| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Editorial |
Department of Surgery, Memorial Sloan-Kettering Cancer Center, 1275 York Avenue, New York, New York 10021
Correspondence: Address correspondence and reprint requests to: Daniel G. Coit, MD; E-mail: coitd{at}mskcc.org.
Since its reintroduction nearly 14 years ago,1 lymphatic mapping with sentinel lymphadenectomy has been widely adopted by the surgical oncology community as a standard of care in the management of patients with primary cutaneous melanoma, without rigorous prospective clinical trials. This technique offered a very attractive alternative to routine elective lymph node dissection at a time when elective lymph node dissection, with its attendant morbidity, could not be shown to improve outcome when compared to nodal observation with therapeutic lymphadenectomy if necessary.2 The initial experience with the technique confirmed that it was robust, reproducible, and reliable in defining the sentinel node as representative of the regional nodal basin. Operational definitions of the concept of the sentinel node were developed and refined. Widespread use of lympho-scintigraphy led to a more sophisticated understanding of the variations in regional lymphatic drainage patterns.3 Subsequent outcomes analyses of large retrospective single-center and multicenter experiences have confirmed the sentinel node status to be a very powerful predictor of outcome in patients with localized melanoma.4 Initial early enthusiasm for identifying patients with subclinical nodal metastases who might benefit from adjuvant high-dose interferon has diminished substantially with the realization that the morbidity of this treatment is not clearly rewarded with a long-term overall survival benefit.5 However, the intrinsic value of sentinel lymph node biopsy as a staging procedure has justified its incorporation into most treatment pathways for melanoma.6 The 2002 American Joint Committee on Cancer melanoma staging system recognized the prognostic significance of the sentinel node by including a separate pathologic stage grouping that includes both microstaging of the primary tumor and pathologic examination of the sentinel node.
Most of the expressed reservations about sentinel lymph node biopsy have focussed on the increased cost and morbidity of a procedure that has not yet been shown to be associated with an improved outcome when compared with nodal observation.7 Although the morbidity of sentinel lymph node biopsy for melanoma is certainly less that that of elective lymph node dissection, it is measurable.8,9 However, two recent independent reports from Europe have raised a more serious concern: that the performance of sentinel lymph node biopsy may in itself alter the pattern of subsequent recurrence of melanoma, thus increasing the incidence of in-transit recurrence.
Estourgie et al.10 compared the incidence of intransit metastasis in 61 patients who underwent completion lymph node dissection for a positive sentinel lymph node with that in 60 patients who underwent therapeutic lymph node dissection for a clinically positive regional node. With comparable follow-up, and looking at in-transit recurrence as a component of all recurrences, they found 14 patients (23%) in the completion lymph node dissection group with in-transit recurrence, compared to only 5 patients (8%) in the therapeutic lymph node dissection group. Although the completion lymph node dissection group had thicker primary tumors and more ulcerated tumors, the authors reported that the only independent predictor of in-transit metastasis was the clinical status of the lymph nodes. Concerned about the possibility that the procedure of sentinel lymph node biopsy might itself be responsible for an increased incidence of subsequent in-transit recurrence, they cautioned that "further data on the benefits and long-term safety of sentinel lymph node mapping are clearly required" and that "it is premature to make removal of the sentinel node part of routine patient management."
Thomas and Clark11 published a review of reported patterns of recurrence after completion lymph node dissection in sentinel node–positive patients. They observed combined local/in-transit recurrences in 43 of 206 (21%) sentinel node-positive patients who underwent completion lymph node dissection, compared to 68 of 1199 (5.7%) sentinel node-negative patients who did not undergo completion lymph node dissection. This latter figure was quite comparable to the 4.4% of a combined 2501 patients who developed local or in-transit recurrence in the context of 3 prospective randomized surgical margin trials in which sentinel lymph node biopsy was not performed. Despite acknowledging that the incidence of in-transit metastasis was related to factors including primary tumor thickness and regional nodal status, the authors concluded that the "incidence of local/intransit recurrence following selective lymphadenectomy in sentinel node-positive patients may be greater than four times expected." Alluding to a possible iatrogenic risk of increased local or in-transit metastases, the authors went on to say that the "the SLNB [sentinel lymph node biopsy] procedure should be regarded as experimental and not performed outside validation trials."
Both of these reports imply that the procedure of sentinel node biopsy may itself alter in some way the pattern of subsequent recurrence of melanoma. Neither report fully addressed the overwhelming influence of tumor biology or patient selection on outcome, as only a prospective randomized trial is able to do. For example, it is not clear that node-positive patients found at sentinel lymph node biopsy are biologically comparable to those selected for therapeutic lymph node dissection in terms of their potential to develop local or in-transit metastases. Two large single-center retrospective series appear in this month’s issue of the Annals of Surgical Oncology and address some of the concerns raised by these European reports.
In the first, Pawlik et al.12 from the M. D. Anderson Cancer Center describe predictors of intransit recurrence after sentinel lymphadenectomy among 1395 patients undergoing sentinel lymph node biopsy for melanoma. As did Thomas and Clark, they embraced the conceptual definition of in-transit as any regional recurrence after adequate local therapy, regardless of the distance from the primary tumor site. They clearly describe the incidence of intransit recurrence as the first and only site (5%), as a component of the pattern of initial recurrence (6.2%), and as a component of the total pattern of recurrence (6.6%). These observations are not statistically different from the 4.4% cited by Thomas and Clark among 2501 patients who underwent wide excision without sentinel node biopsy. This would support the contention that initial treatment with wide excision and sentinel lymph node biopsy is not associated with an increased incidence of in-transit metastasis compared with wide excision and nodal observation. Independent predictors of in-transit disease as a component of the initial pattern of recurrence included age >50 years, increasing Breslow thickness and Clark level, ulceration, lower extremity site, and positive sentinel node status. Among 1136 patients with negative sentinel nodes, 3.5% developed intransit disease as a first site of recurrence. Independent predictors of in-transit recurrence among sentinel lymph node--negative patients included age >50 years, increasing Breslow thickness, and lower extremity site. Among 234 patients with a positive sentinel node, 12% developed in-transit disease as a first site of recurrence—a figure that is lower than the 24% observed by Estourgie et al.10 and the 21% calculated in the review by Thomas and Clark. Reasons for these differences are not clear, because the M. D. Anderson report does not provide specific details on the characteristics of the sentinel node-positive patients. The authors demonstrate the clinical significance of in-transit recurrence after sentinel lymph node biopsy: the 5-year disease-specific survival after in-transit-only recurrence was 55%. Although this is a superb description of the natural history of in-transit disease after sentinel lymph node biopsy, it does not directly address the concern of the Europeans that sentinel lymph node biopsy may in some way alter the pattern of subsequent recurrence.
In the absence of a prospective randomized trial, this concern is addressed more directly by the study from the Sydney Melanoma Unit.13 Querying an extraordinary clinical database of >22,000 patients, van Poll et al.13 were able to define 3 large cohorts of patients: 754 patients undergoing wide excision and sentinel lymph node biopsy, 1035 patients undergoing wide excision and observation with subsequent therapeutic lymph node dissection if necessary, and 229 patients undergoing wide excision and elective lymph node dissection. With a more restricted definition of intransit recurrence as that occurring between the primary excision site and the regional node basin, but excluding local recurrences within 3 cm of the primary tumor site, and with a median follow-up ranging from 35 to 61 months, the authors found no difference in the incidence of in-transit recurrence, either as a first site of recurrence or as a component of all sites of recurrence at any time. With the more restricted definition, they found in-transit disease as a first site of recurrence in 2.4% of patients after wide excision and sentinel node biopsy, in 2.5% of patients after wide excision alone, and in 4.4% of patients after wide excision and elective lymph node dissection. All of these figures are somewhat lower than those observed by the M. D. Anderson group and calculated by the London report. By multivariate analysis, increasing age and increasing Breslow thickness were the only independent predictors of in-transit metastasis as an initial site of recurrence. Although both positive nodes and a lower extremity site were strongly associated with subsequent in-transit disease by univariate analysis, it is not clear that these factors were included in the final multivariate model. However, the principal message of this report is that the incidence of in-transit recurrence does not seem to be associated with the manner in which the melanoma is initially managed.
It is difficult to compare these studies directly and arrive at a clear and unequivocal conclusion. We do not have complete information on patient and tumor characteristics that might predict in-transit recurrence in all patients in all studies, to ensure that the risk profiles of the groups are comparable. The definition of in-transit recurrence varies, either including or excluding that which occurs within 3 cm of the primary excision. Because "local recurrences" after adequate local therapy share the same pathogenesis and prognostic implications as classically defined intransit recurrence, it is appropriate to group them together as regional (nonnodal) recurrences. Finally, it is imperative that authors clearly define whether the incidence of in-transit recurrence cited is that as the first and only site, as a component of the pattern of initial recurrence, or as a component of all sites of recurrence at any time.
Inherent in the design of all retrospective studies is the absolute inability to draw any conclusions about the cause and presence or absence of an observed effect. Limitations of incomplete observations are compounded by innumerable variations of the multivariate models used to arrive at some comprehensible conclusions. Well-designed prospective randomized trials can overcome most of these limitations. The initial report of the seminal Multicenter Selective Lymphadenectomy Trial I, expected later this year, should provide a great deal of insight into the issues of both patterns of recurrence and survival after initial management with sentinel lymph node biopsy compared with nodal observation. Until those results are available, the two very large natural history studies from M. D. Anderson and the Sydney Melanoma Unit should reassure us that the overall incidence of in-transit recurrence after sentinel lymph node biopsy is low and is most likely influenced much more by patient, tumor, and lymph node variables than by any treatment variable. Concerns raised by the two European studies serve to underscore the importance of close follow-up of these patients and of carefully recording events as they occur in the course of that follow-up.
The European reports also prompt reflection on the broader issue of what we do and do not know about the procedure of sentinel lymph node biopsy in melanoma. Although the sentinel node status seems to be highly predictive of outcome in patients with melanoma >1 mm thick, its value in patients with thinner melanoma remains unclear (Wong et al., unpublished data). The practice of completion lymphadenectomy, based on some finite probability of disease detectable by routine pathologic analysis of nonsentinel nodes,14,15 has been embraced as a standard of care in the absence of evidence of improved regional control or survival. Although the Multicenter Selective Lymphadenectomy Trial II directly addresses this issue, those results will not be available for years. Furthermore, although the Multicenter Selective Lymphadenectomy Trial II was designed as an equivalency trial, very little is known about the natural history of patients with a positive sentinel node who do not undergo completion lymph node dissection. The focus of intense pathologic and molecular analysis of the sentinel node has raised the issue of exactly what tumor burden in the sentinel node is clinically relevant.16,17 Whereas initial reports of sentinel node analysis by reverse transcriptase-polymerase chain reaction (RT-PCR) for expression of the tyrosinase gene suggested that patients whose sentinel nodes were negative by both histology and RT-PCR had a uniquely favorable prognosis, at least by univariate analysis, one recent report has shown that with long-term follow-up, single-marker molecular staging adds little prognostic information to conventional histological evaluation of the sentinel node.18 Multimarker quantitative RT-PCR analysis of sentinel nodes may increase the sensitivity of molecular staging, but its prognostic relevance in long-term follow-up remains to be demonstrated.19
The two reports from Europe have served a very constructive purpose by prompting an in-depth look at the natural history of in-transit recurrence after sentinel lymph node biopsy in melanoma from two large single-center clinical databases. This type of evidence-based dialogue with those who raise concerns about commonly accepted practices can only strengthen our understanding of the important interactions among tumor biology, patient selection, and treatment. This type of dialogue will prevent us from ever becoming complacent in our care of cancer patients.
Received for publication February 22, 2005. Accepted for publication March 25, 2005.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
