| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ORIGINAL ARTICLES |
From the Departments of Surgery (HLE, CLS), Radiology (CDT), and Pathology (JWP), University of Virginia Health System, Charlottesville, Virginia; Vermont Cancer Center (DNK, SPH), University of Vermont, Burlington, Vermont; Department of Surgical Oncology (SM), Academisch Ziekenhuis van de Vrije Universiteit, Amsterdam, The Netherlands; Massachusetts General Hospital Cancer Center (KKT, MAG), Harvard Medical School, Boston, Massachusetts; Creighton Cancer Center (BWL), Creighton University Medical Center, Omaha, Nebraska; Baptist Hospital (PWW), Nashville, Tennessee; Charleston Area Medical Center (REK), Charleston, West Virginia; and Lankenau Hospital Cancer Center (NZC), Wynnewood, Pennsylvania.
Correspondence: Address correspondence and reprint requests to: Heather L. Evans, MD, Surgical Infectious Disease Research Laboratory, Building MR-4, Room 3156, Lane Road, Charlottesville, VA 22908; Fax: 434-243-6360; E-mail: hle2r{at}virginia.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Methods: Seventy-six patients, including 18 patients from the University of Virginia and 58 from a multicenter study of SNBx led by investigators at the University of Vermont, who had previous WLE for clinically localized melanoma underwent lymphoscintigraphy with SNBx. Median follow-up time was 38 months.
Results: Intraoperative identification of at least 1 sentinel node was accomplished in 75 patients (98.6%). The mean number of sentinel nodes removed per patient was 2.0. Eleven patients (15%) had positive sentinel nodes. Among the 64 patients with negative SNBx, 3 (4%) developed nodal recurrences in a sentinel node–negative basin simultaneous with systemic metastasis, and 1 (1%) developed an isolated first recurrence in a lymph node.
Conclusions: This multicenter study more than doubles the published experience with SNBx after WLE and provides much-needed outcome data on recurrence after SNBx in these patients. These outcomes compare favorably with the reported literature for patients with SNBx at the time of WLE, suggesting that accurate staging of the regional lymph node bed is possible in patients after WLE.
Key Words: Sentinel node biopsy • Lymphoscintigraphy • Wide local excision • Melanoma
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Studies of SNBx for melanoma have routinely excluded patients who have had a previous WLE for melanoma,8 largely out of concern that surgical disruption of lymphatic channels may render lymphoscintigraphy unreliable. It has been recommended that SNBx not be offered after WLE.2,9 At present, optimal management of regional nodes in this setting has not been defined.10,11 Because accurate staging in these patients is as important as it is in patients who present for SNBx before WLE, it is important to know whether these patients presenting after WLE can be staged accurately with SNBx.
Kelemen et al.12 reported a series of 47 selected patients who underwent SNBx after prior WLE, observing that the incidence of positive sentinel nodes was as high as that seen in SNBx undertaken before WLE. The authors concluded that SNBx after WLE could be used with caution in areas of relatively unambiguous lymphatic drainage. However, no standard has been established for these patients, and rates of regional failure have not been reported. Given the increasing number of patients presenting to tertiary referral centers after WLE for melanoma, further review of the incidence and outcomes of this population is warranted. Through review of multicenter data, we sought to determine whether the practice of SNBx after WLE produces comparable outcomes to SNBx at the time of WLE. Presented in this article are the surgical results and clinical follow-up in 75 patients who underwent SNBx after WLE for primary melanoma in a combined series from 12 medical centers since 1993.
|
PATIENTS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
All patients had biopsy-proven invasive melanoma, clinically negative regional lymph nodes, and no evidence of distant metastases (clinical stage I or II). Entry criteria were similar at the two centers but will be listed separately.
Vermont SNBx Study
Patients were offered SNBx as part of this study regardless of the depth of invasion. SNBx was performed at the same time as the WLE or within 45 days of the WLE. Results of the complete series have been reported.13 All WLE margins were 
2 cm, and there were no prior skin grafts. This study was performed in accordance with the University of Vermont Human Subject Protection Committee and the Radiation Safety Committee of the participating institutions, and patients entered the study by giving informed consent.
University of Virginia
Investigators at the University of Virginia participated in accrual on the University of Vermont study with the approval of the Human Investigation Committee at the University of Virginia. When the study was closed to accrual, patients at the University of Virginia were offered SNBx with the same protocol used in the University of Vermont Study. Patients presenting with a newly diagnosed melanoma with a Breslow thickness of 
1 mm and clinically negative nodes were considered candidates for SNBx if there was no contraindication to adjuvant therapy. Some exceptions were made when patients with thin lesions had adverse prognostic factors (Clark level IV, positive deep margin, or regression) or when patients specifically requested SNBx even when their lesions were somewhat thinner than the stated guidelines. SNBx was offered to patients who had had prior WLEs even when the WLE margins were >5 cm. These patients were accrued from 1997 to 2000, when SNBx had become standard practice in most melanoma centers.14 The surgeon performing these studies had previously abandoned elective node dissection as a routine staging procedure.6 Accordingly, these patients were offered SNBx as the staging procedure of choice. Informed consent was obtained and documented on a standard operative consent form after risks and potential benefits were explained by the surgeon before the procedure.
Lymphoscintigraphy Procedure
Unfiltered 99mTc-labeled colloid was injected intradermally at multiple sites circumferentially around the melanoma excision scar.15,16 The colloid used was either 99mTc sulfur colloid (CIS-US Inc., Bedford, MA, or DuPont Pharmaceuticals, North Billerica, MA) or 99mTc human albumin (Solco Nonocoll; Sorin Biomedica, Vercelli, Italy; or Microlite; DuPont). Injections were made within 1 cm of the scar and were placed approximately 1 cm apart along the full length or circumference of the scar.16 In a few cases with long scars, the injections were limited to the mid portion of the scar. The amount of radiocolloid injected was .2 to 1.4 mCi in a volume ranging from .2 to 1 mL. In the latter half of the study entry period at the University of Virginia, eutectic mixture of local anesthetics cream (lidocaine/prilocaine; AstraZeneca, Wilmington, DE) was applied to the skin around the excision site for 45 minutes to 1 hour before injection, which decreased the discomfort associated with intradermal injection.
Immediately after the radiocolloid was injected, the patient was placed under a gamma camera, and images were taken to identify the sentinel nodes. In the vast majority of cases, a sentinel node was identified within 15 to 20 minutes of the injection time. In many cases, the node was identified immediately on placement under the gamma camera. In unusual cases, there was a delay of up to 2 hours before identification of the draining sentinel node. Gamma camera pictures were taken at the time of initial imaging and before we sent the patient to the operating room. Correlations were made with counts by using a handheld gamma probe (C-trak; Care Wise Medical Products Corp., Morgan Hill, CA) at that time. The probe was set at a threshold of 130 keV and a window of 40 keV.
Injection of Blue Dye
Surgeons had the option to inject vital blue dye in addition to the 99mTc colloid. In those cases, 2 to 5 mL of isosulfan blue dye (Lymphazurin 1%; Hirsch Industries Inc., Richmond, VA) or Patent Blue V 2.5% (Laboratorie Guerbert, Aulney-Sous-Bois, France) was injected intradermally by the surgeon circumferentially around the scar.
Surgical Procedure
The patients were brought to the operating room after lymphoscintigraphy. Usually the radiocolloid was injected the day of surgery, but some patients had injections performed the day before surgery. The patient was placed on the operating table, and the handheld gamma probe was used to confirm the location of the hot spots observed by lymphoscintigraphy. At that time, 10-second counts were taken with the probe to measure the preincision radioactivity of each hot spot, as well as the maximal counts at the injection site and background counts at the liver. All regional node basins and in-transit areas were scanned manually with the handheld probe to rule out additional sentinel nodes not identified on the lymphoscintigram.
The incision, usually 2 to 3 cm long, was made through the skin with a scalpel. With the handheld probe as a guide, dissection was carried straight down to the first hot spot by using sharp dissection and spreading with a tonsil clamp. Once the hot spot was identified, the vessels leading to the sentinel node were clamped and ligated either with a hemoclip or with silk ligatures. The node was then excised sharply. Ten-second counts on the excised node were measured with the handheld probe (ex vivo counts), and residual counts were measured in the node basin. If additional hot nodes were identified or if there are additional hot spots, these were excised in a similar fashion. Once the maximal residual counts in the surgical bed were <10% of the maximal ex vivo counts obtained from the first hot spot node, it was determined that all sentinel nodes had been removed. In any case in which the gamma camera scan identified more sentinel nodes than had been excised, additional dissection occurred until the discrepancy was clarified.
Data Review
A total of 76 patients who presented for SNBx after previous WLE for primary melanoma were identified from the databases of the Vermont Cancer Center and the University of Virginia Health System. Fifty-eight patients included in the series were enrolled in a nonrandomized prospective trial of SNBx for melanoma, sponsored by Dr. David Krag at the University of Vermont. Data entry forms from the trial were reviewed, and patients who had undergone SNBx after WLE for melanoma were identified. Similar patients from the University of Virginia melanoma database, whose data are tracked with Paradox 8.0 software (Corel Corporation, Ottowa, Ontario, Canada), were identified, and data were then stored with Microsoft Excel 2000 software (Microsoft Corp., Redmond, WA). The patient charts were reviewed to obtain additional details. Wherever possible, particulars such as the length of the WLE scar, presence of a rotational flap or skin graft, or length of time between WLE and SNBx were noted. Clinical follow-up data were evaluated and reviewed from the same databases. Descriptive statistics were used to summarize the data.
Histological Evaluation
In the Vermont multicenter trial, nodes were evaluated by one or more sections stained with hematoxylin and eosin (H&E). Routine immunohistochemistry was not performed, but rather was used to confirm areas of suspicion first observed with H&E staining.
The standard protocol adopted at the University of Virginia for evaluation of sentinel nodes in melanoma patients includes H&E section evaluation of three levels of each sentinel node plus immunohistochemical evaluation with S100, HMB45, and tyrosinase antibodies. Evaluation of S100 staining takes into account that all normal lymph nodes contain S100-positive dendritic cells in the paracortical region that can be differentiated morphologically from metastatic melanoma cells, which usually appear in the subcapsular sinus. Thus, a positive finding on S100 staining requires the finding of cells histologically consistent with melanoma that stain with S100. In cases in which differentiation is difficult, staining with CD1a antibodies as a positive marker for dendritic cells may be used. Positive staining of melanoma with HMB45 or tyrosinase antibodies requires that the cells be morphologically consistent with melanoma cells and that they cannot be explained as any benign cell. Occasionally, plasma cells staining positively with HMB45 antibody have been noted, but they are readily differentiated from melanoma cells, and nodes with these findings are defined as histologically negative for melanoma.17
Clinical Management After SNBx
Patients with negative SNBx were observed without additional surgery. Patients with positive SNBx underwent completion lymphadenectomy of all involved nodal basins, except in one case, in which further surgical intervention was refused by the patient. Most patients with positive sentinel nodes in this study underwent surgery before interferon was approved for use by the Food and Drug Administration. Thus, most patients had no additional adjuvant therapy. However, adjuvant therapy options were discussed and offered when appropriate. Of the four University of Virginia patients with positive sentinel nodes, two were treated with high-dose interferon, and one was entered onto an experimental melanoma vaccine trial. Another patient with a thick primary melanoma and negative SNBx was also entered onto an experimental melanoma vaccine trial.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2-cm margins; 29 (50%) of 58 Vermont patients had lesions of 1.5 mm, presumably with margins of 2 cm (data unavailable). Fourteen (78%) of 18 UVA patients had margins 2 cm. Data regarding the scars from prior WLEs, ranging from 4.0 to 15.5 cm, were available for 21 patients. Among the UVA patients,18 the median scar length was 9.4 cm, split-thickness skin grafts were used in five patients, advancement flaps were undertaken in three patients, and a rotational flap was performed in one patient.
|
|
|
1 (15%) sentinel nodes of 11 patients.
|
Outcomes During Patient Follow-Up
Melanoma recurrences were observed in 11 patients overall (14.6%), 8 of whom had negative SNBx (12.5% of negative SNBx). A third of patients in the University of Virginia group have been followed up for at least 2 years. In the Vermont Cancer Center group, after a median follow-up of 3.3 years (mean, 3.1 years), melanoma recurred in 11 patients (19%). The group of patients at the University of Virginia has been without recurrence to date, with a median follow-up interval of 8 months (mean, 1.3 years). Recurrences include tumor found at local, nodal, in-transit, and systemic locations (Table 4). All tumor recurrences appeared in patients with Clarks IV melanoma of at least 1.73-mm thickness at initial presentation, and 6 of 11 tumors recurred in the axilla.
|
Of the 64 patients whose sentinel node was negative for malignancy, 4 developed recurrence in a nodal basin. Of the four patients with lymph node recurrences, two recurrences were in the sentinel nodal basin, and two were in another nodal basin. Three of these four patients presented with systemic disease at the time of their recurrence; only one had an isolated first recurrence in a regional node, but it was in a node basin contralateral to the basin identified by lymphoscintigraphy (and contralateral to the primary lesion). Thus, there were no patients who developed isolated nodal recurrences in lymph node basins that were negative by SNBx. However, isolated recurrence in a contralateral node could be explained by a failure of lymphatic mapping.
There were 11 patients with positive sentinel nodes; therefore, this one nodal recurrence after a negative SNBx would represent an 8% (1 of 12) false-negative rate. The other three patients with nodal recurrences had simultaneous systemic disease. If they are considered false negatives of the SNBx staging (the "true nodal basin failure rate," as described by Kelley et al.10), then a 27% false-negative rate (4 of 15) may be suggested. However, recurrence after systemic dissemination is difficult to interpret as regional failure, especially when the regional recurrence is contralateral. The broadest interpretation of clinical failure possibly attributable to lymphatic mapping would be to include three patients (patients V036, V027, and V047; Table 4). Even when considering these three regional recurrences, the negative predictive value of SNBx in this population is still 95%.
Completion lymphadenectomy was performed in 10 patients (1 patient declined axillary lymphadenectomy because of concern for lymphedema), with a mean of 13 lymph nodes harvested per basin. Two patients had positive nodes harvested in addition to the positive sentinel node; one of these patients had systemic recurrence. Two in-transit recurrences were observed among the patients with positive SNBx and subsequent negative completion lymphadenectomy.
Review of Literature
Several series report the incidence of recurrence after SNBx for melanoma. Seven studies of SNBx at the time of WLE are reviewed, along with this study and two smaller studies of SNBx subsequent to prior WLE (Table 5).12,13,18–23 This summary permits comparison of published data for two groups of patients: patients with SNBx simultaneous with WLE and patients with SNBx subsequent to prior WLE. The mean number of sentinel nodes biopsied in the two groups was 1.7 and 2.0, respectively. The percentage of patients with positive sentinel nodes was 17% when SNBx was performed at the time of WLE and was 21% (despite a lower mean tumor thickness) in patients with delayed SNBx.
|
It should be noted that patients in the Vermont multicenter study had sentinel nodes analyzed with H&E staining alone, without immunohistochemical review, raising the possibility that the false negatives identified may actually represent pathologic misdiagnosis. Pathologic review of original nodal samples was not available. Nonetheless, the results in this study and in the summarized literature on SNBx after WLE are comparable to, or even slightly better than, those reported for patients undergoing SNBx at the time of WLE.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The incidence of positive nodes in this group was 15%, which is in the range that would be expected for this group of patients on the basis of prognostic factors regarding the primary lesion. It should be noted that a single investigator performed all of the SNBx at the University of Virginia (n = 18), whereas multiple surgeons at different institutions with varying degrees of experience in lymphoscintigraphy and sentinel lymphadenectomy performed the procedures from which the Vermont database was built (n = 58). Thus, this report represents a multicenter experience that can be considered applicable to a broad range of surgeons.
This report contains a review of the literature and tabulation of selected studies with substantial numbers of patients, detailed data on outcomes, and/or follow-up data on regional and systemic recurrences. The summary data from those two sets of patients (SNBx at the time of WLE and delayed SNBx) reveal comparable success rates of SNBx (94% vs. 99%), mean number of sentinel nodes excised (1.7 vs. 2.0), percentage of patients with positive SNBx (17% vs. 21%), rates of nodal recurrence in basins with prior negative SNBx (2.8% vs. 2.0%), and the calculated false-negative rate based on nodal recurrence rates (9.6% vs. 6%) for patients with simultaneous WLE and prior WLE, respectively. Thus, the data in this report, with unselected patients from multiple centers and with long follow-up, suggest that SNBx is as accurate in patients who have had prior WLE as in patients who undergo SNBx at the time of WLE.
This study also permits an assessment of success rates with different lymphatic mapping techniques. All patients underwent lymphatic mapping with 99mTc-labeled radiocolloid; however, blue dye was also used in a subset of patients in the Vermont Cancer Center group. In every case in which blue nodes were identified as sentinel nodes, the blue node was also radioactive (data not shown). There were three patients, however, in whom blue dye and radiocolloid were used for localization and in whom tumor metastases were found in radioactive sentinel nodes that were not blue (data not shown). Thus, this experience suggests that the addition of blue dye is not essential for successful mapping, whereas the use of radiocolloid is critical to accurate identification of the sentinel nodes.
The standard approach to calculate false-negative rates in the early SNBx literature was to determine the rate of SN positivity compared with lymph node positivity assessed in a complete elective lymph node dissection (ELND) specimen. However, patients who undergo negative ELND will experience disease recurrence in the negative nodal basin approximately 4% of the time.6 Because ELNDs are positive approximately 15% to 20% of the time, this could be interpreted as a 17% (4 of 24) to 21% (4 of 19) false-negative rate. When phrased this way, that number seems unacceptable, but it is important to realize that this is the result with what has been considered the gold standard for management of the regional nodes. In this article, we have addressed the question of regional nodal failure in several ways and have presented all the available data, in comparison to published comparable data from studies of SNBx at the same time as WLE. There is a risk of lymph node metastases in patients with melanoma, despite negative SNBx, but this does not seem to be increased by delaying lymphatic mapping and SNBx until after the WLE is performed. This risk is comparable to risks of nodal recurrence after complete elective node dissection.6 As such, the occurrence of these nodal metastases should not be interpreted as a justification to abandon SNBx. However, it remains important to monitor regional node basins clinically after a negative SNBx as part of the routine follow-up after a melanoma diagnosis.
Although the data in this report justify the application of SNBx staging to patients who have already had a WLE, it is reasonable to evaluate patients for SNBx with an awareness of the risks of disrupted lymphatic channels. Simple transection of lymphatic channels in a scar does not seem to impair the flow in those lymphatics proximal to the scar. Thus, by injecting radiocolloid around the entire scar, all the lymphatic channels that drain the primary lesion should also be included among the lymphatics draining sulfur colloid from this injection. A primary concern had been that injection of sulfur colloid over a larger area might lead to identification of additional nodes as sentinel because they drained the skin at the ends of the incision, but not the site of the primary lesion. However, the number of sentinel nodes identified in this study is comparable to that reported in numerous series of SNBx.
One of the patients included in this series had a rotation flap, and three advancement flaps were undertaken, with no recurrence observed in any of the four patients, even though one patient had two positive sentinel nodes. Additionally, split-thickness skin grafts were used in five patients with no evidence of local, nodal, in-transit, or systemic recurrence. Patients with split-thickness skin grafts or flap closures have been singled out in the literature as unfavorable candidates for SNBx after WLE for fear of inaccurate staging of nodal basins because of altered drainage patterns.1,9 The number of patients with split-thickness skin grafts or flaps in this study was small; further studies in this particular subset of patients would be useful. However, our data support the application of SNBx techniques even after large skin grafts and simple advancement flaps. Likely, undermining subcutaneous tissue itself does not substantially disrupt lymphatic drainage and does not seem to prevent accurate staging by SNBx.
Because of concern about lymphatic disruption, clinical circumstances remain in which SNBx should not be recommended. Large rotational flaps are likely to disrupt lymphatics and to make SNBx unreliable. One case report of SNBx after prior WLE with rotation flap provides an example. During SNBx, a blue-stained sentinel node was harvested and found to be negative, but an adjacent, nonstaining node was positive for malignancy.18 Another situation arises when the site of the primary tumor lies directly over the nodal basin to which primary drainage is expected. An example would be a melanoma arising in the preauricular area, in which the draining nodes at risk include parotid nodes. If WLE is performed in that setting, it seems probable that the lymphatic channels to the underlying nodes would be disrupted; we do not recommend SNBx in that setting. In most other body sites, however, the lymphatic channels travel through the dermis and subcutaneous tissue for a distance before entering the nodes, thus permitting lymphatic mapping even after a WLE. Melanomas in the head and neck region are potential exceptions, because drainage patterns in those areas may be more ambiguous and less predictable. One patient out of the three with primary head and neck melanomas included in this study (Table 1) had systemic recurrence (Table 4). Although this experience in only a few patients is hardly conclusive evidence, Kelemen et al.12 have stated that the possibility for errors in mapping makes SNBx after WLE moderately risky in this particular subset of patients. Finally, it should be clarified that although delayed SNBx may be used to stage lymphatic drainage after WLE of melanoma, this approach does not represent the optimal management of patients with melanoma and should be reserved for those patients who present after WLE.
In summary, evidence has been presented that WLE for melanoma with margins of 1 to 5 cm, with or without skin grafting, routinely permits subsequent accurate lymphatic mapping by lymphoscintigraphy with 99mTc-labeled colloid. Sentinel node biopsy may not be reliable in those patients who have had large rotation flaps. However, in standard WLE with or without skin grafting, where there has been minimal distortion of lymphatic channels and no transection of the subcutaneous tissue proximal to the melanoma site, SNBx is appropriate and can provide reliable staging information.
|
Footnotes |
|---|
Received for publication May 8, 2002. Accepted for publication December 3, 2002.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  V. H. Stell, H. J. Norton, K. S. Smith, J. C. Salo, and R. L. White Jr. Method of Biopsy and Incidence of Positive Margins in Primary Melanoma Ann. Surg. Oncol., February 1, 2007; 14(2): 893 - 898. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. B. Stitzenberg, P. A. Groben, S. L. Stern, N. E. Thomas, T. A. Hensing, L. B. Sansbury, and D. W. Ollila Indications for Lymphatic Mapping and Sentinel Lymphadenectomy in Patients with Thin Melanoma (Breslow Thickness <=1.0 mm) Ann. Surg. Oncol., October 1, 2004; 11(10): 900 - 906. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
