- Email: [email protected]
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
0022-5347/04/1722-0465/0 THE JOURNAL OF UROLOGY® Copyright © 2004 by AMERICAN UROLOGICAL ASSOCIATION
Vol. 172, 465– 469, August 2004 Printed in U.S.A.
DOI: 10.1097/01.ju.0000129815.91927.85
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 MICHAEL L. BLUTE,* BRADLEY C. LEIBOVICH, JOHN C. CHEVILLE, CHRISTINE M. LOHSE AND HORST ZINCKE* From the Departments of Urology (MLB, BCL, HZ), Pathology (JCC) and Health Sciences Research (CML), Mayo Clinic, Rochester, Minnesota
ABSTRACT
Purpose: We determined the primary pathological features of clear cell renal cell carcinoma that are predictive of positive regional lymph nodes at radical nephrectomy (RN) and developed a protocol for the selective use of extended lymph node dissection. Materials and Methods: We studied 1,652 patients who underwent RN for unilateral pM0 sporadic clear cell renal cell carcinoma between 1970 and 2000. A multivariate logistic regression model was used to determine the pathological features of the primary tumor that were associated with positive regional lymph nodes at RN. Results: There were 887 (54%) patients with no positive nodes (pN0), 57 (3%) with 1 positive node (pN1), 11 (1%) with 2 or more positive nodes (pN2) and 697 (42%) who did not have any lymph nodes dissected (pNx). Nuclear grade 3 or 4 (p ⬍0.001), presence of a sarcomatoid component (p ⬍0.001), tumor size 10 cm or greater (p ⫽ 0.005), tumor stage pT3 or pT4 (p ⫽ 0.017) and histological tumor necrosis (p ⫽ 0.051) were significantly associated with positive regional lymph nodes in a multivariate setting. These features can be used to identify candidates for extended lymph node dissection at the time of RN. For example, only 6 (0.6%) of the 1,031 patients with 0 or 1 of these features had positive lymph nodes at RN compared with 62 (10%) of the 621 patients with at least 2 of these features. Conclusions: The primary tumor pathological features of nuclear grade, sarcomatoid component, tumor size, stage and presence of tumor necrosis can be used to predict patients at the greatest risk for regional lymph node involvement at RN. KEY WORDS: carcinoma, renal cell; lymph nodes; lymph node excision
Renal cell carcinoma (RCC) has a variable natural history and encompasses hematogenous and lymphatic spread.1, 2 Although providing important prognostic and staging information, the survival value of extended lymph node dissection in clinically nonmetastatic kidney cancer in recent series has not been demonstrated.3, 4 In addition, the limits of extended lymph node dissection of RCC have not been adequately evaluated, and currently the debate as to its efficacy remains unresolved.5 Selective use of extended lymph node dissection in patients at high risk for positive lymph nodes based on the easily determined pathological features of the primary tumor determined at the time of surgery would provide a more evidence-based approach to identify those who would potentially benefit. The objective of our study was to develop an intraoperative risk factor protocol to predict the probability of regional lymph node involvement using a cohort of more than 2,000 patients surgically treated for clinically nonmetastatic RCC. In addition, a standardized extended lymph node dissection template for clinically nonmetastatic RCC is proposed.
MATERIALS AND METHODS
Clinical, surgical and pathological features. Upon approval from the Mayo Clinic Institutional Review Board, 2,028 patients whose first surgery between 1970 and 2000 was radical nephrectomy for unilateral pM0 sporadic RCC were selected for study. The clinical features studied included age at radical nephrectomy, sex and symptomatic disease at presentation. Patients with a palpable flank or abdominal mass, discomfort on the ipsilateral or contralateral side, constitutional symptoms including rash, sweats, weight loss and fatigue, gross hematuria, early satiety or decreased appetite or acute onset varicocele were considered symptomatic at presentation. Surgical margin status was recorded as positive or negative based on gross and microscopic examination of the radical nephrectomy specimen. In all cases the renal vein and renal vein margin were examined grossly. If renal vein involvement was identified, the renal vein margin was examined microscopically. All other margins were examined grossly and sampled for microscopic examination if suspicious for involvement. The interface between tumor and perinephric fat was sampled in all cases to evaluate perinephric fat invasion. The microscopic slides from all tumor specimens were reviewed by a urological pathologist (JCC) using a BX40 microscope (Olympus, Optical Corporation, Melville, New York) without knowledge of patient outcome. Histological subtype
Accepted for publication March 19, 2004. Study received Mayo Clinic Institutional Review Board approval. * Correspondence: Department of Urology, Mayo Clinic, 200 First St. SW, Rochester, Minnesota 55905 (telephone: 507-284-3982; FAX: 507-284-4951; e-mail: [email protected]). 465
466
LYMPH NODE DISSECTION USING PATHOLOGICAL FEATURES OF RENAL CELL CARCINOMA
was assessed following the 1997 Union Internationale Contre le Cancer and American Joint Committee on Cancer classification of RCC.6, 7 Tumors were staged using the 2003 TNM classification. Perinephric fat was defined as peripheral fat or pelvic sinus fat. Tumor thrombus level was classified as 0 —limited to the renal vein, detected clinically or during assessment of the pathological specimen, I— extending 2 cm or less above the renal vein, II— extending more than 2 cm above the renal vein but below the hepatic vein, III—at the level of or above the hepatic veins but below the diaphragm and IV— extending above the diaphragm. Nuclear grade was assigned using standardized criteria as previously defined.8 Histological tumor necrosis was defined as the presence of any microscopic coagulative tumor necrosis. Degenerative changes such as hyalinization, hemorrhage and fibrosis were not considered necrosis. A sarcomatoid component was defined as a spindle cell malignancy with histological appearance of a sarcoma. Statistical methods. Comparisons between RCC histological subtype and regional lymph node involvement at nephrectomy (pN0/pNx versus pN1/pN2) were evaluated using chisquare and Fisher’s exact tests. Cancer specific survival was estimated using the Kaplan-Meier method. The associations of regional lymph node involvement at nephrectomy with death from RCC were estimated using Cox proportional hazards regression models, and summarized with risk ratios and 95% confidence intervals. The associations of the clinical, surgical and pathological features with regional lymph node involvement at nephrectomy were assessed using logistic regression models and summarized with odds ratios and 95% CI. Statistical analyses were performed using the SAS software package (SAS Institute, Cary, North Carolina). RESULTS
RCC histological subtype by regional lymph node involvement at nephrectomy. Among the 2,028 patients studied 1,652 (81.5%) patients had clear cell RCC, 254 (12.5%) had papillary RCC, 98 (4.8%) had chromophobe RCC, 5 (0.3%) had collecting duct RCC, 3 (0.2%) had purely sarcomatoid RCC and 16 (0.8%) had RCC not otherwise specified. There were 1,057 (52.1%) patients with no positive regional nodes identified at nephrectomy (pN0), 882 (43.5%) with no nodes resected (pNx), 72 (3.6%) with metastases in a single regional lymph node (pN1) and 17 (0.8%) with metastases in more than 1 regional lymph node (pN2). There was no statistically significant association between RCC histological subtype and regional lymph node involvement at nephrectomy (table 1). Association of regional lymph node involvement at nephrectomy with death from RCC. Of the 1,652 patients with pM0 clear cell RCC 967 died, including 450 (27.2%) from RCC. Average time from radical nephrectomy to death from clear cell RCC was 4.5 years (median 2.7, range 0 to 26). Average time from nephrectomy to last followup for the 685 patients who were still alive at last followup was 9.6 years (median 7.9, range 0 to 33). Of the 1,584 patients with pN0/pNx clear
TABLE 1. RCC histological subtype by regional lymph node involvement at nephrectomy for 2,028 patients with pM0 RCC No. pN0/pNx (%)
No. pN1/pN2 (%)
Total pts 1,939 89 Histologic subtype: Clear cell 1,584 (95.9) 68 (4.1) Papillary 239 (94.1) 15 (5.9) Chromophobe 96 (98.0) 2 (2.0) Collecting duct 2 (40.0) 3 (60.0) Purely sarcomatoid 3 (100.0) 0 (0.0) RCC, not otherwise specified 15 (93.8) 1 (6.3) Clear cell versus papillary p ⫽ 0.193, clear cell versus chromophobe p ⫽ 0.430 and papillary versus chromophobe p ⫽ 0.169.
cell RCC 390 (24.6%) died compared with 60 (88.2%) of the 68 patients with pN1/pN2 clear cell RCC. Cancer specific survival by regional lymph node involvement is shown in the figure. The estimated cancer specific survival rates at 1, 5 and 10 years following radical nephrectomy for patients with pN0/pNx clear cell RCC were 95.5% (SE 0.5%, at risk 1,431), 82.1% (1.0%, 917) and 72.5% (1.3%, 486), respectively, compared with 52.2% (6.1%, 35), 20.9% (5.0%, 14) and 11.4% (4.0%, 6), respectively, for patients with pN1/pN2 clear cell RCC. Regional lymph node involvement at nephrectomy was significantly associated with death from clear cell RCC (risk ratio 7.87, 95% CI 5.98 –10.36, p ⬍0.001). Pathological features by regional lymph node involvement at nephrectomy. A comparison of the clinical, surgical and pathological features between patients with pN0/pNx and pN1/pN2 clear cell RCC is shown in table 2. The univariate associations of these features with regional lymph node involvement at nephrectomy are further summarized in table 3. The results of the multivariate modeling are shown in table 4. Patients with high grade (grade 3 or 4) clear cell RCC were more than 5 times more likely to have regional lymph node involvement at nephrectomy compared with those with low grade (grades 1 and 2) RCC (p ⬍0.001). Patients whose tumors exhibited a sarcomatoid component were approximately 4 times more likely to have regional lymph node involvement compared with those whose tumors did not show evidence of sarcomatoid changes (p ⬍0.001). Patients with tumors 10 cm or greater were more than twice as likely to have pN1 or pN2 disease than those with tumors less than 10 cm (p ⫽ 0.005). Patients with high stage (pT3a, pT3b, pT3c or pT4) RCC were twice as likely to have regional lymph node involvement compared with those with low stage (pT1a, pT1b or pT2) RCC (p ⫽ 0.017). After adjusting for grade, presence of a sarcomatoid component, tumor size and stage, patients with histological tumor necrosis were almost twice as likely to have regional lymph node involvement compared with those without necrosis (p ⫽ 0.051). There were 729 (44.1%) patients with low grade and low stage clear cell tumors less than 10 cm with no evidence of a sarcomatoid component or histological tumor necrosis (ie none of the 5 features in the multivariate model). Only 3 (0.4%) of these patients had regional lymph node involvement at nephrectomy. Among the 621 (37.6%) patients with at least 2 of the features in the multivariate model 62 (10.0%) had regional lymph node involvement at nephrectomy. The proportion of patients with regional lymph node involvement at nephrectomy by the number of features in the multivariate model is further summarized in table 5. DISCUSSION
The incidence of positive lymph nodes among patients treated with radical nephrectomy and lymphadenectomy in historical series ranges from 23% to 35%.1, 2, 5, 9 However, in the era of modern cross-sectional imaging the incidence of positive lymph nodes has decreased to 3% to 5%.2, 5 In this report of more than 2,000 patients over 3 decades the incidence of regional lymph node involvement was 4.4%. In the only prospectively performed phase III trial of extended lymph node dissection versus no lymph node dissection for clinically nonmetastatic RCC the incidence of positive lymph nodes was 3.3%.4 However, computed tomography was required before entry and randomization, and all patients with clinically enlarged lymph nodes were excluded from study. There has been no difference in progression, survival or morbidity between the 2 cohorts in this series although median survival end points have not been reached.4 Contemporary practice does not dictate that a lymph node dissection be performed but, rather, the extent of lymph node dissection is not standardized and is performed on an ad hoc basis. This prevents analysis of the true value of lymph node dissec-
LYMPH NODE DISSECTION USING PATHOLOGICAL FEATURES OF RENAL CELL CARCINOMA
467
Estimated cancer specific survival by regional lymph node involvement (pN0/pNx versus pN1/pN2) for 1,652 patients treated with radical nephrectomy for clear cell RCC.
TABLE 2. Clinical, surgical and pathological features by regional lymph node involvement at nephrectomy Feature
pN0/pNx
pN1/pN2
No. pts 1,584 68 Median age at nephrectomy (range) 65 (24–89) 61 (35–86) Median tumor cm (range) 6.0 (0.8–24.0) 10.0 (1.5–21.0) Sex: No. females (%) 571 (36.1) 26 (38.2) No. males (%) 1,013 (64.0) 42 (61.8) No. symptomatic at presentation (%) 1,110 (70.1) 62 (91.2) No. pos surgical margins (%) 11 (0.7) 2 (2.9) No. tumor 10 cm or greater (%) 318 (20.1) 37 (54.4) No. primary tumor stage (%): pT1a 383 (24.2) 2 (2.9) pT1b 430 (27.2) 8 (11.8) pT2 320 (20.2) 12 (17.7) pT3a 121 (7.6) 11 (16.2) pT3b 310 (19.6) 30 (44.1) pT3c 9 (0.6) 1 (1.5) pT4 11 (0.7) 4 (5.9) pT1 ⫹ pT2 1,133 (71.5) 22 (32.4) pT3 ⫹ pT4 451 (28.5) 46 (67.7) No. tumor thrombus (%) 322 (20.3) 31 (45.6) No. level tumor thrombus (%):* 0 208 (64.8) 17 (56.7) I 40 (12.5) 4 (13.3) II 44 (13.7) 7 (23.3) III 17 (5.3) 1 (3.3) IV 12 (3.7) 1 (3.3) No. nuclear grade (%): 1 177 (11.2) 0 (0.0) 2 771 (48.7) 6 (8.8) 3 555 (35.0) 42 (61.8) 4 81 (5.1) 20 (29.4) 1⫹2 948 (59.9) 6 (8.8) 3⫹4 636 (40.2) 62 (91.2) No. histological tumor necrosis (%) 384 (24.2) 48 (70.6) No. sarcomatoid component (%) 42 (2.7) 16 (23.5) No. cystic architecture (%) 43 (2.7) 0 (0.0) No. multifocality (%) 32 (2.0) 2 (2.9) * Summarized only for patients with tumor thrombus.
tion in RCC. Based on the decreased incidence of lymph node positive disease for contemporary stage T1, it would appear that the value of extended lymph node dissection for the majority of patients would be of little benefit.2, 5, 6, 10 Therefore, it would be of value to identify pathological features from the primary tumor that would place the patient at increased
TABLE 3. Univariate associations of clinical, surgical and pathological features with regional lymph node involvement at nephrectomy Feature
Odds Ratio (95% CI)
Age at nephrectomy 0.88 (0.71–1.09)* Sex: Female 1.0 (reference) Male 0.91 (0.55–1.50) Symptomatic at presentation 4.41 (1.90–10.27) Pos surgical margins 4.33 (0.94–19.95) Tumor 10 cm or greater 4.75 (2.90–7.78) Primary tumor stage: pT1a 1.0 (reference) pT1b 3.56 (0.75–16.88) pT2 7.18 (1.60–32.32) pT3a 17.41 (3.81–79.63) pT3b 18.53 (4.39–78.15) pT3c and pT4 47.87 (8.74–262.15) pT1 ⫹ pT2 1.0 (reference) pT3 ⫹ pT4 5.25 (3.12–8.83) Tumor thrombus 3.28 (2.01–5.37) Tumor thrombus level: 0 1.0 (reference) I, II, III and IV 1.34 (0.64–2.84)† Nuclear grade: 1⫹2 1.0 (reference) 3 11.96 (5.05–28.31) 4 39.01 (15.24–99.88) 3⫹4 15.40 (6.62–35.82) Histological tumor necrosis 7.50 (4.40–12.80) Sarcomatoid component 11.30 (5.96–21.39) Multifocality 1.47 (0.35–6.26) * Odds ratio represents a 10-year increase in age. † Odds ratio was adjusted for patients with tumor thrombus.
p Value 0.232 0.713 ⬍0.001 0.060 ⬍0.001 0.109 0.010 ⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001 0.441 ⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001 0.603
risk for lymph node involvement in the absence of metastatic disease and apply a standardized template dissection in a prospective manner. We analyzed more than 1,600 patients with clear cell RCC from the Mayo Clinic Nephrectomy Registry and performed a multivariate analysis that identified 5 histological features of primary clear cell RCC that placed them at increased risk for positive lymph nodes at radical nephrectomy. As the number of adverse features increases, the risk of positive lymph nodes at nephrectomy increases. The presence of these features (high stage, high nuclear grade, large size, presence of
468
LYMPH NODE DISSECTION USING PATHOLOGICAL FEATURES OF RENAL CELL CARCINOMA
TABLE 4. Multivariate model for regional lymph node involvement at nephrectomy Feature Nuclear grade: 1⫹2 3⫹4 Sarcomatoid component Tumor 10 cm or greater Primary tumor stage: pT1 ⫹ pT2 pT3 ⫹ pT4 Histological tumor necrosis
Odds Ratio (95% CI)
p Value
1.0 (reference) 5.25 (1.99–13.82) 4.11 (2.08–8.12) 2.17 (1.27–3.70)
⬍0.001 ⬍0.001 0.005
1.0 (reference) 2.00 (1.13–3.55) 1.86 (1.00–3.48)
0.017 0.051
TABLE 5. Proportion of patients with regional lymph node involvement at nephrectomy by number of features in multivariate model No. Features Total pts 0 1 2 3 4 5
No. pN0/pNx (%)
No. pN1/pN2 (%)
1,584 726 (99.6) 299 (99.0) 264 (95.7) 183 (87.6) 105 (86.8) 7 (46.7)
68 3 (0.4) 3 (1.0) 12 (4.4) 26 (12.4) 16 (13.2) 8 (53.3)
a sarcomatoid component and presence of histological tumor necrosis), easily attainable at the time of frozen section analysis, identify substantial risk cut points predictive of regional lymph node involvement and can be used to determine the need for an extended lymph node dissection. We have arbitrarily established a risk cut point of 2 features that would place a patient at an approximately 4-fold increased risk of regional lymph node involvement in the absence of these features but up to 50% risk when all 5 features are present (table 5). We believe that this protocol provides a rational approach to performing extended lymph node dissection for clear cell RCC that will improve staging, provide greater prognostic information and improve clinical trial design for these high risk patients. Cancer specific survival is dramatically worse in patients with RCC and regional lymph node involvement. In patients with involved lymph node disease the 5-year survival rate was approximately 35% in the series reported by Robson et al.1 Skinner et al also reported reduced but improved survival in node positive cases in which an extended lymph node dissection at the time of radical nephrectomy was performed.9 In their report the 5-year cancer specific survival rate for patients with pN1/N2 clear cell RCC was 21%. Pantuck et al reported a survival advantage for patients with node positive disease who underwent lymph node dissection and adjuvant immunotherapy versus those with positive nodes who did not undergo lymph node dissection but received immunotherapy.5 In their multivariate analysis patients who did not undergo lymph node dissection were 3 times more likely to die than those who did not, but surprisingly the extent of node dissection did not matter (p ⫽ 0.57). Overall, the 5-year survival for patients with N⫹/M0 disease was 25%.5 Contributing to the lack of knowledge surrounding the benefits of lymph node dissection for RCC is the absence of an accepted standard approach to dissecting the retroperitoneum.11 Lymph node dissection based on the classic anatomical description by Parker involves removal of all nodes anterior and posterior to the great vessels from the crus of the diaphragm to bifurcation of the vessels, including the interaortocaval region on the right side, which results in approximately twice the rate of positive lymph nodes from 10% to 14% to 17% to 29%.12–14 Terrone et al observed the impact of the number of nodes removed in a large series of patients as the incidence doubled from 10% to 21% when 13 or more
nodes were excised.14 The latter figure corresponds to autopsy series of more advanced lesions.13 Therefore, for patients to receive maximum potential benefit from lymph node dissection, it appears that an extended approach should be performed. This approach should include dissection of the vena cava from crus of the diaphragm to bifurcation of the vessels to include precaval, paracaval and retrocaval nodes, and interaortocaval nodes for right tumors, and for left-sided lesions, preaortic, paraaortic and retroaortic nodes from crus to bifurcation. Limitations of this study include its retrospective design and inclusion of patients treated over a 3-decade period, reflecting changes in clinical staging and imaging, and lack of standardized approach to regional lymph node dissection. In addition, there were too few patients with regional lymph node involvement at radical nephrectomy for papillary RCC to develop a protocol for this histological subtype. However, we believe the results of this study warrant additional prospective studies to analyze the true impact of extended lymph node dissection for RCC. CONCLUSIONS
To determine if extended lymph node dissection will result in a survival benefit with acceptable morbidity requires properly performed prospective clinical trials. Since the incidence of lymph node positivity in the era of modern imaging is low and the ongoing results of the only phase III randomized trial reveal equivalent survival and morbidity in performing extended lymph node dissection at the time of radical nephrectomy, the procedure should be used only in those patients who would potentially receive the most benefit.4 The results of our institutional multivariate analysis suggest that patients at greatest risk for regional lymph node involvement at radical nephrectomy for clinically nonmetastatic disease can be identified. A properly performed standardized template dissection for patients with high risk features will hopefully provide the information for staging, prognosis and morbidity. We believe that the proposed protocol and standardized extended lymph node dissection template offer the best approach to this difficult question of RCC tumor biology. REFERENCES
1. Robson, C. J., Churchill, B. M. and Anderson, W.: The results of radical nephrectomy for renal cell carcinoma. J Urol, 101: 297, 1969 2. Ward, J. F., Blute, M. L., Cheville, J. C., Lohse, C. M., Weaver, A. L. and Zincke, H.: The influence of pNx/pN0 grouping in a multivariate setting for outcome modeling in patients with clear cell renal cell carcinoma. J Urol, 168: 56, 2002 3. Giberti, C., Oneto, F., Martorana, G., Rovida, S. and Carmignani, G.: Radical nephrectomy for renal cell carcinoma: long-term results and prognostic factors on a series of 328 cases. Eur Urol, 31: 40, 1997 4. Blom, J. H., van Poppel, H., Marechal, J. M., Jacquim, D., Sylvester, R., Schroder, F. H. et al: Radical nephrectomy with and without lymph node dissection: preliminary results of the EORTC randomized phase III protocol 30881. EORTC Genitourinary Group. Eur Urol, 36: 570, 1999 5. Pantuck, A. J., Zisman, A., Dorey, F., Chao, D. H., Han, K.-r., Said, J. et al: Renal cell carcinoma with retroperitoneal lymph nodes: the role of lymph node dissection. J Urol, 169: 2076, 2003 6. Storkel, S., Eble, J. N., Adlakha, K., Amin, M., Blute, M. L. and Bostwick, D. G.: Classification of renal cell carcinoma: workgroup No. 1. Union Internationale Contre le Cancer (UICC) and the American Joint Committee on Cancer (AJCC). Cancer, 80: 987, 1997 7. Cheville, J. C., Lohse, C. M., Zincke, H., Weaver, A. L. and Blute, M. L.: Comparisons of outcome and prognostic features among histologic subtypes of renal cell carcinoma. Am J Surg Pathol, 27: 612, 2003 8. Lohse, C. M., Blute, M. L., Zincke, H., Weaver, A. L. and Cheville, J. C.: Comparison of standardized and non-
LYMPH NODE DISSECTION USING PATHOLOGICAL FEATURES OF RENAL CELL CARCINOMA
9. 10. 11. 12.
standardized nuclear grade of renal cell carcinoma to predict outcome among 2,042 patients. Am J Clin Pathol, 118: 877, 2002 Skinner, D. G., Vermillion, C. D. and Colvin, R. B.: The surgical management of renal cell carcinoma. J Urol, 107: 705, 1972 Novick, A. C.: Nephron-sparing surgery for renal cell carcinoma. Annu Rev Med, 53: 393, 2002 Guiliani, L., Giberti, C., Martorana, G. and Rovida, S.: Radical extensive surgery for renal cell carcinoma: long-term results and prognostic factors. J Urol, 143: 468, 1990 Parker, A. E.: Studies on the main posterior lymph channels of
469
the abdomen and their connections with the lymphatics of the genitourinary system. Am J Anat, 56: 409, 1935 13. Hultsen, L., Rosencrantz, M., Seeman, T., Wahlquist, L. and Ahren, C.: Occurrence and localization of lymph node metastases in renal cell carcinoma. A lymphographic and histopathological investigation in connection with nephrectomy. Scand J Urol Nephrol, 3: 129, 1969 14. Terrone, C., Guercio, S., De Luca, S., Poggio, M., Castelli, E., Scoffone, C. et al: The number of lymph nodes examined and staging accuracy in renal cell carcinoma. BJU Int, 91: 37, 2003
Vol. 172, 465– 469, August 2004 Printed in U.S.A.
DOI: 10.1097/01.ju.0000129815.91927.85
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 MICHAEL L. BLUTE,* BRADLEY C. LEIBOVICH, JOHN C. CHEVILLE, CHRISTINE M. LOHSE AND HORST ZINCKE* From the Departments of Urology (MLB, BCL, HZ), Pathology (JCC) and Health Sciences Research (CML), Mayo Clinic, Rochester, Minnesota
ABSTRACT
Purpose: We determined the primary pathological features of clear cell renal cell carcinoma that are predictive of positive regional lymph nodes at radical nephrectomy (RN) and developed a protocol for the selective use of extended lymph node dissection. Materials and Methods: We studied 1,652 patients who underwent RN for unilateral pM0 sporadic clear cell renal cell carcinoma between 1970 and 2000. A multivariate logistic regression model was used to determine the pathological features of the primary tumor that were associated with positive regional lymph nodes at RN. Results: There were 887 (54%) patients with no positive nodes (pN0), 57 (3%) with 1 positive node (pN1), 11 (1%) with 2 or more positive nodes (pN2) and 697 (42%) who did not have any lymph nodes dissected (pNx). Nuclear grade 3 or 4 (p ⬍0.001), presence of a sarcomatoid component (p ⬍0.001), tumor size 10 cm or greater (p ⫽ 0.005), tumor stage pT3 or pT4 (p ⫽ 0.017) and histological tumor necrosis (p ⫽ 0.051) were significantly associated with positive regional lymph nodes in a multivariate setting. These features can be used to identify candidates for extended lymph node dissection at the time of RN. For example, only 6 (0.6%) of the 1,031 patients with 0 or 1 of these features had positive lymph nodes at RN compared with 62 (10%) of the 621 patients with at least 2 of these features. Conclusions: The primary tumor pathological features of nuclear grade, sarcomatoid component, tumor size, stage and presence of tumor necrosis can be used to predict patients at the greatest risk for regional lymph node involvement at RN. KEY WORDS: carcinoma, renal cell; lymph nodes; lymph node excision
Renal cell carcinoma (RCC) has a variable natural history and encompasses hematogenous and lymphatic spread.1, 2 Although providing important prognostic and staging information, the survival value of extended lymph node dissection in clinically nonmetastatic kidney cancer in recent series has not been demonstrated.3, 4 In addition, the limits of extended lymph node dissection of RCC have not been adequately evaluated, and currently the debate as to its efficacy remains unresolved.5 Selective use of extended lymph node dissection in patients at high risk for positive lymph nodes based on the easily determined pathological features of the primary tumor determined at the time of surgery would provide a more evidence-based approach to identify those who would potentially benefit. The objective of our study was to develop an intraoperative risk factor protocol to predict the probability of regional lymph node involvement using a cohort of more than 2,000 patients surgically treated for clinically nonmetastatic RCC. In addition, a standardized extended lymph node dissection template for clinically nonmetastatic RCC is proposed.
MATERIALS AND METHODS
Clinical, surgical and pathological features. Upon approval from the Mayo Clinic Institutional Review Board, 2,028 patients whose first surgery between 1970 and 2000 was radical nephrectomy for unilateral pM0 sporadic RCC were selected for study. The clinical features studied included age at radical nephrectomy, sex and symptomatic disease at presentation. Patients with a palpable flank or abdominal mass, discomfort on the ipsilateral or contralateral side, constitutional symptoms including rash, sweats, weight loss and fatigue, gross hematuria, early satiety or decreased appetite or acute onset varicocele were considered symptomatic at presentation. Surgical margin status was recorded as positive or negative based on gross and microscopic examination of the radical nephrectomy specimen. In all cases the renal vein and renal vein margin were examined grossly. If renal vein involvement was identified, the renal vein margin was examined microscopically. All other margins were examined grossly and sampled for microscopic examination if suspicious for involvement. The interface between tumor and perinephric fat was sampled in all cases to evaluate perinephric fat invasion. The microscopic slides from all tumor specimens were reviewed by a urological pathologist (JCC) using a BX40 microscope (Olympus, Optical Corporation, Melville, New York) without knowledge of patient outcome. Histological subtype
Accepted for publication March 19, 2004. Study received Mayo Clinic Institutional Review Board approval. * Correspondence: Department of Urology, Mayo Clinic, 200 First St. SW, Rochester, Minnesota 55905 (telephone: 507-284-3982; FAX: 507-284-4951; e-mail: [email protected]). 465
466
LYMPH NODE DISSECTION USING PATHOLOGICAL FEATURES OF RENAL CELL CARCINOMA
was assessed following the 1997 Union Internationale Contre le Cancer and American Joint Committee on Cancer classification of RCC.6, 7 Tumors were staged using the 2003 TNM classification. Perinephric fat was defined as peripheral fat or pelvic sinus fat. Tumor thrombus level was classified as 0 —limited to the renal vein, detected clinically or during assessment of the pathological specimen, I— extending 2 cm or less above the renal vein, II— extending more than 2 cm above the renal vein but below the hepatic vein, III—at the level of or above the hepatic veins but below the diaphragm and IV— extending above the diaphragm. Nuclear grade was assigned using standardized criteria as previously defined.8 Histological tumor necrosis was defined as the presence of any microscopic coagulative tumor necrosis. Degenerative changes such as hyalinization, hemorrhage and fibrosis were not considered necrosis. A sarcomatoid component was defined as a spindle cell malignancy with histological appearance of a sarcoma. Statistical methods. Comparisons between RCC histological subtype and regional lymph node involvement at nephrectomy (pN0/pNx versus pN1/pN2) were evaluated using chisquare and Fisher’s exact tests. Cancer specific survival was estimated using the Kaplan-Meier method. The associations of regional lymph node involvement at nephrectomy with death from RCC were estimated using Cox proportional hazards regression models, and summarized with risk ratios and 95% confidence intervals. The associations of the clinical, surgical and pathological features with regional lymph node involvement at nephrectomy were assessed using logistic regression models and summarized with odds ratios and 95% CI. Statistical analyses were performed using the SAS software package (SAS Institute, Cary, North Carolina). RESULTS
RCC histological subtype by regional lymph node involvement at nephrectomy. Among the 2,028 patients studied 1,652 (81.5%) patients had clear cell RCC, 254 (12.5%) had papillary RCC, 98 (4.8%) had chromophobe RCC, 5 (0.3%) had collecting duct RCC, 3 (0.2%) had purely sarcomatoid RCC and 16 (0.8%) had RCC not otherwise specified. There were 1,057 (52.1%) patients with no positive regional nodes identified at nephrectomy (pN0), 882 (43.5%) with no nodes resected (pNx), 72 (3.6%) with metastases in a single regional lymph node (pN1) and 17 (0.8%) with metastases in more than 1 regional lymph node (pN2). There was no statistically significant association between RCC histological subtype and regional lymph node involvement at nephrectomy (table 1). Association of regional lymph node involvement at nephrectomy with death from RCC. Of the 1,652 patients with pM0 clear cell RCC 967 died, including 450 (27.2%) from RCC. Average time from radical nephrectomy to death from clear cell RCC was 4.5 years (median 2.7, range 0 to 26). Average time from nephrectomy to last followup for the 685 patients who were still alive at last followup was 9.6 years (median 7.9, range 0 to 33). Of the 1,584 patients with pN0/pNx clear
TABLE 1. RCC histological subtype by regional lymph node involvement at nephrectomy for 2,028 patients with pM0 RCC No. pN0/pNx (%)
No. pN1/pN2 (%)
Total pts 1,939 89 Histologic subtype: Clear cell 1,584 (95.9) 68 (4.1) Papillary 239 (94.1) 15 (5.9) Chromophobe 96 (98.0) 2 (2.0) Collecting duct 2 (40.0) 3 (60.0) Purely sarcomatoid 3 (100.0) 0 (0.0) RCC, not otherwise specified 15 (93.8) 1 (6.3) Clear cell versus papillary p ⫽ 0.193, clear cell versus chromophobe p ⫽ 0.430 and papillary versus chromophobe p ⫽ 0.169.
cell RCC 390 (24.6%) died compared with 60 (88.2%) of the 68 patients with pN1/pN2 clear cell RCC. Cancer specific survival by regional lymph node involvement is shown in the figure. The estimated cancer specific survival rates at 1, 5 and 10 years following radical nephrectomy for patients with pN0/pNx clear cell RCC were 95.5% (SE 0.5%, at risk 1,431), 82.1% (1.0%, 917) and 72.5% (1.3%, 486), respectively, compared with 52.2% (6.1%, 35), 20.9% (5.0%, 14) and 11.4% (4.0%, 6), respectively, for patients with pN1/pN2 clear cell RCC. Regional lymph node involvement at nephrectomy was significantly associated with death from clear cell RCC (risk ratio 7.87, 95% CI 5.98 –10.36, p ⬍0.001). Pathological features by regional lymph node involvement at nephrectomy. A comparison of the clinical, surgical and pathological features between patients with pN0/pNx and pN1/pN2 clear cell RCC is shown in table 2. The univariate associations of these features with regional lymph node involvement at nephrectomy are further summarized in table 3. The results of the multivariate modeling are shown in table 4. Patients with high grade (grade 3 or 4) clear cell RCC were more than 5 times more likely to have regional lymph node involvement at nephrectomy compared with those with low grade (grades 1 and 2) RCC (p ⬍0.001). Patients whose tumors exhibited a sarcomatoid component were approximately 4 times more likely to have regional lymph node involvement compared with those whose tumors did not show evidence of sarcomatoid changes (p ⬍0.001). Patients with tumors 10 cm or greater were more than twice as likely to have pN1 or pN2 disease than those with tumors less than 10 cm (p ⫽ 0.005). Patients with high stage (pT3a, pT3b, pT3c or pT4) RCC were twice as likely to have regional lymph node involvement compared with those with low stage (pT1a, pT1b or pT2) RCC (p ⫽ 0.017). After adjusting for grade, presence of a sarcomatoid component, tumor size and stage, patients with histological tumor necrosis were almost twice as likely to have regional lymph node involvement compared with those without necrosis (p ⫽ 0.051). There were 729 (44.1%) patients with low grade and low stage clear cell tumors less than 10 cm with no evidence of a sarcomatoid component or histological tumor necrosis (ie none of the 5 features in the multivariate model). Only 3 (0.4%) of these patients had regional lymph node involvement at nephrectomy. Among the 621 (37.6%) patients with at least 2 of the features in the multivariate model 62 (10.0%) had regional lymph node involvement at nephrectomy. The proportion of patients with regional lymph node involvement at nephrectomy by the number of features in the multivariate model is further summarized in table 5. DISCUSSION
The incidence of positive lymph nodes among patients treated with radical nephrectomy and lymphadenectomy in historical series ranges from 23% to 35%.1, 2, 5, 9 However, in the era of modern cross-sectional imaging the incidence of positive lymph nodes has decreased to 3% to 5%.2, 5 In this report of more than 2,000 patients over 3 decades the incidence of regional lymph node involvement was 4.4%. In the only prospectively performed phase III trial of extended lymph node dissection versus no lymph node dissection for clinically nonmetastatic RCC the incidence of positive lymph nodes was 3.3%.4 However, computed tomography was required before entry and randomization, and all patients with clinically enlarged lymph nodes were excluded from study. There has been no difference in progression, survival or morbidity between the 2 cohorts in this series although median survival end points have not been reached.4 Contemporary practice does not dictate that a lymph node dissection be performed but, rather, the extent of lymph node dissection is not standardized and is performed on an ad hoc basis. This prevents analysis of the true value of lymph node dissec-
LYMPH NODE DISSECTION USING PATHOLOGICAL FEATURES OF RENAL CELL CARCINOMA
467
Estimated cancer specific survival by regional lymph node involvement (pN0/pNx versus pN1/pN2) for 1,652 patients treated with radical nephrectomy for clear cell RCC.
TABLE 2. Clinical, surgical and pathological features by regional lymph node involvement at nephrectomy Feature
pN0/pNx
pN1/pN2
No. pts 1,584 68 Median age at nephrectomy (range) 65 (24–89) 61 (35–86) Median tumor cm (range) 6.0 (0.8–24.0) 10.0 (1.5–21.0) Sex: No. females (%) 571 (36.1) 26 (38.2) No. males (%) 1,013 (64.0) 42 (61.8) No. symptomatic at presentation (%) 1,110 (70.1) 62 (91.2) No. pos surgical margins (%) 11 (0.7) 2 (2.9) No. tumor 10 cm or greater (%) 318 (20.1) 37 (54.4) No. primary tumor stage (%): pT1a 383 (24.2) 2 (2.9) pT1b 430 (27.2) 8 (11.8) pT2 320 (20.2) 12 (17.7) pT3a 121 (7.6) 11 (16.2) pT3b 310 (19.6) 30 (44.1) pT3c 9 (0.6) 1 (1.5) pT4 11 (0.7) 4 (5.9) pT1 ⫹ pT2 1,133 (71.5) 22 (32.4) pT3 ⫹ pT4 451 (28.5) 46 (67.7) No. tumor thrombus (%) 322 (20.3) 31 (45.6) No. level tumor thrombus (%):* 0 208 (64.8) 17 (56.7) I 40 (12.5) 4 (13.3) II 44 (13.7) 7 (23.3) III 17 (5.3) 1 (3.3) IV 12 (3.7) 1 (3.3) No. nuclear grade (%): 1 177 (11.2) 0 (0.0) 2 771 (48.7) 6 (8.8) 3 555 (35.0) 42 (61.8) 4 81 (5.1) 20 (29.4) 1⫹2 948 (59.9) 6 (8.8) 3⫹4 636 (40.2) 62 (91.2) No. histological tumor necrosis (%) 384 (24.2) 48 (70.6) No. sarcomatoid component (%) 42 (2.7) 16 (23.5) No. cystic architecture (%) 43 (2.7) 0 (0.0) No. multifocality (%) 32 (2.0) 2 (2.9) * Summarized only for patients with tumor thrombus.
tion in RCC. Based on the decreased incidence of lymph node positive disease for contemporary stage T1, it would appear that the value of extended lymph node dissection for the majority of patients would be of little benefit.2, 5, 6, 10 Therefore, it would be of value to identify pathological features from the primary tumor that would place the patient at increased
TABLE 3. Univariate associations of clinical, surgical and pathological features with regional lymph node involvement at nephrectomy Feature
Odds Ratio (95% CI)
Age at nephrectomy 0.88 (0.71–1.09)* Sex: Female 1.0 (reference) Male 0.91 (0.55–1.50) Symptomatic at presentation 4.41 (1.90–10.27) Pos surgical margins 4.33 (0.94–19.95) Tumor 10 cm or greater 4.75 (2.90–7.78) Primary tumor stage: pT1a 1.0 (reference) pT1b 3.56 (0.75–16.88) pT2 7.18 (1.60–32.32) pT3a 17.41 (3.81–79.63) pT3b 18.53 (4.39–78.15) pT3c and pT4 47.87 (8.74–262.15) pT1 ⫹ pT2 1.0 (reference) pT3 ⫹ pT4 5.25 (3.12–8.83) Tumor thrombus 3.28 (2.01–5.37) Tumor thrombus level: 0 1.0 (reference) I, II, III and IV 1.34 (0.64–2.84)† Nuclear grade: 1⫹2 1.0 (reference) 3 11.96 (5.05–28.31) 4 39.01 (15.24–99.88) 3⫹4 15.40 (6.62–35.82) Histological tumor necrosis 7.50 (4.40–12.80) Sarcomatoid component 11.30 (5.96–21.39) Multifocality 1.47 (0.35–6.26) * Odds ratio represents a 10-year increase in age. † Odds ratio was adjusted for patients with tumor thrombus.
p Value 0.232 0.713 ⬍0.001 0.060 ⬍0.001 0.109 0.010 ⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001 0.441 ⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001 0.603
risk for lymph node involvement in the absence of metastatic disease and apply a standardized template dissection in a prospective manner. We analyzed more than 1,600 patients with clear cell RCC from the Mayo Clinic Nephrectomy Registry and performed a multivariate analysis that identified 5 histological features of primary clear cell RCC that placed them at increased risk for positive lymph nodes at radical nephrectomy. As the number of adverse features increases, the risk of positive lymph nodes at nephrectomy increases. The presence of these features (high stage, high nuclear grade, large size, presence of
468
LYMPH NODE DISSECTION USING PATHOLOGICAL FEATURES OF RENAL CELL CARCINOMA
TABLE 4. Multivariate model for regional lymph node involvement at nephrectomy Feature Nuclear grade: 1⫹2 3⫹4 Sarcomatoid component Tumor 10 cm or greater Primary tumor stage: pT1 ⫹ pT2 pT3 ⫹ pT4 Histological tumor necrosis
Odds Ratio (95% CI)
p Value
1.0 (reference) 5.25 (1.99–13.82) 4.11 (2.08–8.12) 2.17 (1.27–3.70)
⬍0.001 ⬍0.001 0.005
1.0 (reference) 2.00 (1.13–3.55) 1.86 (1.00–3.48)
0.017 0.051
TABLE 5. Proportion of patients with regional lymph node involvement at nephrectomy by number of features in multivariate model No. Features Total pts 0 1 2 3 4 5
No. pN0/pNx (%)
No. pN1/pN2 (%)
1,584 726 (99.6) 299 (99.0) 264 (95.7) 183 (87.6) 105 (86.8) 7 (46.7)
68 3 (0.4) 3 (1.0) 12 (4.4) 26 (12.4) 16 (13.2) 8 (53.3)
a sarcomatoid component and presence of histological tumor necrosis), easily attainable at the time of frozen section analysis, identify substantial risk cut points predictive of regional lymph node involvement and can be used to determine the need for an extended lymph node dissection. We have arbitrarily established a risk cut point of 2 features that would place a patient at an approximately 4-fold increased risk of regional lymph node involvement in the absence of these features but up to 50% risk when all 5 features are present (table 5). We believe that this protocol provides a rational approach to performing extended lymph node dissection for clear cell RCC that will improve staging, provide greater prognostic information and improve clinical trial design for these high risk patients. Cancer specific survival is dramatically worse in patients with RCC and regional lymph node involvement. In patients with involved lymph node disease the 5-year survival rate was approximately 35% in the series reported by Robson et al.1 Skinner et al also reported reduced but improved survival in node positive cases in which an extended lymph node dissection at the time of radical nephrectomy was performed.9 In their report the 5-year cancer specific survival rate for patients with pN1/N2 clear cell RCC was 21%. Pantuck et al reported a survival advantage for patients with node positive disease who underwent lymph node dissection and adjuvant immunotherapy versus those with positive nodes who did not undergo lymph node dissection but received immunotherapy.5 In their multivariate analysis patients who did not undergo lymph node dissection were 3 times more likely to die than those who did not, but surprisingly the extent of node dissection did not matter (p ⫽ 0.57). Overall, the 5-year survival for patients with N⫹/M0 disease was 25%.5 Contributing to the lack of knowledge surrounding the benefits of lymph node dissection for RCC is the absence of an accepted standard approach to dissecting the retroperitoneum.11 Lymph node dissection based on the classic anatomical description by Parker involves removal of all nodes anterior and posterior to the great vessels from the crus of the diaphragm to bifurcation of the vessels, including the interaortocaval region on the right side, which results in approximately twice the rate of positive lymph nodes from 10% to 14% to 17% to 29%.12–14 Terrone et al observed the impact of the number of nodes removed in a large series of patients as the incidence doubled from 10% to 21% when 13 or more
nodes were excised.14 The latter figure corresponds to autopsy series of more advanced lesions.13 Therefore, for patients to receive maximum potential benefit from lymph node dissection, it appears that an extended approach should be performed. This approach should include dissection of the vena cava from crus of the diaphragm to bifurcation of the vessels to include precaval, paracaval and retrocaval nodes, and interaortocaval nodes for right tumors, and for left-sided lesions, preaortic, paraaortic and retroaortic nodes from crus to bifurcation. Limitations of this study include its retrospective design and inclusion of patients treated over a 3-decade period, reflecting changes in clinical staging and imaging, and lack of standardized approach to regional lymph node dissection. In addition, there were too few patients with regional lymph node involvement at radical nephrectomy for papillary RCC to develop a protocol for this histological subtype. However, we believe the results of this study warrant additional prospective studies to analyze the true impact of extended lymph node dissection for RCC. CONCLUSIONS
To determine if extended lymph node dissection will result in a survival benefit with acceptable morbidity requires properly performed prospective clinical trials. Since the incidence of lymph node positivity in the era of modern imaging is low and the ongoing results of the only phase III randomized trial reveal equivalent survival and morbidity in performing extended lymph node dissection at the time of radical nephrectomy, the procedure should be used only in those patients who would potentially receive the most benefit.4 The results of our institutional multivariate analysis suggest that patients at greatest risk for regional lymph node involvement at radical nephrectomy for clinically nonmetastatic disease can be identified. A properly performed standardized template dissection for patients with high risk features will hopefully provide the information for staging, prognosis and morbidity. We believe that the proposed protocol and standardized extended lymph node dissection template offer the best approach to this difficult question of RCC tumor biology. REFERENCES
1. Robson, C. J., Churchill, B. M. and Anderson, W.: The results of radical nephrectomy for renal cell carcinoma. J Urol, 101: 297, 1969 2. Ward, J. F., Blute, M. L., Cheville, J. C., Lohse, C. M., Weaver, A. L. and Zincke, H.: The influence of pNx/pN0 grouping in a multivariate setting for outcome modeling in patients with clear cell renal cell carcinoma. J Urol, 168: 56, 2002 3. Giberti, C., Oneto, F., Martorana, G., Rovida, S. and Carmignani, G.: Radical nephrectomy for renal cell carcinoma: long-term results and prognostic factors on a series of 328 cases. Eur Urol, 31: 40, 1997 4. Blom, J. H., van Poppel, H., Marechal, J. M., Jacquim, D., Sylvester, R., Schroder, F. H. et al: Radical nephrectomy with and without lymph node dissection: preliminary results of the EORTC randomized phase III protocol 30881. EORTC Genitourinary Group. Eur Urol, 36: 570, 1999 5. Pantuck, A. J., Zisman, A., Dorey, F., Chao, D. H., Han, K.-r., Said, J. et al: Renal cell carcinoma with retroperitoneal lymph nodes: the role of lymph node dissection. J Urol, 169: 2076, 2003 6. Storkel, S., Eble, J. N., Adlakha, K., Amin, M., Blute, M. L. and Bostwick, D. G.: Classification of renal cell carcinoma: workgroup No. 1. Union Internationale Contre le Cancer (UICC) and the American Joint Committee on Cancer (AJCC). Cancer, 80: 987, 1997 7. Cheville, J. C., Lohse, C. M., Zincke, H., Weaver, A. L. and Blute, M. L.: Comparisons of outcome and prognostic features among histologic subtypes of renal cell carcinoma. Am J Surg Pathol, 27: 612, 2003 8. Lohse, C. M., Blute, M. L., Zincke, H., Weaver, A. L. and Cheville, J. C.: Comparison of standardized and non-
LYMPH NODE DISSECTION USING PATHOLOGICAL FEATURES OF RENAL CELL CARCINOMA
9. 10. 11. 12.
standardized nuclear grade of renal cell carcinoma to predict outcome among 2,042 patients. Am J Clin Pathol, 118: 877, 2002 Skinner, D. G., Vermillion, C. D. and Colvin, R. B.: The surgical management of renal cell carcinoma. J Urol, 107: 705, 1972 Novick, A. C.: Nephron-sparing surgery for renal cell carcinoma. Annu Rev Med, 53: 393, 2002 Guiliani, L., Giberti, C., Martorana, G. and Rovida, S.: Radical extensive surgery for renal cell carcinoma: long-term results and prognostic factors. J Urol, 143: 468, 1990 Parker, A. E.: Studies on the main posterior lymph channels of
469
the abdomen and their connections with the lymphatics of the genitourinary system. Am J Anat, 56: 409, 1935 13. Hultsen, L., Rosencrantz, M., Seeman, T., Wahlquist, L. and Ahren, C.: Occurrence and localization of lymph node metastases in renal cell carcinoma. A lymphographic and histopathological investigation in connection with nephrectomy. Scand J Urol Nephrol, 3: 129, 1969 14. Terrone, C., Guercio, S., De Luca, S., Poggio, M., Castelli, E., Scoffone, C. et al: The number of lymph nodes examined and staging accuracy in renal cell carcinoma. BJU Int, 91: 37, 2003