A Borderland of Neoplasia (Melanocytic System)
The clinical presentation of lesions in the melanocytic system is varied; it ranges from thin, flat lesions to large, polypoid and ulcerated lesions. Clinicopathologic correlations have established that certain histologic patterns are predictive of, and other histologic patterns are unlikely to be associated with, progressive disease. These varied presentations and patterns are manipulatable. Herein, segments of neoplastic continua, comprised of both precursors and fully evolved melanomas, are defined in virtual images.
In the conceptualization of common melanocytic neoplasia, it is necessary to segment neoplastic continua and to provide labels for the segments. The manner in which a neoplastic continuum is segmented impacts upon the structuring of virtual images and, in turn, upon the significance of the histologic diagnosis of melanoma. Herein, two major segments include the dysplasias, and the melanomas. In toto, the spectum to be segmented extends from mild dysplasias to high grade melanomas (including lesions with significant bulk). The major segments are also divisible, but the definitions of the smaller segments are by no means uniform. Having once segmented the continua and labeled the segments, the manipulations of the patterns are subject to individual biases, and to a quality that lends arbitrariness to the selected segments, namely the fluidity of neoplastic processes.
It is the role of the pathologist to define and label segments. In turn, he and others must manipulate the segments and their patterns to provide correlations with clinical observations. Finally, these real and virtual images, in combination with clinical data, are to be manipulated to provide therapeutic guidelines. In these perambulations, the manner in which the continua has been segmented is influential.
Deeply imbued virtual images of the common melanomas have become second nature for all but a few of us. The few actually are compelled by nature to have, and use, them, even if they do not make an acknowledgment. The manipulations, by rote, which are involved in the diagnosis and prognostications of a common melanoma, such as superficial spreading melanoma, have been hardly questioned.
In an examination of the full expression of a spectrum of changes in the common melanocytic neoplastic continua, including those of a temporal nature, there are portions of the neoplastic continua, corresponding to particular periods in the evolution and life history of melanomas, in which the histologic patterns fail to sharply discriminate between precursor and melanoma. For most examples, the troublesome and arbitrary portions (defined segments) of these neoplastic continua are representative of examples of melanocytic neoplasia which are thin in vertical dimensions, with most examples measuring less than 1 mm in height (thickness) .
A store of relevant virtual images is a requisite for the interpretation of a borderland of melanocytic neoplasia in which high grade dysplasias and incipient (thin) melanomas are included. Degrees of cytologic atypia, host immune response, nodular or plaque- like patterns, levels of invasion (modified Clark's criteria), and depth of invasion must be given an accounting in the catergorization of lesions of any defined segments in this borderland of melanocytic neoplasia.
Sets of criteria have been designed and strongly promoted to aid in the partitioning of melanocytic neoplastic continua, but, for each offering (set) in the category of thin melanocytic neoplasia, different morphologic features have been emphasized. The differences are a reflection of the individuality of the virtual images of each of us. They also are a reflection of the imprecision of virtual images in the segmentation of neoplastic continua, of the limitations of language for the characterization of neoplastic phenomena as related to the segmentations, and of the susceptibility of practicing pathologists to public babble and bombast. Offerings of defined segments have been defended by their proponents as if the differences also impact on a patient's well-being. In defining differences between dysplasia and melanoma, the borderland is particularly troublesome. Similar problems with borderline neoplasia have been encountered in other organ systems.
The borderland (a different perspective): In most categories of epithelial neoplasia in a variety of organ systems, the borderland of neoplasia is defined simply on the basis of size and the limited distribution of nests of neoplastic cells in the stroma. The definition depends on the relationship between nests of cells and their stroma, and the manner in which the relationships impact on prognosis.
With squamous cell neoplasia of the uterine cervix as an example, the designations, keratinocytic dysplasias and carcinoma in situ have been largely supplanted by a graded system of "intraepithelial neoplasia." The segment, carcinoma in situ, if selected, is likely to be defended as if it is the sole precursor of invasive carcinoma (the only optional segment). In fact, invasive patterns are common in the immediate vicinity of precursors showing lesser degrees of atypia and architectural derangements. The next obvious step (i.e., qualifying microinvasive carcinoma as keratinocytic neoplasia of indeterminate malignant potential) has not been made.
For some examples in some organ systems, even large neoplasms may qualify as borderline. A borderland of neoplasia has been defined for large ovarian neoplasms. For the borderland of ovarian epithelial neoplasia, the lesions may be considerable in size, even to the point of compromising attempts to adequately sample a given example. Borderline lesions share some of the histologic characteristics of fully evolved malignant ovarian neoplasms but are basically distinguished by an absence of foci in which there is recognizable stromal invasion: large bulky lesions with significant aytpia are qualified as borderline if foci of stromal invasion are not identified on multiple sections. These lesions with characteristics of "atypical cystadenomas,”even though large, are characterized as borderline ovarian neoplasia of low or uncertain malignant potential. In fact, they are borderline ovarian epithelial neoplasias of indeterminate malignant potential. Recently, relationships between isolated nests and stroma have been assigned great significance, but the identification of even a few scattered foci of "invasion" may not greatly alter the factual "borderline" nature of these lesions.
Three additional areas badly in need of a definition of a borderland of neoplasia are represented by encapsulated neoplasms of the thyroid, well differentiated cartilaginous neoplasms of bone, and prostatic neoplasia as encountered on needle biopsy specimens.
In the segment, "melanoma" in situ, neoplastic cells are confined to their site of origin (epithelium: pattern I) and, as their most distinctive characteristic, migrate individually among keratinocytes. "Melanoma" in situ essentially is manifested in single cell patterns; clearly, cellular phenomena are basic to its definition. Paradoxically, proponents of the utility of this segment have offered the corollary that there are "no melanoma cells, only melanomas."
In other schemes, the designation, melanoma, has application only for those segments in which neoplastic cells have migrated out of their native domain, and have joined other similar cells to share in the economy and utility of a community (vertical growth at level III) or have become active migrants in the reticular dermis (diffuse vertical growth at level IV).
Vertical growth at the borderline: The borderland of common melanocytic neoplasia reaches beyond the limits of the dysplasia patterns and into the domain of patterns that identify "melanomas" with a potential for metastasis. The major discriminator for the recognition of the "melanoma" pattern, as distinguished from the "dysplasia" pattern, is a vertical growth component (3,6,7). In vertical growth patterns, the pertinent virtual images depict the emergence of a clone of neoplastic melanocytes with the capacity to survive and multiply in the immunologically hostile environment of the papillary dermis. Such a clone would either be dermal in location and independent of the phenomena at the dermal-epidermal interface or would deliver nests of neoplastic cells from the dermal-epidermal interface into the dermis at a rate exceeding the capabilities of the host immune response. In addition, the capacity to induce a stroma and a blood supply for the neoplastic cells would be embodied in the images.
The segment of the neoplastic continua encompassing thin melanocytic neoplasms (i.e., less than 1 mm in vertical dimensions) in the common premalignant dysplasia- melanorna sequences (a specification embracing common variants, actinic variants, or acral variants) is a borderland; the troublesome patterns are found at the transition from level II to Level III. Included are high grade "dysplasias" and "thin melanomas" measuring less than 1 mm in vertical dimensions. All such lesions are borderline melanocytic neoplasias, a category that is divisible into two dimensional lesions (dysplasia type) and three dimensional lesions (melanoma type). In the final analysis, all the options in regard to the segmentation of melanocytic neoplastic continua have utility. In part the final decision, as to the appropriate segmentation of the neoplastic continua, may have relevancy to the manner in which each segment lends itself to distortions by the legal profession. On the other hand, any scheme, that becomes more or less universal in acceptance, places an onus on practicing dermatopathologists to conform in a general manner.
Borderline dysplasia: Currently, virtual images, in a borderland of melanocytic neoplastic phenomena, must be molded to accommodate a variety of designations. Without regard for cytologic atypia, real images in this borderland provide no clear distinctions between some grades of dysplasia as promoted in some schemes, and lesions that in other schemes have been characterized as "melanoma" at levels I & II. The designation, "melanoma," if selected for the characterization of patterns in this limited portion of a neoplastic spectrum (the borderland of the dysplasias: patterns I & II), easily influences the character of respective virtual images. The relevant virtual images evoked by the examination of lesions in this limited portion would then imbue the aggregate patterns - of what in some schemes is a dysplasia - with the biologic potential of a "melanoma." Practically, prognostications, in this borderland of neoplasia, are meaningless: metastases, if they occur, are too infrequent to lead credence to our common prognostic parameters. A metastasis from a "level I or II dysplasia" or a "level I or II melanoma" would suggest that a vertical growth component has spontaneously regressed, or that it remains in a paraffin block or a formalin bottle.
Borderline melanoma: Herein, in the dysplasia-melanoma sequence, thin melanomas (i.e., at least pattern III) measuring less than 1 mm in vertical dimensions are representatives of a borderland of melanocytic neoplasia (they are prototypic of minimal deviation [incipient] melanomas in this neoplastic system). For them, the usefulness of prognostic parameters is only slightly improved over that for lesions in the borderline dysplasia category (i.e., pattern II). In the diagnosis of thin borderline melanomas (lesions with an indeterminate capacity for metastasis), only in retrospect and in rare instances, will the distinctions between lesions showing patterns I or II, and thin lesions showing patterns III or IV be validated as a distinction between melanoma precursors and melanomas. If a lesion in the borderland of neoplasia (i.e., thin, typical melanoma: a pattern III or IV lesion) has demonstrated a capacity for metastasis, the phenomena might, in retrospect, be characterized as a validation of the concept of minimal deviation melanoma as related to the segment: Thin melanoma, of the common neoplastic continua. The particular segment - thin (i.e., less than 1 mm) melanoma - might be better characterized as borderline neoplasia (melanoma variant) rather than common, thin melanoma. This accommodation finds justification in the observation that all segments of the common melanocytic neoplastic continua that are thin and either pattern II or III are prognostically unpredictable, if the object of the prognostications is to identify specific lesions that have metastasized at the time of initial histologic diagnosis of melanoma.
Thin Melanomas and Homologies:
At extremes of the morphologic spectra of typical melanomas, and their specific precursors, patterns are relatively unambiguous (e.g., the patterns of low grade dysplasias and of large, bulky melanomas), but a limited portion of the spectra, which includes the transition from precursor to melanoma, is morphologically and bioloqically ambiguous: it is difficult to sharply define the exact histologic pattern that discriminates one from the other. Having made a selection from the optional patterns, the "melanomas" of this limited segment provide little threat for the respective patients. For this limited span, terminology must be molded to accommodate either the implications of a precursor stage, or the implications of a fully evolved melanoma, without real assurances that the accommodations are biologically accurate or even significant. Clearly, the ambiguities are such that the contested patterns might be best assigned to a category of borderline melanocytic neoplasia of indeterminate biologic potential.
In the concept of minimal deviation melanoma, there are as many variants (i.e., lesions not of the type) of melanomas as there are "variants” of nevi. Also, there are no melanomas showing patterns I or II: a vertical growth component (i.e., patterns III or IV) is a requisite for the diagnosis of melanoma. In the concept of MDM, the need for special criteria in developing a nosology related to variant patterns, and for defining prognostic implications relative to the variant patterns is acknowledged.
With criteria embodied in the concept of MDM, borderline (thin) melanomas, in the settings of both the common dysplasiamelanoma sequence and MDM of halo nevus type, measure 1 mm or less in vertical dimensions. These thin melanomas in both categories share many qualities. ...........
(Continued NEXT PAGE).