The optimal treatment for synovial sarcoma remains controversial. Treatment, outcome, and prognostic factors in patients treated in a single institution were examined.
The optimal treatment for synovial sarcoma remains controversial. Treatment, outcome, and prognostic factors in patients treated in a single institution were examined.
Synovial sarcoma patients who underwent surgery at the Rizzoli Institute between 1976 and 2006 were identified and analyzed.
Characteristics of the 250 patients (128 female; 122 male) included: median age, 37 years (range, 7-83 years); 177 (71%) with tumors in the lower extremity, 40 (16%) with tumors in the upper extremity, and 33 with tumors in the trunk (13%); primary lesion size >5 cm in 121 patients (55%); and 204 (82%) patients with localized disease and 46 (18%) with metastatic disease at the time of presentation. All patients with localized disease underwent surgery. Twenty-four percent of patients underwent amputation. Adequate surgical margins were achieved in 88% patients. In patients with localized disease, radiotherapy was administered to 103 (50%) patients, and chemotherapy to 98 (48%). With a median follow-up of 5.5 years (range, 1-30 years), the 5-year overall survival rate was 10% for patients with metastatic disease and 76% for patients with localized disease (P = .0001). The 5-year event-free survival was 58% in patients with localized disease. Multivariate analysis indicated that size, age, histologic subtype, and the use of radiotherapy were independent factors for event-free survival.
In those patients with localized disease, a good rate of cure can be achieved. Age, size, histology, and use of radiotherapy influence prognosis, whereas to the authors' knowledge, the role of adjuvant chemotherapy remains unproven. Cancer 2009. © 2009 American Cancer Society.
Synovial sarcoma (SS) comprises approximately 8% of all soft tissue sarcomas (STSs), with the lower limbs being the most common site of primary disease.1 Although relatively rare, SS is the third most common extremity STS. It affects mostly young adults, with a median age of 35 years.2
Three histologic subtypes of SS are described: monophasic, entirely composed of spindle cells; biphasic, composed of both spindle cells and epithelial cells; and poorly differentiated subtypes.3 Synovial sarcoma contains a characteristic translocation (X;18)(p11;q11), representing the fusion of SYT on chromosome 18 with either SSX1, SSX2, or rarely SSX4 (all on chromosome X).4 The resulting gene fusions appear to be mutually exclusive and concordant in primary and metastatic tumors.5
Several prognostic factors, such as age, size, surgical margins, histologic grade, histologic subtype, p53 overexpression, Ki-67 proliferative index, and SYT-SSX fusion type, have been identified.6-12 However, the relative prognostic value of each of these factors remains controversial. Only a large tumor size has been consistently associated with a poor prognosis.6-12
The role of adjuvant chemotherapy in SS is debated. In the metastatic setting, a high response rate to ifosfamide-based therapy has been reported (40%-70%).13, 14 Therefore, adjuvant chemotherapy is currently frequently used for localized disease,10, 15 especially in the pediatric population.10 However, definitive conclusions have been difficult to make in the absence of histology-specific chemotherapy protocols.16-18
The standard treatment of primary tumor is the wide surgical removal of the lesion with or without addition of radiotherapy.19 For patients with recurrent disease, no standard treatment strategies have been defined to our knowledge because of the scanty literature on this specific subject. Nonetheless, Karavasilis et al,14 in a series of 488 advanced STS including SS, conclude that palliative chemotherapy should be regarded as a standard treatment option, with approximately half of patients deriving clinical benefit.
The objective of the current study was to retrospectively examine all SS patients treated at our institution to identify tumor-related and treatment-related factors influencing survival.
All patients diagnosed with SS between 1976 and 2006 were identified from the database of the Pathology Department of Musculoskeletal Oncology at the Rizzoli Institute. Clinical charts and pathology reports were examined, and all the data of interest were collected in a database. In the case of histologic discrepancy (ie, different diagnosis between initial and recurrence histology or histologic subtype missing), slides were re-reviewed by our pathologist (M.A.) and, when required, a polymerase chain reaction was performed to confirm the presence of the specific translocations (X;18)(p11;q11). Tumors were classified according to the World Health Organization classification.3 All patients who were surgically treated at our institution, with histologic diagnosis confirmed by our pathologists and who were followed at our institution for a minimum of 1 year, were included in the analysis. Disease stage was considered localized if there was no evidence of metastases when patients were admitted at the Rizzoli Institute.
After 1986, the staging consisted of a computed tomography (CT) scan and/or magnetic resonance imaging of the primary lesion, and a chest CT scan; other specific tests (bone scan, abdominal CT scan) were performed only in the the case of clinical suspicion. Before 1986, a plain chest x-ray and ultrasound of the lesion were performed. Assessment of the surgical margins was based on both the pathology report and the description of the surgical excision.
Patterns of recurrence for localized patients were defined as follows: local recurrence, when tumor recurrence was confined to the primary tumor area; lung metastases, for lung-only metastases; and multiple–site recurrence (local and lung recurrence or other distant recurrence).
The following parameters were examined for prognostic value in patients with localized disease: patient age and sex, tumor anatomic site, tumor size, surgical margins, histologic subtype, surgical procedure, status at hospital admission, and use of adjuvant chemotherapy and adjuvant radiotherapy.
The following categories were compared: patient age (patients were grouped considering the inclusion criteria of the chemotherapy protocols for soft tissue sarcomas activated in Italy: pediatric protocols enrolled patients aged aged <18 years and protocols for adults enrolled patients ages 18 to 65 years; no specific study protocols were activated in Italy for patients aged ≥65 years); tumor site (upper extremity: at or distal to the shoulder joint; lower extremity: in the groin or leg; and trunk: proximal to the shoulder joint and the groin); tumor size (maximum diameter ≤5 cm vs >5 cm); surgical margins (adequate: wide or radical; inadequate: intralesional, marginal, or contaminated margins, according to the classification of Enneking et al20); histologic subtype (biphasic vs monophasic vs poorly differentiated)3; status at admission (new diagnoses: previously untreated lesions; primary re-excisions: lesions excised with inadequate margins ≤3 months earlier; andlocal recurrences: tumor recurrence >3 months after primary tumor excision); and adjuvant treatments (radiotherapy or chemotherapy performed within 3 months after tumor excision).
We analyzed overall survival (OS), event-free survival (EFS), local recurrence–free survival (LRFS), and postrecurrence survival (PRS) times. OS time was calculated from the time of admission at Rizzoli to death or last follow-up visit. EFS time was calculated from the time of admission at Rizzoli to the occurrence of an event. An event was defined as local recurrence, distant recurrence, or death (disease-related or unrelated). LRFS was calculated from the date of surgical complete remission to the occurrence of local failure. PRS was calculated from the time of first recurrence to death or last follow-up visit.
All time-to-event endpoints were modeled using the method of Kaplan and Meier and analyzed by the log-rank test. Multivariate analysis was performed by means of the Cox proportional hazards regression method using a stepwise procedure. All factors with expected clinical value were included in the multivariate analysis regardless of the statistical significance shown after univariate analysis. The results of the Cox model analysis are reported as relative risks (RRs) and 95% confidence intervals (95% CIs).
A total of 294 consecutive patients with a histologic diagnosis of SS made between 1976 and 2006 were identified. Data on treatment and outcome were incomplete in 44 cases, and therefore 250 patients were included in the study (Table 1).
|Metastatic [n=46], No. (%)||Localized [n=204], No. (%)||P|
|Female||18 (39)||110 (54)|
|Male||28 (61)||94 (46)|
|Site of primary tumor||.04|
|Lower extremity||39 (85)||138 (68)|
|Upper extremity||3 (6)||37 (18)|
|Trunk||4 (9)||29 (14)|
|Monophasic||34 (74)||121 (60)|
|Biphasic||10 (22)||74 (36)|
|Poorly differentiated||2 (4)||9 (4)|
|Median (minimum-maximum)||40 (13-79)||36 (7-83)|
|<18||3 (6.5)||21 (10)|
|18-65||37 (80.5)||170 (83)|
|>65||6 (13)||13 (6)|
|>5||31 (82)||90 (49)|
|≤5||7 (18)||92 (51)|
|Pattern of metastases|
|Lymph nodes ± other||9 (20)||NA|
|Bone + lung||1 (2)||NA|
In 46 (18%) patients, evident metastases were present at diagnosis. All patients with metastatic disease received disparate treatments on a clinical basis, and for this reason a specific analysis of the role of chemotherapy and radiotherapy in this subset of patients was not feasible.
The 204 patients with nonmetastatic disease were admitted to the hospital because of local recurrence in 35 (17%) patients and scar re-excision in 64 (31%) patients, whereas 105 cases (52%) were new diagnoses.
All patients with localized disease underwent surgery, whereas no surgical local treatment was performed in 5 (11%) patients with metastatic disease. An amputation was performed in 59 (24%) patients. The majority of these had a tumor located in the distal extremity: foot/ankle (55%) and hand/forearm (15%). The amputation rate in surgically treated patients was 37% (17 of 46) in patients treated before 1986 and 21% (42 of 199) in those treated in the subsequent period. Margins assessment was available in 238 (95%) patients. Adequate surgical margins were reported in 209 of 238 (88%) of patients.
In 8 (4%) patients with localized disease, no detailed information was available regarding the type of chemotherapy used. Four of them received chemotherapy, but the regimen adopted was unknown. In the remaining 4 patients, it was not possible to know whether chemotherapy was added, and for this reason they were excluded from this analysis.
Overall 102 patients received chemotherapy (Table 2). Chemotherapy was based on the combination of ifosfamide (at a dose of 9 g/m2) and doxorubicin (at a dose of 80 mg/m2) or epirubicin (at a dose of 120 mg/m2) in 77 (75%) patients. Single-agent doxorubicin (at a dose of 80 mg/m2) was delivered to 16 (16%) patients. Five (5%) patients received other combinations (vincristine, dactinomycin, and cyclophosphamide in 3 patients; methotrexate, cisplatin, and bleomycin/dactinomycin/cyclophosphamide in 1 patient; ifosfamide, dactinomycin, and dacarbazine in 1 patient). In 4 (4%) patients, the regimen adopted was unknown. In 42 (41%) patients treated in more recent years, chemotherapy was preoperatively delivered. Ninety-eight (48%) patients did not receive chemotherapy.
|Category||Chemotherapy, n=102 (52%)||No Chemotherapy, n=98 (48%)||P|
|<18||10 (48%)||11 (52%)|
|18-65||90 (54%)||78 (46%)|
|>65||2 (18%)||9 (82%)|
|Male||51 (55%)||42 (45%)|
|Female||51 (48%)||56 (52%)|
|Lower extremity||72 (53%)||64 (47%)|
|Upper extremity||13 (37%)||22 (63%)|
|Trunk||17 (59%)||12 (41%)|
|≤5||31 (34%)||60 (66%)|
|>5||63 (71%)||26 (29%)|
|Biphasic||35 (49%)||37 (51%)|
|Monophasic||61 (51%)||58 (49%)|
|Poorly differentiated||6 (67%)||3 (33%)|
|Status at admission||.0004|
|New diagnosis||66 (63%)||38 (37%)|
|Local recurrence||16 (48%)||17 (52%)|
|Scar re-excision||20 (32%)||43 (68%)|
|Amputation||27 (66%)||14 (34%)|
|Excision||75 (47%)||84 (53%)|
Patients with larger tumors received chemotherapy more frequently compared with those with smaller tumors. Chemotherapy was used less frequently in older patients compared with adults and pediatric patients and in the case of scar re-excision (Table 2).
In 7 (3%) patients, we could not retrieve information regarding radiation therapy use, and for this reason they were excluded from this analysis (Table 3). In 103 (52%) patients, radiotherapy was added to surgery as part of primary therapy; brachytherapy was administered to 11 (11%) patients, external-beam radiotherapy was used in 80 (77%) patients, and combined external-beam radiotherapy and brachytherapy were administered to 12 (12%) patients. The external-beam radiotherapy dose was 44 to 66 grays (Gy) (daily fractions of 200 centigrays) in the case of adjuvant treatment.
|Category||Radiotherapy, n=103 (52%)||No Radiotherapy, n=94 (48%)||P|
|<18||5 (24%)||16 (76%)|
|18-65||94 (57%)||71 (43%)|
|>65||4 (36%)||7 (64%)|
|Male||47 (52%)||44 (48%)|
|Female||56 (53%)||50 (47%)|
|Lower extremity||58 (44%)||75 (56%)|
|Upper extremity||22 (59%)||15 (41%)|
|Trunk||23 (85%)||4 (15%)|
|≤5||40 (45%)||49 (55%)|
|>5||55 (62%)||33 (38%)|
|Biphasic||39 (53%)||34 (47%)|
|Monophasic||56 (49%)||59 (51%)|
|Poorly differentiated||8 (89%)||1 (11%)|
|Status at admission||.6|
|New diagnosis||55 (65%)||43 (44%)|
|Local recurrence||17 (49%)||18 (51%)|
|Scar re-excision||31 (48%)||33 (52%)|
Radiotherapy was more often delivered in patients with larger tumors and in patients with tumors located in the trunk. Pediatric and elderly patients received radiotherapy less frequently (Table 3).
With a median follow-up of 5.5 years (range, 1-30 years), the 5-year OS of the entire cohort of patients was 68%: 10% for patients with metastatic disease and 76% for patients with localized disease (P = .0001) (Fig. 1). Four patients died of disease-unrelated causes: leukemia, pulmonary embolism, gastric cancer, and suicide.
Small tumor size, conservative surgery, and radiotherapy use were factors found to be significantly associated with improved survival in localized patients by univariate analysis (Table 4). We did not observe different prognosis according to the use of chemotherapy.
|Variable||5-year EFS||95% CI||P||5-year OS||95% CI||P|
|Status at admission||.7||.2|
Of the 204 patients with localized disease, 116 were continuously disease-free, with a median follow-up of 77 months (range, 11-300 months). The 5-year EFS rate was 58% (95% CI, 51%-66%).
On univariate analysis, only tumor size was found to be a statistically significant factor for EFS. Age and surgical procedure reached levels close to statistical significance. We did not observe different prognosis according to the use of chemotherapy and radiotherapy (Table 4).
A separate analysis by tumor size (>5 cm vs ≤5 cm) was performed. In addition, the use of chemotherapy was not associated with improved EFS in these subsets (Fig. 2). Furthermore, we did not demonstrate statistical differences (P = .25) according to the chemotherapy regimen adopted: 5-year EFS after single-agent chemotherapy was 43% (95% CI, 22%-64%), compared with 52% (95% CI, 40%-64%) after a combination of ifosfamide (at a dose of 9 g/m2) and doxorubicin (at a dose of 80 mg/m2) or epirubicin (at a dose of 120 mg/m2), and 66% (95% CI, 56%-65%) without chemotherapy. The same analysis, but limited to the subgroups of patients with primary tumor size of >5 cm and >10 cm, did not demonstrate statistically significant differences in 5-year EFS. However, we must emphasize that the latter group included only 21 patients. On the contrary, the use of radiotherapy was found to improve the EFS in patients with larger (>5 cm) tumors (Fig. 2).
After multivariate analysis, adult age and larger tumor size were confirmed as negative prognostic factors. Furthermore, monophasic histology and omission of radiotherapy were found be factors associated with worse EFS, whereas no differences where observed according to the use of chemotherapy (Table 5).
Local failure only as first recurrence was documented in 33 (16%) patients. Lung metastases only developed in 46 (22.5%) patients. Nine (4%) patients had multiple sites of recurrence: 4 had local recurrence and pulmonary metastases (with cardiac metastasis in 1 case), 3 had bone metastases (with lymph node metastasis in 1 case and lung metastases in 1 case), 1 had central nervous system metastases, and 1 had hepatic and pulmonary metastases. The median time to recurrence was 18 months (range, 2 months–265 months). It was longer in the case of local recurrence only (28 months; range, 2 months–265 months) compared with lung only (15 months; range, 2.5 months-115 months) and multiple–site recurrence (16 months; range, 5 months-75 months) (P = .03).
Overall, 18% (37 of 204 patients) patients developed a local failure. The incidence of local recurrence was 32% (12 of 38 patients) in patients treated before 1986, and 15% (25 of 166 patients) in patients treated after that date.
The actuarial 5-year and 10-year LRFS was 81% (95% CI, 75%-87.5%) and 77% (95% CI, 70%-85%), respectively.
On univariate analysis, the quality of surgical margins was found to be the only significant factor influencing the local control: adequate surgical margins, 5-year LRFS of 85% (95% CI, 79%-91%); inadequate surgical margins, 5-year LRFS of 59% (95% CI, 37%-81%) (P = .002).
The use of postoperative radiotherapy improved the local control in the case of inadequate margins (5-year LRFS without radiotherapy, 25%; with radiotherapy, 70% [P = .025]). A higher local control rate in patients receiving radiotherapy was also reported in patients with tumors measuring >5-cm, although the difference was not statistically significant (5-year LRFS without radiotherapy, 67%; with radiotherapy, 85% [P = .2]).
Five-year PRS was 67% for patients with only local recurrence, 21% in the case of lung-only metastases, and 0% at 3 years for patients with multiple–site recurrence (Fig. 3). We did not observe a statistically significant difference (P = .15) in PRS according to the recurrence-free interval (RFI), considering a threshold of 2 years; the 5-year PRS was 32% in the case of a RFI ≤2 years and 53% in the case of a RFI >2 years.
STSs are a heterogeneous group of tumors, and there is an urgent need for histology-specific analysis.21 Nonetheless, it is not simple to conduct prospective and randomized studies, because of the rarity of each entity.
In this perspective, a retrospective analysis represents a useful tool with which to define prognostic factors and the best treatment strategy, accepting the drawbacks of this kind of study design. The inclusion of nonextremity disease, patients who underwent noncurative surgery, patients with recurrent/metastatic disease, and patients treated over a long period of time make the results of this and other retrospective STS studies difficult to interpret. The strength of this study is the large number of patients included, all of whom were treated at a single institution, and all with a histologically confirmed diagnosis of SS.
The results of the current study demonstrate that in localized SS, primary tumor size and patient age at diagnosis are important prognostic factors for EFS.
The prognostic role of size is well known for STS,15-18 and was also demonstrated in several SS studies.4, 6-11, 15, 22 Only in the study by Ladanyi et al was size not found to influence OS in the subgroup of localized SS patients, but it did when metastatic patients were included.12 The finding in the current series that larger tumors were more frequently reported in patients with metastatic disease compared with those with localized disease reinforces the significance of tumor size in SS.
Whether age has a prognostic importance is difficult to address, because to our knowledge only a few studies published to date9, 10, 11, 23 have included both adult and pediatric SS patients, and the age threshold of these studies is very diverse. A significantly better survival was previously demonstrated for patients aged <40 years,10 17 years,11 13 years,22 and 35 years23; the latter study included only patients with metastatic disease. Two other studies found age to be a continuous variable associated with survival.9-15 The results of the current study indicate an independent adverse prognostic role of adult age (ages 18-65 years), compared with pediatric age. This is similar to what has been observed for other sarcomas, such as Ewing sarcoma.24, 25
The role of the histologic subtype is not certain; biphasic histology appears to be associated with better survival,4, 6, 10, 12, 22 but to our knowledge only Guillou et al10 demonstrated a statistically significant difference in disease-specific survival and metastases-free survival when they compared patients with and without a poorly differentiated component. In the current study, we observed an RR of 1.68 (95% CI, 1.02-2.76) for increase of recurrence for monophasic histology when compared with the biphasic subtype.
Radiotherapy was added to surgery in approximately half of all localized patients and only in 24% of pediatric patients. In patients with tumors measuring >5 cm, the use of radiotherapy was 62%. The strategy for local treatment in our institute was then characterized by a less frequent use of radiotherapy compared with that reported in other SS studies, in which >95% of patients underwent adjuvant radotherapy.9, 15
In STS, local recurrence is associated with the development of subsequent metastases and decreased survival.26, 27 In the current study, the 5-year LRFS was 81%. In previous retrospective analyses, the LRFS probability ranged from 65% to 89%.7-9, 11, 15 It is interesting to note that the local control rate improved over the years in our series.
In patients with larger tumors, the local control rate was higher (5-year LRFS of 85%) in those who received radiotherapy compared with patients who were treated only surgically (5-year LRFS of 67%), although the difference was not statistically significant, most likely because of the sample size. Similar to the study by Lewis et al,7 the quality of the surgical excision margins was a significant factor influencing local control in our study. Furthermore, the use of postoperative radiotherapy significantly improved local control in the case of inadequate margins. The role of margins on local control was not confirmed in other studies.9, 15 In these studies, >80% of patients underwent adjuvant radiotherapy, compared with <50% of patients in the current study and the study by Lewis et al.7 As a consequence of these data, after multivariate analysis we observed better EFS for patients who underwent adjuvant radiotherapy. These data indicate the importance of the use of radiotherapy in patients with SS. Radiotherapy is of most benefit with larger tumors and in the case of inadequate margins. This is consistent with other studies,22 and in accordance with recent STS recommendations.19, 28
Whether adjuvant anthracycline-based chemotherapy provides a clinically meaningful benefit for STS patients with localized disease is still debated.18 SS has been considered a chemoresponsive tumor in the metastatic3 and adjuvant setting.29 In the current series, 48% of patients with localized disease received adjuvant or neoadjuvant chemotherapy. Chemotherapy was used less frequently in the case of small tumors, but equally in adult and pediatric patients. Most patients received ifosfamide, and the doses of drugs used were high, similar to recent SS studies,15 but low compared with the pivotal study by Kampe et al.29 Unlike other SS studies,10, 15, 29 the current series did not demonstrate better survival for patients with localized disease undergoing chemotherapy, including when the analysis was limited to larger tumors. However, because of the possible selection bias in this retrospective study, we are not able to comment on the benefit of chemotherapy for patients with localized SS. Other series differ from ours because of the inclusion of more patients with inadequate surgical margins,11 younger age of patients receiving chemotherapy,11 wider use of radiotherapy,14 and higher dose of ifosfamide.28 The small number of patients considered in each study and the retrospective nature of all of them make it difficult to draw any conclusions.
Local recurrence appears to be a later event than distant recurrence (median time to recurrence, 28 months vs 15 months, respectively); this is similar to findings reported by Spillane et al (23 months vs 11 months, respectively)8 and in contrast to other series.9 Overall, the median time to recurrence was 18 months (range, 2 months-265 months). This wide range, with some of the patients developing disease recurrence after 20 years, confirms other SS studies,6, 8 and strongly supports the need for long follow-up in this subset of sarcoma patients.
The 5-year PRS was 67% after local recurrence. This suggests a different biologic behavior compared with bone sarcoma, in which ≤20% of patients are reported to survive after local recurrence.30 The 5-year PRS after lung metastases was 21% in the current series, only 6% in a previous SS study,8 and 35% in an STS study31 that included only patients who were tumor free after surgery. The latter study also reported a worse PRS for patients, with a RFI <2.5 years.31 We observed a worse PRS for patients who developed disease recurrence within 2 years compared with later recurrences, although this difference was not significant.
In conclusion, using a large monoinstitutional series of SS patients, we have demonstrated that adult age, larger tumor size, monophasic histology, and the omission of radiotherapy are independently associated with decreased EFS. Radiotherapy offers the most benefit in the case of larger tumors and inadequate margins.
Prospective studies are warranted to confirm these findings and to fully address the role of chemotherapy.
The authors made no disclosures.