Soft tissue sarcoma is a rare tumor that represents less than 1% of all malignancies ¹ . Complete surgical resection with adequate margin is the mainstay of treatment for patients with this condition. The prognosis of patients with soft tissue sarcoma varies according to histologic subtype, tumor size, tumor location, margin status, and tumor depth ^{2– 6} . Given the heterogeneity of soft tissue sarcomas, it is difficult to determine the value of prognostic factors and the survival outcomes among the different histological subtypes. In term of adjuvant treatment, the impact of adjuvant chemotherapy treatment on survival in soft tissue sarcoma has been investigated in several phase 3 randomized controlled trials ^{7– 10} Although those studies found adjuvant chemotherapy effective for improving outcome in soft tissue sarcoma, the efficacy of this adjuvant treatment continues to be debated. Data from meta-analyses and large randomized trials reported by the European sarcoma group indicated postoperative chemotherapy improved relapse-free survival (RFS) in patients with extremity soft tissue sarcoma, but data regarding improvement in OS were conflicting ^{7– 14}

The aim of this study was to investigate the clinical characteristics, prognostic factors, survival outcomes, and outcomes of treatment in extremity soft tissue sarcoma patients who underwent complete tumor resection at Siriraj Hospital.

The aim of this study was to investigate the clinical characteristics, prognostic factors, survival outcomes, and outcomes of treatment in extremity soft tissue sarcoma patients who underwent complete tumor. We included only patients with high-risk extremity soft tissue sarcoma in order to minimize the heterogeneity of the study population.

Regarding baseline characteristics, the median age of 53.5 years in our study was much higher than the median age of 46 years reported in the European Organization for Research and Treatment of Cancer (EORTC) and Italian Sarcoma Group (ISG) trials ^{9, 13} , but similar to a study by Liu et al. from Taiwan ¹⁸ . The median tumor size of 9 cm found in this study was slightly smaller than the median tumor size of 10 cm reported from the ISG trial ⁹ . However, the median tumor size of 9 cm in both the chemotherapy and non-chemotherapy groups in this study were larger than the median tumor size of 8.6 cm and 7.5 cm in the chemotherapy and non-chemotherapy groups in the EORTC trial, respectively ¹² . The predominant histological subtype in this study was UPS (26/58, 45%), which is higher than the 29% and 18% rates reported from the ISG trial ⁹ , EORTC trial ¹³ , respectively, but similar to 27% rate reported by Roger et al. ¹⁹ Synovial sarcoma in this study was 8.9% (5/58), which is considerably lower than the 25% rate in the ISG and the 14% rate in the EORTC trial ^{9, 13} . A total of 87% (50/58) of patients had negative surgical margin, which was higher than 37%, 38% and 37% reported in the EORTC, Taiwan and Korean studies, respectively ^{13, 18, 20} .

For tumor grading, subjects in our study had more tumor grade 1 than patients in the ISG trial (7/58; 12% vs 0%) ⁹ . In our analysis, we also attempted to identify baseline characteristics that portend an unfavorable prognosis. In previous studies, tumor size and grading were reported to be prognostic markers, especially if the tumor was larger than 10 cm or larger than 5 cm and of a high grade ^{2– 6} . This study revealed that tumor grade 3 >10 cm was associated with significantly shorter DFS. Shorter duration from curative surgery to commencement of chemotherapy administration was found to be a significant favorable prognostic factor for DFS ( p=0.02).

Moreover, the median time from surgery to first chemotherapy administration in non-relapse patients was nearly significantly shorter than in relapse patients (28 vs. 45.5 days; p=0.059). This 4-week period is comparable to the reported 4-week time frame within which chemotherapy was commenced in several previous studies ^{7– 14} . We found tumor grade 3 with size >10 cm showed a non-significant trend toward association with worse OS.

From our survival analysis, the median follow-up period was 51 months, which is similar to durations of follow-up reported in previous studies ^{7– 14, 18– 20} . Overall median DFS was 33 months in our study population. Median DFS in this study was 56 months in the chemotherapy group and 20.5 months in the non-chemotherapy group, which are longer than the 48-month and 16-month DFS durations, respectively, reported in the ISG trial ⁹ . One possible explanation for this difference between studies is that our study consisted of patients with lower risk of tumor grade 3 and size >5 cm for relapse, as compared with 100% of patients in the ISG trial ⁹ . In this study, 8.6% (5/58) of patients had tumors smaller than 5 cm, and 12% (6/50) had tumor grade 1. Only 24% (12/50) of patients in our study had tumor grade 3 that was larger than 10 cm, which is much lower than the 54% reported from the ISG trial ⁹ .

When we compared the median DFS in this study with the 90-month and 78-month DFS of EORTC study in the chemotherapy and non-chemotherapy groups, respectively ¹³ , the shorter survival in our study was probably due to greater heterogeneity among patients in the EORTC study. This heterogeneity included 27% with small tumors, less cases with tumor grade 3 (46% in EORTC vs. 30/50; 60% in our study), various primary sites, and median tumor size (7.5 cm and 8.6 cm in the chemotherapy and non-chemotherapy groups, respectively) that was smaller than the median tumor size by group in our study. The 4-year cumulative OS in the chemotherapy and non-chemotherapy groups in this study was 92% and 81%, respectively, compared to 69% and 50%, respectively, in the ISG trial ⁹ .

The efficacy of adjuvant chemotherapy was also evaluated. We proceeded with our analysis despite potential limitations, which included those inherent to retrospective studies, and unbalanced baseline characteristics and sample sizes between the chemotherapy and non-chemotherapy groups, which could result in some degree of selection bias relative to several clinical characteristics. The median age of patients in the chemotherapy group was lower than that in the non-chemotherapy group (42 vs 55 years, respectively). Moreover, chemotherapy group patients had an ECOG performance status score of 0-1, while all ECOG 2-3 patients were put into the non-chemotherapy group. As such, both age and ECOG score may influence physician decision. A total of 13 patients (22%) in our study received adjuvant chemotherapy compared to 34% reported in Taiwan, which had similar baseline patient characteristics to our study ¹⁸ . On the contrary, a report from Korea demonstrated only 7.3% of patients received adjuvant chemotherapy ²⁰ .

For survival analysis compared between patients who received and who did not receive adjuvant chemotherapy, we found no statistically significant difference for either DFS or OS between groups (56 vs 20.5 months, respectively; p=0.29). It should be noted that this study was a retrospective review. There are, therefore, many confounding factors involving the benefits of chemotherapy in this small sample size of patients that received chemotherapy (13/58; 22%). Secondly, chemotherapy may not improve survival benefit, even in high-risk population as established in previous studies, including a long-term (5-year) follow-up evaluated by an Italian Sarcoma Group study and a Cochrane Review of Sarcoma Meta-analysis Collaboration ^{10, 11, 13} .

The incidence of distant relapse in this study was 34.4% (20/58), as compared to the relapse rate of 47% (49/104) in patients studied in the ISG trial ⁹ . The 4-year cumulative incidence of distant metastasis in ISG study in the chemotherapy and non-chemotherapy groups was 44% and 45%, respectively, compared to 31% and 37% in our study. Post-relapse treatment was also evaluated. In this study, 18/25 (72%) patients were commenced on interventions that included surgery, radiation, and/or chemotherapy.

Compliance with adjuvant chemotherapy was also evaluated and reported. Most patients were given doxorubicin and ifosfamide, except in specific subtypes ( Table 2). The median dose of doxorubicin and ifosfamide was 56 mg/m ²/cycle and 4.8 g/m ²/cycle, respectively, which was lower than reported doses from previous studies ^{7– 14} . We found that 53% (7/13) of patients needed dose reduction or delay in treatment due to hematologic toxicity. Most cases did not use prophylactic growth factor to prevent febrile neutropenia. This may explain the higher rate of dose reduction in our study compared to the dose reduction of 26% of patients in the ISG study ⁹ . All patients in the ISG trial received growth factor for secondary prophylaxis. Non-fatal non-hematologic toxicity was found in 62% (43/69) of patients in this study, which was similar to rates reported in previous studies ^{7– 14} .

This study has some notable limitations. As previously described, the potential for selection bias, incomplete medical records, and disproportionate sample size between the chemotherapy and non-chemotherapy groups made it difficult to evaluate contributing factors. The patients enrolled in this study were also from a single center, which is located in a major urban metropolis. Our findings may, therefore, not be generalizable to other geographic settings within Thailand. Moreover, our center is Thailand’s largest tertiary referral hospital, which means that we are often referred patients with complicated and intransigent. Lastly, tumors with heterogeneous histology subtype with a wide range of intrinsic tumor behavior, such as mitotic rate, grade, and undetermined molecular subtype were included. Further study is warranted in a more specific population.

Tumor grade 3 >10 cm was the only baseline clinical characteristic found to be a significant predictor of unfavorable DFS in univariate analysis. No conclusive benefit of adjuvant chemotherapy was observed relative to DFS and OS.

The data referenced by this article are under copyright with the following copyright statement: Copyright: � 2018 Chantharasamee J et al.

Data associated with the article are available under the terms of the Creative Commons Zero "No rights reserved" data waiver (CC0 1.0 Public domain dedication).

Dataset 1. Complete raw demographic and clinical information for each patient included in this study. DOI: https://doi.org/10.5256/f1000research.15793.d212947 ¹⁷ .