Spindle Cell Melanocytic Neoplasia


 with Desmoplastic Melanoma as an Example

by Richard J. Reed, M.D.
New Orleans, La.     USA
June, 1999

(Note: Click on anything that is underlined to go to the appropriate figures.)


Figure 1 is an attempt to diagrammatically depict the stages of neoplasia in a model melanocytic system. In it, dimensionalities are primalities. Premalignant melanocytic dysplasias are 2 dimensional and represented as a segmented plane. Melanomas are three dimensional and represented as a cone. The cone intersects the plane and extends to the segment of the plane embracing the domain of the moderate dysplasias (moderate atypia); a neoplasm may enter vertical growth without progressing as a dysplasia to the stage of marked dysplasia (atypia). This attribute validates the concept of minimal deviation melanoma (MDM). In the diagram, the common sequences are represented. For special variants of melanoma, new virtual images are required; in turn, the diagram would have to be modified. In the modifications, the pattern characterized as vertical growth takes on new significance and is not always a marker for a melanoma with a predictable potential for metastasis. A borderland is required and is represented by two boundaries defined by measurement of vertical dimensions, one at 1 mm and the other at 1.5 mm.

 Configurations (primary and otherwise) and MDM:

 Most melanomas of the skin are represented in primaryconfiguration; lentiginous and junctional components are represented in the epidermis. The implications of a primary configuration is that the lesion has had its origin from a population of cells in the epidermis. On the other hand, some lesions in a "primary configuration" represent epidermotropic metastases. Some lesions of the melanocytic system of neoplasia appear to have their origin in a dermal population of cells rather than in cellular phenomena at the dermal-epidermal interface. If a new lesion in primary configuration, or even one that is strictly confined to dermis, is manifested in patterns that contribute bulk in three dimensions, is cytologically atypical and mitotically active, and is associated with markers for host immune response, caution should be observed, (if the favored option is to assign such a lesion to some category other than melanoma). The nature of such lesions (i.e., those which have bulk in 3 dimensions) may be masked by a degree of cytologic atypia which is less than that of most vertical growth components of the common melanomas. Some lesions of this type have been characterized as minimal deviation melanomas and more recently have been assigned to the category of nevoid melanoma. Variant patterns which share features with those of the variant "nevi’ (melanocytomas) characterize the nevoid melanomas. In this approach, the concept of nevoid melanoma is simply a disguise of the concept of MDM. In the category of MDM, the nevocyte-like type (including the dermal variants and congenital nevus variants), the Spitz nevus-like variant, the pigmented spindle cell variant, and the blue nevus type have features of the respective "nevi" and are "nevoid." There are other variant melanomas whose features are not readily appreciated as "nevoid." Some of these variants, on the basis of bland cytologic features and deviant patterns, might be characterized as minimal deviation variants, particularly in early stages of their evolution. These lesions, as they progress, acquire the features of a high grade malignancy and relinquish the properties that might be characterized as minimal deviation. In the latter group, desmoplastic melanoma and neurotropic melanoma are examples. Young, thin melanomas are better appreciated as minimal deviation variants than are older lesions. Dedifferentiation is a property of aging (progression is accompanied by changes in grades and by increase in bulk); in the process, patterns become more easily equated with a potential for metastasis. All of this is relative; a significant deviation in patterns may lead an unwary pathologist into the realm of virtual images which have no relevance to the problems at hand. Relativity must be factored into our prognostications.

 The primality of vertical growth components:

 Part of the difficulty in the classification of problematic lesions relates to limitations that have been imposed on our classifications. The limitations relate to emphasis on radial rather than vertical growth components in the assignment of melanomas to categories. Most of us have so studiously accepted Clark’s classification of melanomas that little attention has been paid to the degree of cytologic deviations in vertical growth components. Evaluations of degrees of atypia are neither an exact, nor consistently reproducible, technique. Some difficult melanocytic lesions are remarkably bland. On the other hand, some of us have been blindly focused on melanoma in situ. With the restrictions that are inherent in the concept of melanoma in situ, in which dimensionalities have no relevance for the diagnosis of melanoma, all subsequent patterns, otherwise attributable to neoplastic progressions in steps, are merely "melanoma" with no provisions that might explain the remarkable variability of histologic patterns and degrees of cytologic atypia in vertical growth components; vertical growth and degrees of atypia are irrelevancies for proponents of the concept of "melanoma in situ." Although seldom acknowledged, degrees of atypia are partially assessed if attention is paid to mitotic rate; the two are related. In special melanomas, in which cell to cell arrangements are not epithelioid, difficulties in diagnosis may relate not only to unusual interactions between tumor cells and stroma but also to degrees of cytologic atypia. Acknowledgment of the importance of degrees of atypia is a validation of the concept of MDM.

 Spindle cell melanomas:

 Spindle cell melanomas are relatively common; some of the cells of such lesions commonly are large, and epithelioid or rounded (here, "epithelioid" refers to the configuration of individual cells without attention to the cell to cell relationships). Spindle and epithelioid cell melanoma is not a useful designation in the current scheme of classification and prognostication. Spindle cell patterns in the dermal component of a melanoma are commonly mistaken for the patterns of a Spitz "nevus," hence, the current controversies (i.e., "Spitz nevus/tumor"), and the indecision as to the criteria for the diagnosis of "Spitz nevus." If the designation (abomination), "Spitz nevus/tumor," were to be accepted, it would only confound an already befuddled mess. There are no intentions in the terms, nevus and tumor, either individually or in combination, which connote a borderland of neoplasia in which some lesions minimally deviate, both in pattern and cytologic features, from variant nevi; at least, there is a point of reference in the concept of minimal deviation melanoma, and the admission that some of these troublesome lesions will in fact metastasize. The concept of minimal deviation melanoma is an attempt to supply the intentions and tools required for the manipulation of deceptive patterns.

Spindle cells and lentiginous patterns are interrelated. In some lentigo maligna melanomas, the spindle cells of some vertical growth components are remarkably bland cytologically. In such examples, the cytologic features in vertical growth components (typical pattern III) may resemble cells in the dermal component in pigmented spindle cell "nevus." For such lesions, the distinctions between benign and malignant variants are made in part on the character of the associated lentiginous and junctional components. Such lesions qualify as MDM of the lentigo maligna type
Lentiginous spread in the epidermis is a manifestation of the spread of individual melanocytic cells among basal keratinocytes; it is evidence of the migration of individual cells rather than the movement of clusters (i.e., community) of cells. It is tempting to propose a sequential relationship between cells in lentiginous patterns in the epidermis and the occasional spread of neoplastic melanocytic cells into the dermis as individuals rather than in clusters. In this proposal is the additional implication that the lentiginous melanocytic neoplasias have a propensity for the spread of individual cells.

Most melanomas enter vertical growth in the patterns of nests of cells (i.e., cell to cell arrangements that are "epithelioid"). On this basis most melanomas have epithelioid qualities. Melanomas in which cells migrate into the dermis as individuals are dependent on cellular relationships other than those of direct cell to cell contact. From inception, the cells of such lesions are diffuse migrants; they are independent of a community of neoplastic cells. They induce mesenchymal responses that favor their migrations. In their nature and properties, the individual migrants are closely related to the infiltrating cells often seen at the advancing margin of a level IV, fully evolved, common melanoma. Level IV invasion in common melanomas is generally characterized by the ameboid spread of some of the tumor cells as individuals among collagen bundles. Ameboid spread of tumor cells among collagen bundles of the reticular dermis is a fairly regular feature of the melanocytomas (i.e., Spitz variant, pigmented spindle cell variant, and deep penetrating variant). In addition, the migrants in the Spitz and pigmented spindle cell variants may induce a desmoplastic response in the supporting connective tissue.


 In the concept of MDM, the distinguishing feature of all variants of MDM was a pattern of vertical growth; vertical growth is a primality. On the other hand, the proposition that there is "one melanoma, histologically and biologically" is an hindrance in the search for virtual images that are necessary for the proper assignment of problematic lesions. It promotes the notion that there is no basis for confusion in the evaluation of problematic lesions, such as making a distinction between Spitz melanocytoma and Spitz nevus-like melanoma. The promotion of the proposition, that Spitz "nevi" can be readily distinguished from some spindle cell melanomas, and that the distinctions between "nevi" and melanomas are clearly defined, has lead to arbitrary assignments in which the role of the pathologist is much like that of a child in the game, "pin the tail on the donkey." Blindly, many pathologists are stabbing at silhouettes of spindle cell melanocytic neoplasms (as a blindfolded child might do in his search for the silhouette of the rear side of a donkey), hoping to be the lucky one who, by chance, places the appendage (a designation) closest to the correct location (congruence between diagnosis and biologic behavior). In these imposed struggles with sets of virtual images, the Spitz nevus-like category has become a common default. On a small biopsy, the patterns may be those of migrant melanocytic cells (beyond any community of cells) and desmoplasia (modifications of the fibrous character of the reticular dermis); they may be mistaken for those of a Spitz "nevus," when actually they belong to a variant melanoma.

 Cell to cell contact (the epithelioid qualities of common and many uncommon melanomas):

 The spindle cells of "nevi," dysplasias, and melanomas are likely to cluster in fascicles rather than rounded nests. Fascicles of cells tend to enlarge by extending linearly in continuity; the fascicles, in dissecting along planes among the collagen bundles of the reticular dermis, acquire tortuosity; in the plane of any section, some of the fascicles of tumor cells will be cut parallel to the plane of the section and some will be cut in cross section. For all melanomas, in either nesting or fascicular patterns, clustering of neoplastic cells provides epithelioid qualities (here, in the characterization of the patterns as "epithelioid," attention is given to the close clustering of cells); these qualities are basic to the criteria required for the diagnosis of most melanomas. In all other cellular arrangements, the pathologist must summon special criteria before he is capable of evoking virtual images which are relevant to the diagnosis of unusual nevi or melanomas. Some of these special lesions with deviant morphologic features will be benign, some borderline, and some malignant (i.e., desmoplastic melanoma and myxoid melanoma).

 Individual migrants as opposed to the movement and enlargement of migrant communities (individual migrants and their confounding qualities):

 If the nesting patterns of a common melanoma or the fascicles of the spindle cell variants are not well represented, neoplastic spindle cells, loosely and individually infiltrating the reticular dermis, may be mistaken for neoplastic mesenchymal cells, such as smooth muscle, fibrohistiocytic, or even fibroblastic derivatives. Such patterns and evoked virtual images may greatly complicate the interpretation of some melanomas. They are not generally appreciated as an alternate primality to be evaluated in arriving at a diagnosis of melanoma; an unwary pathologist may not even consider the possibility of melanoma in his differential diagnosis.

 Neurocristic phenotypes (their expression in "melanocytic" neoplasia):

 Melanocytic cells, isolated individually among collagen bundles of the reticular dermis, are encountered in the dermal component of congenital nevi but these nevocytic cells are easily recognized as to type. In both Spitz "nevus" and pigmented spindle cell "nevus," individual tumor cells at the advancing margins may percolate among collagen bundles of the reticular dermis. Often in the latter two disorders, the percolating cells are rounded rather than spindle shaped. They may be associated with desmoplasia. They somewhat resemble common nevus cells of the dermis in congenital nevi.

The common cellular blue nevus classically has organoid qualities. It shows zonal patterns including a peripheral fibroblastic shell, an intermediate melanogenic component, and a relatively amelanotic, neuroid central component. In these zonal patterns, the phenotypic potentials of neurocristically derived dermal cells are manifested; the potentials include mesomorphogenesis (fibroplasia), melanogenesis, and neurosustentacular properties (i.e., schwannian differentiation). If we extend these speculations to melanocytic malignancies, the same neurocristic potentials (a reversion to one or more primitive phenotypes) are variably encountered as phenotypic options in the evolution of neoplasia. In some of these options, the melanocytic origin of tumor cells is masked.

The common melanomas are expressed in melanogenic patterns. The spindle cell and fascicular melanomas are expressions of neurosustentacular potentials. The desmoplastic melanomas are expressions of mesomorphogenesis (fibrosis or sclerosis). Desmoplastic melanoma (i.e., lesions whose vertical growth components are composed of spindle cells) is characterized by individual neoplastic spindle cells which infiltrate the reticular dermis among preexisting collagen bundles; the tumor cells induce hyperplasia of fibroblasts and sclerosis of the affected dermis. This peculiar change in phenotype is seen most often in association with lentiginous patterns in radial growth. It is most common in association with patterns of lentigo maligna but is also an option in the category of acral lentiginous and mucosal variants. It is least common in association with radial growth patterns of superficial spreading melanoma. It may be encountered in the absence of a demonstrable lentiginous and junctional component in the epidermis. In some examples, lacking the latter components, the cell of origin may be a derivative of the sheath cells of peripheral nerves. Alternatively, a lentiginous and junctional component may have been present at the inception of such a lesion but may have spontaneously regressed. Finally, there may be a native neurocristic cell that masquerades in the dermis as a dermal fibroblast and has the potential to undergo neoplastic progressions as a fibroblast, a melanocytic spindle cell, or a Schwann cell-like spindle cell. If the fibroblastic (mesomorphogenic) potentials of neurocristic derivatives are ignored, then desmoplastic melanoma is likely to be assigned to other categories, such as the fibromatoses, the reactive fibroplasias, fibrosarcoma, spindle cell carcinoma, or atypical fibroxanthoma.

The potentials for neoplastic progressions in a line of neurocristically derived cells are such that it may be inappropriate to assume that the phenotype expressed by neoplastic neurocristic cells is predictive of a specific cell of origin. That a neoplastic neurocristic cell has three phenotypic options in regard to patterns of differentiation does not implicate a nonneoplastic, normal cell of similar phenotype as the cell of origin. It might be more appropriate to characterize some neoplasms in this general category as malignant neurocristic neoplasm and then to simply identify the phenotypic expression as either fibroblastic (mesenchymal), melanocytic, or schwannian.

 Derivatives of the neural crest in masquerade:

 Desmoplastic melanomas often are associated with a pattern of a lentiginous premalignant dysplasia in the epidermis, at least focally. Presumably, the epidermal component provides a marker for the precursor from which the desmoplastic lesion had its origin. In biopsies of small, young desmoplastic melanomas, the most obvious atypia in the dermal component is likely to be found in a population of spindle cells, each appearing to be isolated individually in the papillary dermis. Such lesions are not uniformly pandermal in distribution. In such lesions, it is difficult to separate tumor cells in the reticular dermis from activated mesenchymal fibroblasts(figs A1-5). The degree of cytologic atypia in the sclerosing dermal component is variable and in some examples is deceptively bland; such lesions have a minimally deviant quality (figs A1-5). On small superficial biopsies, the limited and inadequate real images of the histologic section may compromise one’s ability to make a distinction between a benign and malignant process. Since many of these minimally deviant processes are also amelanotic, it becomes difficult following the examination of superficial biopsies to distinguish between a melanocytic process and a fibroblastic process. The minimally deviant qualities often are lost in recurrences; progressions are accompanied by greater degrees of cytologic atypia (figs B1-5 & C1).

 New virtual images for the interpretation of desmoplastic patterns:

 Many of the problems in the accurate assessment of desmoplastic melanoma relate to a reliance on sets of virtual images that are adequate for the interpretation of common melanomas but are inappropriate for the interpretation of deviant lesions.

In the category of melanomas, we are predisposed to work with images of neoplastic steps in which dysplasias at the dermal-epidermal interface give way to nidi in the dermis in which nests of cells cluster and often induce their own stroma. In the desmoplastic category, epidermal components are not required for the acceptance of the lesion as having a primary configuration. In young desmoplastic lesions, atypia may be so minimal that distinctions between tumor cells and activated fibroblasts may be inapparent. In fully evolved, pandermal desmoplastic melanomas, spotty lymphoid infiltrates may be a flag but the most significant infiltrates tend to form in the deeper portion of the dermis and may not be a feature on a superficial biopsy. In the face of these complexities, it would be incautious to simply characterize such a lesion, on the basis of the findings from the examination of a superficial biopsy, as a "nevus" or to characterize the fibrosis in the dermis as a peculiar reaction.

An immunoreaction for S-100 protein even in early stages (i.e., young lesions) demonstrates delicate fascicles or files of neoplastic cells in the reticular dermis. In neoplastic progressions, the fascicular qualities become preponderant and the patterns take on a neuroid quality. The fascicles of old lesions are multilayered and tortuous. They often are most prominent in the deeper portion of the lesion. They come to resemble the fascicles of fasciculated epithelioid schwannoma. Neurotropism often becomes a feature of lesions late in their progressions. On this basis, it might be more appropriate to speak of desmoplastic neuroid melanoma rather than a specific and primary desmoplastic melanoma. For lesions in which desmoplasia and neurotropism are combined, the designation, neurotropic melanoma, becomes appropriate only if the neurotropism extends along nerves beyond the domain of the desmoplastic component. If a desmoplastic melanoma in its progressions becomes pleomorphic, such a lesion is likely be misdiagnosed as atypical or malignant fibrous xanthoma or histiocytoma.

 The relief that accompanies progression in desmoplastic melanoma (microstaging and its influence on the ease of histologic diagnosis):

 Desmoplastic melanomas are usually pan-dermal at the time of definitive diagnosis. On the other hand, such lesions may have had multiple earlier biopsies that were not histologically diagnostic (figs A1-5). Often the non-diagnostic specimens will prove to have been superficial biopsies or even superficial lesions that were not pan-dermal in distribution. The non-diagnostic biopsy specimens may have been interpreted as a "nevus," a Spitz "nevus," a cellular scar, a nerve sheath tumor, a fibromatosis, or even a fibrohistiocytic neoplasm (figs B1-5). On review, once a generous portion has been made available for examination, the patterns can then be accommodated in the category of desmoplastic melanoma (figs C1-4). In addition, the problems with the earlier, non-diagnostic specimens will be directly related to the size and depth of the biopsy specimens (figs C1-4). The other obstacles facing the pathologist will be his unfamiliarity with the process and his inability, in attempting to structure a differential diagnosis, to bring the possibility of desmoplastic melanoma to mind; more common lesions tend to overshadow desmoplastic melanomas. If melanoma is not among the virtual images evoked during the examination of a problematic lesion, the pathologist is unlikely to request an immunoreaction for S-100 protein. The possibility of melanoma must be in mind before confirmatory tests can be ordered.

 Differential diagnosis:

 One true example of a nevus which is to be considered in the differential diagnosis of desmoplastic melanoma is cellular blue nevus. Cellular blue nevus has not only overall symmetry from side to side and top to bottom, it usually displays zonal variation in patterns. It is bottom-heavy and tends to be sclerotic, especially in the upper 1/3 of the lesion. Inflammation is generally not a feature of uncomplicated cellular blue nevus. Cellular blue nevus is often pigmented but some examples are relatively amelanotic. Desmoplastic melanoma is more disorganized than cellular blue nevus. Elongated, thin, neuroid fascicles, composed of one or two cell layers and extending among altered collagen bundles of the reticular dermis, are more characteristic of desmoplastic melanoma. Mitoses are uncommon in cellular blue nevus but may be a feature of atypical cellular blue nevus. Malignant cellular blue nevus tends to become more uniformly cellular and less sclerotic. Atypical lentiginous and junctional components are more likely to be encountered in desmoplastic melanoma than in cellular blue nevus.

 Medico-legal pit-falls:

 In medico-legal disputes, if the nature of a particular melanocytic lesion is a problem, the vertical growth component of the problematic lesion often is composed of spindle cells. In many of these disputes, the diagnosis of the defendant commonly has been "nevus" with no additional qualifications. This approach says little for the pathologist’s willingness to commit to a specific diagnosis. It says something about a philosophy in which there is a feeling of security in a benign, noncommittal diagnosis. In the opposite approach, when confronted with uncertain histologic patterns, one’s philosophy might dictate that there is security in assigning such a problematic lesion to a malignant category. The security of the latter approach has been greatly compromised by the possibility that a diagnosis of malignancy may lead to aggressive forms of treatment such as chemotherapy. Although a mistaken diagnosis does not equate with proof of malpractice in all states, causality does become an issue.


 The caveats to place on a poster next to a microscope include:

 1. Spindle cell lesions in "sarcomatous patterns," particularly in actinically damaged skin, should be carefully searched for lentiginous and junctional components.

2. In the differential diagnosis of sarcomatoid spindle cell lesions of the skin, regardless of site and degree of actinic exposure, the possibility of desmoplastic melanoma should always be included.

3. Desmoplastic melanomas are often bottom heavy and show a greater degree of atypia in the deep dermal and subcutaneous components. A superficial shave biopsy may not include any portion of a lentiginous and junctional component; the latter often is spotty and limited in distribution and may not even be represented. The atypia near the dermal-epidermal interface may be insufficient in degree or representation to alert the pathologist to the possibility of desmoplastic melanoma. Fascicles of atypical spindle cells, if identified in single files among both the sclerotic collagen bundles, and hyperplastic "fibroblasts," should sound the alarm.

4. The clinical information supplied by the dermatologist or surgeon often does not include a melanocytic process in the differential diagnosis. Scar or hypertrophic scar is a common clinical impression. Many desmoplastic melanoma are not significantly pigmented.

5. Desmoplastic melanoma may masquerade in the patterns of atypical fibrous xanthoma (AFX). If only H&E sections are available for examination, the diagnosis of AFX should be provisional and the related category should be characterized as generic. Until more definitive studies become available, a diagnosis of AFX includes mesenchymal variants, melanocytic variants, and sarcomatoid spindle cell carcinomas. AFX only acquires specificity in the face of an immunohistochemical profile as follows: vimentin (+), S100 protein (-), and cytokeratin (-). If a lesion, provisionally classified as AFX has the profile: vimentin (+), S100 protein (+), and cytokeratin (-), it should be reevaluated as a possible example of desmoplastic melanoma.

6. Like lentigo maligna melanomas, most desmoplastic melanomas occur in adults past the age of 40. Most desmoplastic melanomas are variants of lentigo maligna melanoma. Age, however, is not a definitive criterion. Examples of desmoplastic melanoma may be encountered in younger age groups and in sites relatively protected from actinic exposure.

7. A Spitz nevus-like variant might be considered in the differential diagnosis on small, superficial shave biopsy specimens; the sclerosis, and atypical spindle cells in a pattern of both thin fascicles and files of a desmoplastic melanoma might be interpreted as the pattern of a dermal, sclerosing Spitz "nevus." Asymmetry of small, thin fascicles favors desmoplastic melanoma.

8. The pathologist, if confronted with atypia and sclerosis in a spindle cell neoplasm, should both alert the clinician to the possibility of desmoplastic melanoma and should also request a more generous biopsy specimen. The clinician must broaden his diagnostic spectrum to include melanocytic neoplasms which are amelanotic, tumoral, and scar-like. Often by the time a definitive diagnosis has been made, a desmoplastic melanoma has acquired significant bulk; it then may be too late for a complete local excision to be curative

9. Some desmoplastic melanomas, especially primary rather than recurrent lesions, are remarkably bland cytologically at all levels. They qualify as minimal deviation desmoplastic examples. In recurrences, cellular atypia is nearly always more advanced. A pathologist who has access to diagnostic patterns on a recurrent lesion should be considerate of the problems faced by any pathologist who had access to less specific material on earlier biopsy material.

10. Neurotropic growth is common in desmoplastic lesions. It may be simplistic to merely dismiss the association as nothing more than an expression of aggressive growth; the affinity of tumor cells for peripheral nerves may be an expression of an alternate phenotype. If desmoplastic melanoma is associated with a neurotropic component there is the potential for spread along peripheral nerves. If scalp or face is the primary site, the lesion may extend along nerves into the cranial cavity. Neuroid patterns are common in desmoplastic melanomas late in their evolution (figs C2-3).

11. Invasion of muscular vessels in the deep portion of the dermis or the subcutis is a common feature of desmoplastic melanomas.

12. In the late stages of desmoplastic melanoma, the lesion may become more cellular and less sclerotic. Fascicular patterns and neurotropism tend to be dominant features. Desmoplastic melanomas may metastasize to regional nodes and by way of the blood stream. Fascicular patterns are often a prominent feature of metastases; the background patterns tend to be overshadowed.

13. Claims of malpractice often ensue from the misdiagnosis of desmoplastic melanoma. Often on review, the original material, at best, will prove to be questionably diagnostic. Plaintiffs and their lawyers are not likely to be sympathetic or understanding. A more general dissemination of the clinical and histologic features as well as the life-history of this peculiar lesion is required. The pathologist and clinician should be constantly aware of the deceptive nature of this aggressive lesion. Recurrences are common. A recurrent "scar" should be viewed with suspicion. A spontaneous "keloid" appearing for the first time in a mature adult should be a source of concern. "Fasciitis" that recurs and grows slowly should be a cause for alarm. Lentiginous and junctional components in actinically damaged skin, in which the component cells have small, somewhat irregular, dark nuclei and may appear as naked nuclei, should alert the pathologist to a variant of lentigo maligna with a potential for desmoplastic transformation in vertical growth.

14. For a patient, there is great risk in the misdiagnosis of desmoplastic melanoma. For the pathologist, there is frustration and injured pride. For the lawyer, there is great financial incentive and a certain amount of vindictiveness. Even in this period of universal distrust, the medical profession has managed to preserve a greater degree of public respect than the legal profession or the politicians.

15. Immunoperoxidase reactions are variable in the category of desmoplastic melanoma. Generally, a reaction for HMB-45 will prove to be negative. S-100 protein positivity is generally a feature of the tumor cells of desmoplastic melanoma. S-100 protein may have utility in evaluating the adequacy of margins of excision, but immunoreactive cells at the margin may be few in number and may be difficult to distinguish from S-100 + dendritic histiocytes which are often increased in number in the dermis adjacent to the tumor.


1. Reed, RJ: Hum Pathol 1999;30:521-524.

Acknowledgment: Bill Weems adapted the original manuscript and photomicrographs for Internet presentation.

Dr. Reed would appreciate comments directed

 Next Page                                                     

 Back a Page                                                    

Down a Page                                                 

Up a Page