The complementary roles of immunohistochemistry and electron microscopy.
The diagnosis of melanoma can be extremely challenging in some situations. The use of immunohistochemistry has become the initial "line of attack" to resolve diagnostic difficulties when surgical pathologists are confronted with poorly differentiated tumors. However, there are no stains that are 100% specific or sensitive in depicting melanomas. In most instances electron microscopy has become relegated to a secondary role and only utilized in a few selected cases when immunohistochemistry does not appear to provide a clear answer. This philosophy, although appearing to some practical and cost-effective, may not be so always.
It must not be forgotten that identification of melanin in the tumor cells either in routine H & E preparations or with the aid of a Fontana stain can easily identify melanotic melanomas. Undoubtedly, speciation of amelanotic melanomas represents the real diagnostic challenge and the information discussed here primarily applies to this category of melanomas.
Melanomas have been referred to as the "great mimickers". The possibility of melanoma should be considered in the differential diagnosis of virtually any poorly differentiated neoplasm. Tumor heterogeneity is exquisitely exhibited by melanomas. The heterogeneity of melanoma is manifested by multiple histologic patterns on light microscopic examination, variability of antigenic expression, and rather board ultrastructural morphologic spectrum. Surgical pathologists should be fully aware of the immunomorphologic variations that can be found in melanoma in order to accurately identify these tumors and not confuse them with many other neoplasms that they mimic closely. The evaluation of melanoma cases must be viewed in at least five different clinico-pathologic contexts.
The first situation represents perhaps the most clear-cut indication to use immunohistochemistry to confirm that the neoplasm represents a metastatic melanoma. If such a tumor is clearly HMB-45 positive (Figure), a diagnosis of metastatic melanoma could be considered confirmed especially if the apparent metastasis is morphologically similar to the primary tumor. This approach is cost effective and the risk involved is minimal.
In contrast, the second situation (A and B) would require a battery of immunohistochemical stains to address the diagnostic problem. Such a panel should include at least wide-spectrum keratin, S-100 protein, and HMB-45 and possibility two or three more stains, depending on the circumstances. In this situation aberrant immunohistochemical results in melanomas may be quite misleading (i.e., keratin positive melanoma or HMB-45 positive non-melanocytic tumor). The expense involved in working up these cases immunohistochemically is often more than if electron microscopy is performed first. In a significant number of these cases immunohistochemistry stains reveal inconclusive results while ultrastructural evaluation, in the hands of experienced electron microscopists, is very rarely inconclusive.
The third and fifth situations can also be tricky ones to define the "best initial plan of attack". I find that electron microscopy is often an easier initial approach to the differential diagnosis but those who have no access to electron microscopy would obviously select the immunohisto-chemistry route. If the immunohistochemical results are not clear, ultrastructural evaluation should be performed.
Finally in regards to cases in the fourth category, immunohistochemistry represents a reasonable starting point for the work-up of these cases.
In general, the use of electron microscopy in fine needle aspirates in which a diagnosis of melanoma is considered is much more productive than immunohistochemistry due to the fact that there are often relatively few cells in the samples and interpretation of immunohisto-chemistry is much more difficult in FNA specimens that surgical biopsies. However, proper collection of specimens from FNA procedures to perform electron microscopy becomes crucial.
A keratin negative, vimentin rich neoplasm, which also immunoreacts with antibodies to S-100 protein and HMB-45 is, with rare exceptions, a melanoma if this diagnosis is tenable from a clinico-pathologic point of view. The interpretation of light microscopic features and immunohistochemical stains must also take into account the full spectrum of immunomorphologic manifestations that may be encountered in melanomas. Similarly, the findings of HMB-45 staining in a tumor does not unequivocally indicate melanocytic differentiation; some non-melanocytic tumors may label with anti-HMB-45 antibodies. Neoplasms that present staining characteristics different from the expected typical melanoma pattern should be evaluated with electron microscopy before a final diagnosis of melanoma is rendered. A subset of melanomas will continue posing a challenge to the surgical pathologist and the correct identification of these must be pursued aggressively. Accurate diagnosis is crucial to appropriate clinical management, not just an academic exercise. In a significant number of these cases, electron microscopy may be the cost-effective, fast and accurate diagnostic modality of choice. Minimum criteria for the definitive diagnosis of melanoma at the ultrastructural level have been proposed to require at least the finding of atypical melanosomal forms in the proper cellular milieu. However, melanoma cells process a number of ultrastructural features that when present together, even in the absence of recognizable melanosomes, can strongly suggest or confirm a diagnosis of melanoma.
One genuine problem with using immunohistochemistry as the initial diagnostic technique, following with electron microscopy when the differential diagnosis of melanoma remains in doubt is that immunohistochemistry sometimes reveals staining patterns that are entirely convincing but erroneous. In such instances a need for ultrastructural evaluation is never appreciated. In general, electron microscopy has a much greater versatility and reliability with a significant lesser risk of leading to diagnostic mistakes. The cost and turn-around time associated with performing electron microscopy is comparable to immunohistochemistry in most cases. Both techniques are truly complementary and it is a shame that current financial pressures preclude suggesting a combined utilization of these two diagnostic modalities in the majority of the cases.