| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Article |
1 Department of Laboratory Medicine and Pathobiology, Banting Institute, 100 College Street, Room 110, Toronto, Ontario, Canada M9G 1L5
2 Department of Pathology and Laboratory Medicine, University Health Network, 620 University Avenue, Room 4-302, Toronto, Ontario, Canada M5G 2M9
3 Department of Surgery, University Health Network, 620 University Avenue, Third Floor, Room 3-130, Toronto, Ontario, Canada M5G 2M9
Correspondence: Address correspondence and reprint requests to: Karen Nicole MacNeill, MD; E-mail: karenmacneill{at}yahoo.com.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Methods: A retrospective analysis of prospectively collected data identified patients with cutaneous melanoma of the head and neck 
.76 mm. Sentinel lymph nodes were identified by using a standardized protocol of preoperative lymphoscintigrams, intraoperative blue dye injections, and handheld gamma probes. Clinical, surgical, and pathologic data were collected and analyzed.
Results: A sentinel lymph node was removed in 41 (94%) of 44 patients. Seven (17%) of 41 had at least 1 positive sentinel lymph node. Three of seven had primary tumors <1 mm (two of the three were not ulcerated). The sites of lymphatic drainage of the primary lesion were discordant, with historical anatomically predicted sites in 24.4% of cases. None of the 34 patients with negative SLNB has had a nodal recurrence (false-negative rate, 0%; sensitivity and negative predictive value, 100%). The mean follow-up is 22.4 months (range, <1–69 months). Seven (17%) of 41 patients had minor complications.
Conclusions:: SLNB in the head and neck area is challenging; however, combined pre-operative, intraoperative, and histological techniques produce a sensitive procedure with a high negative predictive value. The lack of false-negative results obviates the need for prophylactic neck dissections.
Key Words: Sentinel lymph node biopsy • Head and neck • Metastatic melanoma • Cutaneous
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Despite its overall success, SLNB for head and neck cutaneous melanoma is a challenge. The lymphatic drainage of the head and neck can be unpredictable4,6,7,13; there are potential difficulties related to the location of the lymph nodes, isotope blush overlap between the primary tumor and nodes, and juxtaposition of nerves, and the siting of sentinel node incisions can potentially compromise reconstruction planning or subsequent regional node dissection. Finally, the criteria that determine who should have SLNB for melanoma of the head and neck are not well established.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
.76 mm) melanoma of the head and neck were selected. Those with mucosal melanomas or a previous wide local excision were excluded. The standardized protocol included a lymphoscintigram on the day of operation with a four-quadrant intradermal injection of unfiltered 99mTc sulfur colloid 20 to 40 mBq in .5 mL of normal saline. The skin over any sentinel lymph node (SLN) hot spots was marked. At operation, 1 to 3.5 hours subsequently, isosulfan blue dye (.5 mL) was injected intradermally around the primary lesion or biopsy scar. By use of a handheld gamma probe, the operative field was examined, and positive sites were identified and explored. Hot nodes were identified in situ and confirmed on removal, followed by examination of the surgical bed with the gamma probe to ensure that background levels decreased to acceptable levels (<10%). All hot or blue nodes were removed. After SLNB, patients underwent wide excision of the primary tumor with primary reconstructive procedures, including simple closure, interposition, and rotation flaps. SLNs were identified as such on the specimen container and sent to pathology.
Surgical incisions for SLN biopsy were planned so that they could potentially be incorporated into a subsequent modified neck dissection or parotidectomy by using conventional Schobinger or MacFee incisions for neck dissection and a Lilienthal-type incision for parotidectomy. Virtually every SLN in the neck could be reached along the template for one of these incisions without compromise of access for cosmesis secondary to subsequent parotidectomy, neck dissection, or both. In situations in which there were multiple SLNs at multiple levels in the neck, MacFee incisional templating provided access to sites as divergent as facial, submandibular, and posterior triangle while still allowing for subsequent incorporation of the SLN biopsy scars into a versatile neck-dissection incision.
The SLNs submitted were fixed in 10% buffered formalin and quarantined for 24 hours to allow for radioactive decay. The SLN was processed with a protocol modified from that described by Cochrane.14 The node was bisected along its long axis and submitted in toto. Each half was serially sectioned into six levels. Levels 1, 3, and 5 were stained with hematoxylin and eosin, and levels 2, 4, and 6 were stained for the immunohistochemical markers S-100, HMB-45, and, starting in 2001, Melan A. All nodes were examined by a single pathologist. Nodes diagnosed as positive had a synoptic report completed that identified the size of the node, the number and sites of the metastatic foci, the size of the largest focus, the presence of extracapsular extension, and the immunoreactivity to S-100, HMB-45, and Melan A. The wide excision specimen was examined for residual disease and margin status.
Patients with a negative SLNB received no further surgical intervention. Those with a positive SLNB were offered a completion lymph node dissection (CLND). If further positive nodes were identified on CLND or if patients had high-risk histological features of multiple nodes involved, extracapsular extension, or both, then they were referred to medical oncology for consideration of immunotherapy. If the CLND revealed no additional histologically positive nodes, further therapy was not recommended.
All SLNB patients were followed up every 3 months for 2 years and then every 6 months until 5 years. Because we are a regional tertiary care center, some patients travel great distances for surgery. Arrangements were made for these patients to be followed up with the same frequency by the referring physicians or family doctors if the travel distance was too great to come to our facility.
Patient charts were reviewed for personal demographic data, histological features of the primary lesion and SLN, wide excision, and CLND and for follow-up data. Follow-up was also completed by telephone contact with the patient’s family doctor, referring physician, or patient, as necessary.
The results were then analyzed to determine the sensitivity and negative predictive value of the procedure. The results of the SLNB and lymphoscintigram were compared with the historical anatomically predicted drainage patterns of the primary tumor as established by O’Brien et al.15
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
.
|
. A single surgeon performed 28 of 41 cases, 11 of 41 were performed by a second surgeon, and the remaining 2 were performed by 2 others. All were from the same institution.
|
The preoperative lymphoscintigram identified 1 or more areas of nodal activity in 39 (95.12%) of 41 patients. The median number of sites of nodal activity seen was 2 (range, 1–8), and the median number of basins identified was 1, with a range of 0 to 3. Unfortunately, the reports of three preoperative scans did not provide adequate detail of the site or number of nodes identified, and, as a result, they are not included in the analysis.
At operation, at least 1 SLN was identified and removed in 41 (93.1%) of 44 patients. In the three patients in whom an SLN could not be identified, one patient had also had a negative preoperative lymphoscintigram, one had a node identified on the lymphoscintigram that could not be localized during surgery, and the third had an assumed SLN removed, but no lymphoid tissue was identified histologically. As a result, these three patients were excluded from final analysis because the status of their nodal basin could not be established. A single SLN was identified in 19 (46.3%) of 41 cases, 2 SLNs were identified in 9 (22%) of 41 cases, and 3 or more SLNs were identified in 13 (31.7%) of 41 cases. Fifteen (36.6%) of 41 patients had an SLN identified within the parotid gland, and 3 (7.3%) of 41 had bilateral drainage.
To examine the drainage patterns of the primary tumors, we compared the actual SLN sites at operation with the historical anatomically predicted drainage patterns described by O’Brien et al.15 Ten (24.4%) of 41 patients had an SLN identified in a site outside the predicted drainage pattern (Table 3). According to the authors, patients with an anterior scalp melanoma should undergo an END of levels I to IV or I to III and parotid; our single patient with an anterior scalp melanoma had drainage to level V. One of six patients with a posterior scalp melanoma had an SLN identified bilaterally; predicted drainage was II to V and parotid. Finally, 8 of 22 patients with a facial primary tumor had unexpected drainage: 3 had drainage to level V, 2 were bilateral, 3 had drainage to facial nodes, and 1 had drainage to a preauricular node.
|
Positive SLNB
Seven (17%) of the 41 patients had a positive SLNB. Patient characteristics are presented in Table 4. Five patients had a single positive node, and two of seven had two positive nodes. The mean Breslow thickness was 1.99 mm, with a range of .77 to 3.07 mm. Of note is that three of seven patients with a positive SLNB had, by definition, a thin primary melanoma (<1 mm). Three of the seven had ulceration identified in their primary tumor, whereas four did not. None of the three patients with thin melanomas had ulceration in their primary melanoma. All seven had subcapsular metastases, and two of the seven had metastasis limited to the subcapsular region.
|
Three of the five patients received adjuvant immunotherapy (interferon) despite the negative CLND: they were classified as high risk because of the histological characteristics of their primary tumor. Eleven months after CLND, one of the five patients developed a nodal recurrence in the previous operative field. This patient subsequently underwent a second modified radical neck dissection and at the last follow-up at 38 months was disease free. The remaining four patients are all disease free as of the last follow-up (mean, 22.8 months; range, 6–39 months).
Two of the seven patients with a positive SLNB could not undergo a CLND because of rapid disease progression; one developed extensive palpable lymphadenopathy within 1 month of his SLNB, and the other had the procedure abandoned on the day of operation because of palpable disease. One was dead of disease 13 months after his SLNB. The other was alive with disease after radiotherapy at the last follow-up (9 months).
Negative SLNB
Thirty-four patients (83%) had a negative SLNB. Three (7.5%) of 34 had local recurrences excised from the previous operative fields; the earliest recurrence was 2 months after SLNB, and the longest was 45 months. None of the 34 patients have had a nodal recurrence, giving a false-negative rate of 0%. All 34 patients with negative SLNB results are alive without disease at last follow-up (mean, 22.3 months; median, 19 months; range, <1–69 months).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Lymphatic drainage to the head and neck has been shown repeatedly to be complicated and unpredictable. Discordance rates comparing the sites of the actual SLN removed at operation with the sites predicted by historical/anatomical drainage patterns have ranged from a low of 7%13 to a high of 37%.16 The results from our study further support the challenge of accurately identifying the lymphatic drainage of the head and neck region. In our study, when the actual site of the SLN removed at operation was compared with the historical anatomically predicted drainage pattern described by O’Brien et al.,15 an SLN was found outside the predicted drainage pattern in 10 (24.4%) of 41 cases (Table 3
).
To further examine the complexity of lymphatic drainage, we also compared the sites of nodal activity identified on the preoperative lymphoscintigram with the same historic anatomically predicted sites and found a discordance rate of 27.5% (11 of 40). This rate of disagreement was much better than that of 63 % reported by Leong et al.16
We did discover a significantly higher discrepancy rate, however, when we compared the results of the preoperative lymphoscintigram with those at the actual site of the node removed at operation. In this situation, up to 60% of patients had discordant results. In 8 cases, the SLN that was removed at operation was missed altogether by the preoperative scan, in 10 cases the scan identified more SLNs than were identified at operation, and in 6 cases the scan identified fewer SLNs than were identified at operation.
Despite the discordant results among the preoperative scan, surgery, and predicted sites, at least 1 SLN was successfully identified and removed in 41 (93.2%) of 44 patients. Others,4–8 using similar protocols, have identified comparable success rates ranging from 90% to 96.3%. We identified micrometastatic disease in 7 (17%) of 41 patients—a rate comparable to that described by others. Morton et al.5 reported a 15% SLNB-positive rate in a series of 72 patients; Schmalbach et al.,4 Alex et al.,8 and Wells et al.7 reported 18%, 14%, and 11%, respectively; and Jansen et al.6 had a conspicuously higher positive rate of 27 % in a series of 30 patients.
As important as it is to detect patients with micrometastatic disease, identifying those with a negative SLNB is equally important. The persistent concern is whether those with a negative SLNB are truly node negative. We use the false-negative rate as a measure of SLNB failures. A false negative is defined as a recurrence in the nodal basin in a person with a previously negative SLNB. After a mean of 22.4 months’ follow-up, we have had no nodal recurrences in the 34 SLN-negative patients, giving a false-negative rate of 0% and, thus, a sensitivity and negative predictive value of 100%. Morton et al.5 and Wells et al.7 have also reported false-negative rates of 0%. In studies with follow-up similar to ours, false-negative rates have ranged from 4.5% to 10.5%.4,6 We recognize that our follow-up is limited to just under 2 years and that long-term follow-up will be the only means to establish the ultimate success of the procedure.
The results of this study reinforce the use of SLNB in the head and neck region. Despite complicated and sometimes unpredictable lymphatic drainage, the procedure is highly successful and accurate but requires the use of multiple modalities and specialties.
One of the major benefits of the SLNB procedure in the staging of clinically node-negative patients is that it identifies high-risk patients who may benefit from further therapy17 and avoids unnecessary ENDs in patients without metastatic disease. It is known that up to 80% of patients will not benefit from an END; thus, the goal of the SLNB is to identify patients who have micrometastatic disease and, thus, who would benefit from more extensive dissection.14 Our study substantiates this claim. More importantly, an unnecessary END was avoided in 34 (83%) of 41 patients.
Another benefit of SLNB is its lower morbidity when compared with END.10,11 Whereas Schmalbach et al.4 reported essentially no complications in their study of 77 patients, we experienced minor complications in 7 (17%) of 41 patients (Table 2). The most severe were two cases of mild neuropraxia of the mandibular branch of the facial nerve, and both cases resolved. Given that we had 15 (36%) of 41 patients with an SLN removed from the parotid basin, we believe that 2 temporary minor nerve injuries is acceptable even as we recognize that no nerve injuries would have been the ideal outcome.
Research has shown that the risk of metastasis increases with Breslow thickness and that the risk of metastases in thin melanomas (
1 mm) is low.1 As a result, the role of SLNB for these "thin" melanomas is unclear. Several studies, however, have shown that melanomas <1 mm thick do have the potential to metastasize. Lens et al.18 showed that 8% of thin (.76–1.5 mm) melanomas metastasized. Given the low rate, they believed that the international consensus of offering SLNB only to patients with melanoma 1 mm was reasonable but acknowledged that SLNB in thin melanoma may be considered if high-risk histological features (including Clark level >III or ulceration) are present in the primary tumor. Jacobs et al.3 reported that 3% of patients with thin (< 1 mm) melanoma could benefit from SLNB, whereas Lowe et al.10 reported a positive SLNB rate of 7% in thin melanomas ( 1 mm) and, thus, supported the procedure for those with a Clark level of invasion of III or greater. It is this lack of consensus in the literature that makes establishing practice guidelines a challenge.
The results of our study seem to add to the confusion. Our study included four patients with primary tumors <1 mm, which would now be considered thin melanomas. At the time, however, these tumors were staged as intermediate thickness (.76–1.5 mm) according to the old AJCC staging system. The current AJCC recognizes that the Clark level and the presence of ulceration are prognostic in thin melanomas.1 Three (43%) of the seven patients with a positive SLNB had, by definition, a thin melanoma (.77, .95, and .90 mm). All three primary tumors did not show ulceration, and all had Clark level III invasion. Theoretically, they all should have been at low risk of metastasis, yet they accounted for 43% of the positive SLNBs in our series; this is significantly higher than the rates identified by others.10,18 The reason for this is unclear. Perhaps the depth of the original biopsy sample was not accurately measured. Two of the three patients had shave biopsies to diagnose their primary melanoma, and this may have limited accurate determination of the Breslow thickness. In fact, residual tumor was found in both specimens at the time of wide excision. Regardless of the accuracy of the shave biopsy, these patients were offered an SLNB on the basis of the perceived depth (intermediate thickness) of their initial biopsy sample. According to today’s standards, they were referred with what is now considered a thin melanoma. If current criteria of 1 mm had been used, all three patients likely would not have had been offered an SLNB; three patients with metastatic disease would have been missed. Clark level III invasion is the only high-risk feature that might today result in an SLNB being offered. We do not claim that our results are significant given the very small number of patients. Further study is obviously warranted. Patients with thin melanomas of the head and neck who undergo wide excision only should be followed long-term for the appearance of nodal disease. This issue will be pertinent as we are faced with the diagnosis and detection of an increasing number of thin melanomas. The questions to be answered are as follows: to whom should SLNB be offered, and what criteria should determine those at risk of micrometastatic disease?
Finally, the location within the node of the meta-static foci has implications for SLNB. Care must be taken in the removal of all SLNs. Inadvertently cutting through the capsule or excessive use of cautery can interfere with the histological evaluation of the SLN. In our series, all positive SLNs had a subcapsular component; in two of the seven, it was the only component. If the subcapsular area is not preserved during surgery, there is a real risk of missing small metastases, and this would result in patients being incorrectly stratified for treatment.
The results of our study demonstrate that SLNB is a low-morbidity procedure that is highly accurate and successful when multiple modalities are used. It allows accurate stratification of patients who need treatment and limits CLND to those who have documented metastatic disease.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Received for publication November 29, 2004. Accepted for publication April 17, 2005.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
