Differential effect on survival of pelvic lymph node dissection at radical cystectomy for muscle invasive bladder cancer

Available online at www.sciencedirect.com

ScienceDirect EJSO 41 (2015) 353e360

www.ejso.com

Differential effect on survival of pelvic lymph node dissection at radical cystectomy for muscle invasive bladder cancer A. Larcher a,b,*, M. Sun a, J. Schiffmann a,c, Z. Tian a, S.F. Shariat d, M. McCormack e, F. Saad e, N. Fossati b, F. Abdollah f, A. Briganti b, N. Buffi b, M. Graefen c, G. Guazzoni b, F. Montorsi b, P.I. Karakiewicz a,e a

Cancer Prognostics and Health Outcomes Unit, University of Montreal Health Center, Montreal, Canada b Division of Oncology, Unit of Urology, URI, IRCCS Ospedale San Raffaele, Milan, Italy c Martini-Clinic, Prostate Cancer Center Hamburg-Eppendorf, Hamburg, Germany d Department of Urology, Medical University of Vienna, Vienna, Austria e Department of Urology, University of Montreal Health Center, Montreal, Canada f Vattikuti Urology Institute and VUI Center for Outcomes Research Analytics and Evaluation, Henry Ford Hospital, Detroit, MI, USA Accepted 24 October 2014 Available online 21 November 2014

Abstract Purpose: To compare long-term cancer outcomes after radical cystectomy (RC) alone or RC with pelvic lymph node dissection (PLND) according to different age and comorbidities categories. Methods: Using the SEER-Medicare dataset, 3314 patients diagnosed with urothelial carcinoma of the urinary bladder and treated with RC alone or RC with PLND were identified. After propensity score matching to reduce potential selection bias, all cause mortality (ACM)-free and cancer specific mortality (CSM)-free survival rates were estimated. Multivariable regression models (MVA) addressed the effect of PLND on ACM and CSM. Subgroups analyses according to age and comorbidities were performed. Results: After matching, 688 and 688 patients treated with RC alone or RC with PLND remained. The 5-year ACM-free survival rate was 36 after RC alone and 45% after RC with PLND ( p < 0001). In MVA, PLND exerted a protective effect on ACM (HR 0.77, p < 0.001). The 5-year CSM-free survival rate was 54 after RC alone and 65% after RC with PLND ( p < 0.001). In MVA, PLND exerted a protective effect on CSM (HR 0.71, p < 0.001). Similar results were observed in younger (age
75) and healthier (CCI ¼ 0) patients, where PLND exerted a protective effect on ACM (HR 0.64, p ¼ 0.001) and CSM (HR 0.65, p ¼ 0.01). Conversely, in older (age >75) and sicker (CCI 1) patients, PLND was not associated with ACM (HR 0.98, p ¼ 0.8) or CSM (HR 1.01, p ¼ 0.9). Conclusions: RC with PLND is associated with improved all cause and cancer specific survival in younger and healthier RC candidates but not in older and sicker patients. Ó 2014 Elsevier Ltd. All rights reserved.

Keywords: Radical cystectomy; Pelvic lymph node dissection; Bladder cancer; Long-term survival

Introduction

* Corresponding author. Cancer Prognostics and Health Outcomes Unit, 264 Blvd. Rene-Levesque E. Room 228, Montreal, QC H2X 1P1, Canada. Tel.: þ1 514 890 8000x35335; fax: þ1 514 227 5103. E-mail address: [email protected] (A. Larcher). http://dx.doi.org/10.1016/j.ejso.2014.10.061 0748-7983/Ó 2014 Elsevier Ltd. All rights reserved.

According to currently available guidelines, pelvic lymph node dissection (PLND) should be invariably performed at radical cystectomy1,2 (RC) and represents the criteria for comprehensive RC.3 However, there is a paucity of data comparing patients treated with radical cystectomy

354

A. Larcher et al. / EJSO 41 (2015) 353e360

plus PLND vs. radical cystectomy alone,4,5 and more specifically, data evaluating the potential therapeutic role of PLND according to patients age and comorbidity status are still lacking. To address this void, we examined the most contemporary version of the SEER-Medicare database with the intent of comparing long-term cancer outcomes associated with either RC with PLND or RC alone and quantifying the magnitude of the potential benefit of PLND. Our hypothesis stated RC with PLND might have a beneficial effect on urothelial carcinoma of urinary bladder (UCUB) long-term outcomes. Moreover, we hypothesized that such benefit might be consistent among different age and comorbidities status categories. Materials and methods Study source The current study relied on the 1991e2009 SEERMedicare linked database with follow-up updated until December 31, 2011. The SEER registries identify 28% of all cancer cases in the United States. Medicare insures approximately 97% of all Americans aged 65 years. Linkage to the SEER database is complete for approximately 93% of cases.6 Study population Overall 15,080 patients with a primary non-metastatic muscle-invasive (stage T2eT4) UCUB (International Classification of disease for Oncology [ICDeO] site code 67.0, histologic code 8120 or 8130), diagnosed between January 1991 and December 2009 were abstracted. Patient follow-up was available until December 31, 2011. Patients not enrolled in Medicare parts A or B for a minimum of 12 months prior to their first recorded diagnosis and for 6 months after diagnosis were not considered. Patients who had health maintenance organization enrollment in the year prior to diagnosis or for any period following diagnosis were also excluded. To ensure that all subjects had at least 1 year of claims from which comorbidities are derived, only those aged 66 years old were considered. Additional exclusions comprised of those with unknown race (n ¼ 36), and unknown marital status (n ¼ 432). Furthermore, patients treated with surgery 6 months after diagnosis were not considered in the current study, as treatment delay may confound the final results (n ¼ 1185).2 Moreover, patients with T4b or T4 not otherwise specified were omitted from our analyses (n ¼ 477).1 Final exclusions consisted of patients receiving neo-adjuvant chemotherapy or radiotherapy (n ¼ 442), patients who underwent partial cystectomy (n ¼ 566), patients who did not undergo a TURBT (n ¼ 243) and patients without available information

about PLND (n ¼ 603). For the purpose of the study we focused on patients who underwent RC. This resulted in 3314 assessable individuals with T2-T4a, N0, Nþ and Nx UCUB. Study design The study design was a retrospective caseecontrol study, cases were patients treated with RC and PLND and controls were patients treated with RC alone. Covariates Demographic covariates were age at diagnosis, comorbidities derived from the Klabundle’s Charlson comorbidity index (CCI) modification,7 gender, race (white, black, other), marital status (married, unmarried), socioeconomic status (SES; composite variable of income, education, and poverty levels8) and population density status (urban, rural). Cancer-related covariates comprised tumor grade and cancer stage. The latter was coded according to the AJCC staging system as tumor stage (T2, T3, T4). Lymphadenectomy-related covariates comprised nodal stage (N0, Nþ, Nx), number of nodes removed and number of positive nodes. Finally, treatment-related covariate was the administration of adjuvant chemotherapy. Specifically patients who had chemotherapy claims
6 months prior to cystectomy and a claim for RC
6 months before the first chemotherapy claim were considered to have been treated either with neo-adjuvant chemo- or radiotherapy, while other patients were considered to have been treated with adjuvant chemo- or radiotherapy. Outcomes The primary endpoint of the study was to compare all cause mortality (ACM) and cancer specific mortality (CSM) between patients treated with RC alone and patients treated with RC with PLND. The secondary endpoint of the study was to compare all cause mortality (ACM) and cancer specific mortality (CSM) between patients treated with RC alone and patients treated with RC with PLND according to age and comorbidity. Statistical analyses Means, medians and ranges were reported for continuous variables. Frequencies and proportions were reported for categorical variables. Statistical analyses consisted of three steps. First, due to inherent differences among patients included in the two treatment groups (RC alone vs. RC with PLND), adjustment was performed using a 1-to-1 propensity scoreematching ratio.9 Propensity scores were

A. Larcher et al. / EJSO 41 (2015) 353e360

computed by modeling a logistic regression with the dependent variable as the odds of receiving RC alone v. RC with PLND and the independent variable as age of diagnosis, CCI, gender, race, marital status, socioeconomic status, population density status, tumor stage, tumor grade, adjuvant chemotherapy administration and year of diagnosis. Lymphadenectomy-related covariates were not included in the model because lacking by definition in the cohort treated with RC alone. Subsequently, covariate balance between the matched groups was examined.10 Second, we compared the ACM- and CSM-free survival rates between patients treated with RC alone and those treated with RC with PLND. Univariable (UVA) and multivariable Cox regression models (MVA) were fitted to predict ACM and CSM (Cox). KaplaneMeier plots were used to depict graphically the ACM-free survival and the CSM-free survival rates after stratification according to the type of intervention received (RC alone vs. RC with PLND). Finally, all the above-mentioned analyses were repeated after dividing the study population into four sub-cohorts according to median age and median CCI score value: a. Patients aged
75 with CCI ¼ 0; b. Patients aged
75 with CCI 1; c. Patients aged >75 with CCI ¼ 0; d. Patients aged
75 with CCI 1. All statistical tests were performed using R software environment for statistical computing and graphics (Vienna, Austria, version 3.0.1). All tests were 2-sided with a significance level set at p < 0.05. Results Baseline characteristics Overall, 3314 patients were included in the study (Table 1a). Median age at diagnosis was 75 years. Fewer patients (688; 20%) were treated with RC alone while most (2626; 80%) were treated with RC with PLND. Patients treated with RC alone were older, more likely unmarried, more likely of high SES, more frequently diagnosed with T2 UCUB and more frequently diagnosed with low grade UCUB (all p < 0.05). Conversely, patients treated with RC alone less likely to receive adjuvant chemotherapy ( p < 0.001). In patients treated with RC with PLND (Table 1b), lymph node invasion was found in 26% patients, the median number of nodes removed was 9, and the median number of positive nodes was 2. Following propensity score matching for all the covariates except for PLND-related covariates, 688 (50%) and 688 (50%) patients treated with RC alone and RC with PLND remained. The mean standardized differences of patient characteristics between the two groups were <10%, indicating a high degree of similarity in the distribution of all the covariates in both populations. All subsequent analyses were based on the post-propensity matched cohort.

355

Overall population The 5-year ACM-free survival rate (Fig. 1a) was 36% after RC alone and 45% after RC with PLND (HR 0.75, p < 0001). In MVA, PLND exerted a protective effect on ACM after adjusting for all covariates (HR 0.77, p < 0.001; Table 2). The 5-year the CSM-free survival rate (Fig. 1b) was 54% after RC alone and 65% after RC with PLND (HR 0.71, p < 0.001). In MVA, PLND exerted a protective effect on CSM after adjusting for all covariates (HR 0.71, p < 0.001; Table 2). a. Patients aged
75 with CCI ¼ 0 In patients aged
75 with CCI ¼ 0 (n ¼ 394), the 5-year ACM-free survival rate (Fig. 2a) was 45% after RC alone and 60% after RC with PLND (HR 0.64, p ¼ 0.001). In MVA, PLND exerted a protective effect on ACM after adjusting for all the covariates (HR 0.66, p ¼ 0.002; Table 2). The 5-year CSM-free survival rate (Fig. 2e) was 60% after RC alone and 72% after RC with PLND (HR 0.65, p ¼ 0.01). In MVA, PLND exerted a protective effect on CSM after adjusting for all the covariates (HR 0.7, p ¼ 0.05; Table 2). b. Patients aged
75 with CCI 1 In patients aged
75 with CCI 1 (n ¼ 315), the 5-year ACM-free survival rate (Fig. 2b) was 32% after RC alone and 38% after RC with PLND (HR 0.82, p ¼ 0.1). In MVA, PLND exerted a protective effect on ACM after adjusting for all the covariates (HR 0.75, p ¼ 0.03; Table 2). The 5-year CSM-free survival rate (Fig. 2f) was 45% after RC alone and 62% after RC with PLND (HR 0.66, p ¼ 0.01). In MVA, PLND exerted a protective effect on CSM after adjusting for all the covariates (HR 0.62, p ¼ 0.08; Table 2). c. Patients aged >75 with CCI ¼ 0 In patients aged >75 with CCI ¼ 0 (n ¼ 345), the 5-year ACM-free survival rate (Fig. 2c) was 35% after RC alone and 44% after RC with PLND (HR 0.76, p ¼ 0.02). In MVA, PLND was not associated with ACM after adjusting for all the covariates (HR 0.8, p ¼ 0.07; Table 2). The 5-year CSM-free survival rate (Fig. 2g) was 57% after RC alone and 68% after RC with PLND (HR 0.63, p ¼ 0.01). In MVA, PLND exerted a protective effect on CSM after adjusting for all the covariates (HR 0.6, p ¼ 0.08; Table 2). d. Patients aged >75 with CCI 1 In patients aged >75 with CCI 1 (n ¼ 322), the 5-year ACM-free survival rate (Fig. 2d) was 30% after RC alone and 30% after RC with PLND (HR 0.99, p ¼ 0.9). In MVA, PLND was not associated with ACM after adjusting for all the covariates (HR 0.98, p ¼ 0.8; Table 2).

356

A. Larcher et al. / EJSO 41 (2015) 353e360

Table 1a Descriptive characteristics of 3314 patients with muscle-invasive bladder cancer treated with radical cystectomy without (n ¼ 688) or with pelvic lymph node dissection (n ¼ 2626), Surveillance, Epidemiology, and End Results (SEER) Medicare, 1991e2009. Variables

Age, years Mean (median) Range CCI 0 1 2 3 Gender Male Female Race White Black Other Marital status Married Unmarried Socioeconomic status High Low Residency status Urban Rural Tumor stage T2 T3 T4 Tumor grade Low High Adjuvant chemotherapy No Yes Year of diagnosis 1991e1997 1998e2003 2004e2009

Before propensity scored matching RC alone (n ¼ 688)

RC with PLND (n ¼ 2626)

75.6 (76) 66e95

74.9 (75) 66e95

260 102 159 167

(37.8) (14.8) (23.1) (24.3)

1054 (40.1) 387 (14.7) 537 (20.4) 648 (24.7)

479 (69.6) 209 (30.4)

1889 (71.9) 737 (28.1)

608 (88.4) 44 (6.4) 36 (5.2)

2373 (90.4) 122 (4.6) 131 (5)

415 (60.3) 273 (39.7)

1772 (67.5) 854 (32.5)

416 (60.5) 272 (39.5)

1302 (49.6) 1324 (50.4)

623 (90.6) 65 (9.4)

2395 (91.2) 231 (8.8)

434 (63.1) 124 (18) 130 (18.9)

1078 (41.1) 1088 (41.4) 460 (17.5)

70 (10.2) 618 (89.8)

118 (4.5) 2508 (95.5)

619 (90) 69 (10)

2141 (81.5) 485 (18.5)

249 (36.2) 336 (48.8) 103 (15)

381 (14.5) 789 (30) 1456 (55.4)

After propensity scored matching Std. Mean Diff. (%)

RC alone (n ¼ 688)

RC with PLND (n ¼ 688)

75.6 (76) 66e95

75.2 (75) 66e92

260 102 159 167

(37.8) (14.8) (23.1) (24.3)

280 (40.7) 97 (14.1) 137 (19.9) 174 (25.3)

479 (69.6) 209 (30.4)

496 (72.1) 192 (27.9)

608 (88.4) 44 (6.4) 36 (5.2)

606 (88.1) 40 (5.8) 42 (6.1)

415 (60.3) 273 (39.7)

434 (63.1) 254 (36.9)

416 (60.5) 272 (39.5)

406 (59) 282 (41)

623 (90.6) 65 (9.4)

622 (90.4) 66 (9.6)

434 (63.1) 124 (18) 130 (18.9)

434 (63.1) 122 (17.7) 132 (19.2)

70 (10.2) 618 (89.8)

60 (8.7) 628 (91.3)

619 (90) 69 (10)

632 (91.9) 56 (8.1)

249 (36.2) 336 (48.8) 103 (15)

233 (33.9) 351 (51) 104 (15.1)

12

Std. Mean Diff. (%) 7

3

3

5

5

2

5

14

5

22

1

2

3

61

1

18

4

28

6

4

37

Data presented as frequencies and percentages unless otherwise specified. Std. Mean Diff.: Standardized mean difference, after propensity score matching, a value <10% indicates high degree of similarity in the distribution of the variable between the two populations. RC: Radical cystectomy. PLND: Pelvic lymph node dissection. CCI: Charlson comorbidity index.

The 5-year CSM-free survival rate (Fig. 2g) was 56% after RC alone and 53% after RC with PLND (HR 1.02, p ¼ 0.9). In MVA, PLND was not associated with CSM after adjusting for all the covariates (HR 0.98, p ¼ 1.01; Table 2). Discussion We hypothesized that RC with PLND might have a beneficial effect on UCUB long-term outcomes and that such benefit might be consistent among different age and comorbidities status categories. To test these hypotheses, we relied on the most contemporary version of the SEER-Medicare database (1991e2009), which represents the largest North American repository of oncological data.

The results of the current study are several-fold. First, in general, patients treated with RC alone were older, more likely unmarried, more likely of high socioeconomic status, more frequently diagnosed with pT2 UCUB and more frequently diagnosed with low grade UCUB (all p < 0.05). Conversely, patients treated with RC alone were less likely to receive adjuvant chemotherapy ( p < 0.001). These differences justify the use of propensity-score matching for minimizing the potential confounding effect of baseline patient condition on cancer outcomes. Second, in the overall population, the ACM-free survival rate recorded after RC alone was significantly lower than the ACM-free survival rate recorded after RC with

A. Larcher et al. / EJSO 41 (2015) 353e360 Table 1b Characteristics of pelvic lymph node dissection in 2626 patients with muscle-invasive bladder cancer treated with radical cystectomy with pelvic lymph node dissection, Surveillance, Epidemiology, and End Results (SEER) Medicare, 1991e2009. Variables

Before propensity scored matching

After propensity scored matching

RC with PLND (n ¼ 2626)

RC with PLND (n ¼ 688)

Nodal stage N0 1929 (73.5) Nþ 697 (26.5) Number of nodes removed Mean (median) 12.3 (9) Range 1e90 Number of positive nodes Mean (median) 2.9 (2) Range 1e22

Table 2 Univariable and multivariable Cox regression analysis assessing the association between pelvic lymph node dissection and all cause mortality as well as cancer specific mortality in 1376 patients treated with radical cystectomy without (n ¼ 688) or with pelvic lymph node dissection (n ¼ 688), Surveillance, Epidemiology, and End Results (SEER) Medicare, 1991e2009. RC þ PLND vs RC alone e prediction of all cause mortality

11 (9) 1e76 2.6 (2) 1e12

Data presented as frequencies and percentages unless otherwise specified. RC: Radical cystectomy. PLND: Pelvic lymph node dissection.

PLND. Similarly, CSM-free survival rate recorded after RC alone was significantly lower than the CSM-free survival rate recorded after RC plus. These observations support the hypothesis that RC with PLND is beneficial with respect to ACM and CSM. Moreover, it emphasizes the importance of PLND at comprehensive RC. Third, the same benefit was observed when the analysis was restricted to younger and healthier patients. This observation implies that the observed benefit of PLND is well established in younger and healthier RC candidates. Fourth, in older and sicker patients (age >75 with CCI 1), the 5-year ACM-free survival rate recorded after RC alone did not differ from that recorded after RC with PLND. Similarly, the CSM-free survival rate recorded after

Multivariablea analysis

Cohort

Univariable analysis

Overall a. Age
75 and CCI ¼ 0 b. Age
75 and CCI 1 c. Age >75 and CCI ¼ 0 d. Age >75 and CCI 1

0.75 (0.67e0.85) <0.001 0.77 (0.69e0.88) <0.001 0.64 (0.5e0.83) 0.001 0.66 (0.51e0.86) 0.002

HR (95% CI) 554 (80.5) 134 (19.5)

357

p value HR (95% CI)

p value

0.82 (0.64e1.05)

0.1

0.75 (0.58e0.97)

0.03

0.76 (0.6e0.97)

0.03

0.8 (0.63e1.02)

0.07

0.99 (0.78e1.26)

0.9

0.98 (0.76e1.25)

0.8

RC þ PLND vs RC alone e prediction of cancer specific mortality Multivariablea analysis

Cohort

Univariable analysis

Overall a. Age
75 and CCI ¼ 0 b. Age
75 and CCI 1 c. Age >75 and CCI ¼ 0 d. Age >75 and CCI 1

0.71 (0.6e0.84) <0.001 0.71 (0.6e0.85) 0.65 (0.46e0.91) 0.01 0.7 (0.49e1)

HR (95% CI)

p value HR (95% CI)

p value <0.001 0.05

0.66 (0.47e0.93)

0.01

0.62 (0.44e0.89)

0.008

0.63 (0.43e0.91)

0.01

0.6 (0.41e0.88)

0.008

1.02 (0.72e1.44)

0.9

1.01 (0.71e1.43)

0.9

RC: Radical cystectomy. PLND: Pelvic lymph node dissection. HR: Hazard Ratio. CI: 95% Confidence Interval. CCI: Charlson comorbidity index. a Model adjusted for age, CCI, gender, race, marital status, socioeconomic status, residency status, tumor stage, tumor grade, adjuvant chemotherapy and year of diagnosis.

Figure 1. Panel a e KaplaneMeier plot depicting all cause mortality-free survival in the matched population stratified according to type of surgery: radical cystectomy alone vs. radical cystectomy with pelvic lymph node dissection. Panel b e KaplaneMeier plot depicting cancer specific mortality-free survival in the matched population stratified according to type of surgery: radical cystectomy alone vs. radical cystectomy with pelvic lymph node dissection. Time: Months after surgery. C.E.: Cumulative events. N.R.: Number of patients at risk.

358

A. Larcher et al. / EJSO 41 (2015) 353e360

Figure 2. Panels aed e KaplaneMeier plots depicting all cause mortality-free stratified according to type of surgery: radical cystectomy alone vs. radical cystectomy with pelvic lymph node dissection in a. patients aged
75 with CCI ¼ 0; b. patients
75 with CCI 1; c. patients aged >75 with CCI ¼ 0 and d. patients aged
75 with CCI 1. Panels eeh e KaplaneMeier plots depicting cancer specific mortality-free survival stratified according to type of surgery:

A. Larcher et al. / EJSO 41 (2015) 353e360

359

Figure 2. (continued).

RC did not differ from that recorded after RC with PLND. This observation refutes the hypothesis that PLND-related benefit is consistent among different age and comorbidities status categories. This finding implies that the observed benefit of PLND may not be universally applicable to all RC patients. The current report represents the first assessment of survival benefit after RC with PLND with respect to patient age and comorbidities, especially in the context of strict adjustment for potential confounders that might distinguish patients treated with RC and PLND from those who receive RC alone. Previously, Abdollah et al. used a large-scale (n ¼ 11,183) population based cohort in non-randomized study design that directly compared RC with PLND to RC alone.4,5 The authors evaluated the effect of PLND across different tumor stage categories and found that PLND was associated with an improved cancer specific survival in patients with T1 and T2 stage but not in patients with T3 and T4 disease.5 Interestingly, when the same analysis was applied in the current study, the observed findings corroborated the results reported by Abdollah et al., specifically PLND exerted a protective effect on CSM in patients diagnosed with T2 UCUB (HR 0.61; p < 0.001) but not in patients diagnosed with T3-4 UCUB (HR 0.86; p ¼ 0.02). Unfortunately, in the study by Abdollah et al., SEER data were not linked to Medicare records and in consequence, it was not possible to evaluate the effect of patient comorbidities on treatment outcomes. Similarly, adjustment could not be made for use of adjuvant chemotherapy. These limitations are crucial and were properly addressed in the current study, highlighting that baseline comorbidities might heavily affect the therapeutic effect of PLND at RC.

Other studies indicate that PLND extent may have a positive impact on cancer outcomes.11e17 However, none of these directly compared RC with PLND to RC alone. Based on these critical differences that distinguish the current study, our findings cannot be formally or even indirectly compared with previous data. Our data corroborate the concept of the protective effect of lymph node dissection that has already been shown in other primaries than UCUB, such as prostate,18,19 testis,20 penile,21 colon22 or breast cancer.23 The hypothesis that may be proposed to explain why the observed benefit survival is more evident in younger and healthier patients and disappear with increasing age and comorbidities is that the prognosis of older and sicker RC candidates is so poor that is not affected by surgical factors as PLND. The current data have strengths, which include their most contemporary nature and the use of propensityscore matching, which represents a method for minimizing the effect of selection-biases that underlie the choice of RC with PLND over RC alone. However, despite our attempt to adjust for the most comprehensive panel of measured baseline differences that distinguish patients treated with RC with PLND and RC alone, it is possible that unmeasured differences might persist, since only available variables that are recorded within the SEERMedicare can be used to maximally reduce potential selection biases. Unfortunately, many additional variables remain unaccounted for and allow the persistence of potentially significant differences between RC with PLND over RC. For example, the omission of PLND in younger patients might

radical cystectomy alone vs. radical cystectomy with pelvic lymph node dissection in e. patients aged
75 with CCI ¼ 0; f. patients
75 with CCI 1; g. patients aged >75 with CCI ¼ 0 and h. patients aged
75 with CCI 1. Time: Months after surgery. C.E.: Cumulative events. N.R.: Number of patients at risk.

360

A. Larcher et al. / EJSO 41 (2015) 353e360

represent an indicator of poor risks profile that is not comprehensively addressed by the evaluation of tumor stage and CCI. Ideally, a randomized design could eliminate this issue. Moreover, lack of highly detailed information pertaining to patient, surgeon, hospital and tumor characteristics represent other limitations due to the claim-based nature of our source. For example, the lack of information about the anatomical extent of PLND as well as the lack of a central pathology review represents a significant limitation of our study. Finally, the wide time span of case collection (1991e2009) might also be interpreted as limitation, since patters of UCUB care have significantly changed over time. For this reason, adjustment for year of diagnosis was performed in each MVA model. On the other hand, the inclusion of patients treated since 1991 allows a more accurate evaluation of long-term cancer outcomes. In conclusion, our findings indicate that RC with PLND is associated with improved all cause and cancer specific survival relative to RC alone, in younger and healthier RC candidates but not in older and sicker patients. Acknowledgments Study concept and design: Larcher, Sun, Fossati, Karakiewicz. Acquisition of data: Larcher, Sun, Schiffmann, Tian. Analysis and interpretation of data: Larcher, Sun, Schiffmann, Tian, Karakiewicz. Drafting of the manuscript: Larcher, Sun, Karakiewicz. Critical revision of the manuscript for important intellectual content: Shariat, McCormack, Saad, Fossati, Abdollah, Briganti, Buffi, Graefen, Guazzoni, Montorsi, Karakiewicz. Statistical analysis: Larcher, Sun, Tian. Supervision: Sun, Saad, Karakiewicz. Conflicts of interest None. References 1. Clark PE, Agarwal N, Biagioli MC, et al. Bladder cancer. J Natl Compr Canc Netw 2013;11(4):446–75. [Epub 2013/04/16]. 2. Witjes JA, Comperat E, Cowan NC, et al. EAU guidelines on muscleinvasive and metastatic bladder cancer: summary of the 2013 guidelines. Eur Urol 2014. [Epub 2014/01/01]. 3. Tilki D, Brausi M, Colombo R, et al. Lymphadenectomy for bladder cancer at the time of radical cystectomy. Eur Urol 2013;64(2):266– 76. [Epub 2013/05/08]. 4. Abdollah F, Sun M, Shariat SF, et al. The importance of pelvic lymph node dissection in the elderly population: implications for interpreting the 2010 National Comprehensive Cancer Network practice guidelines for bladder cancer treatment. J Urol 2011;185(6):2078–84. [Epub 2011/04/19].

5. Abdollah F, Sun M, Schmitges J, et al. Stage-specific impact of pelvic lymph node dissection on survival in patients with non-metastatic bladder cancer treated with radical cystectomy. BJU Int 2012; 109(8):1147–54. [Epub 2011/09/03]. 6. Warren JL, Klabunde CN, Schrag D, Bach PB, Riley GF. Overview of the SEER-Medicare data: content, research applications, and generalizability to the United States elderly population. Med Care 2002;40(8 Suppl.). IV-3-18. 7. Klabunde CN, Potosky AL, Legler JM, Warren JL. Development of a comorbidity index using physician claims data. J Clin Epidemiol 2000;53(12):1258–67. [Epub 2001/01/09]. 8. Sun M, Abdollah F, Liberman D, et al. Racial disparities and socioeconomic status in men diagnosed with testicular germ cell tumors: a survival analysis. Cancer 2011;117(18):4277–85. [Epub 2011/03/10]. 9. D’Agostino Jr RB. Propensity score methods for bias reduction in the comparison of a treatment to a non-randomized control group. Stat Med 1998;17(19):2265–81. [Epub 1998/11/05]. 10. Ho DE, Imai K, King G, Stuart E. MatchIt: nonparametric preprocessing for parametric causal inference. J Stat Softw 200342. 11. Herr HW, Bochner BH, Dalbagni G, Donat SM, Reuter VE, Bajorin DF. Impact of the number of lymph nodes retrieved on outcome in patients with muscle invasive bladder cancer. J Urol 2002;167(3):1295–8. [Epub 2002/02/08]. 12. Herr HW, Faulkner JR, Grossman HB, et al. Surgical factors influence bladder cancer outcomes: a cooperative group report. J Clin Oncol 2004;22(14):2781–9. [Epub 2004/06/17]. 13. 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(3):946–50. [Epub 2003/02/11]. 14. Wright JL, Lin DW, Porter MP. The association between extent of lymphadenectomy and survival among patients with lymph node metastases undergoing radical cystectomy. Cancer 2008;112(11):2401–8. [Epub 2008/04/03]. 15. May M, Herrmann E, Bolenz C, et al. Lymph node density affects cancer-specific survival in patients with lymph node-positive urothelial bladder cancer following radical cystectomy. Eur Urol 2011; 59(5):712–8. [Epub 2011/02/08]. 16. Tarin TV, Power NE, Ehdaie B, et al. Lymph node-positive bladder cancer treated with radical cystectomy and lymphadenectomy: effect of the level of node positivity. Eur Urol 2012;61(5):1025–30. [Epub 2012/02/22]. 17. van der Poel HG. Words of wisdom: Re: super extended versus extended pelvic lymph node dissection in patients undergoing radical cystectomy for bladder cancer: a comparative study. Eur Urol 2012; 61(6):1266–7. [Epub 2012/05/01]. 18. Joslyn SA, Konety BR. Impact of extent of lymphadenectomy on survival after radical prostatectomy for prostate cancer. Urology 2006; 68(1):121–5. [Epub 2006/06/30]. 19. Masterson TA, Bianco Jr FJ, Vickers AJ, et al. The association between total and positive lymph node counts, and disease progression in clinically localized prostate cancer. J Urol 2006;175(4):1320–4. [discussion 4e5. Epub 2006/03/07]. 20. Motzer RJ, Agarwal N, Beard C, et al. NCCN clinical practice guidelines in oncology: testicular cancer. J Natl Compr Canc Netw 2009; 7(6):672–93. [Epub 2009/06/27]. 21. Djajadiningrat RS, Graafland NM, van Werkhoven E, et al. Contemporary management of regional nodes in penile cancer-improvement of survival? J Urol 2014;191(1):68–73. [Epub 2013/08/07]. 22. West NP, Hohenberger W, Weber K, Perrakis A, Finan PJ, Quirke P. Complete mesocolic excision with central vascular ligation produces an oncologically superior specimen compared with standard surgery for carcinoma of the colon. J Clin Oncol 2010;28(2):272–8. [Epub 2009/12/02]. 23. Wo JY, Chen K, Neville BA, Lin NU, Punglia RS. Effect of very small tumor size on cancer-specific mortality in node-positive breast cancer. J Clin Oncol 2011;29(19):2619–27. [Epub 2011/05/25].