Dermatofibrosarcoma protuberans


  • William M. Mendenhall M.D.,

    Corresponding author
    1. Department of Radiation Oncology, University of Florida College of Medicine, Gainesville, Florida
    • Department of Radiation Oncology, University of Florida Health Science Center, P.O. Box 100385, Gainesville, FL 32610-0386
    Search for more papers by this author
    • Fax: (352) 265-0759

  • Robert A. Zlotecki M.D., Ph.D.,

    1. Department of Radiation Oncology, University of Florida College of Medicine, Gainesville, Florida
    Search for more papers by this author
  • Mark T. Scarborough M.D.

    1. Department of Orthopedic Surgery, University of Florida College of Medicine, Gainesville, Florida
    Search for more papers by this author


In the current review, the authors set out to discuss the natural history and treatment of dermatofibrosarcoma protuberans (DFSP), a rare indolent cutaneous tumor. Approximately 10–15% of all DFSPs contain areas of fibrosarcoma (DFSP-FS), and such cases tend to exhibit more aggressive behavior. The optimal treatment for DFSP is resection with wide margins; the likelihood of local control associated with this procedure exceeds 90%. The probability of regional or distant metastases is ≤ 5%. Patients with positive or close surgical margins have an elevated risk of local recurrence after resection alone; however, postoperative radiotherapy results in local control rates of ≥ 85% in such patients. Postoperative radiotherapy also is indicated in the rare event that a patient has unresectable macroscopic disease. Experience with the use of radiotherapy alone to treat macroscopic disease is limited. Cancer 2004 © 2004 American Cancer Society.

Dermatofibrosarcoma protuberans (DFSP) is a rare monoclonal cutaneous soft tissue sarcoma that was first described by Taylor in 1890.1 Chang et al.,2 reporting on 60 patients with DFSP who were treated at the University of Illinois–Chicago (Chicago, IL) between 1968 and 2001, estimated that approximately 100 patients with soft tissue sarcoma (any type) were treated each year at their institution; thus, DFSP accounted for approximately 1.8% of all soft tissue sarcomas encountered. Approximately 85–90% of all DFSPs are low-grade lesions; the remainder contain a high-grade fibrosarcomatous component, which accounts for ≥ 5% of the tumor volume, and are considered to be intermediate-grade lesions (known as ‘DFSP-FSs’).3 DFSPs arise as pink or violet-red plaques, and the surrounding skin may be telangiectatic.4 These lesions typically are fixed to the dermis but move freely over deeper-lying tissue, and they do not exhibit a nodular growth pattern until late in their course. Fixation to more deeply seated structures often is observed in advanced and/or recurrent cases of DFSP.4

DFSP tends to exhibit an indolent growth pattern, and in many cases, its symptoms are long lasting. Lindner et al.,4 who examined a series of 35 patients with DFSP who were treated at the University of Florida (Gainesville, FL), reported that symptoms ranged in duration from 6 months to 30 years (mean, 6.4 years). DFSP typically arises on the trunk of the body and is more common in men than in women. The following site distribution was observed in a series of 853 patients reported on by Enzinger and Weiss5: trunk, 47%; lower extremity, 20%; upper extremity, 18%; and head and neck, 14%. Rutgers et al.6 reported a male-to-female ratio of approximately 3:2 (males, 57%; females, 43%) in a series of 264 patients with DFSP. DFSP usually presents in the fourth decade of life, although it has been known to arise in patients of widely varying ages.3, 4 In a series of 159 patients treated at the Memorial Sloan-Kettering Cancer Center (New York, NY) between 1950 and 1998, Bowne et al.3 reported a median age of 39 years (range, 12–79 years). Bowne and colleagues also found that the majority of DFSPs measured < 5 cm in maximum diameter and were superficial in nature. Specifically, they reported the following distribution of lesion sizes: < 5 cm, 134 patients (84%); 5–10 cm, 21 patients (13%); and > 10 cm, 4 patients (3%). In addition, they found that DFSPs were superficial in 121 patients (77%) and had invaded deeper structures in only 36 patients (22%); depth of invasion was not reported in the remaining 2 patients (1%).

DFSP arises from the rearrangement of chromosomes 17 and 22 such that the collagen Type Iα1 gene (COL1A1) becomes fused to the platelet-derived growth factor (PDGF) β-chain gene (PDGFB). This rearrangement results in the deregulation of PDGF β-chain expression and leads to continuous activation of the PDGF receptor β (PDGFRβ) protein tyrosine kinase, which promotes DFSP cell growth.7–9 Linn et al.,10 who used array-based comparative genomic hybridization to compare gene expression patterns in 9 DFSPs and 27 other soft tissue tumors, found that DFSP had a distinctive expression profile characterized by the amplification of sequences from chromosomes 17q and 22q, with these sequences being bounded by the COL1A1 and PDGFB genes.

Histologically, DFSP is composed of monomorphic, benign-appearing spindle cells arranged in a matted or storiform pattern, with these cells intersecting at tight right angles around central vessels.4 Early in the course of the disease, there may be a narrow tumor-free zone (known as a ‘grenz zone’) between the lesion and the epidermis.4 Uncommon variants of DFSP include the Bednar tumor, which is characterized by the presence of melanin-containing dendritic cells, and myxoid DFSP.11, 12 In addition, approximately 10–15% of all DFSPs contain a fibrosarcomatous component.3, 13, 14 Goldblum et al.15 reported that out of a series of 18 DFSP-FSs, 15 were previously untreated lesions, whereas the remaining 3 were recurrent tumors. The sarcomatous component, which was fibrosarcoma in 17 of these 18 DFSP-FSs and malignant fibrous histiocytoma in 1, was found to account for 20–80% of the lesion volume (median, 60%). Goldblum and colleagues observed 2–16 mitoses per 10 high-powered fields (HPF) in the sarcomatous component (median, 7 mitoses per 10 HPF), compared with 0–3 mitoses per 10 HPF (median, 1 mitosis per 10 HPF) in the DFSP component. The DFSP component was found to be CD34 positive in all 18 cases, whereas the sarcomatous component had positive CD34 status in only 9 cases.15 Takahira et al.,16 who assessed microsatellite instability (MSI) and p53 mutation status in 44 tumors from a total of 36 patients (DFSP, 27 patients; DFSP-FS, 9 patients) treated at the National Kyushu Cancer Center (Fukuoka, Japan), found that MSI was more common in DFSP-FSs compared with DFSPs (4 of 9 vs. 1 of 27; P = 0.028). In addition, their analysis of p53 mutation status revealed 10-point mutations in 6 of 36 patients (17%). Those investigators postulated that the development of MSI and the mutation of p53 were early and late events, respectively, in the progression of DFSP to DFSP-FS.16

Sasaki et al.17 analyzed MIB-1 labeling index (LI) and p53 expression in 19 patients with DFSP, including 3 who had DFSP-FS, and in 46 patients with dermatofibroma, 11 patients with malignant fibrous histiocytoma, and 4 patients with fibrosarcoma. DFSPs were found to have a higher MIB-1 LI compared with dermatofibromas but a lower MIB-1 LI compared with sarcomas. In addition, MIB-1 LI was found to be higher in recurrent DFSPs than in previously untreated tumors, which were similar to dermatofibromas in terms of this index. DFSP-FSs exhibited a higher level of proliferative activity compared with DFSPs, and increased proliferative activity also was observed in the 16% of DFSPs/DFSP-FSs (3 of 19) that showed evidence of p53 overexpression. None of the dermatofibromas analyzed exhibited p53 overexpression, whereas most of the sarcomas that were examined did.

Diagnostic Evaluation

Tumor extent and mobility generally are assessed on physical examination. DFSPs rarely exhibit lymphatic or hematogenous dissemination; regional lymph nodes are assessed by palpation.

Magnetic resonance imaging (MRI) is useful for ascertaining deep tumor invasion, particularly in patients with large recurrent lesions. Torreggiani et al.18 reported on 10 patients who underwent MRI at the University of British Columbia (Vancouver, British Columbia, Canada) between 1989 and 2002. Conventional T1-weighted images revealed the tumor to be isointense compared with skeletal muscle in 5 of these 10 patients, to be slightly hypointense in 3 patients, and to be hyperintense in 2 patients. All 10 lesions were hypointense compared with subcutaneous fat. Conventional spin-echo T2-weighted images were obtained in six patients, and fast spin-echo T2-weighted images were obtained in four. On these images, five lesions exhibited higher signal intensity and five exhibited similar signal intensity compared with subcutaneous fat.

Computed tomography (CT) is not indicated except in rare cases in which underlying bone involvement is suspected. The occasional patient may experience pulmonary metastases, particularly if the lesion is advanced, recurrent, and/or of intermediate grade. Therefore, chest roentgenography should be performed for all patients, and chest CT should be performed for patients with unfavorable, advanced-stage tumors.

Histologic diagnoses are obtained via core-needle or incisional biopsy. Fine-needle aspiration (FNA) does not yield enough tissue to allow the rendering of an accurate diagnosis for most previously untreated tumors.19, 20 Nonetheless, FNA may be useful in establishing diagnoses in patients with previously treated tumors that are suspected to have recurred.20

At times, it may be difficult to draw a histologic distinction between DFSP and other mesenchymal tumors, such as dermatofibroma and benign fibrous histiocytoma.21, 22 DFSPs tend to stain positively for CD34 and negatively for factor XIIIa, whereas dermatofibromas generally have negative CD34 status and positive factor XIIIa status.21 In addition, dermatofibromas tend to stain positively for CD44, a membrane glycoprotein that is thought to be the cell surface receptor for hyaluronate, the major component of the extracellular matrix. In dermatofibromas, stromal cells typically exhibit faint staining for hyaluronate, whereas DFSPs tend to exhibit strongly positive staining for this marker. Furthermore, unlike dermatofibromas, DFSPs tend to have absent or significantly reduced CD44 immunostaining.21

Fanburg-Smith and Miettinen22 analyzed 1150 tumors, as well as fetal and normal adult tissue samples, for the expression of p75, a low-affinity nerve growth factor receptor. Those investigators reported that a wide variety of mesenchymal and epithelial tumors stained positively for p75, a finding that indicates that p75 expression is not specific to nerve sheath tumors. DFSPs exhibited positive staining for p75 in 69 of 73 cases (95%), whereas benign fibrous histiocytomas typically exhibited negative staining; thus, p75 staining may be useful for distinguishing between these two types of lesions.


DFSPs and DFSP-FSs are staged in accordance with the American Musculoskeletal Tumor Society (MSTS) staging system, which takes into account tumor grade and compartmentalization.23, 24 MSTS Stage IA tumors are low-grade, intracompartmental lesions and can be managed adequately solely with wide excision (i.e., dissection outside of the reactive zone); Stage IB tumors are low-grade lesions that exhibit extracompartmental extension. The ‘Stage II’ classification is indicative of high histologic grade and thus does not apply to DFSPs or DFSP-FSs. Therefore, the MSTS staging system as it applies to DFSPs and DFSP-FSs is as follows: Stage IA, no extension beyond the subcutaneous compartment; and Stage IB, involvement of underlying fascia or muscle.4 The American Joint Committee on Cancer25 has not set forth a system for the staging of DFSPs or DFSP-FSs.


The optimal treatment option for DFSPs and DFSP-FSs is resection with wide margins; the likelihood of local recurrence after this procedure is performed is less than 10%.4 Limited experience with Mohs surgery indicates that this procedure is associated with a high probability of cure provided that the final margins are negative26–30; in contrast, the risk of local recurrence exceeds 50% when the final margins are positive.1 The interval between treatment and the development of recurrent disease is variable. Chang et al.,2 in a series of 60 patients who received surgical treatment for DFSP at the University of Illinois–Chicago, found that 10 patients experienced local recurrence, which arose 1–100 months after surgery (mean, 38 months); 3 of these 10 recurrences (30%) were observed more than 5 years after surgery. Bowne et al.,3 who investigated a series of 159 patients treated at the Memorial Sloan-Kettering Cancer Center, reported a median time of 32 months to the development of local recurrence.

Adjuvant radiotherapy (RT), administered either before or after surgery, significantly reduces the risk of local recurrence in patients who have or who are likely to have close or positive margins.1, 31 The appropriate dose fractionation schedules and treatment techniques are similar to those that are used for other soft tissue sarcomas.24 Data pertaining to the efficacy of RT alone in the treatment of macroscopic disease are scant.1, 31


Of the 159 patients reported on by Bowne et al.,3 156 were treated with surgery alone, whereas only 3 were treated with surgery and RT (Table 1). One hundred thirty-four of these 159 patients (84%) had DFSP, and the remaining 25 had DFSP-FS. Macroscopic total resection was achieved in 157 patients (99%). Fifty-one patients (32%) had positive margins, 15 (10%) had close margins (< 1 mm), and 93 (58%) had negative margins. The observed 5-year local control rates were as follows: DFSP, 81%; DFSP-FS, 28%; and overall, 75%. Multivariate analysis revealed that close or positive margins (P < 0.001) and intermediate histologic grade (P < 0.001) were significantly associated with local recurrence. In addition, two patients with DFSP-FS who had positive margins after surgical resection went on to develop pulmonary metastases.

Table 1. Outcomes after Surgery Alone or Surgery in Combination with Adjuvant Radiotherapy
StudyInstitutionNo. of patientsFollow-upPrevious treatment (%)Surgery (%)Radiotherapy (%)Local controla (%)Metastases (%)
  • MSKCC: Memorial Sloan-Kettering Cancer Center; Univ.: University; CGMH: Chang Gung Memorial Hospital; Mohs: Mohs surgery; NYU: New York University.

  • a

    Interval in parentheses.

  • b

    Crude percentage.

Bowne et al., 20003MSKCC (New York, NY)159Median, 4.8 yrs30100275 (5 yrs)1
Chang et al., 20042Univ. of Illinois–Chicago (Chicago, IL)60Median, 59 mos100576 (10 yrs)
Lindner et al., 19994Univ. of Florida (Gainesville, FL)35Mean, 58 mos491001191b0
Gaynor et al., 199713Mayo Clinic (Rochester, MN)321001666b6
Sun et al., 200012CGMH (Taipei, Taiwan)34Median, 50 mos121002945 (7 yrs)0
Khatri et al., 200326Univ. of California–Davis (Davis, CA)24Median, 4.5 yrs541004100b0
Wacker et al., 200428Univ. of Heidelberg (Heidelberg, Germany)22Mean, 54 mos41100 (Mohs)0100b
DuBay et al., 200429Univ. of Michigan (Ann Arbor, MI)62Median, 4.4 yrs1003100b0
Nouri et al., 200230Univ. of Miami (Miami, FL) and NYU (New York, NY)20Mean, 56 mos100 (Mohs)0100b

Lindner et al.4 reported on 35 patients with DFSP of the trunk or extremities who underwent surgical treatment at the University of Florida between 1975 and 1996. Follow-up for these patients ranged from 1 to 12 years (mean, 58 months), and 1 patient was lost to follow-up. Eighteen patients had previously untreated tumors, and the remaining 17 had recurrent lesions. Thirty-four of the 35 patients initially underwent surgery at an institution other than the University of Florida, and this surgery (either marginal or intralesional resection) was inadequate in all cases. Reexcision left wide margins in 28 patients, marginal margins in 6 patients, and intralesional margins in 1 patient. Local control was achieved in all 28 patients who had wide margins after reexcision. In addition, all 4 patients with inadequate margins who received adjuvant RT (mean dose, 57.5 grays [Gy]; range, 50–65 Gy) experienced local control, which was first documented 4–8 years after treatment. Three other patients who had inadequate margins developed locally recurrent disease, and salvage treatment (namely, additional surgery) was successful for all three.

Sun et al.12 reported on 35 patients with DFSP who were treated with surgery alone (24 patients) or with surgery and RT (11 patients) at Chang Gung Memorial Hospital (Taipei, Taiwan) between 1987 and 1998. Follow-up ranged from 11 to 131 months in duration (median, 50 months). The median RT dose was 54 Gy (range, 46–68 Gy), with RT being administered at 1.8–2.5 Gy per fraction 5 times each week. One patient, who had a 27 cm × 25 cm × 5 cm tumor and received RT before undergoing incomplete resection, experienced local progression. Local control rates for the remaining 34 patients were as follows: close or positive margins, 1 of 3 patients (surgery alone) versus 5 of 6 patients (surgery + RT); and negative margins, 14 of 21 patients (surgery alone) versus 4 of 4 patients (surgery + RT). The 7-year local control rate was 28% after surgery alone and 80% after surgery plus RT. No patient experienced any severe complications.

Data pertaining to the efficacy of RT, and particularly to the efficacy of RT alone in the treatment of macroscopic disease, are limited (Table 2). Suit et al.,1 reporting on 3 patients who were treated with RT alone at total doses of 67–75 Gy, found that all 3 continued to experience local control at 85, 106, and 108 months after treatment, respectively. Fifteen other patients received preoperative (3 patients) or postoperative (12 patients) RT due to the presence of positive (12 patients) or close (3 patients) margins. One of these patients had a 3 mm margin and received 50 Gy, and the remaining 2 patients who had close (≤ 1 mm) or positive margins received 60–78 Gy. The 10-year local control rate after surgery and adjuvant RT was 84%. All three patients who experienced local recurrence underwent salvage surgery, which was successful in two cases. The 12 patients who remained free of disease at the time of the report had follow-up durations ranging from 33 to 119 months (median, 79 months). No patient developed metastases or experienced any severe complications.

Table 2. Outcomes after Radiotherapy Alone or Radiotherapy in Combination with Surgery
StudyInstitutionNo. of patientsFollow-upPrevious treatment (%)Surgical marginsLocal controla (%)
  • MGH: Massachusetts General Hospital; MDACC: M. D. Anderson Cancer Center; CGMH: Chang Gung Memorial Hospital; MSKCC: Memorial Sloan-Kettering Cancer Center.

  • a

    Interval in parentheses.

  • b

    Crude percentage.

Suit et al., 19961MGH (Boston, MA)1839Positive/close, 83%; intralesional, 17%88 (10 yrs)
Ballo et al., 199831MDACC (Houston, TX)19Median, 6 yrsNegative/positive, 95%; intralesional, 5%95 (10 yrs)
Sun et al., 200012CGMH (Taipei, Taiwan)10Positive/close, 60%; negative, 40%80 (7 yrs)
Lindner et al., 19994Univ. of Florida (Gainesville, FL)4Range, 4–8 yrsMarginal/intralesional, 100%100b
Stojadinovic et al., 200014MSKCC (New York, NY)4Positive, 100%75b
DuBay et al., 200429Univ. of Michigan (Ann Arbor, MI)2Positive, 100%100b

Ballo et al.31 investigated 19 patients who were treated with RT alone (1 patient) or with RT and surgery (18 patients) at the M. D. Anderson Cancer Center (Houston, TX) between 1972 and 1995. Patient follow-up ranged from 6 months to 23.5 years in duration (median, 6 years). One patient underwent surgery and thereafter rapidly developed recurrent disease; RT alone (dose, 65 Gy) was used to treat this patient's macroscopic recurrence, but the intervention was unsuccessful, and the patient died with disease 21 months after treatment. Two other patients received preoperative RT (dose, 50 Gy) and surgery, and both experienced local control. The remaining 16 patients underwent surgery before receiving adjuvant RT (mean dose, 59 Gy; range, 50–66 Gy; positive margins, 6 patients; negative margins, 10 patients), and all of these patients also experienced local control. No patient experienced any severe complications.


The optimal treatment option for patients with DFSP or DFSP-FS is resection with wide margins, with the addition of adjuvant RT improving the likelihood of cure in patients with close or positive surgical margins. In addition, although relevant data are limited, it appears that RT alone is reasonably likely to result in cure for the occasional patient with unresectable macroscopic disease. The efficacy of chemotherapy for patients with metastatic DFSP is not well characterized; however, a limited number of clinical reports have indicated that imatinib, a tyrosine kinase inhibitor, can induce regression in patients with unresectable recurrent and/or metastatic DFSP.7 Furthermore, Sjöblom et al.,8 in cultures from four patients with DFSP, demonstrated that treatment with a PDGF antagonist was capable of inducing apoptosis in tumor cells, a finding that could serve as the basis for the development of novel therapeutic options.