| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ORIGINAL ARTICLES |
From the Departments of Surgery, Memorial Sloan-Kettering Cancer Center (BMC, MSB, JJL, DGC) New York, New York; and Royal Melbourne Hospital (BM), Melbourne, Australia.
Correspondence: Address correspondence and reprint requests to: Bryan M. Clary, MD, Box 3247 Duke University Medical Center, Durham, NC 27710; Fax: 919-681-7508; E-mail: clary001{at}mc.duke.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONREFERENCES |
|---|
Methods: All patients who underwent SLND (1991–1998) and ELND (1974–1994) were identified from single institution melanoma databases.
Results: A total of 152 and 329 patients with early-stage melanoma of the extremity underwent SLND and ELND, respectively. Nodal metastases were present in 44 of 329 ELND patients (13%) and in 31 of 152 SLND patients (20%). Early relapse-free and disease-specific survivals were similar for the entire population, although in patients at higher risk for recurrence (age >50 years, thickness >3.0 mm), there was an increased rate of relapse in the SLND group (P = .04). Among all sites of early recurrences, locoregional sites were more common in patients undergoing SLND (72%) compared with ELND (39%, P < .01). SLN-negative patients with nodal recurrence had evidence of metastases on retrospective enhanced pathologic analysis in four of seven cases.
Conclusions: Although overall relapse-free and disease-specific survivals are similar, there is a higher rate of relapse in a subset of SLND node-negative patients who are at high risk for nodal metastases. ELND and SLNB should not be thought of as equivalent approaches until studies with longer follow-up are available.
Key Words: Melanoma • Sentinel lymph node • Recurrence • Elective lymph node dissection
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONREFERENCES |
|---|
In an effort to improve the staging of these patients and allow for a more selective application of a potentially morbid LND, Morton and colleagues described the technique of intraoperative lymphatic mapping and sentinel lymph node biopsy in 1992. This procedure is based on the hypothesis that melanoma metastasizes to the regional lymph nodes in an orderly and defined manner, such that dye or certain radioactive labeled compounds when injected intradermal at the site of the primary, travel first to a specific draining lymph node that represents the status of the entire nodal basin. Sentinel lymph node mapping and biopsy (SLNB) has been shown to be highly accurate in staging nodal basins at risk for regional metastases in patients with primary cutaneous melanoma. In the experience reported by Morton and colleagues in 1992 where nonsentinel nodes (non-SLN) from immediate back-up dissections were subjected to routine examination and immunohistochemical analysis, the sentinel node was positive in 38 of 40 patients found to have any positive node (SLN or non-SLN). In this study, the SLN was reflective of the status of the nodal basins in 192 of 194 clinical stage 1 patients (99%).7 Similar rates of accuracy have also been reported by other groups.8,9
As the technique of intraoperative lymphatic mapping and biopsy is relatively new, reports describing the clinical outcome are limited. Despite this lack of long-term follow-up, a policy of SLNB with selective application of lymph node dissection to those with positive nodes (collectively referred to as selective lymph node dissection or SLND) has been widely adopted as a staging procedure and therapeutic alternative to ELND. The reasons for this include the ability of the pathologist to much more thoroughly evaluate the sentinel node with step-sectioning and immunohistochemistry (compared with the multiple nodes of a complete ELND) and the significantly less morbid nature of the selective lymph node dissection approach. In an effort to address the clinical efficacy of this procedure relative to ELND, the authors present an analysis of a large single institutional experience of patients undergoing ELND and SLND.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONREFERENCES |
|---|
|
Conventional histologic examination (hematoxylin and eosin) was routinely applied to bisected SLNs and non-SLNs. Beginning in late 1997, the SLNs were routinely subjected to hematoxylin and eosin (H&E) staining of serial sectioned specimens and immunohistochemical staining (S-100, HMB-45) in the event that the initial H&E of the bisected specimen did not reveal evidence of metastatic disease. Examination of lymphadenectomy specimens and non-SLNs harvested during SLNB were analyzed by conventional H&E staining of bisected lymph node specimens.
Statistical Analysis
Association between factors was determined by 
2 analysis for categorical variables. Means testing of continuous variables was performed with the Student’s t-test. Disease recurrence and survival curves were constructed using the Kaplan-Meier product-limit method and were analyzed by the log-rank procedure. Multiple covariate analyses of disease-free and disease-specific survival were performed using the Cox proportional hazards regression model incorporating factors that on univariate analysis were statistically significant (P < .05) or demonstrated a trend toward significance.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONREFERENCES |
|---|
|
Histologic Results of SLN and Lymphadenectomy Specimens
From 1991 through late 1997, the SLN was examined with a single H&E of a bivalved specimen. Our current approach to the pathologic review of the sentinel node involves enhanced pathologic analysis defined as serial sectioning with immunohistochemical staining, if the initial H&E of the bisected node does not reveal metastatic disease. This current approach was taken in 55 of 166 patients with a successful SLN identification. The incidence of identified SLN metastases increased from 15% to 26% since this approach was adopted (P = .12). Of the 14 patients with detected SLN metastases since adopting the enhanced pathologic analysis, 6 were noted on the initial H&E. The remaining eight patients were detected through step-sections, immunohistochemistry, or both. All 31 patients with SLN metastases underwent a subsequent completion lymph node dissection. The median number of positive nodes in these patients was 1.0 and the mean, 1.7. A single nodal metastases was noted in 20 (65%), two in 7 (23%), three in 1 (3%), and more than three in 3 (9%). Five of 31 (16%) were found to have metastases to non-SLNs and all had tumor thickness > 1.3 mm.
The median and mean number of nodes harvested during ELND were 13 and 15.4, respectively. Nodal metastases were detected in 44 of the 329 patients in the ELND group (13.4%). The median number of positive nodes in these patients with nodal metastases was 1.0, and the mean was 1.9. A single nodal metastases was present in 24 patients (54%), two in 18%, three in 14%, and more than three in 14%.
On multivariate analysis, tumor thickness was the only significant predictor of nodal metastases (P < .01) (Table 2). The incidence of nodal positivity increased from 7% in patients with tumor thicknesses less than or equal to 2.0 mm to 35% in those whose tumors were > 4.0 mm in thickness. Gender, age, ulceration, Clark level, and the treatment approach (ELND vs. SLND) were not significant.
|
|
|
|
Effect of Treatment in High-Risk Groups
On the basis of the multivariate analysis conducted on the combined population, the effect of treatment in patients at higher risk for recurrence was then studied. Tumor thickness > 3.0 mm, the presence of nodal metastases, and age > 50 years were used to stratify patients into higher risk populations. Treatment differences according to these stratified criteria were then ascertained through the Kaplan-Meier method with log-rank analysis (Fig. 4). The 3-year RFS was 84% and 55% for ELND node-negative and node-positive patients respectively (P < .01). A similar decrement in survival was also noted for SLND node-negative (75%) and node-positive (59%) patients. In patients that were node-positive, there was no significant difference between patients treated with ELND versus SLND, nor was there a difference for node-negative patients, although a trend was suggested. In contrast, patients at higher risk for recurrence as defined by tumor thickness and age treated with SLND had a significantly lower relapse-free survival than patients treated with ELND (Fig. 4). For patients with a tumor thickness > 3.0 mm, the 3-year relapse-free survivals for the ELND and SLND groups were 73% and 51%, respectively (P = .04). The 3-year relapse-free survivals for ELND and SLND patients > 50 years of age were 80% and 65%, respectively (P = .01). Differences between treatment groups were not seen in patients < 50 years of age or in patients with tumor thicknesses less than or equal to 3.0 mm. When only those patients at high risk for recurrence were analyzed, the trend in decreased relapse-free survival seen in the node-negative patients became statistically significant (Fig. 5). The proportion of high risk patients (either age > 50 or thickness > 3.0 mm) who had ulceration was not different when comparing the SLND and ELND populations. Although the proportion of patients with a Clark level of IV or V was greater in the SLND population as it was when patients of all risk levels were examined (Table 1), on univariate and multivariate analysis, ulceration was not a significant variable.
|
|
|
|
The pattern of early first-site recurrence occurring in the SLND group involved a significantly larger proportion of nodal recurrences when compared with the ELND distribution of first-site recurrences (P < .01, chi square). Systemic sites comprised 28% of first-site recurrences in the SLND population, followed by nodal (28%), in-transit (24%), and local (20%). Eight SLND patients experienced a nodal recurrence, seven of which occurred in node-negative patients. All node-negative patients with nodal recurrence had a tumor thickness > 2.0 mm, and six of seven were over the age of 50 years. As demonstrated in Figure 6, the node-positive ELND and SLND groups have a similar pattern of recurrence. The difference in pattern of recurrence evident when looking at all ELND and SLND patients is attributable to the differences seen in the node-negative population. Forty percent of first-site recurrences occurring in the node-negative SLND population occurred in the previously sampled draining nodal basin in comparison to 3% of recurrences in the ELND population.
Pathologic Re-review of Negative SLN
The sentinel lymph nodes in five of the seven node-negative SLND patients experiencing a nodal recurrence were evaluated by H&E at initial review before our current approach of SLN evaluation with IHC and step-sectioning was instituted. The sentinel lymph nodes in these five patients were re-reviewed by step-sectioning and/or PCR-assisted detection of tyrosinase mRNA. Evidence of a metastases was demonstrated in three of the five on IHC/step-sectioning and in one additional patient on detection of tyrosinase mRNA. Two of the seven patients were initially reviewed with IHC/step-sectioning. Thus, four of seven SLN-negative patients with nodal recurrence had evidence of metastases within the sentinel node on enhanced pathologic analysis.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONREFERENCES |
|---|
2B in the management of high risk patients was reported in 1996.12 Although the relatively recent introduction of SLND precludes a long-term survival analysis, reports are beginning to appear addressing the issue of relapse after this technique. The present study is a retrospective investigation, the intent of which was to examine the survival and pattern of early recurrence in a large number of patients with primary cutaneous melanoma of the extremity. The cohorts of patients were well-matched with respect to tumor thickness, gender, and ulceration. Significant differences in the age and Clark level distribution exist. The latter likely reflects a change in the pathologist interpretation over time, as the proportion of ELND patients with a Clark level greater than III in the latter half study time period (1984–1994) is not different from the SLND population (data not shown). The difference in the Clark level distribution was not believed to be important, as this was not an important predictor of relapse in either the SLND or ELND population. Increasing age was statistically associated with relapse, although the hazard ratio would suggest that this effect is quite minimal.
The frequency of node-positive lymph node dissections was significantly greater in the SLND group (20%) than in the ELND group (13%). As tumor thickness is the most important predictor of a node-positive dissection (either SLNB or ELND) and the distribution of tumor thickness was similar between the two groups, part of this difference is likely related to the enhanced pathologic analysis possible with the limited number of nodes submitted in the SLND patients. The node-positive rates in this study are compatible with the reported ELND and sentinel lymph node literature.2,4,11,13–16
As the first measure of efficacy, we first compared the overall disease-specific and relapse-free survivals. With a median follow-up of 24 months, 32 of 152 (21%) of SLND patients have experienced a recurrence. One hundred three of 329 (31%) ELND patients have experienced a recurrence with a median follow-up of 77 months. Half of the recurrences in the ELND population occurred within the first 2 years. Overall disease-specific and relapse-free survivals were similar between the two treatment groups. The short follow-up in the SLND group makes it unlikely that relatively small but potentially significant differences would be detected at this time. Overall disease-specific and relapse-free survivals at 3 years were 91% and 79%, respectively and are consistent with recently reported studies.10,13,15 The treatment received (ELND vs. SLND) was not a predictor of relapse on either univariate or multivariate analysis for the entire population.
We next identified through univariate and multivariate analyses a subset of patients within our population that were at significantly higher risk for relapse and, thus, more apt to demonstrate a difference at this period of time in follow-up. The presence of nodal metastases, increasing age, and increasing tumor thickness were all predictive of relapse on multivariate analysis. When stratified according to the presence of nodal metastases, there was no difference in relapse-free survival between ELND and SLND (Fig. 4) when both high- and low-risk patients were included. This would be expected in the node-positive group as the SLND patients with positive nodes underwent completion lymphadenectomy. If sentinel lymph node accurately identified all basins with nodal metastases, one would also expect there to be no difference in the node-negative group. When only patients at high risk (thickness > 3.0 mm or age > 50 years) were studied though, there was a significant decrease in relapse-free survival after SLND (69%) compared with ELND (83%, P = .04).
In addition to relapse-free survival, the pattern of recurrence after treatment was also investigated as another marker of treatment differences. Early recurrences defined as those occurring within 2 years were studied given the discrepancy in follow-up between the two groups and in Fig. 6 are categorized according to the worst site of first recurrence. There was a significant difference in the pattern of early first-site recurrence observed in the two treatment groups. There was a greater proportion of locoregional recurrences noted after SLND (72%) when compared with ELND (39%, P < .01). Nodal recurrence was an uncommon event after ELND accounting for only 4% of recurrences. In contrast, nodal recurrences occurred in seven SLN-negative patients and composed 28% of recurrences in the SLND group. The pattern of recurrence observed in the SLN-negative group is remarkably similar to that reported by Gershenwald et al.17 and Essner et al.15 who also noted nodal recurrences to constitute 32% and 42% of first-site recurrences, respectively (Table 4). Retrospective enhanced pathologic analysis, including step-sectioning and immunohistochemistry, demonstrated nodal metastases in four of the seven SLN-negative patients with nodal recurrence in the present study. The ability to document metastases in 40% to 80% of SLN-negative patients with documented regional recurrence when the initial pathologic evaluation involved a single H&E section has been reported by other groups.14,15,17 It is unclear, though, whether in a prospective manner these will be identified. This fact is highlighted by Essner and colleagues, who demonstrated metastases within "negative" sentinel lymph nodes of patients with nodal recurrence that were examined initially with IHC.15 Although not used routinely in our patients, the detection of tyrosinase mRNA represents another method of enhanced pathologic analysis. This technique has been demonstrated to significantly upstage patients with primary cutaneous melanoma thought to be node-negative on routine examination. On short follow-up, this seems to confer an intermediate prognosis between those patients with metastases noted on histologic review and those without metastases on histologic review and PCR analysis.18,19
The present study has several limitations, which include its retrospective design, and the relatively short follow-up of the SLND group. Essner and colleagues15 present the only other report in the literature directly comparing SLND with ELND. In this well-designed, matched pair analysis, they studied 534 patients with clinically node-negative primary cutaneous melanoma (all sites). The median follow-up in the SLND group was 45 months. They did not demonstrate a difference in either overall or relapse-free survival. A significantly greater proportion of recurrences within the node-positive SLND group occurred in regional sites compared with node-positive ELND patients. Although this was also noted in the node-negative patients, the small numbers likely precluded statistical significance. Nodal recurrences accounted for 42% of the recurrences within the SLN-negative subset of SLND patients and similarly, on re-evaluation of the sentinel node, metastases were identified in 4 of these 11 patients. This study also noted a decreased relapse-free survival in the node-negative SLND patients compared with node-negative ELND patients, again similar to our findings in high-risk node-negative patients. Essner and colleagues15 conclude from their experience that SLND and ELND are at the least equivalent in the management of patients with primary cutaneous melanoma. Specifically, as in our own study, no differences in overall disease-specific or relapse-free survival were detected, although they did not analyze high-risk patients separately from low-risk patients. In addition, they concluded that SLND resulted in similar local control of the draining nodal basin.
We concur that at this period in follow-up, no differences in overall disease-specific or relapse-free survival are evident for the population as a whole. Our data suggest though that in node-negative patients at higher risk for relapse, these approaches are not equivalent, a finding supported by Essner’s15 observation of a decrease in relapse-free survival in node-negative patients undergoing SLND. Furthermore, there is a higher proportion of nodal recurrences in node-negative SLND patients in our study and others reports9,14,15,17 relative to both our control ELND population and the historical ELND literature.20 Whether these differences in recurrence will be meaningful in the long-term is unknown and will require longer periods of follow-up to determine.
|
CONCLUSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONREFERENCES |
|---|
Received for publication March 17, 2000. Accepted for publication October 16, 2000.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  A. Beavis, M. Dawson, P. Doble, R. A. Scolyer, R. Bourne, L.-X. L. Li, R. Murali, J. R. Stretch, C. L. Lean, R. F. Uren, et al. Confirmation of Sentinel Lymph Node Identity by Analysis of Fine-Needle Biopsy Samples Using Inductively Coupled Plasma-Mass Spectrometry Ann. Surg. Oncol., March 1, 2008; 15(3): 934 - 940. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. F. Thompson, F. S. Hodi, and A. Zembowicz Case 2-2007 -- A 49-Year-Old Woman with a Pigmented Lesion on the Arm N. Engl. J. Med., January 18, 2007; 356(3): 285 - 292. [Full Text] [PDF]
|
|
|
|
|
|
Z. I. Nowecki, P. Rutkowski, A. Nasierowska-Guttmejer, and W. Ruka Survival Analysis and Clinicopathological Factors Associated With False-Negative Sentinel Lymph Node Biopsy Findings in Patients with Cutaneous Melanoma Ann. Surg. Oncol., December 1, 2006; 13(12): 1655 - 1663. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Kretschmer, I. Beckmann, K.-M. Thoms, C. Mitteldorf, H. P. Bertsch, and C. Neumann Factors Predicting the Risk of In-Transit Recurrence After Sentinel Lymphonodectomy in Patients With Cutaneous Malignant Melanoma Ann. Surg. Oncol., August 1, 2006; 13(8): 1105 - 1112. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. L. Wong, D. L. Morton, J. F. Thompson, J. E. Gershenwald, S. P. L. Leong, D. S. Reintgen, H. Gutman, M. S. Sabel, G. W. Carlson, K. M. McMasters, et al. Melanoma Patients with Positive Sentinel Nodes Who Did Not Undergo Completion Lymphadenectomy: A Multi-Institutional Study Ann. Surg. Oncol., June 1, 2006; 13(6): 809 - 816. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. van Poll, J. F. Thompson, M. H. Colman, J. G. McKinnon, R. P. M. Saw, J. R. Stretch, R. A. Scolyer, and R. F. Uren A Sentinel Node Biopsy Does Not Increase the Incidence of In-Transit Metastasis in Patients With Primary Cutaneous Melanoma Ann. Surg. Oncol., August 1, 2005; 12(8): 597 - 608. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
V. S.K. Yee, J. F. Thompson, J. G. McKinnon, R. A. Scolyer, L.-X. L. Li, W. H. McCarthy, C. J. O'Brien, M. J. Quinn, R. P.M. Saw, K. F. Shannon, et al. Outcome in 846 Cutaneous Melanoma Patients From a Single Center After a Negative Sentinel Node Biopsy Ann. Surg. Oncol., June 1, 2005; 12(6): 429 - 439. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. A. Scolyer, J. F. Thompson, L.-X. L. Li, A. Beavis, M. Dawson, P. Doble, V. S. K. Ka, J. Gregory McKinnon, R. Soper, R. F. Uren, et al. Failure to Remove True Sentinel Nodes Can Cause Failure of the Sentinel Node Biopsy Technique: Evidence from Antimony Concentrations in False-Negative Sentinel Nodes from Melanoma Patients Ann. Surg. Oncol., March 1, 2004; 11(3_suppl): 174S - 178S. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. P. L. Leong Paradigm of Metastasis for Melanoma and Breast Cancer Based on the Sentinel Lymph Node Experience Ann. Surg. Oncol., March 1, 2004; 11(3_suppl): 192S - 197S. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 O. Alonso, M. Martinez, L. Delgado, A. De Leon, D. De Boni, G. Lago, M. Garces, F. Fontes, J. Espasandin, and J. Priario Staging of Regional Lymph Nodes in Melanoma Patients by Means of 99mTc-MIBI Scintigraphy J. Nucl. Med., October 1, 2003; 44(10): 1561 - 1565. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. H. Estourgie, O. E. Nieweg, R. A. Valdes Olmos, C. A. Hoefnagel, and B. B. R. Kroon Review and Evaluation of Sentinel Node Procedures in 250 Melanoma Patients With a Median Follow-Up of 6 Years Ann. Surg. Oncol., July 1, 2003; 10(6): 681 - 688. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Ribuffo, A. Gradilone, M. Vonella, S. Chiummariello, E. Cigna, N. Haliassos, R. Massa, I. Silvestri, S. Calvieri, L. Frati, et al. Prognostic Significance of Reverse Transcriptase-Polymerase Chain Reaction-Negative Sentinel Nodes in Malignant Melanoma Ann. Surg. Oncol., May 1, 2003; 10(4): 396 - 402. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V Davids, S H Kidson, and G S Hanekom Accurate molecular detection of melanoma nodal metastases: an assessment of multimarker assay specificity, sensitivity, and detection rate Mol. Pathol., February 1, 2003; 56(1): 43 - 51. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Nomori, H. Horio, T. Naruke, H. Orikasa, K. Yamazaki, and K. Suemasu Use of technetium-99m tin colloid for sentinel lymph node identification in non-small cell lung cancer J. Thorac. Cardiovasc. Surg., September 1, 2002; 124(3): 486 - 492. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
