SEARCH

SEARCH BY CITATION

Keywords:

  • lymph nodes;
  • neoplasm metastasis;
  • prognosis;
  • urinary bladder neoplasms

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONFLICT OF INTEREST
  8. REFERENCES

Study Type – Therapy (case series) Level of Evidence 4

What’s known on the subject? and What does the study add?

Although extranodal extension in node-positive patients may provide prognostic information in certain urological malignancies, contradictory results have been reported in node-positive patients after radical cystectomy for bladder cancer.

We examined whether extranodal extension could have an impact on the outcomes of node-positive patients who underwent radical cystectomy for bladder cancer in a single large tertiary hospital. We have shown that the presence of extranodal extension was not an independent prognostic factor in node-positive patients after radical cystectomy and pelvic lymphadenectomy for bladder cancer.

OBJECTIVE

• To examine whether extranodal extension (ENE) has an impact on the outcome of node-positive patients who underwent radical cystectomy for bladder cancer.

PATIENTS AND METHODS

• Of 543 consecutive patients who underwent radical cystectomy for urothelial carcinoma of the bladder between 1990 and 2007, 112 patients with lymph node metastasis detected on histological examination were evaluated with regard to recurrence-free survival (RFS) and disease-specific survival (DSS) based on ENE status.

RESULTS

• The overall 5-year RFS and DSS rates were 22.3% and 33.8%, respectively. ENE was observed in 41 (36.6%) of the 112 patients.

• The presence of ENE was associated with advanced pathological nodal status (P= 0.004), more positive lymph nodes (P= 0.006), and higher lymph node density (P < 0.001).

• The incidence of positive ENE increased with large positive lymph node diameter (P < 0.001).

• Multivariate analysis showed that lymph node density (hazard ratio (HR) = 2.39, 95% CI 1.09–5.24, P= 0.029; and HR = 3.13, 1.43–6.84, P= 0.004) and use of adjuvant chemotherapy (HR = 1.80, 1.02–3.20, P= 0.041; and HR = 2.07, 1.13–3.79, P= 0.018) were significant predictors of RFS and DSS, respectively.

• After adjustment for other prognostic factors, ENE was not significantly related to RFS (P= 0.825) and DSS (P= 0.961) by multivariate analysis.

CONCLUSIONS

• The presence of ENE was not an independent prognostic factor in node-positive patients after radical cystectomy for bladder cancer.

• Additional prospective studies are needed to determine the independent prognostic role of ENE.


Abbreviations
LND

lymph node density

DSS

disease-specific survival

ENE

extranodal extension

MVAC

methotrexate, vinblastine, doxorubicin and cisplatin

GC

gemcitabine and cisplatin

RFS

recurrence-free survival

MDACC

M.D. Anderson Cancer Center

INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONFLICT OF INTEREST
  8. REFERENCES

The standard surgical treatment for patients with high-grade, muscle-invasive bladder cancer is radical cystectomy with pelvic lymphadenectomy. The most accurate predictors of disease recurrence after radical cystectomy have been shown to be the pathological stage of the primary tumour and regional lymph node status [1]. Regional lymph node involvement after radical cystectomy for bladder cancer has been reported in 14–32% of patients, and patients with positive nodes after radical cystectomy and pelvic lymphadenectomy have 5-year disease-specific or overall survival rates of 30–37%[1–5].

Identification of patients who are at greatest risk for developing distant metastasis and for death from bladder cancer will help to stratify patients into prognostically distinct subgroups. Currently, risk stratification of patients with nodal metastases at the time of radical cystectomy and pelvic lymphadenectomy is based on the 2002 TNM classification system [6]. In this staging system, nodal disease is quantified by the number and size of involved regional lymph nodes. However, lymph node density (LND), defined as the ratio of positive to total nodes removed, may be superior to TNM nodal status in predicting disease-specific survival (DSS) for patients with lymph-node-positive disease after radical cystectomy [2,3,7].

Extranodal extension (ENE) in node-positive patients may provide prognostic information in certain urological malignancies, including kidney and penile cancer [8,9]. Other studies have suggested, however, that ENE is not a significant predictor in patients with prostate and testis cancer [10–12]. To date, contradictory results have been reported in node-positive patients after radical cystectomy for bladder cancer [4,13–15]. We therefore examined whether ENE, defined as microscopic evidence of capsular invasion into perinodal adipose tissue, could have an impact on the outcomes of node-positive patients who underwent radical cystectomy for bladder cancer in a single large tertiary hospital.

PATIENTS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONFLICT OF INTEREST
  8. REFERENCES

From December 1990 to December 2007, a total of 543 patients at our institute underwent radical cystectomy with bilateral pelvic lymphadenectomy for urothelial carcinoma of the bladder. Of these, 142 patients (26.2%) had lymph node metastases on histological examination. The medical records of these patients were reviewed, and information regarding their demographic characteristics and perioperative variables was obtained. Thirty patients were excluded from the study because pathology slides and tissue blocks were not available at the time of analysis. All pathology slides were retrospectively reviewed by a single senior genitourinary pathologist (J.Y.R) blinded to patient outcome. The study protocol was approved by the Institutional Review Board of the Asan Medical Centre, Seoul, Korea.

No patient received preoperative radiation or neoadjuvant chemotherapy. All patients underwent preoperative chest radiography, CT scans of the abdomen and pelvis, and bone scans for tumour staging. None of these patients showed evidence of metastatic disease on physical examination or staging.

All operations were performed by two senior surgeons (HA and CSK). Of the 112 patients, 98 underwent bilateral standard pelvic lymphadenectomy, which included the distal common iliac, external iliac, hypogastric, obturator and perivesical lymph nodes. The remaining 14 patients underwent extended lymphadenectomy, which included the lymph node regions previously described, as well as those at the level of the proximal common iliac, distal aorta and vena cava. The extent of lymph node dissection was determined by each surgeon.

All cystectomy specimens were pathologically staged and reclassified according to the 2002 TNM classification system [6]. Tumour grade, margin status, total number of lymph nodes, number of positive lymph nodes, and the presence of ENE were also assessed. ENE was defined as invasion of the capsule by tumour cells with extranodal growth in perinodal adipose tissue [13]. LND was calculated by dividing the number of positive lymph nodes by the total number of lymph nodes removed.

Adjuvant chemotherapy was routinely recommended for all node-positive patients at our institute and it was initiated within 3 months of surgery in all patients. Our institutional policy includes the use of adjuvant chemotherapy for all node-positive patients; hence, the all-patient cohort was analysed as the intent-to-treat population. Despite our recommendations, patients and their families were often reluctant to receive additional chemotherapy after cystectomy, owing to difficulties associated with frequent hospital visits, advanced age and/or postoperative hyposthenia. Chemotherapy comprised three to six cycles of MVAC (methotrexate, vinblastine, doxorubicin and cisplatin, n= 21) or GC (gemcitabine and cisplatin, n= 52) combination regimens. The chemotherapy regimen was time dependent: from 1992 to 2000 only MVAC regimens were used, whereas from 2001 to 2007 only GC regimens were used. The GC and MVAC regimens have similar survival benefits but the former has a better safety profile and greater tolerability for patients with advanced or metastatic bladder cancer [16].

Following radical cystectomy, patients were evaluated quarterly for the first year, semi-annually during years 2 to 6, and annually thereafter. Follow-up examinations after surgery included history, physical examination, blood chemistry analyses and urine cytology. Patients underwent chest radiography and CT of the abdomen and pelvis at 6 and 12 months, and annually thereafter.

Clinicopathological factors were compared between the groups with and without ENE using the Fisher’s exact test for categorical variables and the Mann–Whitney U-test for continuous variables. Kaplan–Meier survival curves for recurrence-free survival (RFS) and DSS, generated according to the individual potential prognostic variables, were compared using the log-rank test. A Cox proportional hazards regression model was used to estimate the prognostic significance of each variable. RFS and DSS were measured from the date of cystectomy to the date of relapse or death from bladder cancer. The number of lymph nodes removed was analysed as a variable dichotomized around the median. The selected cut-off point for dichotomization of number of positive lymph nodes has been used in other studies [1,13]. The optimal cut-off value for LND to predict RFS and DSS was selected by analysis of receiver operating characteristics curves, with an LND of 18% selected as the cut-off with the best balance between sensitivity and specificity. All statistical tests were two-tailed, with P < 0.05 considered significant. All calculations were performed using the Statistical Package for Social Sciences, version 12.0 (SPSS, Chicago, IL, USA).

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONFLICT OF INTEREST
  8. REFERENCES

Of the 112 patients, 99 were men and 13 were women; their mean age was 61.8 ± 9.9 years (range 37–79 years). Follow-up ranged from 2 to 146 months (median 23 months). At the time of analysis, 71 patients (63.4%) had disease recurrence and 69 (61.6%) had died from their disease. Overall, the median RFS and DSS were 15 months and 29 months, respectively. Kaplan–Meier estimates of 2-year and 5-year RFS rates were 39.6% and 22.3%, respectively, whereas the 2-year and 5-year DSS rates were 58.1% and 33.8%, respectively. The median number of lymph nodes removed was 16 (range 3–94), with a median of three positive lymph nodes (range 1–37). The median largest size of the positive lymph nodes was 1.1 cm (range 0.1–8.0 cm).

ENE was observed in 41 (36.6%) of the 112 patients. Table 1 shows the characteristics of patients stratified by the presence of ENE. The presence of ENE was associated with significantly more advanced pathological nodal status (pN2-3: 85.4% vs 56.3%, P= 0.004), a significantly higher number of mean positive lymph nodes (7.0 vs 3.4, P= 0.006), and a significantly higher mean LND (LND 38.0% vs 19.2%, P < 0.001). Figure 1 shows the distribution of ENE according to the largest dimension of nodal metastasis and the total number of lymph nodes removed. The incidence of ENE increased with large positive lymph node diameter (P < 0.001), but not with the total number of lymph nodes removed (P= 0.845).

Table 1.  Patient characteristics stratified by the presence of extranodal extension
 TotalsENE (−)ENE (+)P-value
  1. ENE, extranodal extension; LND, lymph node density.

Patients, n (%) 11271 (63.4)41 (36.6) 
Sex, n (%)   0.883
 Male 99 (88.4)63 (88.7)36 (87.8) 
 Female 13 (11.6) 8 (11.3) 5 (12.2) 
Mean age ±sd (median, range) 61.8 ± 9.9 (63, 37–79)62.1 ± 9.5 (63, 43–79)61.3 ± 10.6 (63, 37–79)0.397
Grade, n (%)   0.430
 2  5 (4.5) 4 (5.6) 1 (2.4) 
 3107 (95.5)67 (94.4)40 (97.6) 
Soft tissue surgical margin status, n (%)   0.642
 Positive 13 (11.6) 9 (12.7) 4 (9.8) 
 Negative 99 (88.4)62 (87.3)37 (90.2) 
Pathological primary tumour status, n (%)   0.070
 Organ confined 24 (21.4)19 (26.8) 5 (12.2) 
 Extravesicular 88 (78.6)52 (73.2)36 (87.8) 
Pathological nodal status, n (%)   0.004
 1 37 (33.0)31 (43.7) 6 (14.6) 
 2 74 (66.1)40 (56.3)34 (82.9) 
 3  1 (0.9) 0 (0) 1 (2.5) 
Mean positive lymph nodes ±sd (median, range)  4.7 ± 5.6 (3, 1–37) 3.4 ± 3.3 (2, 1–14) 7.0 ± 7.7 (4, 1–37)0.006
Mean lymph node removed ±sd (median, range) 19.2 ± 12.9 (16, 3–94)19.3 ± 13.6 (16, 3–94)18.9 ± 11.8 (15, 3–47)0.801
Mean % LND ±sd (median, range) 26.1 ± 21.7 (20, 3–100)19.2 ± 14.9 (15, 3–70)38.0 ± 26.3 (30, 5–100)<0.001
Adjuvant chemotherapy, n (%)   0.766
 Yes 73 (65.2)47 (66.2)26 (63.4) 
 No 39 (34.8)24 (33.8)15 (36.6) 
image

Figure 1. Incidence of extranodal extension (ENE) in patients with node-positive bladder cancer relative to (A) the largest diameter of positive lymph nodes and (B) the total number of lymph nodes removed.

Download figure to PowerPoint

Univariate and multivariate Cox regression analyses of variables predicting RFS are summarized in Table 2. Univariate analysis showed that significant predictors of decreased RFS were increasing pathological nodal status, presence of five or more positive lymph nodes, and LND >18%. Multivariate analysis hazard ratios (HR) showed that LND (HR = 2.39, 95% CI 1.09–5.24, P= 0.029) and the use of adjuvant chemotherapy (HR = 1.80, 95% CI 1.02–3.20, P= 0.041) were significant predictors of RFS.

Table 2.  Univariate and multivariate analyses of variables predicting recurrence-free survival
VariablesNo. of patientsNo. of recurrencesMed RFS (months)UnivariateMultivariate
HR95% CIP-valueHR95% CIP-value
  1. RFS, recurrence-free survival; HR, hazard ratio; 95% CI, 95% confidence intervals.

Primary tumour status
 T1–22415171.001.00
 T3–48856151.150.65–2.040.6311.370.74–2.540.319
Pathological N stage
 N13718301.001.00
 N2–37553122.201.28–3.890.0041.450.66–3.160.353
No. positive lymph nodes
 <57544171.001.00
 ≥53727111.621.00–2.630.0500.830.43–1.600.569
No. lymph nodes removed
 <165535151.001.00  
 ≥165736161.060.66–1.680.8241.200.69–2.100.508
Lymph node density
 0–18% 5426251.001.00
 >18%5845112.531.54–4.15<0.0012.391.09–5.240.029
Extranodal extension
 No7142231.001.00
 Yes4129121.580.98–2.550.0611.060.63–1.800.825
Adjuvant chemotherapy
 Yes7349171.001.00
 No3922121.390.84–2.310.2051.801.02–3.200.041

Table 3 shows the univariate and multivariate analysis of variables predicting DSS. By univariate analysis, pathological nodal status, number of positive lymph nodes, LND and the presence of ENE were all significant predictors of DSS. By multivariate analysis, only LND >18% (HR = 3.13, 95% CI 1.43–6.84, P= 0.004) and absence of adjuvant chemotherapy (HR = 2.07, 95% CI 1.13–3.79, P= 0.018) were significant predictors of DSS. After adjusting for other prognostic factors, ENE did not retain significance in the multivariate analysis (P= 0.961).

Table 3.  Univariate and multivariate analyses of variables predicting disease-specific survival
VariablesNo. of patientsNo. of deathsMed. DSS (months)UnivariateMultivariate
HR95% CIP-valueHR95% CIP-value
  1. DSS, disease-specific survival; HR, hazard ratio; 95% CI, 95% confidence intervals.

Primary tumour status
 T1–22415341.001.00
 T3–48854290.980.55–1.740.9451.160.61–2.190.658
Pathological N stage
 N13714 1.001.00
 N2–37555193.111.71–5.65<0.0011.920.84–4.350.120
No. positive lymph nodes
 <57541341.001.00
 ≥53728371.831.13–2.960.0150.750.39–1.470.405
No. lymph nodes removed
 <165534271.001.00  
 ≥165735341.120.70–1.800.6441.320.74–2.360.356
Lymph node density
 0–18%5422631.001.00
 >18%5847163.281.96–5.49<0.0013.131.43–6.840.004
Extranodal extension
 No7139351.001.00
 Yes4130171.661.03–2.680.0391.010.59–1.730.961
Adjuvant chemotherapy
 Yes7347341.001.00
 No3922151.440.87–2.400.1612.071.13–3.790.018

When the prognostic relevance of ENE was investigated according to nodal burden (categorized by the LND), we found that RFS and DSS did not differ between patients with and without ENE. In patients with low LND (0–18%), the presence of ENE was not significant for RFS (P= 0.409) or DSS (P= 0.316); similar results were observed in patients with high LND (>18%) for RFS (P= 0.659) and DSS (P= 0.877). To identify prognostic factors in patients who received adjuvant chemotherapy we performed subgroup analysis. When including 73 patients, LND was an independent prognostic factor for RFS and DSS. However, the presence of ENE was not significant for RFS or DSS (Table 4).

Table 4.  Multivariate analysis of variables predicting recurrence-free and disease-specific survival in 73 patients who received adjuvant chemotherapy
VariablesP-value 
  1. RFS, recurrence-free survival; DSS, disease-specific survival.

 RFSDSS
Primary tumour status0.8380.616
Pathological N stage0.8650.845
No. positive lymph nodes0.8530.362
No. lymph nodes removed0.3370.086
Lymph node density0.018<0.001
Extranodal extension0.2930.493

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONFLICT OF INTEREST
  8. REFERENCES

We found that 36.6% of patients who had bladder cancer with lymph node metastasis had ENE. This incidence of ENE was similar to the 41% reported at the M.D. Anderson Cancer Center, Houston, TX, USA (MDACC) but lower than the 58% reported at the University of Berne, Switzerland [13–15]. Although the presence of ENE was correlated with higher pathological nodal status, higher number of positive lymph nodes, and higher LND, the presence of ENE was not associated with patient outcome, as assessed by RFS and DSS.

To date, conflicting results have been reported on the significance of ENE in bladder cancer. In a Swiss study of 101 patients with lymph node metastasis at the time of radical cystectomy, ENE was reported to be an independent prognostic factor for RFS. In that study, however, primary tumour stage, number of positive lymph nodes and adjuvant chemotherapy were not independent prognostic factors for both RFS and overall survival [13,15]. In the MDACC study of 108 patients with positive lymph nodes at radical cystectomy, ENE was not an independent prognostic factor for DSS; although primary pathological stage and adjuvant chemotherapy were independent predictors of DSS [14].

Our results confirmed that pathological N stage and LND were significantly associated with RFS and DSS on univariate analysis. In addition, the number of positive lymph nodes removed and ENE were significantly associated with DSS. ENE, however, did not retain statistical significance on multivariate analysis after adjustment for other variables known to be prognostic for node-positive bladder cancer patients. Only LND and the use of adjuvant chemotherapy retained statistical significance by multivariate analysis.

There may be several possible explanations for the differences in the prognostic value of ENE. First, harvest of a large number of lymph nodes during pelvic lymphadenectomy may increase the number examined during pathological processing, resulting in an increased incidence of ENE. In our study, the median number of lymph nodes removed was 16, fewer than the 22 in the Swiss series but more than the 12 from MDACC [13,14]. We found, however, that the incidence of ENE did not increase according to the total number of lymph nodes removed, making it unlikely that differences in the total number of harvested lymph nodes altered the incidence of ENE.

Second, we found that the incidence of ENE increased with increasing largest diameter of positive lymph nodes. Although we observed a similar proportion of patients with pN2-3 status (67%) as in the Swiss (68%) and MDACC (57%) cohorts, a difference in mean largest diameter of positive lymph nodes among studies may have affected the incidence of ENE.

Third, differences in the prognostic value of other confounding variables may have affected the prognostic significance of ENE in node-positive patients. Although the only independent predictor of RFS in the Swiss series was ENE, the MDACC study showed that primary pathological tumour stage and adjuvant chemotherapy were significantly associated with DSS by multivariate analysis [13,14]. In contrast, our multivariate analysis showed that LND (0–18% vs >18%) and adjuvant chemotherapy were independent predictors for both RFS and DSS. We were not surprised to find that LND was an independent predictor of both RFS and DSS, because a recent large study of patients at the MDACC and the Memorial Sloan-Kettering Cancer Center, New York, NY, USA, showed that LND was superior to current TNM nodal status in predicting DSS, even after controlling for adjuvant chemotherapy [2]. Our joint analysis of ENE and LND also showed the inability of ENE to predict RFS and DSS in different LND subgroups.

Although primary tumour status has been reported to be prognostic in patients with node-positive bladder cancer [2,5,17], we found no survival differences between patients with organ-confined and extravesicular disease. Our findings were similar to the results of other studies, in which primary tumour stage was not an independent prognostic factor of DSS in node-positive patients [13,18]. It may have resulted from the fact that most patients with node-positive bladder cancer had extravesicular disease. Also, we observed that survival did not differ significantly between node-positive patients with <16 or ≥16 lymph nodes removed; similar results were seen when the total number of lymph nodes removed was analysed as a continuous variable. At present, it is unclear whether the removal of more lymph nodes may be associated with a survival advantage in lymph-node-positive patients. Although retrospective studies have shown that the removal of more lymph nodes was associated with improved prognosis, even in node-positive patients [7,17,19,20], others have failed to show a prognostic significance for the total number of lymph nodes removed in node-positive patients [2,3,5,14].

This study has several potential limitations, including those inherent to any retrospective study. In addition, the selection criteria for adjuvant chemotherapy may have led to selection bias. Because the proportion of men who received adjuvant chemotherapy did not differ between those with and those without ENE, it was unlikely that any potential selection bias from adjuvant chemotherapy may have affected our results. Second, the extent of lymphadenectomy changed during the study period, with about 10% of our patients undergoing extended pelvic lymphadenectomy and so it was not analysed as a covariate in survival analysis. Instead, because the total number of lymph nodes removed (either dichotomized around the median or as a continuous variable) did not affect disease recurrence or survival, it was unlikely that selection bias from the different extent of lymphadenectomy affected the prognostic value of ENE.

In conclusion, we have shown that the presence of ENE was not an independent prognostic factor in node-positive patients after radical cystectomy and pelvic lymphadenectomy for bladder cancer. Additional prospective studies to determine the independent prognostic role of ENE are needed.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONFLICT OF INTEREST
  8. REFERENCES