3.1

Breslow thickness

Since its description by Breslow in 1970, the correlation of regional nodal spread with thickness of the primary tumor has been firmly established in the melanoma literature, and the current AJCC staging system bears out this fact. The Breslow thickness is defined as the distance measured from the epidermal granular cell layer to the deepest layer of invasion below the epidermis.

The risk for occult nodal metastasis in intermediate-thickness melanoma (1–4 mm) is well established and significant, and the decision to perform SLNB solely on the basis of a Breslow thickness > 1 mm is accepted and considered the standard of care. The threshold for recommending SLNB on the basis of thickness in thin melanomas is somewhat less clear. In perhaps the largest study of thin melanomas, Faries and colleagues retrospectively evaluated 2211 patients with thin melanoma treated solely by wide local excision for the presence of regional nodal basin recurrence. This study found that regional recurrence increased as a function of increasing Breslow thickness, with incidences of nodal recurrences for lesions < 0.25 mm, 0.26–0.50 mm, 0.51–0.75 mm, and 0.75–1.0 mm being 0%, 1.1%, 4.3%, and 8.5% respectively.

In a systematic review of 24 studies investigating predictors of nodal metastasis in thin melanoma, Andtbacka and Gershenwald observe that the incidence of nodal metastasis in primary lesions < 0.75 mm in thickness is quite small, approximately 2.7%. For lesions exceeding 0.75 mm, however, they found that the average rate of SLN positivity was 6.2%.

For lesions less than 0.75 mm, the risk for occult nodal disease is low, and many studies have suggested that it is difficult to justify SLNB in this patient subset . Current NCCN guidelines, however, recommend SLNB for all T1b melanomas. Thus, lesions < 0.75 mm with ulceration or a mitotic rate (MR) ≥ 1/mm ² , would technically qualify for SLNB according to guidelines. This has been questioned in a recent article by Sabel , who aptly addresses whether mitotic rate alone is sufficient to justify SLNB. While the management of these lesions remains controversial, the use of SLNB in any lesion < 0.75 mm should be highly selective and based on additional pathologic and clinical features thought to increase a patient’s risk.

For any lesion between 0.75 and 1 mm, the current literature suggests that SLNB may be justified solely on the basis of depth, even in the absence of other high-risk features. In a 2006 study by Wong et al. at Memorial Sloan Kettering Cancer Center, a study of 223 patients with thin melanomas revealed that no clinicopathologic factors were associated with positive SLNB, with the exception of Breslow thickness greater than 0.75 mm. This study, among others, stimulated inclusion of Breslow depth > 0.75 mm as an “adverse feature” predictive of nodal metastasis in T1a lesions.

3.2

Clark level

Prior to the updated 2009 AJCC staging system, the presence of a Clark level IV depth of invasions was used to upstage thin lesions from T1a to T1b. Since that time, support for Clark level as an independent predictor of nodal metastasis has faded. In the systematic review performed by Andtbacka and Gershenwald , they reported that only 2/10 studies found that Clark level had a statistically significant bearing on nodal disease. Similarly, Faries et al. found that Clark level was not predictive of subsequent nodal recurrence. Historically, however, Clark level IV/V has been used as a marker for more aggressive disease and has been found in one large study to predict occult nodal metastasis .

3.3

Ulceration

Since the sixth edition of the AJCC staging system set forth in 2002, ulceration at the primary site has been recognized as a predictor not only of occult nodal disease, but also of poor prognosis and decreased survival. Given its generalized acceptance as a negative prognostic variable, there exists little controversy regarding performing SLNB for thin lesions in the presence of ulceration. The problem, however, lies in the fact that very few thin melanomas have ulceration. The incidence of ulceration in thin melanoma specimens has been shown to be only 4%–9% .Thus, while ulceration likely represents a clear indication for SLNB, it is limited in its use as a sole selection criterion.

3.4

Anatomic site

The incidence of melanoma is disproportionate with respect to its body site distribution. The head and neck region accounts for merely 9% of total body surface area and yet harbors nearly 18%–35% of all cutaneous melanomas. Additionally, these lesions seem to be associated with a poorer prognosis than melanomas at other sites .In a large-scale review of nearly 6300 cutaneous melanomas at all sites at Duke Medical Center, CMHNs were associated with the lowest 10-year survival rate (54%), compared to truncal (61%) and extremity (76%) melanomas . Additionally, a recent study by Callender et al. on 2500 intermediate-thickness melanomas found that the location of the primary in the head and neck region was associated with decreased disease-free and overall survival, and decreased incidence of SLN positivity.

The reason for this trend is unclear. Given that the presence of subclinical nodal metastasis is the most important prognostic factor with melanoma at any site, the question naturally arises as to whether CMHNs, based on anatomic location alone, have a higher propensity toward nodal metastasis. Though intriguing, this has not been verified in the literature, in fact head and neck melanomas have been found in one study to have lower rates of SLN positivity . Faries et al. found that head and neck primary location was not a statistically significant predictor of regional disease, though head and neck primaries had the absolute highest rate of regional recurrence at 3.6%. Subsequent studies have also failed to demonstrate any direct correlation between the anatomic location of the primary and the rate of SLN positivity . However, the false-negative rate (FNR) and regional recurrence rate for SLNB in the head and neck region are higher relative to other sites . A systematic review by de Rosa et al. reported a median FNR of 20.4%, vastly exceeding that associated with truncal and extremity melanomas. This elevated FNR and rate of regional recurrence have been thought to result from the unpredictable nature of lymphatic drainage in the head and neck and the technical difficulties associated with SLNB in certain areas of the head and neck.

A more recent series of 353 patients with CMHN who underwent SLNB, however, demonstrated the safety and efficacy of this technique in the head and neck region . This study demonstrated that the negative predictive value of a negative sentinel lymph node was 95.8%. In other words, patients with local control and a negative SLN failed in the regional basin in 4.2% of cases. The rate of positive SLN was 20% and the status of the SLN was found to be the most important predictor of poor outcome. The comparable rate of SLN positivity to truncal/extremity melanoma and the low FNR in this large single-institution series suggest that SLNB is accurate and reliable, and its results hold prognostic significance in CMHN. It is important to note, however, that all cases in this study were performed by surgeons specializing exclusively in the head and neck.

Little retrospective data exist, specifically, on thin melanomas of the head and neck. However, a recent article by Jaber et al. described 49 patients with thin melanoma of the head and neck and recommended performing SLNB for lesions exceeding 0.75 mm, regardless of the histopathologic aspects of the tumor. Additionally, they found no evidence of nodal disease in any patient with a lesion < 0.75 mm, regardless of pathologic features.

3.5

Tumor mitotic rate (TMR)

The relevance of dermal melanocyte proliferation to tumor invasiveness has been a relatively recent development in melanoma . Gimotty and Guerry at the University of Pennsylvania have demonstrated the overwhelming importance of mitogenicity in the prognosis of patients with melanoma. This 2004 study of 884 patients with thin melanoma found that only mitogenicity (mitotic rate > 1 per mm ² ) was an independent predictor of both nodal micrometastasis and adverse prognosis. In an innovative prognostication scheme based on mitotic rate, vertical growth phase and sex, these authors stratified patients into four distinct risk groups that had a greater predictive power for 10-year mortality than the current AJCC staging system . Men with mitogenic tumors with vertical growth phase were the risk group most likely to develop micrometastatic disease, at a rate of nearly 31%. In this schema, a mitotic rate < 1 per mm ² was consistently associated with 10-year rates of nodal metastasis less than 5%. More recently, a study of over 3600 patients from the Sydney Melanoma Unit also found mitotic rate > 1 per mm ² associated with a significantly decreased melanoma-specific survival . In this study, mitotic rate was second only to Breslow thickness with regard to prognostic significance. Interval increases in mitotic rate above 1 per mm ² have not been associated with a trend of decreasing survival. The impressive results of these and additional studies prompted the AJCC to revise staging in 2009 to include mitogenicity as a defining factor of T1b thin melanomas. These studies suggest that mitotic rate > 1 per mm ² is sufficient to warrant SLNB at any Breslow depth and in the absence of any additional adverse features.

3.6

Tumor regression

Primary tumor regression is the histologic finding of tumor loss associated with variable stromal inflammation. While the underlying biologic basis is unclear, the finding of regression in the biopsy specimen may indicate an inadequate evaluation of the true Breslow thickness. Interestingly, the incidence of tumor regression has been found to vary inversely with Breslow thickness, with thin lesions having the highest documented rate of regression at 46% . Thus far, no study has convincingly found regression to be an independent predictor of SLN positivity or poor prognosis . A recent study by McClain et al. found no survival difference between patients with partially or completely regressed primaries who underwent wide local excision (WLE) and SLNB and those who underwent WLE alone. Unfortunately, given the tremendous variation among institutions and pathologists in the histologic diagnosis of regression, its prognostic significance is difficult to study. As a result, tumor regression remains a controversial prognostic factor vis-à-vis SLNB, and several centers continue to use extensive regression (defined as > 75% of the primary tumor) as criteria for performing SLNB.

3.7

Tumor-infiltrating lymphocytes

The presence of tumor-infiltrating lymphocytes (TILs) is, in theory, advantageous in that it represents some degree of anti-tumor immunity. Historically, this response has been described as brisk, non-brisk, or absent. While the available data are mixed regarding the implications of TILs, Gimotty et al. found that the presence of brisk TILs was associated with a significantly reduced rate of 10-year metastasis in thin melanoma. Similarly, Kruper et al. found that the absence of TILs was an independent predictor of nodal metastasis in patients with stage I and II melanoma, and afforded a threefold increase in subclinical nodal disease at diagnosis. Though the absence of TILs has yet to be universally accepted as an adverse feature predictive of nodal metastasis, these studies suggest a possible future role as an additional biomarker for regional progression.

3.8

Gender

Several studies to date suggest that men are more likely to die from melanoma than women. In patients with thin melanoma, Faries et al. found male sex to be a significant predictor of occult nodal metastasis. In over 1700 patients with thin melanoma, the incidence of occult nodal disease was 4.4% in all male patients and only 1.4% in females. Gimotty and Guerry found male gender to be predictive of both elevated 10-year metastasis rates and decreased overall survival.

3.9

Age

Younger age has been found in several studies to be a significant predictor of nodal metastasis .The current theory regarding this observed phenomenon is that lymphatic function decreases with age and renders older individuals at a lower risk of regional disease . Faries et al. found that patients between the ages of 30 and 39 years had the highest incidence of SLN positivity, and that nodal disease decreased almost linearly with age. Sondak and colleagues found age < 35, especially in the presence of an elevated tumor mitotic rate, to be significantly associated with an increased risk of micrometastatic disease. Also, Wright et al. found age < 50 to be independently associated with SLN positivity. In their review, Andtbacka and Gershenwald cite an age of < 40 years as a reasonable threshold for offering SLNB in thin melanoma.

3.10

Management of positive sentinel nodes in thin CMHN

Just as the selection of patients for SLNB is complex in the setting of thin lesions, so too is the management following identification of a positive SLN on pathologic analysis. The current standard of care entails a complete therapeutic lymph node dissection (cTLND). According to a recent survey of 337 surgeons across 25 countries, 92% reported routinely performing cTLND in the setting of a positive SLNB . The rationale for cTLND following positive SLNB stems from the results of the Multicenter Selective Lymphadenectomy Trial , which found a survival benefit and improved regional control in patients undergoing SLNB and subsequent cTLND with a positive SLN compared to those who underwent simple WLE, observation, and subsequent salvage cTLND for regional failure. While this study has driven the current standard of care, its results have nevertheless been subject to controversy. Some recent studies have called into the question the survival advantage seen in this study, finding a lack of survival difference between patients with a positive SLNB vs. false negative . The final analysis of MSLT-1 will hopefully clarify what survival advantage, if any, SLNB plus cTLND provides patients.

Also controversial is the fact that the overall incidence of positive non-sentinel nodes in cTLND specimens is relatively low, and morbidity of a neck dissection can be considerable. The incidence of positive non-sentinel nodes in cLND after positive SLNB has been found to be 14%–17% . Thus, for a large proportion of patients with positive SLNB, a cTLND may represent overtreatment and an unnecessary risk of morbidity. Given that lymph nodes removed during cTLND are not scrutinized to the same extent as the SLN; however, the incidence of positive non-sentinel nodes may very well be underestimated. Clinicopathologic characteristics of the SLN such as microscopic tumor burden, Breslow thickness > 2 mm and the number of SLN harvested have all been found to be significant predictors of non-sentinel lymph node status . In patients with thin lesions who undergo complete sentinel lymphadenectomy, therefore, the risk for non-sentinel node positivity should be exceedingly low. The currently ongoing Multicenter Lymphadenectomy Trial-2, which randomizes patients to cTLND versus close sonographic nodal observation for a positive SLN, should quantify the true risk for non-sentinel node positivity. While the trial includes melanomas with a thickness of 1.2 mm or greater, it also includes melanomas Clark’s level IV or V or ulcerated melanomas regardless of Breslow thickness. Thus, it may provide some insight into which of these patients with thin melanomas might be spared the morbidity of a cTLND.