In both, variable degrees of atypia and markers for a host immune response are features. The lymphoid response in the halo nevus variant has band-like qualities, with regular spacing of nests of melanocytic cells in the dermis in close association with the lymphoid infiltrates. Both neoplasms often show areas of regression. In either category, prognostic evaluations of thin lesions - lesions measuring less than 1 mm in vertical dimensions - are unlikely to identify representatives that have a significant potential for metastases. For thin melanomas in the categories of minimal deviation melanomas arising in melanocytic dysplasias, minimal deviation melanomas of the halo nevus type, thin lentigo maligna melanoma, thin acral lentiginous melanoma, and even thin superficial spreading melanoma at level III (all measuring less than 1 mm in height), it would be more accurate to assign these lesions to a category of borderline melanocytic neoplasia of indeterminate malignant potential, and to then qualify each as a neoplastic variant on the basis of any other distinguishing characteristics.
From a different perspective, patterns (i.e., patterns I & II), other than those of vertical growth (i.e., patterns III & IV), have been accepted as delimiters of "melanoma." "Melanoma" in situ, and thin "melanomas" at level II (i.e., measuring less than 1 mm in height and including examples of "superficial spreading melanoma at thin level II:" microinvasive melanomas) share histologic and cytologic features with melanocytic dysplasias (i.e., patterns I & II). They are distinguished merely by the willingness of an observer to arbitrarily assign a problem example to one or another category. From the perspective that only lesions in vertical growth are melanomas (i.e., concept of MDM), all dysplasias in the dysplasia-melanoma category (including lesions that otherwise would be included in the above three categories) showing moderately severe to severe atypia (i.e., high grade dysplasias) would also qualify as melanocytic neoplasias of indeterrninate malignant potential (dysplasia variants I & II: Two dimensional variants, patterns I & Il) .
Histologic diagnosis as a presumption of malignancy: In the conduct of general surgical pathology, the assignment of a neoplasm to a malignant category on the basis of its histologic patterns is simply an accommodation, structured on the supposition that the lesion has already metastasized or, if left undisturbed, will eventually produce progressive disease with metastases andlor death.
Many histologic attributes have been emphasized in the definition of cutaneous melanoma. The passion with which some observers defend a set of selected attributes - those which they have chosen to then be used in formulating diagnoses of melanoma - has politicized all the relevant proposed collections of attributes. A muddle, in which a choice of criteria for the segmentation of melanocytic neoplastic continua rests more on the credentials of the chief proponents than on the relevancy of strict cytologic and histologic, criteria to biologic potentials, is the specter facing practicing pathologists.
In all honest presumptions of malignancy, if based solely on the interpretation of the histologic patterns of a primary tumor in its primary site, absoluteness, as the attribute defining the appropriateness of all prospective linguistic assignments (i.e., assignments that could anticipate progressive disease, metastases and eventual death), is belied by inherent uncertainties, not the least of which is the capriciousness of neoplastic phenomena.
In cutaneous malignant melanoma (as in most malignancies), the appropriateness of the designation, melanoma (i.e., malignancy), depends upon many variables, the most predictable of which are the presence, and the vertical dimensions, of a vertical growth component (with the stipulation that the responsible observer from an examination of the limited two dimensional real images of a histologic section has the wherewithal to translate real images into virtual images, and to then structure a three dimensionality.
Note: some proponents of the concept of "one melanoma" readily admit their inability to manipulate virtual images.
For melanomas measuring greater than 3.5 mm in vertical dimensions ("thickness"), the validity of attributes and relationships - having been imposed by the manipulation of real and virtual images, and then amalgamated as dimensionalities- are easily confirmed. In lesions with such impressive physical attributes, the almost invariable associations of the amalgams with progressive disease lends credence to the embodied darkness of the designation, "melanoma." At the opposite end of the physical spectrum - the end corresponding to the first emergence of vertical growth as an expression of neoplastic progressions (i.e., thin pattern III or IV), the criteria for the recognition of a vertical growth component become arbitrary, particularly for lesions showing variant vertical growth, and even more so for lesions in which thin vertical growth components are formed of cells with rather bland, uniform cytologic features. The likelihood of progressive disease after the removal of a thin (less than 1 mm in height), cytologically bland "melanoma" (minimal deviation melanoma of the most common type: a thin lesion in vertical growth in the setting of a dysplastic nevus) is remote. If the definition of melanoma is extended to include lesions with no vertical growth components (lesions at levels I & II), predictions of the likelihood for metastasis will have been eliminated as a significantly valid correlate. On the other hand, pathologists, who are often engaged by lawyers as experts, will, in the examination of material pertaining to claims of malpractice, commonly encounter such lesions. To designate thin lesions in vertical growth, as "melanoma" is at best a compromise, if the desideratum is to identify lesions which have a significant potential for metastasis.
Efforts continue to be wasted in the search for the earliest stage of true melanoma in the sequence leading from the common premalignant melanocytic dysplasias to the common melanomas. In the concept of minimal deviation melanoma, lesions in this segment of the neoplastic sequence were afforded special significance. Vertical growth, as an embodiment of a set of virtual images defining three dimensionality, was basic to the definition of the common variant of minimal deviation melanoma. In essence in defining this limited segment, an attempt was made to extrapolate from patterns of vertical growth in "real" (i.e., thick) melanomas to the minimal requisites for characterizing a cluster of nests, or fascicles, of neoplastic cells as vertical growth. In the concept of minimal deviation melanoma as related to the dysplastic nevusmelanoma sequence, two strata each containing 2-3 closely spaced nests, or fascicles, of neoplastic cells (5 or 6 nests in two strata) were deemed sufficient. As a corollary, it was stipulated that both atypia of at least moderate degree, and markers for host immune response were required. In the presence of severe atypia, all examples of thin melanoma in this category were, by default, assigned to the category of superficial spreading melanoma (i.e., a common final pathway). In this approach, the requisites were:
1) size (i.e., <1 mm),
2) atypia (features less than those of the common final pathway), and
3) biologic potential (e.g., metastases uncommon and unpredictable). Herein, it is proposed that all thin typical melanomas might be best assigned to a borderline category.
Melanomas commonly are pleomorphic in the region of the vertical growth component. In part, the different clones (as manifested in cytologic variations) are markers for the genetic instability of melanocytic cells in the vicinity of a vertical growth component. With rare exceptions, the cells at the dermalepidermal interface usually are the most atypical and, by inference, the most advanced neoplastically (this inference is based on the assumption that by nature a malignant cell is predisposed to progressive deterioration of genetic controls, and that the most advanced stages of neoplasia are correlated with the most bizarre cytologic aberrations). In this approach, the most atypical cells at the dermal-epidermal interface are the youngest, and the generations that are most advanced neoplastically; the least atypical of the neoplastic cells (generally to be found at the deep margin of the vertical growth component) are the oldest and least deviant generations (minimally deviant clones). Nests, situated between the two extremes, may also show cytologic variations with the implication that clusters of cytologically similar cells are genetically similar and, if locally clustered to form a vertical growth component or a portion of such a component, provide a locus that is genetically homogeneous. Polymorphic lesions are commonly stratified in regard to degree of cytologic atypia and, by inference, in regard to genetic heterogeneity. In thick, typical vertical growth components, pleomorphism may not be confined to the dermal-epidermal interface.
Prognostications: Parcelling of collections of real and virtual images in the borderland of melanocytic neoplasia is basic to current efforts at prognostication. The patterns of nevi, premalignant dysplasias, and thin or incipient (minimal deviation) melanomas, all in a borderland of confusing virtual images, are parcelled by pathologists as if, in the performance, the assigned designations come to define the biologic nature of stages in a neoplastic system. The images of "melanoma" in situ (4), of thin minimal deviation melanoma as manifested in the setting of the dysplastic nevus-melanoma syndrome (3,6,7), of severe melanocytic dysplasia (a "common final pathway") (3), of "borderline" dysplasias and melanomas (7), and of thin superficial spreading melanoma, as well as thin acral lentiginous melanoma and lentigo maligna rnelanorna (i.e., pattern II lesions ) do not exist in nature as distinct entities: nature has no need for them. They are merely figments embodied in our amalgams of real and virtual images, and incorporated in our language as linguistic labels. In the borderland of melanocytic neoplasia, observers, in citing a preference for one or another of these conceptualized parcels of borderline neoplastic phenomena, must also admit their inability to reliably predict the malignant potential of a large group of lesions, all of which are representative of an early and common neoplastic pathway. Practically, the importance of the selection of a suitable parcel, and the assignment of an appropriate designation loses significance in the realization that all these problematic lesions are amenable to a common form of treatrnent (therapeutic commonness). For either a thin (i.e., less than 1 mm in height), problematic lesion in the premalignant melanocytic dysplasiamelanoma category, a melanoma in situ, or a superficial spreading melanoma at level II and measuring less than 1 mm height (Breslow's criteria) (8) - if the adequacy of the original biopsy procedure is problematical - the question of which borderline category is most appropriate for a final assignment is of little consequence, if the assignment carries with it recommendations for both a conservative re-excision of the area, and follow-up. In fact, all these designations are encompassed in the proposed category of borderline melanocytic neoplasia of indeterminate malignant potential.
Prognostic categories: The prime prognostic categories have relevance to qualities which are discernible in evaluations of vertical growth components. They are based primarily on a measurement of the distance the deepest nest of atypical melanocytes is from the granular level of the epidermis ( a measurement that generally is perpendicular to the surface of the lesion). The commonly emphasized categories include:
1) lesions up to 0.75 mm vertical dimensions;
2) lesions in the range of 0.76 to 1.50 mm in vertical dimensions; and
3) lesions greater than 1.50 mm in vertical dimensions.
In giving credence to the value of these parameters, an observer, with repetition of the technique, will likely come to presume that his measurements and prognostications have relevance in regard to the biologic potential of any selected individual lesion, a presumption easily played upon by plaintiff's attorneys, and easily discredited by careful observations of the natural history of melanomas, especially thin melanomas.
In defining a borderland of melanocytic neoplasia, the dimensions of the commonly emphasized categories have been modified:
Category I lesions are representatives of a borderland of thin lesions measuring less than 1 mm in vertical dimensions. Included in this category are high grade dysplasias (i.e., lesions showing patterns I or II), and thin melanomas (i.e., lesions showing patterns III or IV). Lesions in category I are unlikely to metastasize. With rare exceptions, such lesions are biologically innocent. To characterize all pattern I, category I lesions as unqualified melanomas, irrespective of the presence or absence of a vertical growth component, unfairly stigmatizes a large group of patients.
In practice, thin lesions in the categories of high grade dysplasias, melanomas in situ, superficial spreading melanomas, lentigo maligna melanomas, acral lentiginous melanomas, and melanomas arising in atypical nevi (exceptional lesions having evolved from common premalignant melanocytic dysplasias but not manifesting the features of the common final pathway) rarely metastasize even if the limits for the definition of category I are extended to 1 mm. Such thin lesions are best characterized as borderline melanocytic neoplasias of indeterminate malignant potential and qualified as either dysplasia (patterns I & II), or melanoma (patterns III & IV) variants.
In the modified range of 1 mm to 1.5 mm, only lesions in vertical growth (i.e., patterns III & IV) have relevance to the diagnosis of melanoma. Lesions in this size range comprise category II. Melanomas in category II are only slightly more predictable; the likelihood of predicting metastases from any individual lesion in this category is sufficiently remote to justify the characterization of category II lesions as unpredictable, thin melanomas.
Lesions in category III (beyond 1.5 mm in vertical dimensions) qualify as real melanomas.
In the above scheme in which a neoplastic continuum is segmented on the basis of patterns in the dermis, and the vertical dimensions of any vertical growth components, there is a borderline category and two categories of malignant melanoma. The two categories of melanoma are defined both on the basis of vertical dimensions, and on the relationships between dimensions and the likelihood of predicting metastasis (Table I).
The three categories as developed in the preceding section find utility in the relationships between physical attributes and prognosis. Arbitrariness is inherent in the designation, melanoma (particularly for category I lesions); an inability to identify those rare members of the category I lesions with a capacity for metastasis introduces relativity into the histologic diagnosis of melanoma.
Cytologic disparities in epidermal and dermal components of melanomas as they impact on taxonomy and prognostications: The character and pleomorphism of melanoma in vertical growth components generally has little impact on the selection of a specific designation for the assignment of respective lesions. For example, a vertical growth component may be composed of small, rather bland amelanotic cells in closely spaced nests of relatively uniform size. These cells may be difficult to distinguish from common nevus cells, if a remnant of a nevus is available for comparison. If, in such a lesion, the patterns at the dermal- epidermal interface are those of the common final pathway (as manifested in the radial growth component of a fully evolved superficial spreading melanoma) then the minimal deviation character of the vertical growth component will likely be ignored and, in turn, the lesion will likely be assigned to the category of superficial spreading melanoma (SSM). On the other hand, in providing prognostications for such a lesion, the vertical growth component, with no attention to the character of its small, bland cells, will be the prime determinant. To place emphasis on a population of cells (i.e., the radial growth component) which is mostly a new growth of limited expanse at the dermal-epidermal interface, and to ignore the character of the vertical growth component in characterizing most melanomas is an anachronism.
Estimates of nuclear (tumor) grade in the vertical growth component may have utility in the classification of melanoma. In this approach, most lentigo maligna melanomas would be low or intermediate grade (minimal deviation), and all superficial spreading melanomas with a degree of atypia in the vertical growth component that is comparable to the degree of atypia in the radial component (pattern of common final pathway) would be high grade melanomas. A polymorphic lesion with the pattern of the common final pathway at the dermal-epidermal interface but with a preponderance of bland (minimal deviation) nuclear features in the vertical growth component would be evaluated by the nuclear grade of the latter component and be characterized as being of low or intermediate grade.
Subjectively, a certain elegance and utility, that is lacking in cold evaluations of nuclear grade, reside in our "archaic" designations. In practice, a combination of the two options would appear to be the best compromise.
Analysis of neoplasms (comparisons with cryptanalysis and transmission analysis): In the practice of histopathology, patterns are analyzed for the purpose of rendering a diagnosis, and to provide therapeutic guidelines. Real images, having been observed, must be compared with evoked virtual images (as they relate to a variety of processes), and then all the images must be amalgamated. In these endeavors, the intention is to find matching pairs among the real and virtual images and to structure from them a scenario.
In the analysis of a histologic section of a neoplasm, the biologic potential is anticipated in predictions rather than assurances. In an observer's interplay of images, virtual images once imposed on real images, may modify or alter the appreciation of real images in a detrimental manner. In such an event, a diagnosis (with its dependency on the relevancy of imposed virtual images) may prove to be inappropriate.
In one portion of the spectrum of melanocytic neoplasia, the nature of respective neoplasms cannot be sharply defined by manipulations of real and virtual images. Designations such as severe junctional or compound dysplasia, "melanoma" in situ”, and thin "melanoma" at level II are assigned relevance in the interpretation of certain stages of neoplasia, but are merely linguistic symbols that translate amalgams of real and virtual images into language; in the process, a neoplastic continuum is segmented . The manner of segmentation is variable among different observers. If the qualities of the variable segments are critically compared, there are likely to be more overlaps than distinctions. To defend one of these arbitrary segments to the exclusion of all the other options promotes controversies. In a limited portion of the neoplastic continuum (lesions less than 1 mm in vertical dimensions), the ensuing controversies cannot be resolved by manipulations of virtual or real images, or modified in degree of contentiousness by manipulations of the related linguistic symbols. Any arbitrary selection can only be defended on the basis of personal preferences.
The concept of borderline melanocytic neoplasia seems to be objectionable to some workers, particularly the molecular "pathologists." They would contend that there is a role for cell markers in an unraveling of the imprecision that is inherent in the concept of borderline neoplasia. They would have us believe that controversies in neoplasia can be resolved by appropriate marker studies. If cell markers, as demonstrated with labeled monoclonal antibodies, prove to be inadequate, some technologists, having failed with one technique, will then direct us to more sophisticated techniques, such as polymerase chain reactions, etc.
The problems in the molecular approach to the analysis of neoplasms are comparable to those of cryptographers in wartime. It is the role of cryptographers to anticipate the plans and intentions of the enemy. Their efforts may be of great assistance, if coded messages are intercepted and decoded (i.e., cryptanalysis). If successful in their endeavors, intentions and ambitions, having been exposed, can be thwarted by effective countermeasures.
If the code has not been, or cannot be, broken, or if broken but subsequently changed, then the mundane task of simply monitoring messages, even without an ability to decipher the actual meanings, and evaluating other known parameters, such as frequency of transmissions as they relate to assumed activities - such as increased activities of destroyers in anticipation of a battle, or the identification of the location of aircraft carriers which are under radio silence by the identification of cruisers which are not under radio silence but are known to be part of a fleet in common with the carriers - will often provide useful information. This technique is characterized as transmission analysis.
Molecular pathologists would have us believe, that by a modification of crypt-analysis, they have or will decode messages from borderline lesions. They then will be able to clearly resolve the conflict of whether a lesion is a precursor, or a fully evolved malignancy. Their success would then make an analysis of histologic patterns in the borderland a moot endeavor. In fact, most of what the molecular pathologists accomplish is transmission analysis. The borderline neoplasm is by nature ambiguous, and its genetic messages will prove to be equally ambiguous. A fully evolved malignancy will transmit messages that are decipherable but, in such examples, histologic patterns are equally informative.
Surgical pathologists in their attempts to grade neoplasms to provide prognostic and therapeutic guidelines must also deal primarily with transmission analysis. The associations between grades of neoplasia, mitotic rate, and degrees of differentiation on the one hand, and clinical behavior on the other are only hinted at in histologic evaluations. In the evaluation of the potential of many neoplasms, tentatively considered to be malignant, time alone will prove to be the final arbiter.
Historical perspective as it impacts on the interpretation of borderline neoplasia: The conceit accompanying historical perspective compromises any attempt to arrive at an unbiased evaluation of the performance of pathologists, who have earlier reviewed identical material, but did so without the luxury of historical perspective. The following scenario is an example of the evocation and manipulation of fanciful virtual images: A thin melanocytic lesion (i.e., borderline lesion) has been excised with histologically free margins. With subsequent follow-up, regional lymph node metastases, but no local recurrences, are discovered. In the evaluation of this sequence, virtual images of a lesion that had metastasized prior to the excision of the primary lesion most often are evoked. Histologically, the primary lesion may have been diagnosed originally as nevus, melanocytic dysplasia, or melanoma. The virtual images evoked by an evaluation of the now defined clinical sequence (i.e., a lesion that with time has demonstrated a capacity for metastasis), even if evoked without any regard for the histologic patterns, or the vertical dimensions of the primary lesion, would be the assumption that the primary lesion was a melanoma (a lesion in vertical growth). On review of the primary lesion, in amalgamating the real images of the "primary" lesion with the evoked virtual images - which have been conditioned by a knowledge of the existence of a subsequent metastasis - the real images are likely to be conditioned in a manner favoring a diagnosis of melanoma. In this sequence of impositions, with any markers for a dysplastic “nevus” in the "primary" site, it becomes tempting, in retrospect, to assign the "primary" lesion to the melanoma category. It may even be possible to list "features" (real images) of the "primary" lesion in confirmation of this diagnosis. A special liberty, available mainly to those who subscribe to the concept of "melanoma" in situ and superficial spreading melanoma at level II, blurs the distinctions between dysplasia and melanoma. Such subscribers, in the face of documented metastases, are especially prone to assign a problematic lesion to the category of "melanoma." The manner in which virtual images, that are related to historical perspective, may have modified the interpretation of a suspected "primary" lesion seldom will have been formally acknowledged: most often the respective observer may not appreciate that even his "real images," when subjected to the influences of historic perspective, are susceptible to transmogrifications. If the thin "primary" lesion had been originally interpreted as a benign process, but later, in the face of metastases, reviewed and interpreted as malignant by a different observer, the capabilities of the original pathologist, who may have struggled in the interpretation of the deceptive cytologic features of a thin problematic lesion, then become suspect. That the original pathologist, in his initial interpretation of the "primary" lesion, may have been betrayed by inadequacies in his store of virtual images will be a likely presumption. In the above scenario, but with the knowledge of subsequent clinical findings, and under the influence of the imposed virtual images, the real images of an excised, primary lesion will have been reappraised and then reinterpreted histologically as a melanoma. The reappraisal eventually may prove to have provided an inaccurate depiction of events. In fact, there may have been another melanoma that was the source of the metastasis, but had been removed (and may even have been misdiagnosed). As another alternative, the actual primary melanoma may have spontaneously regressed. The patient may have no recall of such events. In the approach to problematic cases, the fallibility of not only virtual images, but even real images, must be in the forefront of deliberations.
Some examples of superficial spreading melanoma at level II with purported metastases may be representative of the distortions produced by fanciful virtual images. It behooves the pathologist to investigate the clinical setting, and to inquire regarding other possible primary sites when faced with the challenge of accepting a metastasizing superficial spreading melanoma at level II (i.e., "metastasizing pattern II lesion") With few exceptions, a true metastasizing melanoma, initially interpreted as showing only level II invasion, will, on additional sections, show an area of acceptable vertical growth, or an area of regression.
Prognostications and Grade of Neoplasia: In general, prognostications are based on the identification of features that relate to grades of neoplasia. In common with the grading of most malignant neoplasms, features in melanomas, such as nuclear and nucleolar size, distribution and content of nuclear chromatin, nuclear-cytoplasmic ratio, cohesiveness of cells, and mitotic rate, are evaluated in arriving at a diagnosis, even if they are not given great significance in routine prognostications. In these evaluations of histologic features, learned responses might seem to afford an avenue leading to uniformity and agreement among observers. In practice, learned responses are accessible only as virtual images and, thereby, are fickle perceptions.
(continued on NEXT PAGE)