| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Article |
1 Sydney Melanoma Unit and Melanoma and Skin Cancer Research Institute, Sydney Cancer Centre, Gloucester House, Royal Prince Alfred Hospital, Missenden Road, Camperdown, 2050, New South WalesAustralia
2 Discipline of Surgery, Faculty of Medicine, University of Sydney, Sydney, 2006, New South WalesAustralia
3 Department of Surgery, University of Calgary, 1331 29th Street NW, Calgary, Alberta T2N9 4N2Canada
4 Department of Anatomical Pathology, Royal Prince Alfred Hospital, Missenden Road, 2050, Camperdown, New South WalesAustralia
5 Nuclear Medicine and Diagnostic Ultrasound, RPAH Medical Centre, 100 Carillon Avenue, Newtown, 2042, New South WalesAustralia
6 Discipline of Medicine, University of Sydney, Sydney, 2006, New South WalesAustralia
Correspondence: Address correspondence and reprint requests to: John F. Thompson, MD Sydney Melanoma Unit and Skin Cancer Research Institute, Sydney Cancer Centre, Gloucester House, Royal Prince Alfred Hospital, Missenden Road, Camperdown, 2050, New South Wales, Australia; E-mail: thompson{at}smu.org.au.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Methods: Survival and prognostic factors were analyzed for 836 SNB-negative patients. All patients with a node field recurrence were reviewed, and sentinel node (SN) tissue was reexamined.
Results: The median tumor thickness was 1.7 mm, and 23.8% were ulcerated. The median follow-up was 42.1 months. Melanoma specific survival at 5 years was 90%, compared with 56% for SN-positive patients (P < .001). On multivariate analysis, only thickness and ulceration retained significance for disease-free and disease-specific survival. Five-year survival for patients with nonulcerated lesions was 94% vs. 78% with ulceration. Eighty-three patients (9.9%) had a recurrence. Twenty-seven patients developed recurrence in the regional node field, and in 22 of these, it was the first recurrence site. Six developed local recurrence, 17 an intransit metastasis, and 58 distant disease. The false-negative rate was 13.2%. SN slides and tissue blocks were further examined in 18 patients with recurrence in the node field, and metastatic disease was found in 3 of them.
Conclusions: This large, single-center study confirms that patients with a negative SNB have a significantly better prognosis than those with positive SNs. In those with a negative SNB, primary tumor thickness and ulceration are independent predictors of survival. Incorrect pathologic diagnosis contributed to only a minority of the false-negative results in this study.
Key Words: Melanoma • Sentinel lymph node • Metastasis • Survival
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
In the space of a few years, this procedure has become the most frequently performed treatment in most major melanoma treatment centers worldwide, despite the absence of evidence for any benefit to patient survival. Pending the results of ongoing prospective studies, the rationale for its widespread use is the possible benefit of early rather than late therapeutic node dissection and the undisputed ability of SNB to improve staging. The accuracy of SNB has been confirmed in a few studies in which the remainder of the node field underwent a full completion node dissection and in which it was shown that no metastases were present in non-SNs when none existed in the SN.3–5 These studies are unlikely to be repeated,6 because elective lymph node dissection (ELND) has largely been abandoned after randomized trials failed to demonstrate an overall survival benefit.
The other measure of accuracy for SNB has been long-term follow-up of patients who undergo SNB and have negative results. These patients would be expected to have improved survival over the SNB-positive group and, furthermore, to have very few regional recurrences in the mapped node fields. Nodal recurrence after a negative SNB would suggest a limitation of either the concept or the technique of the procedure. Nodal recurrence rates have been reported to date in only a few series, some with very limited follow-up.7–19
The Sydney Melanoma Unit (SMU) experience of SNB is one of the world’s largest, and there is now a sufficient duration of follow-up to report recurrence rates and calculate 5-year survival. The purpose of this study was to review the outcome of 846 patients who had a negative SNB for cutaneous melanoma. Furthermore, an analysis was undertaken of those patients who experienced a regional recurrence to determine possible reasons for the false-negative result.
|
PATIENTS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
1 mm thick, Clark level IV or V, or ulcerated and there was no clinical evidence of metastatic disease. In the early 1990s 157 patients underwent synchronous ELND, most as part of an initial validation study,5 and were excluded from analysis. At least 1 SN was identified and retrieved in 991 (97.9%) of 1012 patients. Patients who had no SN identified were excluded from analysis. Nodal metastases were identified histopathologically in 145 patients (14.6%) and were absent in 846 patients (85.4%).
SN Mapping Technique
After histologic confirmation of the diagnosis of cutaneous melanoma, lymphoscintigraphy was performed on all patients except three. Details of the protocol used at the SMU have been published previously.1 Briefly, four intradermal injections of 99mTc antimony sulfide colloid (particle size, 5–40 nm), each in .05 to .1 ml, were given around the primary melanoma site. Early and delayed imaging was then performed, and the location and depth of each SN was marked and recorded. The SNs were biopsied within 24 hours of the isotope injection. Approximately 15 minutes before the procedure, 1.0 to 2.0 ml of Patent Blue V dye (Guerbert, Aulney-Sous-Bois, France) was injected into four to six sites around the excision-biopsy scar. A careful search for each SN was then conducted, aided by blue dye visualization of afferent lymphatics, and the identity of each SN was confirmed by observing its blue staining and noting the presence of at least one blue-stained afferent lymphatic entering it. From May 1995 onward, SN identity was also checked by using a handheld gamma probe (Neoprobe 1000; Neoprobe Corporation, Dublin, OH). Any node that had a radioactivity count at least three times the residual count in the node field was accepted as a SN.20 In 236 patients who had their SNB before May 1995, blue dye alone was used for SN identification. In the remaining 776 patients in the series, SN identification involved both visual recognition of blue dye staining and gamma probe recording of nodal radioactivity. All patients were treated by wide excision of the primary tumor site at the time of the SNB. Patient data and follow-up information were entered into a prospectively collected database.
Histologic Assessment of SNs
SNs from all patients were cut along their longitudinal axis in 3-mm slices and embedded entirely in paraffin blocks after tissue processing. Four sequential 5-mm-thick tissue sections were cut from each block and stained with hematoxylin and eosin (H&E) on sections 1 and 4 and with immunohistochemical markers for S-100 protein and HMB-45 on sections 2 and 3.
Analysis After Recurrence
The case records and histology of all patients with recurrence in the node field from which a SN reported to have been negative had been removed were reviewed in detail. All histologic slides were reassessed by a single pathologist and, if negative, had further sections cut from each tissue block. The additional sections consisted of four serial sections performed at each of two levels (50 µm apart). The sections from each level were stained with H&E and immunohistochemical stains for S-100 protein, HMB-45, and Melan-A. Both the original slide and additional slides were examined microscopically by scanning the entire sections at a magnification of x100.
Statistical Analysis
Standard statistical tests were used for analysis. Survival curves were constructed by using the Kap-lan-Meier method, and differences were determined by using the log-rank test. Disease-specific and disease-free survival times were calculated from the time of diagnosis of the primary melanoma. Categoric variables were analyzed with the 
2 test. Variables that were significant on univariate analysis were tested for independent significance by using the Cox proportional hazards regression model.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. There were 513 men and 333 women. Median and mean ages were 55.4 and 54.5 years, respectively. Median and mean tumor thicknesses were 1.7 and 2.2 mm, respectively (range, .27–12.5 mm). Ulceration was seen in 23.8% of the primary lesions for which this pathologic feature was recorded.
|
. Actuarial survival at 5 years was 90%. Also shown for comparison is the survival of node-positive patients (56% at 5 years). The difference in survival between patients with positive and negative SNB results was highly significant (P < .001). Disease-free survival for the same groups is shown in Fig. 1B; it was 85.5% for the SNB-negative group and 43.5% for the SNB-positive group (P < .001).
|
and 3. Disease-specific and disease-free survival for patients with and without ulceration are shown in Fig. 2. Five-year disease-specific survival for patients whose melanomas were ulcerated was 78% vs. 94% for patients with nonulcerated lesions. Disease-free survival at 5 years was 74% and 89% for patients with ulcerated and nonulcerated lesions, respectively. Disease-specific survival analyzed by thickness ranged from 95% to 82% at 5 years, as shown in Fig. 3A. Disease-free survival by thickness is shown in Fig. 3B and ranged from 94% for patients with melanomas 
1 mm thick to 75% for patients with lesions > 4.0 mm thick.
|
|
|
|
Several patients first developed recurrent disease simultaneously in more than one site. Table 4 lists the number of recurrences and their sites. Twenty-seven patients (3.2%) developed recurrence in the regional node field, and in 22 of these it was the first site of recurrence. Six (7%) developed local recurrence, and 16 (2%) developed in-transit recurrence. The most common form of recurrence was distant disease. Fifty-eight patients (6.9%) developed distant disease, and in 45 of them, it was their first recurrence.
|
The original histologic slides of the SN were available for review in 18 of the 22 patients with a false-negative result. One showed evidence of meta-static melanoma on review of the original slides. Two demonstrated evidence of metastasis on pathologic examination of additional sections cut from the archival tissue block(s). These three cases were clas-sified as failures of pathologic analysis.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Since it was first described in 1992, the SNB technique has been progressively refined. Originally, only the technique of intradermal blue dye injection around the melanoma site was used to identify lymphatics leading to SNs, and a success rate of 81.8% was reported.3 The addition of preoperative lym-phoscintigraphy and the use of an intraoperative gamma probe to the blue dye technique resulted in much higher SN identification rates,21 which now approach 100%. It is clear from the very small number of cases in which there was failure to identify at least one SN that the technique used at the SMU is reliable and reproducible. The next important issue to address is whether the SNs are being sampled and analyzed correctly.
Recent data have shown that the number of met-astatic nodes and the tumor burden of nodal metastases are the most significant predictors of outcome in stage III melanoma patients.22 This information was used to guide revision of the American Joint Committee on Cancer staging system for melanoma.23 It is already abundantly clear that patients with a positive SNB have a worse prognosis than those with no evidence of microscopic metastatic disease. However, some patients reported to have a negative SNB will be found to have locoregional and distant metastases on subsequent follow-up. Of special interest are those who develop recurrence in the mapped node field, which indicates a failure of the SNB to accurately predict regional node involvement. We therefore examined the outcome and recurrence patterns of 846 SMU patients with negative SNBs.
Positive Versus Negative SNBs
The Kaplan-Meier melanoma-specific survival at 5 years in this series was 90% for those with negative SNBs and only 56% for those with at least one positive node. Others have reported similar finding.13,14,18 These data confirm that SN status is a highly significant prognostic factor that identifies a subset of patients who are at high risk of recurrence. The corollary of this is that the melanoma patient who has a negative SNB has a very good prognosis.
Given a negative SNB result, what are the next most important prognostic features? The results of analysis of prognostic factors in this series are consistent with other reports.12,14,17 In the node-negative population, thickness and ulceration remained the most important prognostic determinants for disease-free and disease-specific survival on multivariate analysis. Mitotic rate, site, age, and sex were all nonsignificant on Cox proportional hazards analysis. Disease-specific survival at 5 years by Kaplan-Meier analysis for patients with ulcerated lesions was only 78%, compared with 94% for those with nonulcerated lesions. These data confirm that regardless of the primary tumor thickness or node status, patients with ulcerated lesions are at high risk, merit particularly careful follow-up, and should be considered for adjuvant therapy trials.
Patterns of Recurrence
Of 83 patients who experienced a recurrence, the first site of metastasis was distant in 43 cases. Local or in-transit recurrence was uncommon and occurred in only 2.7% of patients. It seems logical to report these forms of recurrence as a combined total, because it is likely that most local recurrences are in fact intralymphatic metastases,24,25 and the distinction between local and in-transit recurrence is therefore arbitrary. For the same reason, the classification of local recurrences as those within 5 cm of the primary tumor site was arbitrary; however, it was convenient because the information in the SMU database had been collected by using 5 cm as the cutoff between local and in-transit recurrence. Nodal disease was the first site in only 22 (26.5%) of 83 recurrences. Other reports of this rate range from 19% to 42%.13,14,17,26 The overall rate of regional failure was 27 (3.2%) of 836. These data may also be compared with other authors’ experience of patterns of recurrence after a negative ELND. These rates have been published by Essner et al.,13 who reported recurrence in 35 of 225 patients, 5 of them in the regional node field. It would seem that although SNB has not been shown to be therapeutic, the chance of regional failure is very low and comparable to the rate after ELND if a negative result is obtained.
False-Negative Rate
There are several ways to define the false-negative rate of SNB.6 The first is to complete a dissection of the same node field to ensure that other nodes are not positive when the SN was negative. Several groups have reported this rate in relatively small series.3–5 It is unlikely that any other studies like these will be undertaken because of the morbidity of ELND and its apparent lack of benefit. This method is also dependent on close scrutiny of all nodes. An examination of the SN or other nodes with H&E staining of sections from a bivalved lymph node specimen will have a lower rate of positivity than multiple step sections with immunohistochemical staining, and this method in turn will have a lower rate than testing with reverse transcription-polymerase chain reaction to detect the presence of tyrosinase.27 In most centers, it is logistically and economically impossible to examine more than a single H&E section from each node in an ELND specimen.
The other method of detemining the false-negative rate, as reported here, is to follow up patients reported to be SNB negative for a sufficient length of time to observe recurrence in the regional node field. The inherent inaccuracies of this method are the failure to detect late recurrence because of insufficient follow-up or the death of the patient from distant disease before occult regional disease is detected. The former difficulty can be dealt with only by longer follow-up. The latter would require autopsy studies, yet to be reported. It is also possible that a false-negative SN was the only node containing metastatic tumor cells, and, therefore, after its removal, there could be no subsequent recurrence in that node field.
It might be predicted that the positive SNB rate added to the regional node field recurrence rate should be very close to the regional recurrence rate reported in studies in which lymph node fields were simply observed, with no immediate surgical intervention. The Intergroup Melanoma Surgical Trial, for example, demonstrated regional recurrence rates in the observed group who underwent no nodal dissection of 13%, 27%, and 33% for patients with lesions 1 to 2 mm, 2 to 3 mm and 3 to 4 mm thick, respectively.28 In the present series, in which the median tumor thickness was 1.7 mm and the median follow-up was 42 months, 172 (17%) of 991 patients had developed nodal disease at the time of reporting. This is entirely consistent with the proposal outlined previously, although, of course, patient groups may not have been comparable.
We suggest that the false-negative rate should be patient based rather than node field based. It should include all nodal recurrences that are either first recurrences or recurrences synchronous with recurrent disease elsewhere. Most authors follow this convention, which is useful when comparing different series.6–10,14,26 However, the false-negative rate may underestimate the actual "miss rate" of SNB in detecting regional disease. For example, in the present series, two patients first developed in-transit disease and were subsequently found to have nodal disease. There is no way of knowing with certainty at which site metastatic disease developed first. It is entirely possible that these patients may have harbored disease in their lymph nodes that remained undetected from the outset. Alternatively, occult microscopic local or in-transit disease may have been responsible for the subsequent appearance of metastatic disease in regional node fields from which a truly negative SN had previously been removed. If this occurred, the false-negative rate would have been overestimated, because the effect of such biologic processes on subsequent nodal failure was not considered. It is also possible that patients who developed non–node basin recurrences after a reportedly negative SNB may have had undetected occult metastases in their SNs. Because it seems more likely that such recurrences are related to systemic metastasis rather than spread from regional node fields, they have not been included in the determination of SN false-negative rates in most previous studies. In accordance with these conventions, the false-negative rate in this series was 13.2% (22 of 167 patients), and the sensitivity was 86.8% (where the denominator is true positives plus false negatives). If one assumes that regional recurrences that occur after the appearance of local, systemic, or in-transit disease are still a failure of the SNB technique, then the sensitivity is slightly different. In that case, there were 27 false negatives, and the sensitivity rate decreases to 84.3%. It is important to note that this rate is quite different from the failure rate, as defined by the number of regional recurrences divided by the total number of SNB procedures. That figure is 22 (2.6%) of 836. Data from the present series are compared with data from other series in Table 5.
|
Pathologic Reanalysis
We attempted to determine how often the cause of regional recurrence in the negative-SNB patients was an incorrect pathologic interpretation of SN status (with consequent failure to perform a completion regional node field dissection). Of the 18 patients whose first recurrence was nodal and whose tissue was available for further assessment, only 3 were reinterpreted as being positive, despite the examination of additional sections cut from the tissue blocks and immunohistochemical staining for S-100 protein, HMB-45, and Melan-A. Similarly, Doting et al.,16 in their recent small series, found no tumor cells in the reanalysis of the SNs of six patients who developed a recurrence in a previously mapped node field that had been reported to be negative. In contrast, Gershen-wald et al.14 reported 10 patients who developed regional recurrence after a negative SNB, and 8 of the 10 demonstrated metastatic nodal disease on pathologic reevaluation.14 Essner et al.13 reported that 4 of 11 patients with regional recurrences had tumor in the SNs on re-examination. Clary et al.11 demonstrated metastatic nodal disease in six of seven regional recurrences after the initial report was interpreted as negative. The fact that there were relatively few pathologic failures in the present series is likely to be due to the routine initial assessment of all SNs by examination of multiple (four) sections, with routine immunohistochemical staining. Other groups have relied on standard histologic assessment of H&E-stained sections in most cases, without the routine use of immunohistochemical staining. It should be understood that the retrospective pathologic evaluation of archival material is not comparable to prospective evaluation, no matter how careful. For example, Clary et al.11 reported no change in the rate of detection of positive SNs before and after the introduction of step-sectioning and immunohistochemistry. The finding of metastases in a SN after a recurrence does not necessarily mean that it could have been found on the initial evaluation, even with the same technique. A prospective measure of the true false-negative rate may become available when the long-term follow-up results of the Sunbelt Melanoma Trial29 are analyzed. This prospective randomized trial uses a standardized pathologic assessment including both immunohistochemistry and reverse transcription-polymerase chain reaction.
When our entire series is considered, only 3 of the 991 patients who had successful SNB procedures could be identified in whom there was an incorrect pathologic diagnosis of their SNs, albeit with limited follow-up. It is difficult to see how this figure can be decreased to the point where it would have a signifi-cant clinical effect on regional recurrence simply by performing more extensive pathologic assessment of the SN at the time of biopsy. It is more likely that other factors are predominantly responsible for the failure of the SNB technique to detect disease in the regional node field. The procedures in this series date from 1992, shortly after the procedure was introduced. In early cases, a gamma probe was not used to identify the appropriate nodes to remove. A learning curve also exists for the vital contribution that nuclear medicine makes to the SNB technique. Although there has been a trend to fewer false-negative results in the past few years, it is impossible to determine yet whether this is due to improved technique or simply to shorter follow-up. It is likely, however, that both nuclear medicine and surgical factors may have contributed to the false-negative rate.30 Further studies that we have performed have shown that antimony levels (originating from the 99mTc antimony sulfide colloid used for lymphoscintigraphy in our institution) were low in some patients with a false-negative SNB; this suggests that the wrong node was removed.31 Ultimately, it should be remembered that any effect on survival of a false-negative report resulting in a delayed node dissection is speculative. This conclusion awaits the results of ongoing randomized trials.
Distant metastasis with no clinical evidence of lo-coregional recurrence was the most common form of recurrence in this series. Gershenwald et al.14 have suggested that this may represent true hematogenous spread without lymphatic spread.14 However, it is possible that nodal disease remains occult in these patients. Detailed histopathologic review of the SNs of such patients, not performed in this study, could provide evidence in support of this hypothesis. Another possibility is that regional disease existed at one time but was eliminated by an immune response, analogous to the regression seen in primary cutaneous lesions. This proposal seems impossible to prove or disprove in any retrospective study. Because thickness and ulceration remain important prognostic factors in patients with negative SNs, consideration should be given to using these features to select patients for adjuvant therapy trials, with the aim of decreasing the systemic recurrence rate.
|
ACKNOWLEDGMENTS |
|---|
Received for publication March 20, 2004. Accepted for publication January 11, 2005.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  I. Satzger, B. Volker, A. Meier, A. Kapp, and R. Gutzmer Criteria in Sentinel Lymph Nodes of Melanoma Patients that Predict Involvement of Nonsentinel Lymph Nodes Ann. Surg. Oncol., June 1, 2008; 15(6): 1723 - 1732. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
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]
|
|
|
|
|
|
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. Caraco, U. Marone, E. Celentano, G. Botti, and N. Mozzillo Impact of False-Negative Sentinel Lymph Node Biopsy on Survival in Patients with Cutaneous Melanoma Ann. Surg. Oncol., September 1, 2007; 14(9): 2662 - 2667. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. F. Uren, R. Howman-Giles, D. K. V. Chung, R. L. Morton, and J. F. Thompson The Reproducibility in Routine Clinical Practice of Sentinel Lymph Node Identification by Pre-operative Lymphoscintigraphy in Patients with Cutaneous Melanoma Ann. Surg. Oncol., February 1, 2007; 14(2): 899 - 905. [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]
|
|
|
|
 |
|
 A. van Akkooi, J. de Wilt, C Verhoef, P. Schmitz, A. van Geel, A. Eggermont, and M Kliffen Clinical relevance of melanoma micrometastases (<0.1 mm) in sentinel nodes: are these nodes to be considered negative? Ann. Onc., October 1, 2006; 17(10): 1578 - 1585. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. J. Lejeune What is the impact of sentinel node biopsy in the management of cancer? Ann. Onc., August 1, 2005; 16(8): 1217 - 1218. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
