1.01 (0.83e1.22) 1.09 (0.92e1.30) 1.11 (0.92e1.34)

0.94 0.33 0.29

1.01 (0.88e1.15) 1.02 (0.91e1.14) 1.11 (0.97e1.26)

0.91 0.80 0.12

As a secondary analysis we assessed 241 patients (9%) who underwent extended LND, defined as the removal of 13 or more lymph nodes. After multivariable adjustment extended LND was not significantly associated with the development of distant metastases (HR 1.08, 95% CI 0.86e1.35, p ¼ 0.50), CSM (HR 1.07, 95% CI 0.83e1.39, p ¼ 0.60) or ACM (HR 1.11, 95% CI 0.90e1.37, p ¼ 0.35). In a separate secondary analysis we evaluated the association of LND extent with oncologic outcomes. Among patients who underwent LND the number of LNs removed as a continuous variable was not significantly associated with the development of distant metastases (HR 1.07, 95% CI 0.95e1.22, p ¼ 0.26), CSM (HR 0.99, 95% CI 0.85e1.15, p ¼ 0.90) or ACM (HR 1.04, 95% CI 0.92e1.18, p ¼ 0.54) after multivariable adjustment.

DISCUSSION In this study of a large international cohort we observed no survival benefit to LND in patients with nonmetastatic RCC. In particular even among patients at high risk, including those with preoperative radiographic lymphadenopathy, or across increasing disease risk (ie the threshold probability of pN1 disease) LND was not associated with oncologic outcomes. Similarly neither extended LND nor the extent of LND was associated with improved survival. These results reinforce prior retrospective studies that did not demonstrate a survival benefit for LND5,7,16 as well as the only randomized trial, EORTC 30881, to examine this clinical question.6 The greatest limitation of prior studies, including EORTC 30881,6 was a low incidence of LN involvement. Since these are the patients who in principle may benefit from LND, the proponents of LND have contended that LND may benefit this small patient subset.11,12 In support of this hypothesis studies in patients with surgically resected pN1 M0 disease have consistently identified a subset with longterm survival.8e10,17 Given the rarity of LN involvement in nonmetastatic RCC, it would be difficult to accrue a randomized trial to evaluate this question and single institutional series may likewise be underpowered. Indeed, the impetus for the current study arose from our recent examination of LND in nonmetastatic RCC, in which we found no benefit to LND even in high risk patient groups.7 Like other retrospective investigations,5 that study was limited by a relatively few patients at high risk. To this end the current collaboration provides the largest international cohort in which to evaluate the oncologic efficacy of LND in patients with high risk, nonmetastatic RCC.

NEPHRECTOMY WITH OR WITHOUT LYMPH NODE DISSECTION FOR RENAL CELL CANCER

1147

Table 3. Lymph node dissection associations with oncologic outcomes in patients at increased risk for pN1 disease adjusted by propensity score quintiles, stabilized IPW reweighting and multivariable adjustment of all covariates Distant Metastasis PS Technique

LND vs no LND HR (95% CI)

Ca Specific Mortality p Value

LND vs no LND HR (95% CI)

All Cause Mortality p Value

LND vs no LND HR (95% CI)

p Value

PS quintiles cN1 pN1 probability: 0.10 or Greater 0.20 or Greater 0.30 or Greater 0.40 or Greater 0.50 or Greater 0.50 or Greater*

0.62 (0.32e1.19)

0.15

0.94 (0.42e2.10)

0.88

1.12 (0.59e2.14)

0.74

1.19 1.25 1.34 0.91 0.43 0.57

(0.94e1.50) (0.87e1.80) (0.84e2.13) (0.50e1.65) (0.19e0.94) (0.26e1.25)

0.15 0.22 0.22 0.75 0.03 0.16

1.05 1.22 1.59 1.12 0.72 0.95

(0.82e1.35) (0.84e1.79) (0.95e2.66) (0.57e2.20) (0.30e1.72) (0.39e2.28)

0.68 0.30 0.08 0.75 0.46 0.90

1.30 1.50 1.67 1.08 0.73 0.83

(1.06e1.59) (1.10e2.05) (1.12e2.48) (0.61e1.89) (0.35e1.55) (0.39e1.76)

0.01 0.01 0.01 0.80 0.42 0.62

0.59 (0.32e1.12)

0.11

1.01 (0.47e2.20)

0.97

1.09 (0.59e1.99)

0.78

1.21 1.28 1.13 0.79 0.43 0.62

(0.98e1.49) (0.91e1.80) (0.73e1.77) (0.45e1.41) (0.19e0.95) (0.28e1.41)

0.08 0.15 0.58 0.43 0.04 0.25

1.05 1.27 1.41 1.06 0.54 0.76

0.65 0.19 0.19 0.86 0.15 0.55

1.12 1.23 1.21 0.81 0.32 0.37

0.22 0.18 0.36 0.43 0.003 0.01

0.97 (0.53e1.77)

0.92

1.29 (0.61e2.74)

0.50

1.37 (0.75e2.50)

0.31

1.22 1.42 1.29 1.33 1.04

0.09 0.05 0.26 0.33 0.92

1.02 1.25 1.39 1.36 1.31

0.90 0.24 0.20 0.36 0.49

1.24 1.39 1.33 1.15 1.17

0.04 0.04 0.15 0.61 0.65

IPWs cN1 pN1 probability: 0.10 or Greater 0.20 or Greater 0.30 or Greater 0.40 or Greater 0.50 or Greater 0.50 or Greater*

(0.84e1.32) (0.89e1.83) (0.85e2.35) (0.56e2.00) (0.23e1.26) (0.32e1.83)

(0.93e1.36) (0.91e1.67) (0.81e1.80) (0.47e1.38) (0.15e0.67) (0.18e0.79)

Multivariable cN1 pN1 probability: 0.10 or Greater 0.20 or Greater 0.30 or Greater 0.40 or Greater 0.50 or Greater

(0.97e1.54) (1.00e2.04) (0.83e2.02) (0.75e2.36) (0.51e2.09)

(0.79e1.30) (0.86e1.83) (0.84e2.31) (0.71e2.60) (0.61e2.85)

(1.01e1.51) (1.02e1.89) (0.90e1.96) (0.68e1.94) (0.60e2.26)

* Evaluated after further adjustment for covariates with evidence of residual imbalance between treatment groups with significance level of 0.10 in patient subset with 0.50 or greater predicted probability, including surgery year, ECOG performance status and histological subtype.

Several potential mechanisms may explain the apparent lack of benefit to LND even in patients at high risk. Perhaps most importantly RCC is rarely associated with LN involvement in the absence of concurrent distant metastases.1,12,18,19 Indeed, clinical metastases develop in most patients with clinically isolated pN1 disease within 12 months of surgery.9 The largely hematogenous dissemination route of RCC may reflect tumor biology or anatomical considerations since direct lymphovenous communications bypassing the retroperitoneal LNs were described in several anatomical studies.20 Despite the absence of robust data to support an oncologic benefit to LND, it may still have an important staging role in RCC management. Lymph node metastases have been associated with poor prognosis in nonmetastatic and metastatic RCC.1,9,21,22 Accordingly LND may provide actionable staging information to select patients for more intensive surveillance protocols, consideration of adjuvant systemic therapy23 or enrollment in a clinical trial. There are several limitations to the current study. It is retrospective, and residual selection bias and unmeasured confounding remain a potential issue despite attempts to adjust for a nonrandomized design. Evidence of reverse cause in

several subset analyses may reflect these considerations. Moreover, the decision to perform LND and the boundaries of LND were not standardized, and they varied in and across institutions. A randomized trial of LND in high risk RCC would be ideal to overcome such limitations of observational study design but, given the rarity of pN1 M0 disease, it may be difficult to perform. In addition, we were unable to further categorize cN1 status according to the size of radiographic lymphadenopathy or the number of enlarged LNs. Furthermore, use of a multi-institutional cohort introduced heterogeneity in patient selection and clinical practice. Finally, it is still possible that there is a high risk subset which we did not examine in this study that may derive a survival benefit from LND.

CONCLUSIONS The current analysis of a large, international cohort indicates that LND is not associated with improved oncologic outcomes among patients who undergo RN for M0 RCC. This includes patients at increased risk for LN metastases such as those with radiographic lymphadenopathy (cN1) or across increasing probability thresholds for pN1 disease.

1148

NEPHRECTOMY WITH OR WITHOUT LYMPH NODE DISSECTION FOR RENAL CELL CANCER

REFERENCES 1. Pantuck AJ, Zisman A, Dorey F et al: Renal cell carcinoma with retroperitoneal lymph nodes: role of lymph node dissection. J Urol 2003; 169: 2076. 2. Vasselli JR, Yang JC, Linehan WM et al: Lack of retroperitoneal lymphadenopathy predicts survival of patients with metastatic renal cell carcinoma. J Urol 2001; 166: 68. 3. Herrlinger A, Schrott KM, Schott G et al: What are the benefits of extended dissection of the regional renal lymph nodes in the therapy of renal cell carcinoma. J Urol 1991; 146: 1224. 4. Schafhauser W, Ebert A, Brod J et al: Lymph node involvement in renal cell carcinoma and survival chance by systematic lymphadenectomy. Anticancer Res 1999; 19: 1573. 5. Feuerstein MA, Kent M, Bazzi WM et al: Analysis of lymph node dissection in patients with 7-cm renal tumors. World J Urol 2014; 32: 1531. 6. Blom JH, van Poppel H, Marechal JM et al: Radical nephrectomy with and without lymphnode dissection: final results of European Organization for Research and Treatment of Cancer (EORTC) randomized phase 3 trial 30881. Eur Urol 2009; 55: 28. 7. Gershman B, Thompson RH, Moreira DM et al: Radical nephrectomy with or without lymph node dissection for nonmetastatic renal cell carcinoma: a propensity score-based analysis. Eur Urol 2016; 71: 560. 8. Delacroix SE Jr, Chapin BF, Chen JJ et al: Can a durable disease-free survival be achieved with surgical resection in patients with pathological

node positive renal cell carcinoma? J Urol 2011; 186: 1236.

suspected adenopathy before or during surgery? BJU Int 2001; 88: 169.

9. Gershman B, Moreira DM, Thompson RH et al: Renal cell carcinoma with isolated lymph node involvement: long-term natural history and predictors of oncologic outcomes following surgical resection. Eur Urol 2017; 72: 300.

17. Canfield SE, Kamat AM, Sanchez-Ortiz RF et al: Renal cell carcinoma with nodal metastases in the absence of distant metastatic disease (clinical stage TxN1-2M0): the impact of aggressive surgical resection on patient outcome. J Urol 2006; 175: 864.

10. Karakiewicz PI, Trinh QD, Bhojani N et al: Renal cell carcinoma with nodal metastases in the absence of distant metastatic disease: prognostic indicators of disease-specific survival. Eur Urol 2007; 51: 1616. 11. Margulis V and Wood CG: The role of lymph node dissection in renal cell carcinoma: the pendulum swings back. Cancer J 2008; 14: 308. 12. Capitanio U, Becker F, Blute ML et al: Lymph node dissection in renal cell carcinoma. Eur Urol 2011; 60: 1212. 13. D’Agostino RB Jr: Propensity score methods for bias reduction in the comparison of a treatment to a non-randomized control group. Stat Med 1998; 17: 2265. 14. Xu S, Ross C, Raebel MA et al: Use of stabilized inverse propensity scores as weights to directly estimate relative risk and its confidence intervals. Value Health 2010; 13: 273.

18. Giuliani L, Giberti C, Martorana G et al: Radical extensive surgery for renal cell carcinoma: longterm results and prognostic factors. J Urol 1990; 143: 468. 19. Chandrasekar T, Klaassen Z, Goldberg H et al: Metastatic renal cell carcinoma: patterns and predictors of metastasesda contemporary population-based series. Urol Oncol 2017; 35: 661.e7. 20. Karmali RJ, Suami H, Wood CG et al: Lymphatic drainage in renal cell carcinoma: back to the basics. BJU Int 2014; 114: 806. 21. Pantuck AJ, Zisman A, Dorey F et al: Renal cell carcinoma with retroperitoneal lymph nodes. Impact on survival and benefits of immunotherapy. Cancer 2003; 97: 2995.

15. Terrone C, Guercio S, De Luca S et al: The number of lymph nodes examined and staging accuracy in renal cell carcinoma. BJU Int 2003; 91: 37.

22. Gershman B, Thompson RH, Moreira DM et al: Lymph node dissection is not associated with improved survival among patients undergoing cytoreductive nephrectomy for metastatic renal cell carcinoma: a propensity score based analysis. J Urol 2017; 197: 574.

16. Minervini A, Lilas L, Morelli G et al: Regional lymph node dissection in the treatment of renal cell carcinoma: is it useful in patients with no

23. Ravaud A, Motzer RJ, Pandha HS et al: Adjuvant sunitinib in high-risk renal-cell carcinoma after nephrectomy. N Engl J Med 2016; 375: 2246.