Pigmented Atypical Fibroxanthoma, a Dermatofibroma Variant?

To the Editor: We read with great interest the article by Diaz-Cascajo, Weyers, and Borghi entitled “Pimented Atypical Fibroxanthoma. A Tumor That May Be Easily Mistaken for Malignant Melanoma.”1 First of all we want to congratulate the authors and the editor for publishing such high-quality photographs, which give the critical reader an excellent impression of what the authors saw and meant. Moreover, the setting of their figures allows authenticity between low- and high-power photomicrographs, which is frequently not seen in publications. Yet, what the 4 photomicrographs of the above publication 1 show, is in our interpretation not a pigmented variant of atypical fibroxanthoma (AFX), but a pigmented variant of dermatofibroma, first described by Hurt and Santa Cruz as aneurysmal or angiomatoid fibrous histiocytoma. 2 Synonyms for this entity are hemosiderrhotic (fibrous) histiocytoma or “elusive” dermatofibroma, 3 the latter term as elusive as the entity sometimes itself seems to be. This variant of DF characteristically shows prominent hemorrhage, which may even mimic aneurysms or ectatic vessels, therefore the name. Moreover, these lesions are usually rich in siderophages, which frequently leads to the clinical miscue of a melanoma; in other instances vascular lesions and in particular nodular Kaposi sarcoma is suspected. Not rarely, sudden painful increase in size of a long-standing lesion is reported by the patient. Apart from hemorrhage, histology reveals characteristics of a dermatofibroma: an ill-defined dermal lesion with a moderately acanthotic epidermis—no erosion or exulceration as in “classic AFX”!—a variable mixture of fibrocytes and macrophages (siderophages!), which usually dominate, some fibrosclerotic collagen, and a variable amount of lymphocytes within and around the lesion. There may even be numerous bizarre giant cells and in early stages an increased number of mitoses. Some of the lesions may be strongly pigmented so that even nodular melanoma is suspected histologically, and in other instances with more spindle cells nodular Kaposi sarcoma may be imitated. Histochemistry, which reveals iron deposition, and immunohistochemistry, which is negative for S100 protein and vascular markers such as CD31 or factor VIII, easily exclude these diagnostic pitfalls. In contrast, variable reactivity is seen for factor XIIIa, as also described in the present series, 1 and with KiM1p. Moreover, in the photomicrographs provided 1 we could not recognize significant atypia and/or numerous, irregular mitoses, which are characteristics of “classic AFX”. In contrast, the lesions appear ill defined, with some fibrosclerotic collagen at the periphery and even collarette formation is mentioned by the authors in 3 cases, a finding that frequently is missing or only mildly present in this variant of dermatofibroma. In short, pigmented AFX is not a variant of AFX, but a variant of dermatofibroma, and thus expands the number of entities covered under the heading of AFX. In our experience the pleomorphic storiform pattern of AFX can be seen in a variety of diseases. On one hand this includes malignant neoplasms such as squamous cell carcinomas, melanomas, or leiomyosarcomas; when immunoreactivity for keratin, melanocytic, or smooth muscle markers is seen, correct interpretation is easily established. Yet, we have seen rare cases of melanoma and leiomyosarcoma that missed these markers, but subsequently developed recurrences and/or metastases positive for such markers. In squamous cell carcinomas masquerading as AFX, loss of immunoreactivity for keratins in our experience is a consistent finding. 4 Squamous cell carcinomas, type AFX, behave similar to epithelial malignancies from internal organs such as thyroid gland, stomach, colon, or kidney, where loss of immunoreactivity for keratin markers in dedifferentiated lesions is well established. 5–8 The simultaneous acquisition of reactivity for vimentin seems to mirror the fact that intermediate filaments for vimentin form the core of the longer keratin filaments, which obviously get lost by tumor progression. 9 Careful histology frequently reveals numerous hints of keratogenous differentiation such as contiguity or even continuity with a squamous cell carcinoma, type actinic keratosis, overlap with other types of more conventional (“classic”) squamous cell carcinoma or keratoacanthoma, parakeratotic horn pearls, dyskeratoses, or connection of neighboring tumor cells by multiple spines. 4 Clinically, these lesions are found in sun-exposed areas of patients in their seventh to eighth decade, are frequently diagnosed as squamous cell carcinomas, and there frequently is a history of a previous squamous cell carcinoma at the location, or there are numerous other types of squamous cell carcinoma, in particular actinic keratoses, in the neighborhood. Metastases from carcinomas or sarcomas may also rarely imitate AFX. They also grow rapidly, frequently occur on the head, exulcerate, and show a nodular growth pattern. There are 2 basic types of metastases: on one hand well-demarcated, nodular lesions with good differentiation, which cause no difficulty in separation from AFX; on the other hand ill-defined, diffusely growing lesions with bad or no signs of differentiation and anaplastic morphology. The latter characteristically show a radial growth pattern (originating from a central vessel), which leads to a sunbeam-like appearance at the periphery. This is due to strands of tumor cells (“Indian filing”) with molding of crowded nuclei imitating a serial car crush accident. In both types of metastases one can not uncommonly find intravascular neoplastic cells, which from a practical point of view are never seen in AFX. The history of a preexisting neoplasia elsewhere and coexistence of other metastases can be additional, helpful clinical features. Moreover, the further course of disease is completely different: wide-spread metastases will frequently characterize metastatic disease, while AFX in the vast majority of cases is cured by complete excision; only a few exceptional cases have been reported to progress as a recurrence, local lymph node metastases or, in exceptional cases, metastases to the lung. 10 The pattern of AFX may also be imitated by inflammatory disorders; one is reticulohistiocytoma, a variant of solitary xanthogranuloma with characteristic ground glass giant cells 11; others are variants of dermatofibroma such as those with monster cells, 12 the atypical (“pseudosarcomatous”) type 13, or, as in the present paper, aneurysmal/angiomatoid fibrous histiocytoma. 2 It seems that the same principle by which CDM Fletcher a decade ago was able to demonstrate that malignant fibrous histiocytoma is no special entity, but a potpourri of different, markedly dedifferentiated entities, 14 is true for AFX including its pigmented variant. Accordingly, AFX as well as its pigmented variant should be given the diagnosis they deserve: squamous cell carcinoma (in the vast majority of cases) and aneurysmal or angiomatoid fibrous histiocytoma, respectively. Bernhard Zelger, MD