Overall survival in patients with metastatic renal cell carcinoma and clinical N1 disease undergoing cytoreductive nephrectomy and lymph node dissection

Urologic Oncology: Seminars and Original Investigations ] (2017) ∎∎∎–∎∎∎

Original article

Overall survival in patients with metastatic renal cell carcinoma and clinical N1 disease undergoing cytoreductive nephrectomy and lymph node dissection Izak Faiena, M.D.a,b,*, Amirali Salmasi, M.D.a,b, Andrew T. Lenis, M.D., M.S.a, Nicholas M. Donin, M.D.a,b, David C. Johnson, M.D., M.P.H.a,b, Kinan Bachour, B.S.a, Alexandra Drakaki, M.D. Ph.D.b,c, Arie S. Belldegrun, M.D.a,b, Allan J. Pantuck, M.D., M.S.a,b, Karim Chamie, M.D., M.S.H.S.a,b a

Department of Urology, David Geffen School of Medicine at University of California, Los Angeles, CA b Institute of Urologic Oncology, University of California, Los Angeles, CA c Department of Hematology and Oncology, David Geffen School of Medicine at University of California, Los Angeles, CA Received 27 June 2017; received in revised form 14 August 2017; accepted 9 October 2017

Abstract Background: Patients with metastatic renal cell carcinoma (mRCC) have limited treatment options. Cytoreductive nephrectomy (CN) in select patients has been associated with improved survival. We aim to assess the survival in patients with mRCC and cN1 disease who underwent CN with and without lymph node dissection (LND). Methods: Data were abstracted from the National Cancer Database for patients diagnosed with mRCC and cN1 from 2003 to 2014. Using propensity matching, we compared overall survival (OS) in patients who underwent a LND. Kaplan-Meier survival analysis and multivariable Cox proportional hazards modeling were used. We performed a logistic regression to assess predictors of LND. Results: We identified 1,780 patients in the matched cohort, of which 71% underwent a LND. Patients undergoing LND were younger (P ¼ 0.01) and had similar size tumors (5 cm; P ¼ 0.31). Increased LN yield was associated with LND at an academic center (odds ratio ¼ 1.91; 95% CI: 1.51–2.42; P o 0.01). LND was associated with worse OS on KM analysis (log rank; P ¼ 0.01). However, on multivariable analysis, we found no significant difference in OS (hazard ratio ¼ 1.10; 95% CI: 0.94–1.29; P ¼ 0.22). However, when adjusting for number of positive LN removed, an increase in LN yield was associated with improved OS (hazard ratio ¼ 0.97; 95% CI: 0.95–0.99; P ¼ 0.01). Conclusion: We demonstrate that patients with mRCC and cN1 disease undergoing LND did not have a survival benefit when compared with patients undergoing CN. However, lymph node yield showed an increase in survival when adjusting for the number of positive lymph nodes. Further research and validation of the ideal number of LN removed that may benefit patients is warranted. r 2017 Elsevier Inc. All rights reserved.

Keywords: Metastatic renal cell carcinoma; Cytoreductive nephrectomy; Lymph node dissection

1. Introduction The role of lymph node dissection (LND) in renal cell carcinoma (RCC), particularly metastatic RCC (mRCC), remains an uncertainty. It has been established that there is no role for LND in low-stage nonmetastatic disease [1], Corresponding author. Tel.: þ1-310-794-8248; fax: þ1-310-794-3513. E-mail address: [email protected] (I. Faiena). *

http://dx.doi.org/10.1016/j.urolonc.2017.10.009 1078-1439/r 2017 Elsevier Inc. All rights reserved.

although retrospective institutional series have demonstrated a limited role in locally advanced node-positive RCC [2]. Recent studies have demonstrated that appropriately selected patients with mRCC who undergo a cytoreductive nephrectomy (CN) have an improved overall survival (OS) compared with observation in the targeted therapy era [3]. The benefit of LND during CN for patients with clinically mRCC remains controversial given conflicting

2

I. Faiena et al. / Urologic Oncology: Seminars and Original Investigations ] (2017) ∎∎∎–∎∎∎

published data. In a study assessing patients who underwent CN and LND before planned IL-2 therapy, Pantuck et al. [2] demonstrated a statistically significant improvement in survival . Conversely, Gershman et al. [4] recently reported no marginal improvement from a LND in cancer-specific or all-cause mortality among patients undergoing CN. This was also the case in patients who were predicted to have a high likelihood of N1 disease. However, it remains unclear whether LND benefits patients who present with radiographic evidence of nodal disease (cN1). In our study, we aim to use a large national cohort to assess the incremental survival benefit, if any, in patients with cN1 disease undergoing CN and LND for mRCC. Fig. 1. CONSORT diagram of patient selection in NCDB.

2. Methods 2.1. Data source The National Cancer Database (NCDB) is a hospital registry-based database compiled from more than 1,500 Commission on Cancer-accredited centers, and is sponsored jointly by the American College of Surgeons and the American Cancer Society. The NCDB captures more than 70% of newly diagnosed cancers in the United States and represents more than 34 million historical records [5]. 2.2. Study population We identified patients with mRCC who underwent CN (International Classification of Disease for Oncology, Third Edition; primary site codes C64.9, and surgical procedure of primary site codes for partial and radical nephrectomy) between 2004 and 2013. Patients with missing stage or lymph node data were excluded. Patients with a second primary malignancy were also excluded (Fig. 1). Patients with only cN1 status were analyzed. 2.3. Study covariates Patient covariates included age (o55, 55–65, 66–75, and 475), sex, race (white, black, Hispanic, and other), insurance status (Medicare, Private, Medicaid, and other governmental support, none, unknown), residence location (urban, rural, and metropolitan), and percentage of individuals in the patient’s area of residency without a high-school diploma. Comorbidity was categorized using the CharlsonDeyo classification (CDCC). Hospital characteristics included facility location (Atlantic, New England, East Central, West Central, and West), and facility type (community center, academic center, and integrated cancer network). Tumor characteristics included histologic subtype (Supplementary Table 1), tumor grade (low [1 or 2] vs. high [3 or 4] based on NCDB grade variable), tumor size, clinical and pathologic T-stage (T1–4), clinical and pathologic N-stage (cN1, pNx, and pN0–1), and surgical margins

(positive, negative, and unknown) (AJCC TMN sixth and seventh editions). Surgical characteristics included year of surgery (before and after approval of targeted therapies) [3], concurrent metastectomy (defined as surgery to distant lymph nodes or distant sites), and performance of a lymphadenectomy. LND was defined using a combination of the variables for scope of LN surgery and number of nodes examined by the pathologist. Extended LND was defined as removal of 413 LNs [6]. Our primary outcome was OS in patients with cN1 disease who underwent CN with or without LND. Subset analysis included OS by lymph node yield, and extended LND. 2.4. Statistical analysis Categorical and continuous data are presented as number (frequency) and median (interquartile range), as appropriate. Chi-square test was used to compare categorical variables between groups. Student t-test and Wilcoxon rank-sum test were used to compare parametric and nonparametric continuous variables, respectively. To balance preoperative confounding variables in our study, we calculated propensity scores predicting likelihood of LND in cN1 disease and performed a kernel matching using inverse propensity weights between those who did and did not undergo LND. We used the percent standardized differences to assess covariate balance in baseline characteristics between groups. A standardized difference o10% was defined as a good balance [7]. We then conducted multivariable modeling using logistic regression analysis on the unmatched cohort to calculate the odds ratio (OR) of receipt of LND by preoperative covariates. To evaluate the effects of our covariates on OS, we computed Kaplan-Meier survival curves on the matched cohorts and compared them using the log-rank test. Using the multivariable Cox model for OS we adjusted for postoperative confounding variables (Table 4) as well as using unique Facility ID variable in the database to adjust for clustering among hospitals. Furthermore, missing data

I. Faiena et al. / Urologic Oncology: Seminars and Original Investigations ] (2017) ∎∎∎–∎∎∎

3

Table 1 Baseline characteristics of study cohort matched on preoperative variables No LND, n ¼ 514, (%)

Age, median (IQR) 62 (54–69) Gender Male 364 (71) Female 150 (29) Race White 399 (78) Black 34 (7) Hispanic/Other 45 (8) Unknown 36 (7) Income o$38,000 63 (12) $38,000–47,999 90 (18) $48,000–62,999 149 (18) 4$63,000 212 (41) Location Atlantic 196 (38) New England 36 (7) East Central 112 (22) West Central 93 (18) West 77 (15) Patient location Rural 18 (3) Metropolitan 419 (82) Urban 63 (12) Missing 14 (3) Insurance status Medicare 199 (39) Private 247 (48) Medicaid 33 (6) None/Unknown 35 (7) Facility type Community 256 (50) Academic 209 (40) Integrated network 49 (10) Population without high-school degree in patient area (%) 421% 81 (16) 13%–20.9% 116 (23) 7%–12.9% 140 (27) 177 (34) o7% Year 2004–2005 79 (15) 2006–2013 435 (85) CDCC 0 368 (72) 1 121 (24) 42 25 (4) Clinical stage cT1–2 209 (41) cT3a 122 (24) cT3b–c 63 (12) cT4 68 (13) cTx 51 (10) Missing 1 (0.2) IQR ¼ interquartile range. a SD ¼ |P1 - P2| / √(P1[1 – P1] þ P2[1 – P2] / 2).

LND, n ¼ 1,266, (%)

Standardized difference, %a Before matching

After matching

−21

−2.6

869 (69) 397 (31)

– 4.9

– 0.7

997 100 109 60

(78) (8) (9) (5)

– 5.0 −0.5 −10

– 0.7 1.7 1.0

171 229 366 500

(14) (18) (29) (40)

– 1.6 −0.5 −3.4

– −0.8 0.5 −0.3

414 65 327 244 216

(33) (5) (26) (19) (17)

– −7.8 9.6 3.1 5.8

– −1.1 −3.2 2.6 3.8

29 981 232 24

(2) (78) (18) (2)

– −10 17 −5.5

– −3.2 3.7 −0.4

403 692 89 82

(32) (55) (7) (6)

– 13 2.5 −1.3

– 0.5 0.5 −1.4

458 (36) 688 (54) 120 (10)

– 28 −0.8

– 1.3 −1.0

206 289 327 444

(16) (26) (26) (35)

– 0.3 −3.1 1.5

– −0.3 2.1 −1.8

139 (11) 1,127 (89)

– 13.5

– −1.8

950 (75) 253 (20) 63 (5)

– −8.5 0.6

– 1.5 −2.9

414 401 210 167 71 3

– 18 12 −0.0 −17 0.9

– 3.6 0.8 0.1 −1.8 2.4

60 (53–67)

(33) (32) (17) (13) (6) (0.2)

I. Faiena et al. / Urologic Oncology: Seminars and Original Investigations ] (2017) ∎∎∎–∎∎∎

4 Table 2 Postoperative characteristics

No LND, n ¼ 514 (%) Tumor size (cm), median (IQR) Histology Clear cell Papillary Sarcomatoid Other aggressive Missing Grade 1/2 3/4 Missing Pathologic stage pT1–2 pT3a pT3b-c pT4 pTx Pathologic node stage pN0 pN1 pNx Missing Lymph node yield Median (IQR) Number of LN positive Median (IQR) Surgical margins Negative Positive Missing Concurrent metastectomy No Yes Missing

5 (6–9) 349 (68) 31 (6) 58 (11) 42 (8) 34 (7)

LND, n ¼ 1,266 (%) 5 (6–9) 810 98 168 111 79

(64) (8) (13) (9) (6)

P value 0.31a 0.45

o0.01 77 (15) 378 (74) 59 (11) 41 (9) 263 (51) 78 (15) 59 (11) 73 (14) N/A N/A 514 (100) 0

104 (8) 1,066 (84) 96 (8) 101 656 255 212 42

(8) (52) (20) (17) (3)

o0.01

163 1,036 41 26

(13) (82) (3) (2)

o0.01

N/A

4 (1–8)

–

N/A

1 (1–3)

–

313 (61) 153 (30) 48 (9)

807 (64) 401 (32) 58 (5)

449 (87) 53 (11) 12 (2)

1.137 (90) 123 (10) 6 (0.5)

0.01

o0.01

IQR ¼ interquartile range. a Wilcoxon rank-sum test.

variables were created and included in all the models to account for possible bias in the missing data. To assess performance of the Cox model, we evaluated the tests for proportionality and they were not violated. Two-tailed P o 0.05 was considered statistically significant. The institutional review board at our institution exempted our study. Statistical analyses were performed with STATA statistical software version 14 (StataCorp, College Station, TX).

(60 vs. 62 y, P ¼ 0.01), more likely to have their surgery at an academic facility (P o 0.01), performed after 2006 (P o 0.01), but had similar tumor sizes (5 cm, P ¼ 0.31). In patients with cN1 disease who underwent LND, 82% had pN1 on pathologic staging (Table 2). Of the patients that received a LND, only 255 (20%) underwent an extended LND. There were 1,266 patients (71%) who underwent CN with LND, and 514 (29%) who did not undergo LND. Furthermore, most of the cohort did not undergo concurrent metastectomy (Table 2).

3. Results

3.2. Predictors of LND

3.1. Baseline characteristics

For every year increase in age at diagnosis, the odds of receiving LND decreased by 2% (OR ¼ 0.98; 95% CI: 0.97–0.99; P ¼ 0.01) (Table 3). Interestingly, there was no difference in the odds of receiving LND in patients with more significant comorbidities (CDCC 42) and in those those without (OR ¼ 1.01, 95% CI: 0.61–1.65; P ¼ 0.98).

The study cohort consisted of 1,780 patients with clinically node positive disease, and their baseline characteristics were matched based on preoperative variables (Table 1). Those undergoing CN and LND were younger

I. Faiena et al. / Urologic Oncology: Seminars and Original Investigations ] (2017) ∎∎∎–∎∎∎

respectively, when compared with those undergoing surgery in the Atlantic region.

Table 3 Logistic regression for predictors of LND OR Age 0.98 Sex Male Ref. Female 1.15 Race White Ref. Black 0.99 Hispanic/other 0.95 Unknown 0.67 Insurance Medicare Ref. Private 1.12 Medicaid 0.95 None/unknown 0.82 Facility Community Ref. Academic 1.91 Integrated network 1.39 Location Atlantic Ref. New England 0.70 East Central 1.41 West Central 1.30 West 1.41 Urban/rural Rural Ref. Metropolitan 1.23 Urban 1.86 Missing 1.01 Year 2004–2005 Ref. 2006–2013 1.68 Income o$38,000 Ref. $38,000–47,999 0.95 $48,000–62,999 0.96 4$63,000 0.87 Population without high-school degree 421% Ref. 13%–20.9% 0.99 7%–12.9% 1.03 o7% 1.20 CDCC 0 Ref. 1 0.87 ≥2 1.01 Clinical stage cT1–2 Ref. cT3a 1.78 cT3b–c 1.71 cT4 1.32 cTx 0.72

5

95% CI

P value

0.97

0.99

0.01

0.93

1.18

0.22

0.63 0.64 0.43

1.55 1.42 1.05

0.99 0.80 0.09

0.83 0.83 0.50

1.51 1.56 1.36

0.47 0.83 0.46

1.51 0.95

2.42 2.05

o0.01 0.09

0.44 1.06 0.96 1.02

1.13 1.88 1.79 1.97

0.14 0.02 0.09 0.04

0.64 0.94 0.39

2.36 3.69 2.61

0.53 0.08 0.98

1.223

2.31

o0.01

0.62 0.63 0.55 (%)

1.42 1.45 1.38

0.79 0.83 0.56

0.68 0.70 0.78

1.43 1.53 1.85

0.94 0.86 0.42

0.67 0.61

1.13 1.65

0.31 0.98

1.36 1.21 0.94 0.48

2.33 2.40 1.86 1.08

o0.01 o0.01 0.11 0.11

CN performed at an academic center was associated with a 91% increased odds of receiving a LND (OR ¼ 1.91, 95% CI: 1.51–2.42; P o 0.001). Location was also shown to be significant in predicting whether patients receive LND. In both the West and East Central, patients had a 41% increase in the odds of a LND (P ¼ 0.04) and (P ¼ 0.02),

3.3. Survival analysis In the overall cohort, median follow-up time was 17.3 months (interquartile range: 7–35 mo) and median OS was 11 months (interquartile range: 4.9–26.0). In patients who underwent LND (Fig. 2) median OS was 10.8 months compared with 12.6 months in patients who did not undergo LND (log rank; P o 0.01). The 1-, 3-, and 5-year OS for patients who underwent LND were 47%, 17%, and 9%, respectively. Patients who did not undergo LND had a 1-, 3, and 5-year OS of 51%, 23%, and 13%, respectively (log rank P o 0.001). In the multivariable Cox proportional hazards model, LND in cN1 disease was not associated with a difference in OS (hazard ratio ¼ 1.10; 95% CI: 0.94– 1.29; P ¼ 0.22). Lymph node yield, however, showed an increase in survival when adjusting for the number of positive lymph nodes (Table 4). In this model, for every 1 lymph node removed there was a 3% increase in survival (P ≤ 0.01). Given that the median LN yield for this dataset was only 4 nodes, we conducted a subset analysis on patients who received an extended LND vs. no LND and found no difference in survival (hazard ratio ¼ 0.92; 95% CI: 0.70– 1.19; P ¼ 0.50) (Table 4). When stratifying by pathologic node stage, patients with pN1 disease on pathologic specimen had worse OS compared with pN0 and pNx patients (median OS 10 mo for pN1 vs. 18 mo for pN0 vs. 12 mo for pNx; log-rank test P o 0.001) (Fig. 3). 4. Discussion Despite detection of renal masses at earlier stages of disease, approximately 30% of patients present with metastatic disease [8]. In this cohort of patients, CN has been shown to be associated with improved survival, however, the utility of concurrent LND remains unclear as patients with nodal disease on pathologic specimen have a worse prognosis regardless of other confounding factors [9]. In the cytokine era, there were a number of studies that demonstrated a survival benefit to LND, particularly in patients cN0M1 undergoing CN and LND [2,10,11]. Contemporary studies, however, failed to show a significant improvement in survival when performing a CN and LND in all patients [4]. We hypothesized that performing a LND in patients who present with nodal disease may confer a survival benefit given that the vast majority of patients who presented with nodal disease and had LND were confirmed to have nodal metastases on pathologic examination (86%), and removal of metastatic disease might confer a benefit to these patients. In testing this hypothesis, we note 2 principal findings. First, we report on a population-level analysis demonstrating that LND is not associated with OS in patients with mRCC who are cN1. Furthermore, when

I. Faiena et al. / Urologic Oncology: Seminars and Original Investigations ] (2017) ∎∎∎–∎∎∎

0.75

1.00

Overall Survival cN1 with eLND

0.50

Log-rank p=0.5

0.00

Percent Survival

0.75 0.25

0.50

Log-rank p=0.01

0.00

Percent Survival

1.00

Overall Survival cN1

0.25

6

0

10

20

30

40

50

60

70

80

90

100

0

10

Follow-up time (months) Number at risk cN1 no LND 514 284 cN1 with LND 1266 648

173 353

110 192

74 117

45 74

28 43

17 19

no LND

20

30

40

50

60

45 16

28 8

Follow-up time (months) 8 12

7 8

5 3

Number at risk No LND 514 Extended LND 255

284 133

LND

173 86

110 50

74 27

No LND

Extended LND

Fig. 2. Overall survival in patients with cN1 disease who underwent LND (A) and overall survival in patients who underwent extended LND vs. no LND.

HR

95% CI

1.10 0.92 0.97

0.94 0.70 0.95

P value

0.75

1.00

Overall Survival by pN

0.25

0.50

Log-rank p=<0.001

0.00

Table 4 Multivariable Cox proportional hazards model adjusting for postoperative covariates in the matched cohort

vs. 15 mo). Again, this highlights that nodal disease in the metastatic setting is a harbinger of more aggressive disease, and perhaps a LND in this setting does not have any therapeutic value [16]. The benefits of an extended LND are also debated. Prior studies have shown that removal of 13 nodes or more was the optimal number for adequate nodal staging as there was a higher rate of detecting node-positive disease in confined and locally advanced disease [6]. A recent study reported that performing an extended LND was not associated with improved cancer-specific mortality or OS in a cohort of N0 and N1 patients [4]. This further supports the role of LND as a prognostic marker rather than a therapeutic one. However, Whitson et al. showed that an increase in lymph node yield improved CSS in patients with positive lymph nodes in a nonmetastatic setting. They found that for every 10 lymph nodes removed in patients with a single positive lymph node, there was a 10% absolute increase in CSS at 5 years (P o 0.01) [17]. This is similar to our findings in mRCC in the cN1 setting, whereby adjusting for number of lymph nodes positive, for every lymph node removed there was a 3% improvement in OS. Another study suggested

Percent Survival

adjusting for lymph nodes positive, lymph node yield was found to be associated with improved survival. Our study results are in concordance with other recently published data showing no survival benefit to performing a LND during CN for mRCC [9,12,13]. In a population-based study using Surveillance, Epidemiology, and End Results registry data, Patel et al. report that LND was associated with worsened OS, but omitted variables (lack of information on comorbid conditions and systemic therapy) may have significantly biased their findings. The decrease in OS for patients who underwent LND was also attributed to a decrease in cancer-specific survival indicating that the presence of positive lymph nodes is ultimately an indicator of aggressive disease, thus contributing to worsened survival. The detriments in survival in those undergoing LND were no longer significant when 43 nodes were removed [14]. Trinh et al. [13] have also shown in a Surveillance, Epidemiology, and End Results cohort that increase in number LN positive portends poor prognosis in a similar cohort . However, there is a paucity of data on LND in patients with clinically node positive disease. In a study by Vasselli et al. [15], patients presenting with cN1 disease had shorter survival compared with patients without cN0 disease (median 9 vs. 15 mo) . In an effort to determine whether complete resection of nodal disease improved survival, they found that these patients had no difference in OS when compared with patients not undergoing a LND (median 9

0 a

LND vs. no LND Extended LND vs. no LNDa Lymph node yieldb

1.29 1.19 0.99

0.22 0.5 0.009

HR ¼ hazard ratio. Separate Cox model adjusting for histologic subtype, surgical margins, concurrent metastectomy, pathologic tumor stage, grade, and tumor size. b As above with the addition of pathologic nodal stage, and number of LN positive.

12

24

36

48

60

20 57 55 1

13 29 29 0

Follow-up time (months) Number at risk pN0 pN1 pNx Missing

163 1036 555 26

102 445 276 12

66 206 155 10

32 97 93 4

a

pN0

pN1

pNx

Missing

Fig. 3. Overall survival in patients with cN1 disease by pathologic nodal stage.

I. Faiena et al. / Urologic Oncology: Seminars and Original Investigations ] (2017) ∎∎∎–∎∎∎

extended LND may be beneficial for patients with advanced local disease (T3–T4, high Fuhrman grade, large tumor, sarcamotoid features, or coagulative tumor necrosis) [16]. However, it is unclear what other important characteristics of these patients are not observable that are contributing to survival, such as type and duration of systemic therapy. Most importantly, with emerging effective therapies, it is essential to define the role of LND. Perhaps template LND are more beneficial when done in a systematic way [18]. Therefore, it is essential to further evaluate this in a clinical trial setting. Our findings must be interpreted within the context of several known limitations. First, this is a retrospective analysis with inherent biases and confounders that we attempted to control using propensity score matched analysis and covariate adjustment. However, we could not address unmeasured confounders, which include absence of performance status. Second, the NCDB only contains data from Commission on Cancer-accredited hospitals and therefore our findings may not be generalizable to nonparticipating sites. Third, several variables known to be associated with OS were unavailable in this database, such as perioperative laboratory results, type of systemic therapy, and surgical and medical postoperative complications. Another possible limitation is that the survival benefit seen in patients with lymph node yield may be indicative of surgical quality, and not necessarily lymph nodes removed themselves. In other words, removing a large number of lymph nodes may be a surrogate for the overall quality of the surgical procedure. This is further exacerbated in this dataset as the median LN yield was low indicating a potentially inadequate LND. For this reason, we conducted a subset analysis in patients who received an extended LND and found no difference in survival. Another limitation is that it was unknown whether patients with higher lymph yield had more favorable metastatic characteristics that could explain a more aggressive LND. Finally, only OS could be evaluated as information regarding cause of death of progression is not available, precluding analysis of cancer-specific and progression outcomes. Despite these limitations, our study is significant because we are able to assess the role of LND among patients with mRCC and cN1 disease in a large database. This is pertinent, as we are able to account for comorbidities, lymph node yield, and lymph nodes positive, which added precision to our findings. Moreover, our study is strengthened by adjustment and matching using propensity score analyses that accounted for measured confounders in variables at patient presentation that may influence the decision to perform a LND. Lastly, although we show, as in other studies, that a LND in cN1 disease had similar OS compared with those undergoing CN alone, we noted an improvement in survival with increasing lymph node yield in patients with cN1 disease when adjusting for the number of positive lymph nodes in the specimen.

7

5. Conclusion In this study, we confirm that LND in patients with cN1 disease does not confer a survival benefit in patients with mRCC similarly in patients who underwent extended LND. However, we do find that increasing LN yield on pathologic specimen is associated with improved OS. Further validation of this finding, elucidation of the optimal number of LNs removed and its effects on survival is warranted. Acknowledgment A special thanks to Jeffery Gornbein, DrPH for his statistical support and guidance. Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j. urolonc.2017.10.009.

References [1] Blom JH, van Poppel H, Marechal JM, et al. Radical nephrectomy with and without lymph-node dissection: final results of European Organization for Research and Treatment of Cancer (EORTC) randomized phase 3 trial 30881. Eur Urol 2009;55:28–34. [2] Pantuck AJ, Zisman A, Dorey F, et al. Renal cell carcinoma with retroperitoneal lymph nodes: role of lymph node dissection. J Urol 2003;169:2076–83. [3] Hanna N, Sun M, Meyer CP, et al. Survival analyses of patients with metastatic renal cancer treated with targeted therapy with or without cytoreductive nephrectomy: a National Cancer Data Base study. J Clin Oncol 2016;34:3267–75. [4] 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;197:574–9. [5] Boffa DJ, Rosen JE, Mallin K, et al. Using the National Cancer Database for outcomes research: a review. J Am Med Assoc Oncol 2017, http://dx.doi.org/10.1001/jamaoncol.2016.6905. [6] 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;91:37–40. [7] Austin PC. An introduction to propensity score methods for reducing the effects of confounding in observational studies. Multivariate Behav Res 2011;46:399–424. [8] Lam JS, Leppert JT, Belldegrun AS, Figlin RA. Novel approaches in the therapy of metastatic renal cell carcinoma. World J Urol 2005;23:202–12. [9] Lughezzani G, Capitanio U, Jeldres C, et al. Prognostic significance of lymph node invasion in patients with metastatic renal cell carcinoma: a population-based perspective. Cancer 2009;115:5680–7. [10] Blute ML, Leibovich BC, Cheville JC, Lohse CM, Zincke H. A protocol for performing extended lymph node dissection using primary tumor pathological features for patients treated with radical nephrectomy for clear cell renal cell carcinoma. J Urol 2004;172: 465–9. [11] Margulis V, Wood CG. The role of lymph node dissection in renal cell carcinoma: the pendulum swings back. Cancer J 2008;14:308–14.

8

I. Faiena et al. / Urologic Oncology: Seminars and Original Investigations ] (2017) ∎∎∎–∎∎∎

[12] Feuerstein MA, Kent M, Bernstein M, Russo P. Lymph node dissection during cytoreductive nephrectomy: a retrospective analysis. Int J Urol 2014;21:874–9. [13] Trinh QD, Sukumar S, Schmitges J, et al. Effect of nodal metastases on cancer-specific mortality after cytoreductive nephrectomy. Ann Surg Oncol 2013;20:2096–102. [14] Patel HD, Gorin MA, Gupta N, et al. Mortality trends and the impact of lymphadenectomy on survival for renal cell carcinoma patients with distant metastasis. Can Urol Assoc J 2016;10:389–95. [15] Vasselli JR, Yang JC, Linehan WM, White DE, Rosenberg SA, Walther MM. Lack of retroperitoneal lymphadenopathy predicts

survival of patients with metastatic renal cell carcinoma. J Urol 2001;166:68–72. [16] Capitanio U, Becker F, Blute ML, et al. Lymph node dissection in renal cell carcinoma. Eur Urol. 2011;60:1212–20. [17] Whitson JM, Harris CR, Reese AC, Meng MV. Lymphadenectomy improves survival of patients with renal cell carcinoma and nodal metastases. J Urol 2011;185:1615–20. [18] Crispen PL, Breau RH, Allmer C, et al. Lymph node dissection at the time of radical nephrectomy for high-risk clear cell renal cell carcinoma: indications and recommendations for surgical templates. Eur Urol 2011;59:18–23.