| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ORIGINAL ARTICLES |
From the Sydney Melanoma Unit, the Melanoma and Skin Cancer Research Institute, the Department of Surgery of the University of Sydney (AD, WHM, JFT); the Department of Anatomical Pathology, Royal Prince Alfred Hospital (RAS), Sydney, New South Wales, Australia; and the Daniel den Hoed Cancer Center (JHWdW), Rotterdam, the Netherlands.
Correspondence: Address correspondence and reprint requests to: J.F. Thompson, MD, Sydney Melanoma Unit, Sydney Cancer Centre, Royal Prince Alfred Hospital, Camperdown, NSW 2050, Australia; Fax: 61-2-9550-6316; E-mail: thompson{at}smu.org.au
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Methods: A retrospective cohort study was conducted among patients with American Joint Committee on Cancer (AJCC) stage II disease treated at a single center between 1983 and 2000 with either SNB (n = 672) or ELND (n = 793). Multivariate analyses were performed using the logistic regression model for nodal staging accuracy and Cox’s proportional hazards regression model for survival.
Results: Patient factors that influenced nodal positivity included age, Breslow thickness, ulceration, head or neck primary, and operation type (SNB or ELND). SNB was superior to ELND in the detection of micrometastases (odds ratio 1.23, 95% CI, 1.06 – 1.43) but operation type did not influence survival (P = .24).
Conclusions: Sentinel node biopsy identified more nodal micrometastases than ELND but did not influence survival, although complete regional node dissection was performed in all patients who were SNB positive. This increase in staging accuracy likely results from the reliable identification of the appropriate lymph node field by preoperative lymphoscintigraphy, along with more detailed pathologic examination of the nodes removed by SNB.
Key Words: Melanoma • Elective lymph node dissection • Sentinel lymph node • Staging
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
An alternative to ELND is the sentinel node biopsy (SNB) procedure. A sentinel node (SN) is best defined as any node receiving direct lymphatic drainage from the primary tumor site.4 A SNB procedure involves removal and examination of SNs, usually identified by preoperative lymphoscintigraphy and by intraoperative blue staining after blue dye injection at the primary tumor site, with greater reliability achieved by use of a hand-held gamma probe to detect "hot" SNs. Only patients found to have metastatic disease in their SNs are then subjected to complete dissection of the affected node field. Patients not found to be harboring subclinical metastases are spared the morbidity associated with a complete regional lymphadenectomy.
Whether the use of SNB, with subsequent complete regional lymphadenectomy in patients with metastatic disease in a SN, confers any benefit in terms of patient survival remains unclear. Results of large prospective, randomized trials such as the Multicenter Selective Lymphadenectomy Trial5 and the Sunbelt Melanoma Trial6 must be awaited before this technique can be recommended as a standard therapeutic procedure.
The SNB procedure can be technically difficult. Accuracy depends on the experience and expertise of the surgeon, nuclear medicine physician, and pathologist, resulting in the correct identification and evaluation of sentinel nodes in 70% to 100% of patients.7–9
Although more regional lymph nodes are provided for evaluation by ELND, SNs are routinely examined pathologically in far greater detail. Whereas it is usual for each node in an ELND specimen to be evaluated microscopically by examining one section from it, stained with hematoxylin and eosin (H&E), SNs are normally examined much more thoroughly using multiple sections and additional immunohistochemical stains, in a focused search for small micrometastatic foci. Lymphoscintigraphy, an integral part of the SNB technique, has also enabled the reliable identification of the regional node fields to which lymphatic drainage occurs for melanomas located in axial sites.
In patients with breast cancer, the SNB technique has been shown to improve staging when compared with an historical group of patients treated with standard axillary lymphadenectomy.10 Improved staging, however, has not been demonstrated in patients with melanoma. One study did report that SNB was more efficient than ELND in identifying nodal metastases in a subset of patients with melanoma (with intermediate thickness tumors), but not for the group overall.11
We sought to test the hypothesis that SNB would be superior in the detection of clinically occult metastatic disease in patients with melanoma when compared with ELND, and to assess whether this improved staging, if demonstrated, had an effect on survival.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
1.5 mm at the SMU between 1983 and 2000 were identified from the database and their records were reviewed. Excluded were patients with multiple primary lesions and those with metastatic disease diagnosed before their SNB or ELND. Also excluded were patients who had both an SNB and an immediate ELND (n = 118) as part of an initial SNB validation study.12 In all, 44 patients (3%) were excluded because follow-up information was incomplete. A total of 793 ELND and 672 SNB patients then remained for analysis.
From 1983 onward, ELND was the standard treatment at the SMU for patients with primary melanomas 1.5 mm in thickness. ELND performed for patients with clinically negative lymph nodes involved standard techniques for regional lymphadenectomy.13–15 This treatment policy continued until 1992. In March 1992, the first SMU patients were treated using the SNB technique, and in 1993 ELND was abandoned as the standard treatment recommendation, with only occasional, selective use thereafter. SNB was offered in the context of a clinical trial5 from April 1994, if a primary tumor was 1 mm thick or had invasion to Clark level IV or V and no clinical sign of metastasis. One third of enrolled patients were randomized to an observation-only control arm.
For patients undergoing SNB, preoperative lymphoscintigraphy was performed routinely after 1993, having been performed in most but not all patients having SNB at the SMU before this time. In those having ELND, lymphoscintigraphy was performed only occasionally, for trunk melanomas, in cases of clinical uncertainty about the direction of lymphatic drainage. The lymphoscintigraphy technique that was used has been described.4 In brief, multiple small-volume intradermal injections of technetium 99m antimony trisulfide colloid (99mTc-Sb2S3) were placed around the primary melanoma excision site on the day of or the day before surgery. Scanning was carried out immediately to identify the major lymphatic channels, followed by a delayed scan approximately 2.5 hours later. Anteroposterior and lateral views were obtained to localize the SNs precisely, and allow their positions to be marked on the overlying skin. A SN was defined as any draining node that was seen on dynamic imaging to receive direct lymph draining from the melanoma site.16 On the delayed images, these were usually the most intensively radioactive nodes on the scan.
Immediately preoperatively, all patients having SNB received interdermal injections of patent blue dye at the melanoma site to provide blue staining of the SN for the surgeon. A hand-held gamma-probe was used from 1995 onward to confirm SNs identity at the time of surgery.12 Almost all SNs were both "hot" and blue.
Positive SNs were identified pathologically by the presence of metastatic melanoma cells in tissue sections stained with H&E, immunohistochemical markers, or both. The SN were examined according to the following protocol. Each SN was cut into 3-mm thick slices in its longitudinal axis; these were embedded in entirety in paraffin blocks following tissue processing. Four sequential 5-µ thick tissue sections were cut from each block and stained with H&E, for S100 protein, for HMB45, and with H&E, respectively. The sections were examined microscopically, initially at a scanning magnification of 100x, by histopathologists experienced in the assessment of melanoma pathology. If melanoma metastases were detected in a SN, the patient subsequently had a complete regional node dissection of the affected node field.
Removed non-SNs and all nodes in regional lymph node dissection specimens were embedded whole in paraffin (or sliced when the nodes were >3 mm in diameter) and H&E stained sections were examined for the presence of melanoma metastases.
Analysis and Statistical Methods
All analyses were performed using the S-PLUS software package (Insightful Corporation, Seattle, WA) and Microsoft Excel (Microsoft Excel, Version 2000; Microsoft, Redmond, WA). Survival analysis on the selected patient data was carried out using Kaplan-Meier product-limit estimations of survival probabilities and Cox’s proportional hazards regression model, where appropriate.17–18 Survival distributions of product-limit estimates were compared using the log-rank test. Death from melanoma was considered one endpoint in this study. Survival time was defined as the time between the initial histologic diagnosis of primary melanoma and the date of last follow-up or death from melanoma. Patients with causes of death other than melanoma were censored at date of death. Seven variables were modeled. Five of these were categorical: metastatic nodes found (yes, no); operation performed (SNB, ELND); gender (male, female); ulceration of the primary lesion (present, absent); and its location (head and neck, trunk, extremity). Two variables were continuous: age at diagnosis (in years) and Breslow thickness of the primary lesion (log [value in mm]). The distribution of thickness was positively skewed. To produce normality, it was log transformed. Interactions between variables were evaluated and included in a stepwise variable selection procedure when model building. The appropriateness of using any of the variables in the proportional hazards model was investigated by –2 log(–logS(t)) plots. No variable violated the proportional hazards assumptions.
To determine which procedure (SNB or ELND) better identified nodes containing metastasis, the logistic regression model was used.19 Multivariate logistic regression analysis was performed with the following variables: gender (male, female), age (in years), thickness (in mm), ulceration (present, absent), and tumor location (head and neck, trunk, extremity).
Association between factors was determined by 
2 analysis for categorical variables, and using the Student t test and Wilcoxon rank sum test for testing of continuous variables. All P values were two-sided and were considered statistically significant if the P value was <.05.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
2 = 28.57, df = 2, P < .001). Although no difference was seen in the sites of regional fields in which procedures were performed, the SNB group contained significantly more patients who had nodes removed from two or more fields (2 = 37.9, df = 2, P < .001). Most in the SNB group had a prior lymphoscintigram (90.6%), compared with a small minority of those in the ELND group (9.6%). Subclinical nodal metastases were found in 18% of patients in the SNB group, but in only 14.6% of those in the ELND group (2 = 2.82, df = 1, P = .09).
|
|
2 = 2.2, df = 1, P = .14).
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Elective lymph node dissection was performed in patients with primary melanoma who were considered to be at high risk for locoregional recurrence to identify and remove nodal metastases, in the hope of preventing subsequent spread to distant sites. Several randomized clinical trials of ELND have failed to show an overall survival advantage.1–3 In fact, of all patients with AJCC/Union Internationale Contre le Cancer (UICC) stage I and II melanoma who go on to develop metastatic disease, regional nodes will be the site of first recurrence in only 50% and another 25% will develop distant metastases without any prior regional involvement.20
In this study, we compared the efficacy of SNB and ELND in identifying regional node metastases using a large retrospective cohort of 1465 patients. The two groups had significant differences in several known prognostic variables. Patients having ELND had primary lesions, which had a poorer prognosis because they were thicker and more were ulcerated. No difference in gender proportions or primary site was detected, but patients having SNB were significantly older. Multivariate analysis of the data, therefore, was required.
Disease-specific survival was similar for patients having SNB or ELND using Kaplan-Meier estimates. Overall survival of patients who had a SNB was 82.4% after 5 years. This is comparable with other studies, where similar 5-year survival rates have been reported.8,11,21–22 Patients having SNB showed no survival advantage when cases were partitioned according to nodal status. Because SNB identifies more patients who are node-positive, fewer false–negative findings would be expected in the node-negative SNB subgroup compared with the node-negative group having ELND. Although the survival curves for patients in the node-negative group did appear to separate with time, the difference did not reach statistical significance.
After controlling for known variables of prognosis in a Cox proportional-hazard analysis, operation type (SNB or ELND) was not an independent prognostic factor. Length of follow-up was significantly different between the two groups but this was controlled for by the statistical methods used. The difference in length of follow-up was related to changes in the treatment recommendations for patients with primary melanomas treated at the SMU, where SNB had largely replaced ELND by 1994.
Essner et al.11 compared the efficacy of lymphatic mapping, SNB, and selective complete regional lymphadenectomy with ELND in a matched pair analysis of 534 patients. They found that only in patients with intermediate thickness melanomas (1.51 to 4 mm) was SNB more effective than ELND in identifying nodes harboring clinically occult metastatic melanoma deposits. Our investigation used a much larger patient set composed of 1465 individuals. Significant differences were seen in the distribution of prognostic factors (e.g., Breslow thickness, ulceration, and age) between the two patient groups and, consequently, multivariate logistic regression analysis was performed to control for these. This analysis indicated that SNB was a significantly more accurate method than ELND in identifying nodal metastases in all patients.
The reason that SNB is superior to ELND in identifying melanoma metastases probably relates to the different histopathologic protocols used to examine the lymph nodes microscopically in the specimens obtained from each of these procedures. As pathologists are usually presented with only one or two SNs, it is possible to examine these in greater detail than it is to examine the many lymph nodes present in ELND specimens. As ELND usually provides more than 10 times the number of nodes than SNB, the time and expense involved in examining each node using the same histopathologic protocol used to examine SNs make it impractical to do so.23 Such a practice would involve examining an average of 80 sections per ELND specimen compared with an average of 6 for SN assessment using our standard protocol.
Preoperative lymphoscintigraphy used as part of the SNB procedure to identify with certainty each regional node drainage field may also have improved the detection rate of melanoma metastases from primary lesions located on the trunk, and also from those on the lower limb that drained directly to external iliac or obturator SNs.
Independent variables predicting disease-specific survival may not have the same ability to predict nodal metastases. In the present study, Breslow thickness and ulceration of the primary tumor were significant prognostic variables in the analysis of patient survival. As expected, they correlated also with an increased probability of lymph node metastases. This is consistent with several previous studies evaluating factors predictive of positive sentinel nodes.11,21,24–31
An unexpected finding, however, was that younger patients were more likely to develop nodal metastases. This was surprising because increasing age is known to be associated with reduced survival32 and, indeed, in this particular group of 1465 patients, older age was associated with a reduced survival probability. Most previous studies of predictive factors for SNB positivity have indicated that age has no effect on the metastatic prediction.11,21,26,27 Nevertheless, in four studies that analyzed large series of patients by multivariate logistic regression analysis, patient age was reported to be an independent prognostic factor for the prediction of nodal metastases in SNB.24,28,29,31 In the largest of these studies, reporting results of the Sunbelt Melanoma Trial,28 the likelihood of SN positivity decreased by approximately 2% per annum as age increased. In the present study, an identical finding was made, with increasing age being significantly associated with fewer nodal metastases. For each 10 years of additional age, a decrease of 20% was noted in the probability that nodal metastasis would be found. If older patients have higher death rates but are less likely to develop lymph node metastases, a possible explanation would be that they are more likely to first develop distant metastases via hematogenous spread32 than nodal metastases. This might be caused by age-related changes in lymphatic flow or density of lymphatic vessels.33 This matter needs to be explored further in future studies.
In the present report, primary tumor location was categorized as head and neck, trunk, or extremity. Patients with head and neck, and trunk primary lesions are known to have inferior survival in comparison with those who have primary tumors located on the extremities,32 and this was observed in these patients. It is paradoxical, however, that head and neck location was associated with a significantly lower rate of nodal metastasis identification in this group of 1465 patients having ELND and SNB. One possible explanation for this observation is that the site of the first detectable melanoma metastases may be dependent on the location of the primary lesion in these patients with stage II disease. This was examined by Meier et al.,20 who studied the development of metastatic pathways in primary melanoma. In half of their patients, the first metastases detected were in regional lymph nodes, but this proportion was influenced by location of the primary lesion. Primary melanomas of the head and neck were less likely to develop first metastases in regional nodes compared with other body sites (trunk, upper extremity, lower extremity). Furthermore, primary melanomas of the head and neck were shown to produce in transit and satellite metastases in greater proportions than melanomas in other locations.20
Primary location, however, had no predictive value in multivariate models for SNB positivity in several previously published studies.11,24–26 In these studies, primary location was classified as "axial" or "extremity" and found to have no predictive value, although some other studies reported that primaries in axial sites were associated with higher rates of metastasis in SN.21,27
In summary, the present study found that SNB was superior to ELND in the identification of occult nodal metastases in patients with melanoma and, consequently, provides better staging. This is likely because of more rigorous microscopic examination of multiple H&E and immunochemically stained sections of the nodal tissue and more reliable identification of the appropriate lymph node field by preoperative lymphoscintigraphy in patients having SNB. Patients who had a SNB, however, had no demonstrable survival advantage compared with those patients who received an ELND.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Nodal staging accuracy following treatment by sentinel node biopsy or elective lymph node dissection was compared in 1465 American Joint Committee on Cancer stage II melanoma patients. Sentinel node biopsy identified more micrometastases than elective lymph node dissection but did not influence survival.
Received for publication January 28, 2004. Accepted for publication June 8, 2004.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  Z. I. Nowecki, P. Rutkowski, and W. Michej The Survival Benefit to Patients with Positive Sentinel Node Melanoma After Completion Lymph Node Dissection May Be Limited to the Subgroup with a Primary Lesion Breslow Thickness Greater Than 1.0 and Less Than or Equal to 4 mm (pT2-pT3) Ann. Surg. Oncol., August 1, 2008; 15(8): 2223 - 2234. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Murali, R. N. Sharma, J. F. Thompson, J. R. Stretch, C. S. Lee, S. W. McCarthy, and R. A. Scolyer Sentinel Lymph Node Biopsy in Histologically Ambiguous Melanocytic Tumors With Spitzoid Features (So-Called Atypical Spitzoid Tumors) Ann. Surg. Oncol., January 1, 2008; 15(1): 302 - 309. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. R. Scoggins, M. I. Ross, D. S. Reintgen, R. D. Noyes, J. S. Goydos, P. D. Beitsch, M. M. Urist, S. Ariyan, B. S. Davidson, J. J. Sussman, et al. Prospective Multi-Institutional Study of Reverse Transcriptase Polymerase Chain Reaction for Molecular Staging of Melanoma J. Clin. Oncol., June 20, 2006; 24(18): 2849 - 2857. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Xi, W. Gooding, K. McCarty, T. E. Godfrey, and S. J. Hughes Identification of mRNA markers for molecular staging of lymph nodes in colorectal cancer. Clin. Chem., March 1, 2006; 52(3): 520 - 523. [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]
|
|
|
|
|
|
K. M. McMasters Editorial: What Good is Sentinel Lymph Node Biopsy for Melanoma if it does not Improve Survival? Ann. Surg. Oncol., September 1, 2004; 11(9): 810 - 812. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
