Advanced dermatofibrosarcoma protuberans: an updated analysis of cases from an Indian sarcoma clinic

Dermatofibrosarcoma protuberans (DFSP) is a rare tumor, accounting for less than 0.1% of all cancers and approximately 1% of all soft-tissue sarcomas. It is usually a low-grade malignancy, although 10–15% cases may be classifiable as intermediate grade due to a high-grade sarcomatous component (usually a fibrosarcoma [FS], DFSP-FS) [1]. It was first recognized as a distinct entity by Darier and Ferrand-Drake in 1924, and termed ‘dermatofibrosarkoma protuberans’ by Hoffman in 1925 [2,3]. It has an annual incidence of 0.8–4.5 per million and most commonly occurs in the fourth and fifth decades of life [4–6]. The most common primary site is the trunk, followed by the arms, the legs, and the head and neck region [6]. DFSP usually starts as an asymptomatic, slow-growing indurated plaque confined to the dermis, ranging in color from blue–red to violaceous or flesh colored, and eventually progresses to form a firm, protuberant nodular growth.

DFSP has a distinctive morphology, with a uniform population of monomorphic neoplastic spindle cells arranged in a predominantly storiform pattern over a background of fibrous stroma. There are no specific immunohistochemical markers for DFSP, but the tumor cells are usually positive for CD34 and vimentin on immunohistochemistry [7]. Chromosomal translocations involving 17q22 and 22q13 giving rise to a COL1A1-PDGFB fusion gene can be identified by fluorescence in situ hybridization (FISH) or real-time PCR in more than 90% of cases [8]. The COL1A1 breakpoint can vary from exon 6 to exon 49, while the breakpoint in the PDGFB gene always occurs in the intron preceding exon 2 [8]. The specific position of the COL1A1 breakpoint does not have any clinical or therapeutic implication [8,9]. The resultant fusion gene overcomes all negative controls on the translation and transcription of the PDGFB gene, causing an autocrine stimulation of DFSP tumor cell growth [10,11].

Localized tumors have a good prognosis. In a review of data from the SEER program for 1973–2002, the cause-specific 15-year survival rate was 99.7% (95% CI: 99.4–99.9%) [6]. The treatment for localized tumors is wide local excision (WLE) or Mohs micrographic surgery (MMS) with negative margins. There are no randomized trials comparing WLE versus MMS, and the choice between the two often comes down to availability, expertise, cost, the size of the tumor and its location. Achieving negative margins is often difficult due to the infiltration of tumor tissue into the subcutaneous planes, including fascia, muscle or bone, especially in recurrent cases. Higher age (>50 years), close (<1 mm) or positive surgical margins, DFSP-FS variant, high mitotic rate and increased cellularity are associated with decreased survival [1]. Local recurrence rates vary from 2 to 53% depending on the surgical techniques used for excision, but distant metastases are very rare (1–4%) [12–15]. In a series of 57 patients with DFSP, the mean time to the first recurrence after primary simple excision was 3.27 years for ordinary DFSP versus 1.87 years for DFSP-FS (p = 0.038) [16].

The US FDA approved imatinib for use in the management of unresectable, recurrent or metastatic DFSP in 2006 [17]. The largest prospective data for imatinib in advanced DFSP come from the pooled analysis of two Phase II trials conducted by the European Organisation for Research and Treatment of Cancer (EORTC) and the Southwest Oncology group (SWOG), with a total of 24 patients with locally advanced/metastatic DFSP given doses of imatinib ranging from 400 and 800 mg, respectively [18]. We had earlier published our experience with seven consecutive patients with metastatic/unresectable DFSP [19]. This updated analysis, after an interval of 3 years, includes seven additional patients with longer follow-up and also includes our experience with post-imatinib therapies in DFSP.

Methods

We retrospectively reviewed cases from a prospectively collated database of patients who had presented to our medical oncology clinic between January 2016 and January 2020, with follow up till July 2020. We included all cases of metastatic or locally advanced DFSP. Cases where surgery was not feasible due to the site, extent of disease or risk of functional impairment, were considered to be locally advanced. All cases were discussed in a multidisciplinary clinic. The dose of imatinib was not standardized and depended on the physician’s discretion, clinical response and the tolerance of the patient.

Data including age, sex, primary site, histopathology, metastatic lesions, prior surgeries, prior radiotherapy, dose of imatinib, response rate, post-imatinib therapies and outcomes were extracted from hospital records. Data analysis was done using the IBM SPSS (Statistical Package for the Social Sciences) software for Windows, version 26.0.0 (IBM Corp., NY, USA). Baseline characteristics were assessed using descriptive statistics. Nominal data are presented as percentages and continuous data as median (range). Progression-free survival (PFS) was defined as time from the date of the first clinical visit at our center to the date of documented progressive disease or death from any cause. PFS was calculated using the Kaplan–Meier method. For univariate analysis, comparisons were made using the log-rank test.

Results

Patients

We identified a total of 14 patients who fit the inclusion criteria, with a median age of 39 years (range: 19–60 years). All cases were diagnosed based on histopathology and immunohistochemistry alone, by experts at a tertiary-level hospital (Figure 1). Molecular confirmation of the diagnosis was precluded by the unavailability of the tests in India. Patient characteristics are detailed in Table 1. Most of the patients (78.6%) were males. Ten of the 14 cases (71.4%) had sarcomatous transformation. The most common primary site was the trunk (12/14, 85.7%). One patient had a tumor involving the forehead and the adjoining medial orbital wall, and one had a tumor on the nape of the neck. Eleven cases had metastatic disease and the remaining three were locally advanced/unresectable. Among the metastatic cases, the median number of metastatic sites was 2 (range: 1–6), implying a relatively high burden of disease. The lungs were the most common site of metastases, being involved in nine of the 11 metastatic cases (81.8%). In the two cases with isolated lung metastases, the metastatic lesions were confirmed by biopsy. In the remaining patients, the timing of development of the lung lesions/their synchronicity with the appearance of lesions at other metastatic sites, and their response to imatinib was consistent with a clinical diagnosis of metastatic DFSP. Bone metastases were seen in four patients (36.3%) and soft-tissue metastases in three (27.3%). The median time from baseline diagnosis to development of metastasis was 45 months (range: 17–204 months). Our patients had undergone a median of three surgeries before being registered at our center (range: 0–5 surgeries) and half the cases (50%) had received radiotherapy previously.

Treatment outcomes

Three patients in our series had locally advanced disease and had excellent responses to imatinib. Patient 1 had presented to us with soft-tissue swellings over his forehead and the medial wall of the right orbit. He was started on imatinib at a dose of 400 mg once daily (OD) and achieved maximal response at 3 months of therapy with near complete resolution of the disease. A further increase in dose to 800 mg OD did not add to the response. The response achieved in this case with imatinib alone made surgery unnecessary.

Patient 8 had presented with an inoperable, recurrent lesion over her lower back, with a size of 15 × 15 cm. She was started on imatinib 400 mg OD, with which she had partial response after four months of therapy. The dose was then increased to 800 mg OD and after eight months of therapy, she was able to undergo WLE with split-skin grafting (SSG), followed by postoperative RT (PORT) at a dose of 60 Gy in 30 fractions over six weeks. She remains disease free 22 months after the surgery and is on regular follow-up. Patient 9 presented with a seven-year history of a pigmented, violaceous, plaque-like lesion over the left anterior chest wall that had gradually progressed to a size of 7 × 8 cm. He was given imatinib 400 mg OD for four months, with which he had partial response. He then underwent WLE with negative margins, followed by SSG and PORT. He has been on observation for 22 months since without disease recurrence.

Among the metastatic cases, one patient with a history of multiple recurrences, presented with two pulmonary metastases and no disease at the primary site. He successfully underwent metastatectomy with R0 resection and remains disease free 9 months later without any adjuvant therapy. Among the remaining ten metastatic cases, seven patients initially had at least a partial response (PR) with imatinib, one had stable disease (SD) and two experienced disease progression (PD) while on imatinib. Thus, the initial response rate to first-line imatinib was 76.9% (10/13) and overall disease control was achieved with imatinib in 84.6% (11/13). The median time from baseline diagnosis to the development of metastatic disease was 45 months (range: 17–204 months).

Patient 6 had developed metastatic disease within 3 years of initial presentation, and had received three cycles of ifosfamide-etoposide (IE) and vincristine-adriamycin-cyclophosphamide (VAC) at another center. He then had disease recurrence and lung metastases within two years, and underwent WLE followed by PORT (60 Gy in 30 fractions over 6 weeks), followed by pazopanib 400 mg OD, to which his best response was SD. The disease progressed after three years on pazopanib, at which time he presented to us. He was then started on imatinib 400 mg OD and had partial response. His disease progressed after 17 months on imatinib (Figure 2), at which point he was put back on pazopanib. His disease has now been stable on pazopanib for the past 42 months. Patient 5 (Figure 3) remains progression-free on imatinib at 53 months of therapy.

As shown in Figure 4A, the overall median PFS was 15 months (95% CI: 5.9–24.1 months) at a median follow-up of 17.5 months (range: 1–59). As shown in Table 2, of the eight cases who had progressive disease on imatinib, three were not fit for further therapy and three progressed within a month of second-line therapy: one each on dacarbazine, pazopanib and doxorubicin. Patient 2 received one cycle gemcitabine as third-line therapy, but the disease continued to progress, and he was only fit for best supportive care (BSC). Patient 4 progressed after two cycles of dacarbazine in the third line, and was re-exposed to imatinib at a dose of 800 mg. The disease remained stable on imatinib for eight months, after which the patient had PD and was advised BSC.

As shown in Figure 4B & C, on univariate analysis, cases with nonmetastatic disease (Chi-square: 4.67; p = 0.031) or DFSP pathology (Chi-square: 5.61; p = 0.018) were associated with a more favorable PFS. The median PFS on imatinib for the metastatic cases was 11 months (95% CI: 2.6–19.4 months) and the median PFS for the DFSP-FS cases was 7 months (95% CI: 1.0–15.3 months). There was no significant association between PFS and either imatinib dose (Chi-square: 3.145; p = 0.076) or gender (Chi-square: 0.097; p = 0.76).

Discussion

The median age of patients in our study was 39 years (range: 19–60 years) compared with a median age of 47.4 years among the 24 patients in the pooled analysis of trials in advanced DFSP patients from the EORTC/SWOG groups [18]. The younger age at the onset of inoperable disease may be simply due to the younger demographic of the Indian population. In one series of 214 patients by David et al., the mean age at which patients first noticed the lesions of DFSP was 29.6 years and the median diagnostic delay was 4 years [20]. The majority of the patients in our study were male (78.6%), compared with 58.3% in the EORTC/SWOG pooled analysis and 50% in the B5222 study [18,21].

FISH or PCR for COL1A1-PDGFB can be used for confirming a diagnosis of DFSP. Of the 50 patients with DFSP enrolled in the GENSARC study, molecular testing lead to a revised diagnosis in eight patients (16%) [22]. In the same study, three cases that had initially been diagnosed as being benign and two others with sarcomas with a different initial diagnosis were found to be DFSP on molecular analysis, with significant effects on disease management. Thus, molecular confirmation of the diagnosis in all our cases would have been ideal, but was not available. However, the good initial responses, we had in most of these cases attests to the validity of the diagnosis.

The overall incidence of fibrosarcomatous transformation in DFSP is reported to be 7–16% [1,23]. The incidence of fibrosarcomatous transformation in advanced/metastatic DFSP in the EORTC/SWOG and B2225 studies was 43–52%. In our series, ten of the 11 metastatic cases had DFSP-FS (90.9%). This may be explained by the high number of prior recurrences among the patients in our series. In a series of DFSP-FS cases by Mentzel et al., the local recurrence rate was 58%, and 14.7% of cases developed metastatic disease [24]. The series of patients presented in this study include only those with advanced or metastatic disease. Metastatic disease is usually found in patients with a history of repeated surgeries. Among the DFSP-FS cases in our series, the median number of prior surgeries was three (range: 1–5). Seven of the 11 metastatic cases (63.6%) had also received radiotherapy for earlier recurrences. The high proportion of DFSP-FS among our metastatic cases has also been observed in other studies [25]. Fibrosarcomatous transformation is believed to contribute to an increased risk of recurrence and distant metastasis, and is considered by some to be a form of tumor progression [26]. The median PFS for the DFSP-FS cases was only 7 months (95% CI: 1.0–15.3 months), compared with the overall median PFS of 15 months.

Neoadjuvant imatinib is used at some centers to reduce tumor size and improve surgical outcomes, especially when the cosmetic results are important. It has had a response rate of 45–57% in Phase II trials [18,27]. Neoadjuvant doses of 400, 600 and 800 mg have been reported to have similar efficacy and tolerance profiles [18,27,28]. All three cases in our series with unresectable, advanced disease had good responses to imatinib 400 mg OD and higher doses did not bring increased benefits. The overall partial response rate to first-line imatinib in our series was 76.9% (10/13). The overall response rate to imatinib has been shown to be 46–90% in various studies, with shorter-lasting responses in fibrosarcomatous variants [18,21,29].

Imatinib has been shown to be effective in metastatic DFSP and is used as the first-line therapy. Among the ten metastatic cases exposed to imatinib in our series, seven had partial response and one had stable disease. The median PFS on imatinib for the metastatic cases in our series was 11 months (95% CI: 2.6–19.4 months). The overall median PFS was 15 months (95% CI: 5.9–24.1 months) at a median follow up of 17.5 months (range: 1–59 months), compared with a median time-to-progression of 1.7 years in the EOTRTC/SWOG series [18]. The lower PFS may be attributed to the high proportion of FS variants in our series.

The patient with limited metastases to the lungs, who is currently disease free after pulmonary metastatectomy, represents an interesting case study. However, the role of surgery versus continued imatinib therapy in limited metastatic disease is still under review. Perhaps, the absence of fibrosarcomatous transformation in this patient’s histology was also advantageous. Two of our patients had primary resistance to imatinib, and six cases had progressive disease after initial response, indicating secondary resistance. The mechanisms behind the development of this secondary resistance are still uncertain. In our series, six of 13 cases (46.2%) developed secondary resistance to imatinib after a median treatment duration of 6 months. In a series of 22 patients with locally advanced or metastatic DFSP, five of 22 patients (22.7%) developed secondary resistance to imatinib after a median treatment duration of 15 months [30]. Multiple case reports have documented the used of ifosfamide/adriamycin, dacarbazine, gemcitabine-docetaxel, pazopanib, sunitinib, sorafenib, nilotinib and methotrexate postprogression on imatinib, but none of these options have shown clear success. As such, DFSP is thought to be chemoresistant [31,32]. A multicentric Phase II trial of pazopanib in patients with unresectable or metastatic DFSP has reported an objective response rate of 30% [33], but further studies are required.

Conclusion

The necessity of combined tumor boards (including medical, surgical and radiation oncologists) as well as multimodality treatment for the management of advanced DFSPs is clearly illustrated in our study. Our series had a high proportion of patients with sarcomatous transformation, in contrast to their relative rarity in western studies. Our patients had good outcomes with imatinib, and high initial response rate, both in locally advanced and metastatic disease. However, the results were dismal with subsequent therapies after progression on imatinib.

Future perspective

Therapy for unresectable or metastatic carcinomas is currently mostly limited to imatinib alone. Although many drugs have been tried in the second line, there are only anecdotal reports of success with various agents. Prospective clinical trials, preferably with molecular testing, are required to identify better therapeutic options for these patients after progression on imatinib.