Outcome after surgical treatment of dermatofibrosarcoma protuberans: Is clinical follow-up always indicated?
Abstract
Background
Dermatofibrosarcoma protuberans (DFSP) is a rare soft tissue sarcoma for which clinical examination up to 10 years is recommended. The objective of this study was to identify prognostic factors for recurrences and metastases that can be used to evaluate the validity of follow-up schedules after treatment for DFSP.
Methods
Patients with DFSP who received treatment between 1991 and 2016 at 3 tertiary centers were included. Cox regression analyses were conducted to identify variables associated with the primary endpoints.
Results
In total 357 patients were included, with a median age of 38 years (age range, 2-87 years) and a median follow-up of 60 months (interquartile range, 24-115 months). Eighty-one patients developed recurrent disease (22.7%), and the median time to recurrence was 55.5 months (interquartile range, 20-90 months). Of these, 50 tumors (61.7%) were identified by patient self-examination, whereas 3 recurrences (3.7%) were identified at clinical surveillance. For the remaining 28 tumors, no information was available on how the recurrences were identified (34.6%). Fibrosarcomatous change (hazard ratio, 21.865; P < .001), and positive resection margins (hazard ratio, 14.645; P < .001), were independent prognostic factors for recurrence. Metastases occurred in 4 patients (1.1%). All tumors were identified by imaging after patients presented with symptomatic metastases. Fibrosarcomatous change (P < .001) and tumor size >5 cm (P = .014) were associated with the development of metastases.
Conclusions
Disease recurrence after resection of DFSP remains a significant issue, whereas metastases are uncommon. The majority of recurrences are identified by patient self-examination. Consideration should be given to individualized follow-up schedules based on risk factors for recurrences and metastases.
Introduction
Dermatofibrosarcoma protuberans (DFSP) is a superficial soft tissue sarcoma with an incidence of 4 in 1,000,000.1, 2 DFSP often presents as a slow-growing, nodular, dermal lesion, with unexpectedly wide, impalpable, subcutaneous tumor extensions. It is known for its locally aggressive growth pattern and high local recurrence rates after incomplete resection (range, 2%-21%), but it has a low potential to metastasize (range, 1%-4%).3 DFSP is most frequently identified in men and women between ages 20 and 40 years.1, 2 The mainstay of treatment is wide local excision, preferably with a surgical margin of at least 3 cm.4-7 In 3% to 20% of patients with DFSP, fibrosarcomatous (FS) change is observed on histopathology.3 DFSPs with FS change are considered to be even more aggressive, with a higher rate of local metastasis and a particularly high rate of distant metastasis.3 Because DFSP is a rare disease, larger studies examining outcomes and prognostic factors are scarce. The available studies have focused on epidemiology, survival, factors related to survival, and treatment modalities.3, 8-11 However, few studies have evaluated the effectiveness and usefulness of clinical follow-up of these patients after initial treatment. Currently, follow-up schedules vary widely among different countries and institutions. Most follow-up recommendations are based on those for more common sarcomas, although DFSP is a distinct entity with different biologic behavior.11
European guidelines recommend clinical examination every 6 months for the first 5 years, and at yearly intervals thereafter for up to 10 years. In addition, imaging should be preserved for recurrent DFSP or DFSP with FS change. However, no clarification is provided on the interval or type of imaging.4 The National Comprehensive Cancer Network guideline on DFSP recommends follow-up visits every 6 to 12 months, with regular imaging for patients who have high-risk features.6 Unfortunately, this guideline does not specify these high-risk features, and it does not state the length of follow-up. The objective of follow-up in general is to identify recurrences or metastases at an early stage to improve the final outcome, although it has never been proven that follow-up visits detect these events earlier. Also, there is no evidence that earlier detection leads to a better outcome.
The objectives of the current study were to identify prognostic factors for local and distant recurrences and to evaluate the current follow-up strategy. The subsequent goal was to design a more personalized and efficient follow-up schedule based on these data.
Materials and Methods
Patients
All patients with pathologically confirmed DFSP who presented to 1 of 3 tertiary referral centers for treatment or a second opinion between 1991 and 2016 were included. The 3 tertiary centers included were the Royal Marsden Hospital (London, UK), the Netherlands Cancer Institute Antoni van Leeuwenhoek (Amsterdam, the Netherlands), and Erasmus Medical Center-Cancer Institute (Rotterdam, the Netherlands). Whenever feasible, surgical radial margins of at least 3 cm were attained according to a European-based guideline, the National Comprehensive Cancer Network guideline, and the World Health Organization classification of soft tissue tumors.4, 7 Underlying fascia was always resected en bloc with the tumor whenever feasible to achieve better deep margins. In all 3 centers, a sarcoma pathologist performed the pathologic assessment. Patient characteristics and clinical data were obtained either from a prospective maintained database or from patient files, including data regarding the identification of recurrences or metastases. The age of each patient was noted as the age at first presentation either in the tertiary center or in another hospital. A resection was defined as R0 resection when pathologically clear resection margins measuring >1 mm were obtained, as determined by expert sarcoma pathologists. Resection margins <1 mm or the microscopic presence of DFSP in the inked resection margin were defined as R1 resections.
Follow-Up
The length of follow-up was measured from the date of first presentation to the date of the last clinic visit or telephone call. Follow-up visits were stratified into 4 categories: clinical examination, clinical examination with routine chest x-rays, clinical examination with other imaging (ultrasound or magnetic resonance imaging studies of the local tumor after reconstruction), or unknown follow-up routine. The follow-up was unknown in patients who had no documentation on the method of follow-up, or if the follow-up was done in another hospital. Follow-up usually consisted of clinical visits every 3 months shortly after resection, followed by clinical visits every 6 to 12 months after 1 or 2 years.
Statistical Analysis
All data are noted as median values with ranges. For normally distributed, continuous variables, the Student t test was used to compare means; and, for nonparametric data, the Mann-Whitney U test was used. For categorical or ordinal data, the chi-square test or the Fisher exact test was used to compare means, as appropriate. Survival analyses were conducted using the Kaplan-Meier method, in which statistically significance was tested using the log-rank test. To analyze covariates, univariate and multivariate Cox regression were used to calculate hazard ratios and 95% confidence intervals. All variables in the univariate analysis with a P value < .10 were included in the multivariate analysis. P values ≤ .05 were considered significant. IBM SPSS Statistics (version 24; IBM Corporation, Armonk, NY) was used for statistical analyses.
Results
Patient Characteristics
In total, 357 patients were identified. The median patient age was 38 years (range, 2-87 years), with a median follow-up of 60 months (interquartile range [IQR], 24-115 months). Although most patients were referred for primary tumors, 62 patients presented with recurrent DFSP at the time of referral to the tertiary centers (17.4%) (Table 1). Of all patients, 79 were referred for the primary excision of DFSP (22.1%), 155 were referred for re-excision of primary DFSP (43.4%), 41 were referred for local recurrence (11.5%), 21 were referred for re-excision of a local recurrence (5.9%), and 61 were referred for a second opinion (12%). The patient characteristics are listed in Table 1. Most patients underwent wide local excision, although a small subset of patients (n = 41; 11.5%) underwent Mohs micrographic surgery. The median clinical tumor size was 3 cm (range, 0.3-15 cm). In total, 41 patients had histopathologic FS changes (11.4%). Patients with FS changes had significantly larger tumors (P = .015) and more often received radiotherapy as a treatment modality (P = .013) (Table 1). The different follow-up methods are depicted in Figure 1A. Most patients were followed with clinical examination and routine chest x-rays (n = 144) or with clinical examination only (n = 130). For 69 patients, the method of follow-up was unknown (19.2%). The 5-year and 10-year overall survival rates for patients with DFSP were 100% and 95%, respectively, and the 10-year disease-specific survival rate was 99%.
| Characteristic | No. of Patients (%) | |||
|---|---|---|---|---|
| Total | FS changesa | P | ||
| No, n = 305 | Yes, n = 40 | |||
| Sex | .082b | |||
| Male | 178 (49.9) | 146 (47.9) | 25 (62.5) | |
| Female | 179 (50.1) | 159 (52.1) | 15 (37.5) | |
| Age at diagnosis: Median [range], y | 38 [2-87] | 38 [2-87] | 39 [18-77] | .512c |
| Disease stage at presentationd | .199b | |||
| Primary DFSP | 262 (98.9) | 231 (79.9) | 26 (63.4) | |
| Recurrent DFSP | 70 (21.1) | 58 (20.1 | 11(29.7) | |
| Follow-up: Median [range], mo | 60 [0-590] | 59 [0-493] | 60 [1-320] | .982e |
| Tumor size | .015b | |||
| ≤5 cm | 134 (37.5) | 130 (42.6) | 13 (32.5) | |
| >5 cm | 49 (13.7) | 38 (12.5) | 11 (27.5) | |
| Unknown | 174(48.7) | 137 (44.9) | 16 (40) | |
| Disease site | .454f | |||
| Head and neck | 38 (10.6) | 34 (11.1) | 1 (2.5) | |
| Shoulder | 51 (14.3) | 41 (13.4) | 6 (15) | |
| Arm | 24 (6.8) | 21 (6.2) | 3 (7.5) | |
| Torso | 103 (28.8) | 131 | 17 (42.5) | |
| Hip/groin | 34 (9.6) | 22 (7.2) | 4 (10) | |
| Leg | 44 (12.3) | 43 (14.1) | 8 (20) | |
| Unknown | 1 (0.002) | 0 (0) | 1 (2.5) | |
| Resection margin status | .802b | |||
| R0 | 234 (65.5) | 203 (66.6) | 26 (65) | |
| R1 | 40 (11.2) | 36 (11.8) | 4 (10) | |
| Unknown | 83 (23.2) | 66 (21.6) | 10 (25) | |
| Radiotherapy | .013f | |||
| Primary | 11 (3.1) | 9 (3) | 2 (5) | |
| Recurrence | 10 (2.8) | 3 (1) | 7 (17.5) | |
| Metastases | 2 (0.6) | 0 (0) | 2 (5) | |
| None | 334 (93.6) | 293 (96.1) | 29 (72.5) | |
- Abbreviations: DFSP, dermatofibrosarcoma protuberans; FS changes, fibrosarcomatous changes.
- a In 12 patients, FS change status was unknown.
- b This P value was determined with the chi-square test.
- c This P value was determined with the t test.
- d In 25 patients, disease stage at presentation was unknown.
- e This P value was determined with the Mann-Whitney U test.
- f This P value was determined with the Fisher exact test.
Recurrence
In total, 81 patients developed a local recurrence (22.7%), and the median time to recurrence was 55.5 months (IQR, 24-115 months). In most patients (n = 50), recurrences were identified by patient self-examination (61.7%), which resulted in scheduling a clinic appointment. Only in 3 patients (3.7%) the recurrence was identified during routine follow-up visits, although, in 28 patients (34.6%), data were lacking about how the recurrence was identified. It is noteworthy that all DFSPs identified on clinical examination at routine follow-up visits were in difficult anatomic sites for self-examination (ie, buttock, shoulder, and vulva). In 1 patient, a tiny nodule at the posterior limit of the scar was identified during clinical examination; in the second patient, an asymptomatic nodular mass was observed; and the last patient had multiple recurrences, of which 1 was identified during clinical examination. The 2-year and 5-year recurrence-free survival (RFS) rates for the whole group were 94% and 82%, respectively, and the median RFS was 156 months (95% confidence interval, 122-190 months). No difference in recurrence rates were observed between patients who underwent Mohs micrographic surgery and those who underwent wide local excision (P = .519). RFS was significantly prolonged in patients without FS changes (P = .025), in those who achieved R0 resections (P < .001), and in those who had tumors ≤5 cm (P = .012) (Fig. 2A,B). Univariate and multivariate analyses of the factors associated with recurrence are presented in Table 2. Only FS change (P < .001) and R1 resection (P < .001) were correlated with higher recurrence rates in multivariate analysis. Patients who had the combination of tumor size ≤5 cm, R0 resection, and no FS change did not develop any recurrences in this cohort (n = 107) (Fig. 2 C).
| Variable | Univariate Analysis | Multivariate Analysis | ||||
|---|---|---|---|---|---|---|
| HR | 95% CI | P | HR | 95% CI | P | |
| Sex | ||||||
| Male | 1.54 | 0.98-2.41 | .061 | 0.62 | 0.18-2.08 | .436 |
| Female | Ref | Ref | ||||
| Age in years at diagnosis | 0.99 | 0.97-0.01 | .179 | |||
| Tumor size >5 cm | ||||||
| Yes | 2.23 | 1.05-4.72 | .037 | 1.38 | 0.42-4.55 | .601 |
| No | Ref | Ref | ||||
| Disease site | .281 | |||||
| Head and neck | 0.59 | 0.27-1.30 | .189 | |||
| Extremity | Ref | |||||
| Trunk | 0.99 | 0.57-1.72 | .960 | |||
| Resection for primary DFSP | 0.17 | 0.09-0.32 | < .001 | 0.35 | 0.09-1.33 | .123 |
| Resection margin status | ||||||
| R0 | 0.11 | 0.05-0.25 | < .001 | 0.03 | 0.01-0.19 | < .001 |
| R1 | Ref | Ref | ||||
| Radiotherapy for primary DFSP? | ||||||
| Yes | Ref | |||||
| No | 1.11 | 0.35-3.59 | .856 | |||
| FS changes | ||||||
| Yes | 1.97 | 1.08-3.61 | .028 | 21.87 | 3.06-156.2 | < .001 |
| No | Ref | Ref | ||||
- Abbreviations: CI, confidence interval; HR, hazard ratio; FS changes, fibrosarcomatous changes; Ref, reference category.
Of all patients who developed recurrent disease, in total, 60 had only 1 recurrence (74.1%), and 21 had ≥2 recurrences (25.9%). All recurrences were treated successfully with resection. In all, 6 patients received neoadjuvant imatinib, and 9 received postoperative radiotherapy.
Metastasis
Only 4 patients developed distant metastases (1.1%), and the median time to developing metastases was 68 months (IQR, 41-111 months). Of these 4 patients, 3 had FS changes; and, for 1 patient, the presence or absence of FS changes was unknown. All metastases were identified on imaging studies after patients presented with symptoms. None of the metastases were identified during routine clinic visits or follow-up chest x-rays. The characteristics of these 4 patients and how their metastases were identified are detailed in Table 3. Factors that were correlated significantly with developing metastases were FS change (P < .001) and size >5 cm (P = .008). Because of the small number of patients who had metastases, a multivariate analysis on this issue would be unreliable and thus was not performed.
| Sex | Age, y | Tumor Location (Size, cm) | Metastases Location | Time to Metastases, mo | Finding of Metastases | FS Changes | Status | FU, mo |
|---|---|---|---|---|---|---|---|---|
| Woman | 45 | Foot | Lung, pancreas, liver, bone | 85 | Jaundice | Yes | Dead | 109 |
| Woman | 29 | Shoulder (7) | Lung | 50 | Unknowna | Unknowna | Alive | 194 |
| Man | 43 | Foot (11) | Lung | 38 | Unrelated bowel obstruction | Yes | Alive | 55 |
| Man | 40 | Chest (15) | Lung, liver | 120 | Chest pain | Yes | Dead | 123 |
- Abbreviations: FS changes, fibrosarcomatous changes; FU, follow-up.
- a How metastases were identified and whether the dermatofibrosarcoma protuberans had fibrosarcomatous changes were unknown.
Discussion
The objectives of the current multicenter study of 357 patients were to identify risk factors for local recurrence and metastases of DFSP after surgical resection and to evaluate the usefulness of clinical follow-up after initial treatment. The results indicate a recurrence rate of 22.7% and a metastasis rate of 1.1% after initial treatment. Most local recurrences were identified by patient self-examination between clinical follow-up visits, and not on clinical examination by physicians. It is noteworthy that none of the patients without any of the identified risk factors (ie, histologic FS change, R1 resection, and tumors >5 cm) developed local recurrences. Not a single metastasis was identified during clinical follow-up or on routine chest x-ray screening. The few metastases that were identified in this study were symptomatic and thus were presented by the patients themselves. To the best of our knowledge, this is the first study evaluating the usefulness of clinical follow-up for patients with DFSP.
The risk factors for recurrences identified in the current study with multivariate analysis (FS change and R1 resection) are in line with previously published data.8, 10, 12-18 The median time to recurrence in previous studies ranged from 18.5 to 40.6 months10, 16, 19-21; whereas, in our study, the median time to recurrence was 55.5 months. This demonstrates that recurrences can occur even many years after the primary tumor, possibly leading to an underestimation of the median time to recurrence in studies with shorter follow-up.
In this study, FS change was identified as an independent prognostic factor for recurrence, as also described in most published reports.3, 22, 23 Only Fields and colleagues did not identify FS change as a factor that influenced recurrence.24
Patients who had tumors >5 cm were significantly more prone to develop recurrences compared with those who had smaller tumors in the univariate analysis, but not in the multivariate analysis. The prognostic value of tumor size also is debated in other reports.9, 10, 24 Criscito et al. reported that survival was influenced by tumor size, but they did not analyze the relation between recurrence and size. Other studies did not report a significant influence of tumor size on recurrences.10, 24
In the current study, only 4 patients (1.1%) developed distant metastases, indicating that larger DFSPs and those with FS changes are more likely to develop metastases versus smaller DFSPs and those without FS changes. The risk of metastases in patients who have DFSP with FS changes has been described previously, but series on this subject are scarce and consist of limited numbers of patients or case reports.3, 10, 15, 25, 26 Two reviews reported metastases rates of 13.4% and 14.4% in patients with FS changes and 1.1% in patients without FS changes.3, 25 One of those reviews indicated that 83.4% of metastases were preceded by a recurrence.25 Three smaller studies reported metastases rates ranging from 0% to 7.4%; and, in 1 study, metastases were reported only in tumors >5 cm.10, 15, 26
No metastases were identified in patients without FS changes in this series. However, metastases of DFSPs without FS changes have been previously described.3, 26 Metastases most often are located in the lung but, in this cohort, also were observed in other locations, such as the pancreas, liver, and bone.3, 10, 25, 26A review by Liang et al indicated that metastases are located mostly in the lungs but also in bone, soft tissue, liver, kidney, the gastrointestinal system, and lymph nodes.3 All metastases in this series were identified by symptomatic presentation and not by routine imagining, which is in line with the report by Kim et al, who also observed that all metastases were symptomatic.10
Apart from the retrospective nature, the main limitation of this study is that all patients were referred to tertiary referral centers, which resulted in a heterogeneous patient population. Of all patients, 62 already had recurrent disease at the time of presentation (17.4%). This heterogeneous population may explain the recurrence rate of 22.7%, which is higher than that in the recently published literature on DFSP.3-5, 27 In addition, the study population probably had larger and more aggressive tumors, or they may have been an over-representation of patients who initially underwent resection with positive margins in primary centers. Investigating the more aggressive variant of DFSP (DFSP with FS change), this study identified FS changes in 40 DFSPs (11.2%), comparable to the 3% to 22% described in literature.3, 12, 18, 19, 26 This suggests that, in this study population, DFSP with FS change was not over-represented. Conclusions related to metastases in this setting should be interpreted with caution because, only 4 patients developed distant metastases.
The current study is unique in assessing the utility of clinical follow-up appointments and imaging for identifying DFSP recurrences. In total, only 3 recurrences were identified on clinical examination. None of the routine chest x-rays diagnosed metastases. The role of magnetic resonance imaging scans to follow the local resection region remains controversial; and, in this study, no recurrences were identified on imaging studies of the local region, even after plastic reconstructions. With the identification of clear risk factors in this study, it seems reasonable to further personalize follow-up strategies for patients who have DFSP based on the results.
Conclusion
The current results indicate that an adjusted follow-up regime after surgical treatment for DFSP (Fig. 1B) could be considered. Given the very low risk of recurrences in patients who are without risk factors for local recurrences (ie, tumors ≤5 cm, R0 resection, and no FS change), these patients could be given clear instructions for self-examination, with advice to revisit the clinic only when they suspect a potential recurrence. However, in patients for whom self-examination is not considered possible given anatomic location, or previous reconstruction, or when patients are not considered capable of self-examination, routine clinical examination during clinical visits will still be useful. Patients who have DFSP with high-risk features could be followed for 10 years given the high median time to local recurrence or metastases. The ideal frequency of clinical visits remains unclear. Routine chest imaging during follow-up does not seem to be useful in patients with DFSP, except for those who have FS changes, given their relatively high risk of metastases.
Funding Support
No specific funding was disclosed.
Conflict of Interest Disclosures
Alexander C. J. van Akkooi reports grants from Amgen and Novartis and personal fees from Amgen, Bristol-Meyers Squibb, Novartis, MSD-Merck, and Merck-Pfizer, outside the submitted work. The remaining authors made no disclosures.
Author Contributions
Eva A. Huis in ‘t Veld: Conceptualization, formal analysis, methodology, visualization, validation, writing–original draft, and writing–review and editing. Frits van Coevorden: Conceptualization and writing–review and editing. Dirk J. Grünhagen: Writing–review and editing. Myles J. Smith: Writing–review and editing. Alexander C. J. van Akkooi: Writing–review and editing. Michel W. J. M. Wouters: Writing–review and editing. Andrew J. Hayes: Writing–review and editing. Cornelis Verhoef: Writing–review and editing. Dirk C. Strauss: Conceptualization, visualization, and writing–review and editing. Winan J. van Houdt: Conceptualization, supervision, methodology, validation, visualization, writing–review and editing.
