Numerous studies have validated the sentinel lymph node (SLN) biopsy procedure as a safe, effective method for providing patients who have melanoma with highly accurate staging and prognostic information to aid in decision-making.1, 2 The results of this procedure guide several important clinical decisions, such as the need for complete lymph node dissection, adjuvant therapy, and eligibility for clinical trials.3 Although controversy remains concerning the role of SLN biopsy as a therapeutic procedure,4, 5 the value of the prognostic information is great, even in the setting of a negative SLN. However, because the majority of new melanomas are diagnosed when they are extremely thin (65% measure <1 mm in thickness),6 numerous patients who undergo this procedure will have negative SLNs and an excellent prognosis. Not surprisingly, there has been a recent explosion of literature attempting to define the attributes related to a favorable outcome in patients with thinner melanomas who undergo SLN biopsy.6-9 However, a persistent and very real criticism of those studies is the issue of selection bias, which may influence the choice of patients for the procedure in this low-risk group—virtually none of the studies involved truly “unselected” patients who had thin melanomas.
In this issue of Cancer, Mays et al10 report on a subset of patients from the Sunbelt Melanoma Trial, a multicenter, prospective trial involving 79 centers and >3600 patients who underwent SLN biopsy11 that commenced enrollment in 1997 and was closed to enrollment in 2003. Their current report focuses on 1110 patients from that trial who had melanomas that measured between 1.0 mm and 2.0 mm in thickness, all of whom underwent SLN biopsy as part of that trial (patients with melanomas <1.0 mm and those aged >70 years were excluded from the trial). Thus, by studying this group of patients, the authors avoided some of the selection bias that has complicated attempts to define prognostic factors for SLN positivity in thinner primary melanomas. In their study, Mays et al analyzed several clinical and pathologic factors in an attempt define a subset of patients at especially low risk for SLN involvement and to assess the impact of a positive SLN on disease-free and overall survival.
Their report illustrates how ongoing analyses of the large group of patients diagnosed and prospectively treated in a uniform manner on the Sunbelt Melanoma Trial continues to generate informative data—much of it confirming previous findings from institutional reviews. The 12% incidence of SLN positivity in their subset of the Sunbelt Melanoma Trial population (melanomas between 1.0 mm and 2.0 mm) certainly validates the current “standard practice” of offering SLN biopsy to all healthy patients who have tumors ≥1 mm in depth. They also demonstrated, even in this selected group of patients, that Breslow depth, patient age, ulceration, and SLN status each had a significant impact on disease-free and overall survival in multivariate analysis. In their current report, Mays et al confirmed the significance of Breslow depth and lymphovascular invasion in predicting lymph node positivity, whereas regression and Clark level demonstrated no impact on the status of the SLN. By using a recursive partitioning method, a breakpoint depth of 1.60 mm was identified that defined 2 distinct subsets of patients with very different risks of lymph node involvement: 8.7% for melanomas between 1.00 mm and 1.59 mm compared with 19.3% for melanomas between 1.60 mm to 2.00 mm. In the group with thinner melanoma, age >59 years and the absence of lymphovascular invasion further delineated a subset of patients with an even lower risk of lymph node metastasis (4.9%).
Mays et al conclude—rightfully in our opinion—that these data do not define a group of patients with melanomas between 1.0 mm and 2.0 mm who have such a minimal risk of lymph node metastasis that foregoing the SLN procedure is justified. Furthermore, the fact that, in the Sunbelt Melanoma Trial, a significant portion of each SLN was procured for polymerase chain reaction analysis raises the possibility that an even higher rate of lymph node positivity might have been the case if some of the metastases had been missed by this sampling. That being said, the study validates the current standard practice of offering SLN biopsy to all healthy patients who have tumors >1 mm; the 8.7% positive lymph node rate for melanomas between 1.00 mm and 1.59 mm justifies the authors' continuing to advocate SLN biopsy in this group but emphasizes the potential role for improved selection criteria.
Future investigations aimed at defining lower risk subsets need to include an assessment of mitotic rate; the absence of these data in the Sunbelt Melanoma Trial is unfortunate, especially considering the role mitotic rate plays in the upcoming revision of the American Joint Committee on Cancer (AJCC) melanoma staging system. It is interesting to note the similarity between the cutoff of 1.59 mm used by Mays et al and the 1.50 mm delineation between stage I and II melanomas in past (1997 and earlier) iterations of the AJCC melanoma staging system, which raises the very real possibility that 1.5 mm or 1.6 mm may be a biologically significant cutoff point more suitable for staging purposes than 1.0 mm. It is entirely possible that the group of patients with melanoma between 0.76 mm and 1.5 mm may be more alike than different, especially when the mitotic rate is factored in, and this bears further consideration. With regard to the AJCC, that system excludes 1.0 mm melanomas from the T2 category, yet the current study and virtually all melanoma surgeons offer SLN biopsy to patients with melanomas that measure precisely 1.0 mm in depth even if they do not offer it to those with melanomas <1.0 mm. In our opinion, it is time for future versions of the AJCC system to reflect current practice and to categorize 1.0 mm melanomas as T2 instead of T1, as currently classified, if, indeed, 1.0 mm is going to be retained as the cutoff point for T1 versus T2.
An additional observation in this study lends support to the premise that younger patients have a higher rate of positive SLNs, which was a finding in several previous large institutional series.9, 12 Hopefully, analysis of the mature Sunbelt Melanoma Trial data as a whole will shed light on the paradox of this phenomenon, because this study and many others have equated improved survival with younger age at diagnosis.13 It is noteworthy that Mays et al observed an inconsistent statistical impact of a positive SLN on outcome in the cohort of patients who had melanomas between 1.0 mm to 1.59 mm, with significantly worse overall survival but not disease-free survival. Whether this is a statistical aberration or a true biologic difference should be investigated further in a larger cohort of patients; currently, it is more plausible to considered it related to post hoc subsetting of an already post hoc subset.
In conclusion, as we continue to learn from the large group of melanoma patients who enrolled in the Sunbelt Melanoma Trial, our analyses continue to raise questions. These questions should serve as a roadmap for future investigations and revisions to the melanoma staging system:
How well will the Sunbelt Melanoma Trial findings in patients with melanomas ≥1.0 mm mirror those in patients with thinner tumors? Is it time for a larger prospective trial of SLN biopsy specifically in thin melanomas?
What other pathologic factors should be analyzed routinely, especially in large prospective studies, to evaluate their role in prognostication: Tumor-infiltrating lymphocytes? Vertical versus radial growth phase?
Should we cease to evaluate regression altogether? By now, it should be accepted that there is no indication for using regression to select patients with thin melanoma for SLN biopsy.
What is the right threshold for SLN biopsy? A 5% expectation of finding a positive lymph node (which, by the way, is our current threshold for healthy patients aged ≤70 years)? 10%? Should the decision be “personalized,” taking into consideration patient age and overall health as well as the risks of the procedure and the likelihood the patient will pursue adjuvant therapy if the SLN is positive?
When the SLN is positive, are there features of the primary melanoma or the SLN metastasis itself that predict different outcomes and could influence the choice of adjuvant treatment? In particular, are there any lymph node “micrometastases” so minute that they are clinically insignificant, allowing patients to safely avoid completion lymph node dissection and adjuvant systemic therapy?
Because thinner melanomas constitute such a large percentage of all newly diagnosed melanomas, and because SLN micrometastases constitute such a large percentage of all newly diagnosed stage III melanomas, the “impact” of answering these questions for future melanoma patients is extraordinarily high. The road ahead is clearly marked, and it seems the time is right for further prospective trials to carry us on this journey of discovery.