An increasingly common scenario that we may encounter is illustrated by the following hypothetical case:

A 12-year-old girl has a dome-shaped papule biopsied from an extremity. The histologic features are unusual but not overtly malignant, and the diagnostic differential includes an atypical Spitz nevus vs. a melanoma with spindle and epithelioid cell features. Consultation does not resolve the diagnostic dilemma. The patient is referred to a plastic surgeon who insists that a sentinel lymph node biopsy be performed as an adjunctive diagnostic tool.

A recent review article by Busam and Pulitzer1, ‘Sentinel lymph node biopsy for patients with diagnostically controversial Spitzoid melanocytic tumors?’ cogently explores the literature surrounding the above scenario. We will touch upon some of their important points and expand the discussion with additional references.

The featured review article begins with the original commentary by Kelley and Cockerell in 20002 in which two reasons were proposed for sentinel lymph node (SLN) biopsy in difficult melanocytic tumors: one that if the tumor is indeed a melanoma, then SLN would be an appropriate procedure, and two, the presence of nodal positivity may help validate the diagnosis of melanoma. The ensuing discussion illustrates the pros and cons of SLN. Distinction between an atypical Spitz nevus and a melanoma can be difficult, as some lesions present with significant morphologic overlap. Even among experienced dermatopathologists, there is poor interobserver agreement in classifying these diagnostically problematic Spitzoid melanocytic lesions.3 As such, the uncertainty in the prediction of biologic behavior and difficulty in morphologic classification have led some to adopt SLN biopsy as both a prognostic and diagnostic adjunct in these patients. Additional support for this is in the often quoted statement that ‘a melanocytic neoplasm that metastasizes is melanoma’, by Mones and Ackerman.4

The practice of assigning histopathologically ambiguous Spitzoid lesions as malignant based on SLN positivity is increasingly being reported, especially in the surgical literature. In one recently described case, a 7-year-old girl with a ‘melanocytic lesion of uncertain malignant potential’ was deemed to have a melanoma after a positive SLN was found; the patient subsequently underwent axillary clearance with removal of 27 additional lymph nodes.5 The case illustrates the potential management quagmire that we may find ourselves in when accepting SLN biopsy as a means of diagnosing melanoma. Finding a lymph node with melanocytes may lead to a complete nodal dissection. If we do not know the risk (i.e. biologic behavior) of a ‘melanocytic lesion of uncertain malignant potential’, then how can we make an evidenced-based decision to do a complete lymphadenectomy?

Does nodal presence of melanocytes equate metastasis? As mentioned by Busam and Pulitzer and others, lymph nodes are known to harbor foci of many normal cell types from tissues that they drain, including breast epithelium, salivary gland epithelium, mesothelium, Mullerian epithelium, decidua, thyroid and urothelium.1,6 Small nodal deposits of benign tumors, such as Warthin’s tumor, pleomorphic adenoma of salivary gland and metanephric adenoma, have also been seen and rarely pose diagnostic problems.6 Nodal melanocytic nevi associated with benign common acquired or congenital melanocytic nevi have also been described.6,7 Benign nodal nevi have been found in staging lymph node biopsies for other carcinomas in patients without history of melanoma before or after lymph node removal as well as in lymph nodes of melanoma staging cases.8,9

Busam and Pulitzer1 also review and emphasize three criteria used to distinguish nodal nevi from metastatic melanoma: location within lymph node, cytology and immunohistochemistry. The location of melanocytes within the lymph node (i.e. intraparenchymal, capsular) has been used in the determination of benign vs. malignant metastasis; however, there are exceptions. Nodal nevus cells are usually encountered within the fibrous capsule and trabeculae, as opposed to melanoma metastases, which are generally found intraparenchymally and only rarely in the capsule.1 Biddle et al.7 reported 13 cases of nevus cells seen in the cortical and/or medullary parenchyma of lymph nodes from three patients without known malignancy, one with breast carcinoma, one with cutaneous adnexal carcinoma, one with tonsillar squamous cell carcinoma, in addition to seven patients with melanoma. All 13 cases of intraparenchymal nevi had histologically identical, accompanying aggregates within the capsule or trabeculae. The nodal nevus cells showed no mitoses, prominent nucleoli nor lymphovascular invasion and were morphologically different when compared with the primary tumor. The nevus cell collections ranged from only a few cells to aggregates of 2.1 mm. Immunohistochemical studies revealed S-100 and/or MART-1 positive but HMB-45 negative staining. Less than 1% of the nevus cells were Ki-67 positive as opposed to a co-existent melanoma metastasis, which showed increased HMB-45 and Ki-67 positivity.7 Biddle et al.7 as well as Busam and Pulitzer1 acknowledge that due to variance in cytology it can be impossible to exclude the possibility of metastatic melanoma and stress the importance of comparing the morphology of the nodal component to the primary lesion.1,7,9 The pitfalls of immunostains are also discussed; specifically the limitations when nodal aggregates are very small, as is often the case, and the need to exercise caution when interpreting HMB-45 results,1,7 which can help support the diagnosis of melanoma, but can by no means exclude it. A recent publication by Brennick and Yan10 cautioned about MART-1 false positivity when examining lymph nodes for melanocytes because non-melanocytic cells containing cross-reacting antigens or phagocytes containing melanocyte antigens may also stain positively.

Five studies in the past 6 years report the experience of 71 patients undergoing SLN biopsies for diagnostically problematic Spitzoid tumors.11–15 In summary, the results of the five studies show that in toto 26 (36.6%) of 71 patients with atypical Spitz tumors who underwent SLN biopsies had positive nodal melanocytic aggregates.1 Except Breslow’s depth, there were no histological or clinical features that distinguished the nodal positive and negative groups.11–15 The size of the nodal melanocytic aggregates ranged from 0.1 to 5 mm. 11–15 The location of nodal melanocytic aggregates included multiple areas such as the subcapsular, subcapsular sinus and/or parenchyma.11–15 Two non-Spitzoid capsular nevi were observed.14 The high rate of SLN positivity is, at first glance, alarming, and Urso et al.13 state that such a high rate of metastasis is incompatible with the belief that atypical Spitz tumors belong to a group of tumors with relatively low metastatic potential. Urso et al.13 also conclude that their eight nonmetastasizing cases exhibiting the same histological characteristics as their metastasizing ones should be considered malignant in addition to the four cases with nodal deposits. Mooi and Krausz argue that Spitz nevus cell aggregates in the lymph node may resemble a melanoma metastasis more than benign nevus nodal aggregates. This would make the lymph node deposits in atypical Spitz tumors difficult to interpret and ‘directly challenge(s) the appropriateness of a sentinel node biopsy to clinch a diagnosis of Spitzoid melanoma’.6

When nodal melanocytes do appear overtly malignant with mitoses, necrosis and near replacement of the node, there is little question about diagnosing the primary tumor as melanoma. But not all cases are so obvious. In a classic cover case showing a lymph node with subcapsular large oval melanocytes, LeBoit16 asks, ‘What do these cells prove?’ In the absence of data on the natural behavior of classic Spitz nevi, LeBoit16 questions whether we can be sure of the significance of melanocytic aggregates. Urso et al.13 suggest that pathologists may be under diagnosing melanomas and that criteria for malignancy within ‘atypical Spitz tumors’ need to be reevaluated. Mooi and Krausz6 state, ‘Just like Spitz nevus, in the distant past, was generally over-diagnosed as melanoma, we may now be making a similar mistake if we interpret any lymph node deposit consisting of Spitzoid melanocytes as formal proof of metastasis. If the phenomenon of spread to regional lymph node of a benign nevus is a regular occurrence, it might similarly happen in a Spitz nevus’. Conversely, what does a negative SLN biopsy prove in atypical Spitz tumors? Absence of nodal metastasis does not exclude a malignant primary tumor, as metastasis from melanomas can be seen long after diagnosis. Lymph node biopsies are for staging purposes and not for diagnosis. What is the sensitivity and specificity of the diagnostic test of SLN biopsy? How can we even determine the positive predictive value without knowing the prevalence of ‘Spitzoid’ melanomas? For any diagnostic test, there is the gold standard to which we must compare the test. Unfortunately, time or retrospect is our current gold standard.

Ultimately, patients with difficult melanocytic lesions need to be managed, despite the lack of a cogent evidence-based approach. Is it for fear of the unknown that would argue the safest route would be to use SLN as a diagnostic tool, for fear of missing a diagnosis or for fear of not knowing the biologic potential of an atypical Spitz nevus? In doing so, are we unduly labeling individuals with an erroneous diagnosis and subjecting them to unnecessary procedures and chemotherapy? New modalities, such as comparative genomic hybridization (CGH) and fluorescence in situ hybridization (FISH) may offer objective means to assess the malignant potential of melanocytic lesions. However, these methods have yet to be validated by widespread use or substantial outcome data. Busam and Pulitzer propose an algorithm utilizing CGH and FISH in the evaluation of these tumors, suggesting that atypical Spitz tumors in the absence of specific CGH or FISH findings associated with melanoma should forgo SLN procedures. This approach could conceivably avert spurious SLN biopsies yielding information that, at best, we do not fully understand and at worst may lead to harm to our patients. As an editorial by Wick17 states of SLN in atypical Spitz tumors, ‘Compounding uncertainty with more uncertainty is never a good idea’.


  1. Top of page
  2. References