SEARCH

SEARCH BY CITATION

Keywords:

  • lymph node metastases;
  • nephrectomy;
  • renal cell carcinoma;
  • survival

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Conflict of interest
  9. References

Objectives

To examine cancer-specific mortality in patients with nodal metastases relative to patients without nodal involvement at nephrectomy for non-metastatic renal cell carcinoma in a population-based cohort.

Methods

A total of 11 374 non-metastatic renal cell carcinoma patients who underwent a lymph node dissection at nephrectomy were identified using the Surveillance, Epidemiology and End Results database (1988–2008). The 5-year cancer-specific mortality-free survival rates were examined according to the presence or absence of nodal involvement within the entire cohort, and stratified according to pathological tumor stage (pT1 vs pT2 vs pT3 vs pT4) and Fuhrman grade (I vs II vs III vs IV). Cox regression analyses for prediction of cancer-specific mortality were modeled to assess the effect of nodal metastases versus no nodal involvement in the entire population. Finally, separate Cox regression models were fitted within each pathological stage and grade.

Results

Overall, 1260 (11%) patients had nodal metastases at nephrectomy. The overall 5-year cancer-specific mortality-free survival rates were 38.4 versus 83.8% in patients with nodal metastases and without nodal metastases, respectively. In multivariable analyses, amongst pT1, pT2, pT3 and pT4, patients with nodal metastases were 6.0-, 3.6-, 3.2- and 2.0-fold, respectively, more likely to die after nephrectomy (all P < 0.001). Similarly, amongst Fuhrman grade I, Fuhrman grade II, Fuhrman grade III and Fuhrman grade IV, patients with nodal metastases were 3.9-, 3.5-, 3.1- and 2.7-fold, respectively, more likely to die of cancer-specific mortality (all P < 0.001).

Conclusions

Nodal involvement is an important determinant of higher cancer-specific mortality after nephrectomy. The detrimental effect of nodal metastases is particularly strong amongst patients with low-stage or low-grade non-metastatic renal cell carcinoma.


Abbreviations & Acronyms
CSM

cancer-specific mortality

FG

Fuhrman grade

HR

hazard ratio

IQR

interquartile range

LND

lymph node dissection

pN0

without nodal metastases

pN1

with nodal metastases

RCC

renal cell carcinoma

Ref

referent category

SEER

Surveillance, Epidemiology and End Results

Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Conflict of interest
  9. References

The role of LND in the setting of RCC remains a topic of debate. Given the increased ability to diagnose tumors at an earlier stage in recent years, the prevalence of nodal involvement has decreased substantially when compared with more historical data (from 30% to 3%).[1-3] Indeed, the rate of nodal metastases varies according to stage and grade distribution. For example, a previous multi-institutional report showed that the rate of pN1 was 1.2% in pT1 patients versus 4.5% and 12.3% in those with pT2 and pT3, respectively (n = 3507).[3] In a similar context, the rate of nodal metastases increases with more advanced FG according to data from a single institution: 8% in FG I–II versus 35% in FG III–IV (n = 900).[2, 3] Interestingly, while the prevalence of nodal metastases is lowest in low-stage tumors, the detrimental effect of the presence of nodal metastases on survival is also the strongest in low-stage tumors, relative to patients with higher-stage disease.[3]

To date, no population-based study has assessed the risk of CSM in a stage-for-stage and grade-for-grade manner. Our goal was to provide generalizable rates of node metastases amongst those who underwent a LND at nephrectomy, and to provide the estimates of the number of LND that would be required to identify one patient with nodal metastases (pN1) in a stage- and grade-specific manner. Finally, we examined the rates of CSM according to presence (pN1) or absence (pN0) of positive lymph node metastases.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Conflict of interest
  9. References

Data source

The study cohort consisted of individuals diagnosed with RCC (International Classification of Disease for Oncology [C67.0–C67.9]) from the SEER database reported by the National Cancer Institute statistics program between 1988 and 2008. The SEER routinely collects patient demographics, and publishes cancer incidence and survival data from population-based cancer registries covering approximately 26% of the USA population.

Study population

Patients diagnosed with non-metastatic RCC of all stages treated with a nephrectomy between the years 1988 and 2008 were abstracted (n = 80 525). Of those, only patients who received a LND or who had a known status of nodal involvement were considered, which resulted in 13 885 individuals. Patients with an unknown status of pathological tumor extent (n = 82) and tumor size at surgery (n = 348) were omitted. In as well, patients without a clear cell, papillary or chromophobe histological subtype (n = 1329) or those who were younger than 18 years-of-age (n = 752) were excluded. This resulted in a total of 11 374 assessable patients.

Description of covariates

Baseline characteristics included patient age, sex (male, female), race (white, black, other), tumor size, histological subtype (clear cell, papillary, chromophobe) and year of surgery. Nodal involvement status was stratified as pN0 and pN1. Pathological tumor extent was defined as pT1, pT2, pT3 and pT4. Tumor grade was categorized as FG I, II, III, IV and unknown.

Statistical analyses

Frequencies and proportions, as well as means, medians and interquartile ranges were reported for categorical and continuously coded variables, respectively. The χ2-test and Kruskal–Wallis test were used to compare proportions and medians, respectively. First, the 5-year CSM-free survival rates were computed according to nodal involvement status (pN0 vs pN1) within the entire population. Similarly, the 5-year CSM-free survival rates were also computed according to each stage and grade. Furthermore, to quantify the difference between pN0 and pN1 regarding the risk of CSM, the absolute risk reduction was calculated.

Subsequently, univariable and multivariable Cox regression analyses were carried out for prediction of CSM according to nodal involvement status. Covariates comprised of patient age, sex, race, tumor size, histological subtype, pathological tumor extent, FG and year of surgery. Finally, separate stage-for-stage and grade-for-grade Cox regression analyses for prediction of CSM were fitted.

All tests were two-sided with a statistical significance set at P < 0.05. Analyses were carried out using the statistical package for R (version 2.13.1; R Foundation for Statistical Computing, Vienna, Austria).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Conflict of interest
  9. References

Overall, 11 374 non-metastatic RCC patients who underwent LND at nephrectomy were identified (Table 1). The average age was 60 years (median 61 years). The mean tumor size was 7.4 cm (median 7.0 cm). The pathological tumor stage was 42, 22, 34 and 2.3% for pT1, pT2, pT3 and pT4, respectively. FG was 10, 33, 23 and 6% for I, II, III and IV, respectively. The majority of patients had clear cell histological subtype at nephrectomy (91%).

Table 1. Descriptive characteristics of 11 374 non-metastatic renal cell carcinoma patients who underwent lymph node dissection at nephrectomy, SEER 1988–2008
CharacteristicsOverallLymph node statusP-value
pN0pN1
  1. a

    Independent-Samples Mann–Whitney test. Except for first column, percentages are presented in rows.

No. patients11 374 (100.0)10 114 (88.9)1260 (11.1)
Age (years)   0.023a
Mean (median)60 (61)60 (61)61 (62)
IQR51–7051–7052–70
Sex   <0.001
Male7 220 (63.5)6 354 (88.0)866 (12.0)
Female4 154 (36.5)3 760 (90.5)394 (9.5)
Race   <0.001
White9 604 (84.4)8 583 (89.4)1021 (10.6)
Black992 (8.7)829 (83.6)163 (16.4)
Other778 (6.8)702 (90.2)76 (9.8)
Tumor, size (cm)   <0.001a
Mean (median)7.4 (7.0)7.2 (6.5)9.3 (9.0)
IQR4.5–9.54.3–9.26.5–11.8
Histological subtype   <0.001
Clear cell10 319 (90.7)9 241 (89.6)1078 (10.4)
Papillary667 (5.9)512 (76.8)155 (23.2)
Chromophobe388 (3.4)361 (93.0)27 (7.0)
Pathological stage   <0.001
pT14 750 (41.8)4 565 (96.1)185 (3.9)
pT22 466 (21.7)2 253 (91.4)213 (8.6)
pT33 898 (34.3)3 128 (80.2)770 (19.8
pT4260 (2.3)168 (64.6)92 (35.4)
FG   <0.001
I1 099 (9.7)1 064 (96.8)35 (3.2)
II3 775 (33.2)3 530 (93.5)245 (6.5)
III2 571 (22.6)2 129 (82.8)442 (17.2)
IV720 (6.3)504 (70.0)216 (30.0)
Unknown3 209 (28.2)2 887 (90.0)322 (10.0)
Year of surgery   0.028
1988–19962 999 (26.4)2 664 (88.8)335 (11.2)
1997–20012 865 (25.2)2 554 (89.1)311 (10.9)
2002–20053 083 (27.1)2 705 (87.7)378 (12.3)
2006–20082 427 (21.3)2 191 (90.3)236 (9.7)

A total of 1260 patients (11%) had nodal metastases at nephrectomy. The distribution of pN1 varied significantly according to sex, race and histological subtype. For example, the rate of pN1 was 12.0 versus 9.5% amongst males and females, respectively (P < 0.001). Furthermore, the rate of pN1 was higher in persons of black race compared with those of white race (16.4 vs 10.6%, P < 0.001). In as well, the rate of pN1 was 23.2% in papillary RCC versus 10.4% in clear cell and 7.0% of chromophobe RCC, respectively (P < 0.001).

The rate of nodal involvement also varied according to pathological tumor stage and grade (Table 1). For example, amongst patients with pT1, pT2, pT3 and pT4, the rate of pN1 was 3.9, 8.6, 19.8 and 35.4%, respectively (P < 0.001). These rates translated into 25, 11, 5 and 3 LND required to detect one patient with positive lymph node metastases for the same groups, respectively. Amongst patients with FG I, FG II, FG III and FG IV, the rate of pN1 was 3.1, 6.5, 17.2 and 30.0%, respectively (P < 0.001). These rates translated into 33, 14, 6 and 3 LND required to detect one pN1 patient for the same groups, respectively. For each pathological stage, the rate of nodal involvement increased with advancing grade. For example, amongst pT1 RCC patients, the rate of pN1 was 1.4, 2.5, 8.1 and 17.7% for FG I, FG II, FG III and FG IV, respectively (Fig. 1). Amongst pT4 RCC, the rate of pN1 was 0.0, 20.0, 39.3 and 41.7% for FG I, FG II, FG III and FG IV, respectively. Finally, the rate of pN1 was the lowest in the most contemporary year of surgery quartile (2006–2008) at 9.7%.

figure

Figure 1. Distribution of patients with pN1 amongst pathological tumor stage, according to FG: (a) pT1, (b) pT2, (c) pT3 and (d) pT4.

Download figure to PowerPoint

The overall 5-year CSM-free survival rates for patients with pN1 versus pN0 were 38.4% (95% CI 35.4–41.6%) vs 83.8% (95% CI 82.9–84.6%, log–rank P < 0.001; Fig. 2). For the same groups, the 5-year CSM-free survival rates were 60.2 versus 94.0% amongst pT1, 52.9 versus 85.4% amongst pT2, 31.1 versus 69.8% amongst pT3, and 16.9 versus 39.5% amongst pT4 patients, respectively (all log–rank P < 0.001; Fig. 3a–d). Similarly, the 5-year CSM-free survival rates were 75.3 versus 93.8% in FG I, 52.7 versus 88.0% in FG II, 36.2 versus 76.4% in FG III, and 24.6 versus 60.7% in FG IV, respectively (all log–rank P < 0.001; Fig. 4a–d).

figure

Figure 2. Overall Kaplan–Meier plot showing CSM-free survival rates according to presence (pN1) or absence (pN0) of nodal metastases (log–rank P < 0.001).

Download figure to PowerPoint

figure

Figure 3. Kaplan–Meier plots showing CSM-free survival rates according to presence (pN1) or absence (pN0) of nodal metastases amongst (a) pT1, (b) pT2, (c), pT3 and (d) pT4. All log–rank P < 0.001.

Download figure to PowerPoint

figure

Figure 4. Kaplan–Meier plots showing CSM-free survival rates according to presence (pN1) or absence (pN0) of nodal metastases amongst (a) FG I, (b) FG II, (c) FG III and (d) FG IV.

Download figure to PowerPoint

The overall absolute risk reduction for CSM between pN0 and pN1 was 45%, which translates into a number needed to treat of 2. The absolute risk reduction according to stage was 34, 33, 38 and 22% for pT1, pT2, pT3 and pT4, respectively (Fig. 5a). In comparison, the absolute risk reduction according to grade was 18, 35, 40 and 36 for FG I, FG II, FG III and FG IV, respectively (Fig. 5b).

figure

Figure 5. Absolute risk reduction rates of the 5-year CSM-free survival probabilities between patients with presence (pN1) and absence (pN0) of nodal metastases according to (a) pathological tumor stage and (b) FG. image, pN1; image, pN0.

Download figure to PowerPoint

After adjustment for all other covariates (Table 2), patients with nodal metastases were more likely to die of RCC than patients without nodal metastases (HR 3.40, 95% CI 3.09–3.73, P < 0.001). In stage-for-stage analyses, pT1, pT2, pT3 and pT4 patients with nodal metastases at nephrectomy were 6.0-, 3.6-, 3.2- and 2.0-fold more likely to die of CSM than pN0 patients (all P < 0.001, Table 3). Finally, in grade-for-grade analyses, FG I, FG II, FG III and FG IV patients with nodal metastases at nephrectomy were 3.9-, 3.5-, 3.1- and 2.7-fold more likely to succumb to CSM after nephrectomy relative to their pN0 counterparts (all P < 0.001; Table 4).

Table 2. Cox regression analyses for prediction of cancer-specific mortality (CSM) within the entire cohort (n = 11 374)
VariablesUnivariableMultivariable
HR (95% CI)PHR (95% CI)P
Lymph node status    
pN0Ref Ref 
pN15.68 (5.20–6.21)<0.0013.42 (3.11–3.71)<0.001
Age1.01 (1.01–1.02)<0.0011.01 (1.01–1.02)<0.001
Sex    
MaleRef Ref 
Female0.89 (0.82–0.97)0.0071.01 (0.93–1.10)0.8
Race    
WhiteRef Ref 
Black1.00 (0.87–1.16)1.01.11 (0.96–1.29)0.2
Other0.81 (0.68–0.96)0.020.80 (0.67–0.95)0.01
Tumor size1.01 (1.00–1.01)<0.0011.01 (1.00–1.02)<0.001
Histological subtype    
Clear cellRef Ref 
Papillary0.97 (0.80–1.18)0.80.91 (0.74–1.12)0.4
Chromophobe0.29 (0.19–0.46)<0.0010.37 (0.23–0.58)<0.001
Pathological stage    
pT1Ref Ref 
pT22.44 (2.13–2.79)<0.0011.87 (1.62–2.16)<0.001
pT35.95 (5.31–6.66)<0.0013.70 (3.26–4.18)<0.001
pT415.52 (12.89–18.68)<0.0017.83 (6.43–9.52)<0.001
FG    
IRef Ref 
II1.92 (1.55–2.38)<0.0011.55 (1.25–1.93)<0.001
III4.22 (3.41–5.22)<0.0012.38 (1.91–2.95)<0.001
IV8.44 (6.71–10.61)<0.0013.50 (2.77–4.44)<0.001
Unknown2.74 (2.22–3.38)<0.0011.86 (1.50–2.30)<0.001
Year of surgery0.98 (0.97–0.99)<0.0010.97 (0.96–0.98)<0.001
Table 3. Multivariable Cox regression analyses predicting CSM, stratified according to pathological tumor stage
VariablespT1pT2pT3pT4
  1. a

    P < 0.001. Covariates comprised of patient age, sex, race, tumor size, histological subtype, FG and year of surgery.

Lymph node status    
pN0RefRefRefRef
pN16.03 (4.61–7.88)a3.55 (2.80–4.50)a3.21 (2.85–3.60)a2.04 (1.46–2.84)a
Table 4. Multivariable Cox regression analyses predicting CSM, stratified according to FG
VariablesFG IFG IIFG IIIFG IVFG X
  1. a

    P < 0.001. Covariates comprised of patient age, sex, race, tumor size, histological subtype, tumor stage and year of surgery.

Lymph node status     
pN0RefRefRefRefRef
pN13.87 (2.11–7.10)a3.53 (2.83–4.41)a3.09 (2.61–3.66)a2.66 (2.10–3.37)a4.03 (3.40–4.77)a

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Conflict of interest
  9. References

Nodal involvement represents a major determinant for prognosis in patients with RCC.[4, 5] The estimated 5-year CSM-free survival rates after nephrectomy range between 21–38% amongst pN1 RCC patients.[3, 4] The prevalence of nodal metastases averages between 13% to over 32%, and increases with more advanced tumor stage and grade.[2, 3] The objective of the current study was to examine the prevalence of nodal metastases in those who were treated with a LND at nephrectomy in a contemporary population-based cohort according to RCC stage and grade. Additionally, we determined the number of LND required to find one individual with positive lymph node metastases. Finally, we compared CSM rates in patients with and without nodal involvement in a stage-for-stage and grade-for-grade fashion.

Our findings are severalfold. First, the prevalence of pN1 among individuals who were treated with LND at nephrectomy decreased during the study span (from 11.2% in years 1988–1996 to 9.7% in years 2006–2008, P = 0.28). This observation might be a direct result of the downward stage- and grade migration in contemporary years,[6, 7] which ultimately lowers the utilization of LND and the prevalence of nodal involvement, given that these patients are at a substantially lower risk of harboring nodal metastases at nephrectomy. Indeed, historical data have reported a rate of nodal metastases of >30% versus 3% in more contemporary years.[1-3]

That said, the prevalence of lymph node invasion has been shown to vary in a stage- and grade-specific manner. For example, in a previous study the rate of pN1 was 1.2% in pT1 patients versus 4.5 and 12.3% in those with pT2 and pT3, respectively.[3] Similarly, the rate of pN1 was 28.5 versus 71.5% for FG I–II versus FG III–IV patients, respectively. In the current study, the rate of pN1 also showed a direct correlation with increasing stage and grade. Specifically, nodal metastases rates were 3.9% in pT1 and 35.4% in pT4 RCC (P < 0.001). In as well, the pN1 rate was 3.2% in FG I and 30.0% in FG IV (P < 0.001). Furthermore, the rate of nodal metastases also increased in a grade-for-grade fashion within each tumor stage. For example, amongst patients with pT1 RCC, the rate of pN1 was 1.4% in FG I versus 17.7% in FG IV. Similarly, amongst patients with pT3 RCC, the rate of pN1 was 9.2% in FG I versus 34.0% in FG IV. This trend was invariably consistent throughout each tumor stage.

Third, the presence of nodal metastases was invariably associated with worse prognosis relative to patients with the absence of nodal metastases at nephrectomy (5-year CSM-free survival rate 38.4 vs 83.8%, P < 0.001). This effect was consistent when CSM-free survival rates were compared between patients with the presence or absence of nodal metastases, across each stage and grade.

Finally, following the consideration of all available covariates, pN1 patients were 3.4-fold more likely to succumb to RCC-specific mortality than their pN0 counterparts (P < 0.001). Interestingly, the effect of nodal involvement was systematically less pronounced across increasing pathological tumor stage (HR 6.0–2.0) and tumor grade (HR 3.9–2.6). These observations suggest that given a low-stage and/or grade disease, the presence of nodal metastases is particularly detrimental on survival, whereas in the context of high-stage and/or grade disease, although not inconsequential, its effect becomes less pronounced with respect to CSM. These findings are in line with the results originating from a previous multi-institutional report. Using retrospective data that comprised of 3507 non-metastatic RCC patients, Capitanio et al. found that low-stage patients were at the highest increased risk of CSM (HR 28.9, P < 0.001).[3]

Taken together, the present results show that although the prevalence of nodal metastases is lowest in low-stage/grade disease, the prognostic value of the presence of nodal metastases is the highest in these patients. These findings pose a clinical problem, because the likelihood of finding positive lymph nodes in low-stage/grade patients is extremely low (<5%), yet it is in these exact patients that the prognosis of harboring positive lymph nodes will be the worst.

Do these findings suggest that LND might be particularly worthwhile in low-stage/grade patients? Given a potential morbidity associated with a LND at nephrectomy, the extremely low rate of finding a positive lymph node in such patients, the increasing rate of nephron-sparing procedure and other minimally invasive procedures, as well as the lack of a proven survival benefit – probably not. However, what the current findings do suggest is the importance of being able to discriminate those who do and those who do not harbor positive lymph nodes in the setting of low-stage/grade RCC in a preoperative manner. In that regard, preoperative nomograms for prediction of positive lymph nodes at nephrectomy exist.[4, 8, 9] In addition, the identification of molecular markers, which can help differentiate patients harboring positive lymph nodes might also be of primordial importance.[10] Finally, the current findings might also be relevant during clinical counseling and postoperative follow-up management.[11]

In contrast, should a LND at nephrectomy be recommended as a standard of care in high-risk patients? Although official guideline recommendations favor a LND in patients with palpable or computed tomography-detected enlarged lymph nodes,[11] this question remains unclear in the context of RCC. Although a number of non-randomized retrospective data have suggested a potential survival benefit in patients who undergo a LND at nephrectomy over patients who do not,[2, 4, 12, 13] the final results of the only randomized phase III trial originating from the European Organization for Research and Treatment of Cancer Genito-Urinary Group showed no survival advantage between carrying out a LND at nephrectomy and nephrectomy alone.[1] Unfortunately, the lack of inclusion of high-risk patients in the phase III trial renders the results inapplicable to such patients. Indeed, over 70% of patients harbored T1–2 RCC, which also explains the low incidence of positive lymph nodes in that report (4%).[14] In consequence, it might be argued that LND can still be considered in high-risk patients for the purpose of staging, and perhaps for a therapeutic purpose, until proven otherwise.

The present study was not devoid of limitations. First, the SEER-limited database does not contain any information of individuals’ baseline comorbid conditions or performance status (e.g. Eastern Cooperative Oncology Group, Karnofsky performance status). Other factors associated with RCC outcomes after nephrectomy, such as tumor necrosis, LND templates, as well as the lack of knowledge of clinical nodal involvement, might have undermined the findings. Furthermore, some degree of nodal status misattribution might have been present. Previous studies that also relied on the SEER-limited data were also affected by such limitations.[8, 15] In as well, there was a lack of information and adjustment for adjuvant therapy, which might have affected patients’ prognosis. Finally, the current study was unable to assess the real prevalence of nodal metastases at surgery, as only patients who received a LND at nephrectomy were included in the analyses.

Nodal involvement has a high prognostic value in patients with low-stage or low-grade RCC. However, given the low prevalence of harboring positive lymph node metastases, there is no justification to undertake a LND at nephrectomy in such patients.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Conflict of interest
  9. References

Pierre I Karakiewicz is partially supported by the University of Montreal Health Centre Urology Specialists, Fonds de la Recherche en Santé du Québec, the University of Montreal Department of Surgery and the University of Montreal Health Centre (CHUM) Foundation.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Conflict of interest
  9. References