Fibrohistiocytic tumors of intermediate malignancy originally included only dermatofibrosarcoma protuberans and the closely related giant cell fibroblastoma. This category now embraces other lesions, such as plexiform fibrohistiocytic tumor and soft tissue giant cell tumor. Angiomatoid fibrous histiocytoma, formerly considered a fibrohistiocytic tumor of intermediate malignancy, is better considered a translocation-associated mesenchymal tumor of uncertain type (see Chapter 32 ). All the fibrohistiocytic tumors of intermediate malignancy are characterized by a significant risk of local recurrence but a limited risk of regional and distant metastasis, differing from fully malignant sarcomas in this important respect. They also occur in a decidedly younger population than do most sarcomas; indeed, some seem to occur almost exclusively in children. Although there is a general consensus that these lesions do not display true histiocytic differentiation, the term “fibrohistiocytic” remains a useful and widely understood descriptive term for this group of neoplasms. Dermatofibrosarcoma protuberans and its juvenile counterpart, giant cell fibroblastoma, seem to be most closely related to fibroblasts, and indeed the presence of CD34 immunoreactivity in these two lesions provides a linkage to the CD34-positive dendritic cells that populate the dermis. On the other hand, plexiform fibrohistiocytic tumor seems to most closely approach the spirit of “fibrohistiocytic,” inasmuch as it has a bimodal population of cells, one of which contains large numbers of histiocytes and the other resembling fibroblasts or myofibroblasts.
Dermatofibrosarcoma Protuberans
Dermatofibrosarcoma protuberans (DFSP), first described in 1924 by Darier and Ferrand as “progressive and recurring dermatofibroma,” is a nodular cutaneous tumor characterized by a prominent storiform pattern. Over the years it has been considered a fibroblastic, histiocytic, and neural tumor. DFSP bears some histologic similarity to cellular benign fibrous histiocytoma; on this basis, along with its pigmented counterpart (Bednar tumor), DFSP was classified with the fibrohistiocytic neoplasms. In contrast to fibrous histiocytoma, DFSP grows in a more infiltrative manner and has a greater capacity for local recurrence. Moreover, in unusual circumstances it metastasizes, although distant metastasis is usually a late event.
Clinical Findings
DFSP typically presents during early or middle adult life as a nodular cutaneous mass. Although early studies reflected its rarity in children, it is increasingly reported in the pediatric age group. In fact, given the indolent growth and long preclinical duration, it is likely that many begin during childhood and become apparent only during young adulthood. Males are affected more frequently than females. Although these tumors occur at almost any site, they are seen most frequently on the trunk and proximal extremities ( Table 11.1 ). Dermatofibrosarcomas involving the hands and feet are rare. Antecedent trauma, reported in about 10% to 20% of cases, is probably coincidental.
| Anatomic Location | No. of Cases | % |
|---|---|---|
| Head and neck | 124 | 14.5 |
| Upper extremity | 155 | 18.2 |
| Trunk | 404 | 47.4 |
| Lower extremity | 170 | 19.9 |
| total | 853 | 100 |
In most cases this tumor is characterized by slow but persistent growth over a long period, often several years. The clinical and gross appearances are determined to a great extent by the stage of the disease. The initial manifestation is usually the development of a firm, plaquelike lesion of the skin, often with surrounding red to blue discoloration. These lesions have been compared with the morphea of scleroderma or morphea-like basal cell carcinoma. Rarely, the lesions appear as a depressed area of atrophy (“atrophic variant”) or as subcutaneous masses without obvious connection to the overlying dermis. Subcutaneous dermatofibrosarcomas involving the breast may clinically closely mimic mammary carcinoma. Less often, multiple small subcutaneous nodules appear initially rather than a plaque. The plaque may grow slowly or remain stationary for a variable period, eventually entering a more rapid growth phase and giving rise to one or more nodules. Thus, only in the fully developed lesion is the typical “protuberant” appearance manifested. Neglected tumors may achieve enormous proportions and have multiple satellite nodules. Despite the large size of many of these tumors, patients appear surprisingly well and lack the signs of cachexia associated with malignancies.
Gross Findings
Most of these tumors are biopsied during the nodular stage; therefore the specimen consists of a solitary, protuberant, gray-white mass involving subcutis and skin ( Figs. 11.1 to 11.3 ). The average size at surgery is approximately 5 cm. Multiple discrete masses are usually not seen in the original tumor but are more characteristic of recurrent lesions ( Fig. 11.4 ). The skin overlying these tumors is taut or even ulcerated. Skeletal muscle extension is uncommon except in large or recurrent lesions. Rarely, dermatofibrosarcomas are centered in the subcutis, with only subtle dermal involvement. Occasionally, areas of the tumor have a translucent or gelatinous appearance corresponding microscopically to myxoid change. Hemorrhage and cystic change are sometimes seen, but necrosis, a common feature of undifferentiated pleomorphic sarcomas, is rare.
Microscopic Findings
Despite the apparent gross circumscription of these lesions, the tumor diffusely infiltrates the dermis and subcutis ( Fig. 11.5 ). The tumor may reach the epidermis or leave an uninvolved zone of dermis just underneath the epidermis. In either event, the overlying epidermis does not usually display the hyperplasia that characterizes some cutaneous fibrous histiocytomas ( dermatofibromas ). The peripheral portions of the tumor have a deceptively bland appearance partly caused by the marked attenuation of the cells at their advancing edge. This is especially true in superficial areas, where the spread of slender cells between preexisting collagen is easily mistaken for cutaneous fibrous histiocytoma ( Fig. 11.6A ). In deep regions the tumor spreads along connective tissue septa and between adnexae ( Fig. 11.7 ), or it intricately interdigitates with lobules of subcutaneous fat, creating a lacelike or honeycomb effect ( Fig. 11.6B ).
The central or main portion of the tumor is composed of a uniform population of slender fibroblasts arranged in a distinct, often monotonous, storiform pattern around an inconspicuous vasculature ( Figs. 11.8 and 11.9 ). There is usually little nuclear pleomorphism and only low to moderate mitotic activity. Secondary elements such as giant cells, xanthoma cells, and inflammatory elements are few in number or absent altogether. In this respect, DFSP displays remarkable uniformity compared with other fibrohistiocytic neoplasms. Although most tumors are characterized by these highly ordered cellular areas, occasional tumors contain myxoid areas ( Fig. 11.10 ). These myxoid areas occur in both primary and recurrent lesions and are characterized by the interstitial accumulation of ground substance material. As myxoid change of the stroma becomes more pronounced, the storiform pattern becomes less distinct and the vascular pattern more apparent. By virtue of these features, such tumors can resemble myxoid liposarcoma ( Fig. 11.10B ). Confident diagnosis of highly myxoid dermatofibrosarcomas usually requires identification of more typical areas.
Giant cells, similar to those in giant cell fibroblastoma, can be identified in a small percentage of otherwise typical dermatofibrosarcomas. An unusual feature of DFSP is the myoid nodule ( Fig. 11.11 ). Originally construed as evidence of myofibroblastic differentiation, these structures seem to be centered in some cases around blood vessels, and likely represent an unusual nonneoplastic vascular response to the tumor. Infrequently, DFSP contains areas that are indistinguishable from fibrosarcoma ( Figs. 11.12 to 11.14 ). Characterized by long fascicles of spindle cells with more nuclear atypia and mitotic activity, these areas usually sharply abut conventional low-grade areas. Mitotic activity usually averages more than 5 mitotic figures per 10 high-power fields (hpf), in contrast to areas of conventional DFSP, which usually have fewer than 5 mitotic figures/10 hpf. In exceptional instances, DFSP contains areas resembling undifferentiated pleomorphic sarcoma ( Figs. 11.15 and 11.16 ). Fibrosarcomatous areas were originally believed to be more common in recurrent lesions, but studies have documented that the contrary is true. The biologic significance of sarcomatous areas in DFSP is discussed later. Metastatic deposits from this tumor occur most often in the lung and then in regional lymph nodes, where they may resemble the parent tumor or may appear more pleomorphic ( Fig. 11.17 ).
Immunohistochemical Findings
DFSP is characterized by the almost consistent presence of CD34 ( Fig. 11.18 ), the human progenitor cell antigen, in a significant proportion of its cells. Although CD34 expression is widespread in mesenchymal cells and tumors, its presence in DFSP suggests a close linkage to the normal CD34-positive dendritic cells of the dermis, including those that ensheath the adnexae, nerves, and vessels. The nearly consistent expression of this antigen has also proved useful for distinguishing DFSP from benign fibrous histiocytoma, especially when dealing with small biopsies. Although cutaneous fibrous histiocytomas may contain CD34-positive cells, these frequently account for only a minority of the tumor cell population and are often found at the deep border of the tumor. Caution should be used when interpreting CD34 immunostains in spindle cell tumors of the skin, being certain that positively staining cells are neoplastic, not entrapped normal dermal dendritic cells. Apolipoprotein D may also be of value in the distinction of dermatofibrosarcomas from fibrous histiocytomas, although this antibody is used in relatively few laboratories.
Cytogenetic and Molecular Genetic Features
Both dermatofibrosarcoma and giant cell fibroblastoma are characterized either by a supernumerary ring chromosome, consisting of low-level amplification of sequences from chromosomes 17 and 22 , and infrequently 8, or by linear translocation derivatives. The presence of a ring versus a linear translocation may be related to age. Specifically, adult cases typically possess the ring chromosome, whereas pediatric cases have the linear translocation derivative. Either event fuses exon 2 of the platelet-derived growth factor β-chain gene ( PDGFB ) to various exons of the collagen type 1 α 1 gene ( COL1A1 ), resulting in a fusion transcript that places PDGFB under the control of the COL1A1 promotor. The fusion protein is processed to an end product that is indistinguishable from the normal PDGFB product. Overproduction of PDGFB by dermatofibrosarcoma results in autocrine stimulation and cell proliferation, a sequence of events that can be interrupted by specific tyrosine kinase inhibitors (see later). Approximately 8% of dermatofibrosarcoma cases are fusion negative; it remains to be seen if these rare cases contain cryptic rearrangements of COL1A1 and PDGFB or altogether different genetic abnormalities. PDGFB amplification and rearrangement in dermatofibrosarcoma can be detected by fluorescence in situ hybridization (FISH) in formalin-fixed, paraffin-embedded tissue (FFPE) sections.
Differential Diagnosis
The most common problem in the differential diagnosis is distinguishing this tumor from other fibrohistiocytic neoplasms. DFSP has a more uniform appearance, more distinct storiform pattern, and fewer secondary elements (i.e., giant cells, inflammatory cells) than either a benign fibrous histiocytoma or a superficially located undifferentiated pleomorphic sarcoma. The distinction between benign fibrous histiocytoma and dermatofibrosarcoma occasionally proves difficult when only the superficial portion of the dermatofibrosarcoma is present in a biopsy specimen, because these areas appear so well differentiated ( Table 11.2 ). Under these circumstances, knowledge of the size and configuration of the lesion in question suggests the diagnosis, and biopsy of a deeper portion confirms it. In addition, because CD34 is almost always expressed by dermatofibrosarcoma and is much less often positive in benign fibrous histiocytoma, CD34 is an extremely useful antigen for solving this problem. Undifferentiated pleomorphic sarcoma is not often confused with this tumor because it is characterized by far greater pleomorphism, mitotic activity, and necrosis. Moreover, its typical deep location in muscle and more rapid growth are at variance with the indolent course of this tumor. Rarely, one encounters DFSP with areas of undifferentiated pleomorphic sarcoma (see Figs. 11.15 and 11.16 ). As indicated earlier, when such areas represent more than just a microscopic focus, they should be diagnosed as “sarcoma arising in dermatofibrosarcoma protuberans.”
| Parameter | Benign Fibrous Histiocytoma | DFSP |
|---|---|---|
| Common locations | Extremities | Trunk; groin |
| Size | Usually small | Small to large |
| Growth pattern | Short fascicles, haphazard | Monotonous storiform |
| Cell population | Plump spindle cells often admixed with inflammatory cells, siderophages, giant cells | Slender spindle cells with few if any secondary elements |
| Hemorrhage | Occasional | No |
| Subcutaneous extension | Occasional and limited | Consistent and extensive |
| CD34 | Focal staining in occasional cases; cuff of CD34-positive cells at deep border | Diffuse and extensive staining in most cases |
| Local recurrence | 5%–10% | 20%–50% |
| Metastasis | Anecdotal cases only | Rare in conventional form; potentially higher if fibrosarcoma present with inadequate local control |
| Malignant transformation | Anecdotal cases only | Fibrosarcoma in occasional cases |
A second common problem is the confusion of this tumor with benign neural tumors, specifically a diffuse form of neurofibroma . This is most likely to occur when dermatofibrosarcoma is in the plaque stage or when a biopsy is done on only the periphery of the tumor. However, neurofibroma usually contains tactoid structures or other features of neural differentiation, and it lacks the highly cellular areas with mitotic figures that characterize the central portion of a dermatofibrosarcoma. The presence of S-100 protein and SOX10 in virtually all neurofibromas and their absence in dermatofibrosarcoma is an additional point of contrast.
Highly myxoid forms of dermatofibrosarcoma may resemble myxoid liposarcoma by virtue of the prominent vasculature and bland stellate or fusiform cells. However, the superficial location, gross configuration, CD34 immunoreactivity, and complete absence of lipoblasts should raise serious questions concerning the diagnosis of liposarcoma. In such cases, additional sampling of the tumor or review of the original material in a recurrent lesion may reveal the diagnostic cellular areas.
Superficial acral fibromyxoma (cellular digital fibroma), another CD34-positive spindle cell tumor of the dermis, typically occurs in the fingers and toes (unusual locations for dermatofibrosarcomas), contains wiry collagen, and lacks a storiform growth pattern.
Plaque-like CD34-positive dermal fibroma (“medallion-like dermal dendrocyte hamartoma”) is a very rare CD34-positive spindle cell tumor that most often involves the neck and upper extremities. The histologic distinction of this lesion from dermatofibrosarcoma may be extremely challenging, although plaquelike CD34-positive dermal fibroma tends to show a parallel arrangement of tumor cells and sparing of adnexal structures. FISH for PDGFB amplification has been reported to be negative in a small number of studied cases, potentially assisting in the distinction of this lesion from “atrophic” examples of dermatofibrosarcoma.
Discussion
Unlike benign fibrous histiocytoma, DFSP is a locally aggressive neoplasm that recurs in up to one-half of patients. The high recurrence rate in part reflects the extensive infiltration of the tumor compared with fibrous histiocytoma and failure to appreciate this phenomenon at surgery. It is clear that prompt wide local excision (2-3 cm), the standard of practice for this lesion, can greatly alter the recurrence rate. Recurrence rates for patients treated by wide local excision range from 10% to 20%, compared with 43% when the excision was undefined or conservative. In addition, recurrence rates in patients treated primarily at large referral centers are low (1.75%–33%), again suggesting that adequate initial surgery is essential for minimizing recurrences. The risk of local recurrence also correlates well with the extent of the wide excision. If the excision margin is 3 cm or more, the recurrence rate is 20%, compared with 41% if the margin is 2 cm or less. If local recurrence develops, it is usually within 3 years of the initial surgery, although about one-third of patients will develop recurrences after 5 years, attesting to the need for long-term follow-up. In patients who develop multiple recurrences, progressively shorter intervals between successive recurrences have been noted.
Mohs micrographic surgery has been met with growing enthusiasm for treatment of this disease. Advocates of this approach note that DFSP occasionally grows in an asymmetric fashion from its epicenter such that a traditional wide local excision fails to remove all tumor in a subset of cases. Mohs surgery offers the potential to achieve clear margins with minimum removal of normal tissue, an advantage particularly attractive for sites such as the head and neck. Local recurrence rates following Mohs surgery are less than 10% and in some studies approach 0%.
Despite its locally aggressive behavior, this tumor infrequently metastasizes and therefore should be clearly distinguished from conventional sarcomas. The incidence of metastasis is difficult to assess because of the bias introduced when selectively reporting metastasizing tumors, the inability to determine if sarcomatous areas were noted in a subset of reported cases, and the lack of uniform treatment. In general, however, for ordinary DFSP uncomplicated by areas of fibrosarcoma, metastasis appears to be an uncommon event. In one large study of 115 patients, no metastases were observed, whereas, in two other studies, 5 of 86 patients and 4 of 96 patients developed metastases. In the latter study the follow-up period was 15 years. Thus, long-term follow-up may reveal higher metastatic rates than previously reported. Of the 471 patients reported in the literature, 16 (3.4%) developed metastatic disease. About three-fourths of patients with metastases have hematogenous spread to the lungs, and one-fourth have lymphatic spread to regional lymph nodes. Metastases to other sites, such as the brain, bones, and heart, have also been documented. Although some metastasizing cases have clearly originated from tumors with areas of “sarcoma,” some have not.
Metastasizing lesions share some common clinical features. They are almost always recurrent lesions, and there is usually an interval of several years between diagnosis and metastasis. The low incidence of regional lymph node metastasis and the negative findings in a small series of blind lymph node dissections do not warrant routine node dissection. Resection of isolated pulmonary metastases has been advocated because of the overall low-grade behavior of the tumor. Radiotherapy has been recommended for large, unresectable tumors or postoperatively for margin-positive tumors. Molecular targeting of DFSP with imatinib mesylate in patients with advanced or metastatic disease resulted in significant reductions in tumor burden.
Sarcoma Arising in Dermatofibrosarcoma Protuberans (Fibrosarcomatous Variant of Dermatofibrosarcoma Protuberans)
It has become increasingly clear that a small subset of dermatofibrosarcomas contain areas indistinguishable from conventional fibrosarcoma (and rarely undifferentiated pleomorphic sarcoma). This has led to use of the term fibrosarcomatous variant of dermatofibrosarcoma and the suggestion that these lesions pursue a more aggressive course, although the risk of distant metastasis is debated. These tumors share the same general clinical properties as ordinary DFSP, and in most cases the sarcomatous foci are noted in the original tumor.
To diagnose sarcoma arising in DFSP, we have generally used a constellation of features. The “sarcomatous” foci should constitute at least 5% to 10% of the tumor, in contrast to simply a rare to occasional microscopic focus. These zones are characterized by a fascicular (rather than storiform) architectural pattern and are composed of plump spindle cells of high nuclear grade. Mitotic activity is increased in these areas, whereas CD34 immunoreactivity is often (but not always) diminished (see Fig. 11.18B ) compared to the surrounding dermatofibrosarcoma. In addition, fibrosarcomatous areas are also characterized by a higher MIB-1 labeling index and increased p53 immunostaining than the classic areas. Although we have never required an absolute level of mitotic activity to diagnose sarcomatous change, mitotic activity within these sarcomatous areas averages 7 to 15 mitotic figures/10 hpf, compared with 1 to 3 mitotic figures/10 hpf in dermatofibrosarcoma.
The significance of sarcomatous forms of DFSP has been well studied. Although it seems logical that high-grade areas within a low-grade lesion would affect behavior adversely, early studies failed to confirm this hypothesis in a statistically meaningful fashion. Ding et al. suggested that fibrosarcomatous areas within DFSP are associated with a higher local recurrence rate and thus a more aggressive course, but the status and adequacy of surgical excisions in these cases was not made clear. Connelly and Evans reported no difference in the local recurrence rate or in time to recurrence compared with conventional DFSP but noted that two of their six patients with fibrosarcomatous areas developed metastatic disease. Two large studies by Mentzel et al. and Pizarro et al., on the other hand, revealed higher local recurrence rates (58% and 42%, respectively) and metastatic rates (14.7% and 33.0%, respectively) ( Table 11.3 ). In neither study, however, did it appear that wide local excision with clear margins was achieved in most cases. A study of 41 patients documented a 10% incidence of metastasis in fibrosarcomatous DFSP, but again, patients apparently did not receive optimal surgical therapy. All received local excisions that often left positive margins, and three of four patients with metastasis also experienced local recurrences, again suggesting inadequate primary surgery. Because the ability to eradicate tumor locally arguably affects the risk of subsequent dissemination, none of these studies addressed the behavior of this neoplasm in the context of the current standard of practice. In the study of Goldblum et al. of 18 patients treated by wide local excision and clear margins and with a minimum follow-up of 5 years, local recurrence rates were essentially identical to those of typical DFSP, with no reported metastases. Similarly, culling cases from the literature of sarcomatous dermatofibrosarcoma in which wide local excisions were performed, only one instance of metastasis was noted ( Table 11.4 ). To clearly resolve the issue of metastasis in fibrosarcomatous DFSP, follow-up of a larger cohort of patients treated with adequate surgery and having negative margins is needed.
