Introduction

Sarcomas are a rare form of cancer, representing only ~1 % of all adult malignancies [1]. There are over 50 different histologic subtypes of sarcoma [2], emphasizing that sarcomas are multiple diseases with only a questionable basis for inclusion in the same diagnostic category. There are ~13,000 cases of adult sarcomas per year in the USA [3]. Unfortunately ~50 % of sarcoma patients will still die from their disease, despite appropriate and effective local treatment [1]. The primary management of these tumors is wide local surgical excision [4]. The addition of radiation therapy in high-risk patients, such as those with higher-grade tumors or positive resection margins, is beneficial and considered a standard of care [5–7], with a definite increase in local disease control. Neoadjuvant radiation compared to adjuvant radiation has similar efficacies, though differing side-effect profiles, and both are acceptable treatment modalities [8]. Additionally, adjuvant chemotherapy is used in many cases of trunk and extremity soft tissue sarcomas, though the data for efficacy are limited [9–11]. Interpretation of these studies is hampered by the heterogeneous nature of the soft tissue sarcomas. The soft tissue sarcomas included in these studies have varied in size, grade, location, and histologic subtype. These factors may influence the effect of adjuvant chemotherapy, and lead to confounding variables that may be a basis for overall negative results in larger studies [9]. In reality, there may be certain sarcoma populations that benefit from adjuvant chemotherapy, as has been observed in small studies. This benefit may be more in terms of improvement in progression-free survival as the data are limited for improvement in overall survival.

Large-size tumors, higher-grade tumors, and tumors located in the extremity are all tumor characteristics which have been associated with progression-free survival and to some extent an overall survival benefit with adjuvant chemotherapy [9]. These factors are by themselves prognostic for soft tissue sarcomas. Additionally, studies have shown that age, sex, depth, histologic subtype, and extent of surgical resection are all prognostic factors in soft tissue sarcomas [12–24]. For retroperitoneal sarcomas, sex, grade, histology, and extent of surgical resection are the most important prognostic factors [25–33]. Unfortunately, despite advances in surgical resection, neoadjuvant or adjuvant radiation therapy, and adjuvant chemotherapy, the overall prognosis for soft tissue sarcomas has not changed over the last 20 years, and the 5-year overall survival for stage III soft tissue sarcomas is still ~50 % [34]. Clearly, new strategies or techniques are needed to improve this outcome. One possible methodology is the use of neoadjuvant chemotherapy.

The theoretical benefit of neoadjuvant therapy is multifaceted. First, neoadjuvant therapy may help to shrink the tumor and improve limb salvage rates after primary surgical resection. In addition, neoadjuvant therapy leads to the meaningful downstaging of the extent of surgery, resulting in sparing of nerves and muscle groups. Both the reduction in limb amputations and improvement in limb function would result in improved functional outcomes for patients; significantly improving their quality of life. A second potential benefit of neoadjuvant therapy is an improvement in the rate of achievement of negative surgical resection margins, as positive margins are associated with increased risk of local recurrence and worse survival [15, 17–19, 23, 35–37]. Third, neoadjuvant therapy may result in improved local and distant recurrence-free survival as well as improved disease-specific and overall survival by eradication of micrometastatic disease. Since at least 50 % of high-grade soft tissue sarcomas recur, despite adequate treatment at the primary site, recurrence must be the result of undetectable metastatic disease at diagnosis, and earlier initiation of chemotherapy may be more effective in this “low-volume” disease, ideally increasing the chance of cure. Finally, pathologic examination of tumors after neoadjuvant chemotherapy can determine tumor response to therapy. A significant response to neoadjuvant therapy may correlate with outcomes, and be useful as an additional prognostic factor. This methodology is currently useful in osteosarcomas [38]. If there is a poor response to neoadjuvant therapy, at the very least this indicates that a change in the type of chemotherapy in an adjuvant or post-recurrence setting is required. This strategy is currently under investigation in pediatric osteosarcomas [39]. Overall, there are many theoretical benefits of neoadjuvant therapy. However, many difficulties remain.

This article will review the issues and difficulties that exist in the current interpretation of data in the use of neoadjuvant therapy in soft tissue sarcomas. The current state of evidence for the use of neoadjuvant therapy, chemotherapy alone and chemotherapy in combination with radiation therapy in soft tissue sarcomas will be reviewed. Recommendation for future trial and research will be made.

Materials and methods

An in-depth literature search was conducted using Ovid Medline and PubMed using the search terms of sarcomas, soft tissue sarcoma, neoadjuvant, adjuvant, chemotherapy, radiation, concurrent chemoradiation, hyperthermia, doxorubicin, ifosfamide, limb perfusion, intra-arterial, AIM, MAID, and epirubicin.

Issues with neoadjuvant therapy

Patient selection

There are a variety of factors to consider when including a patient in neoadjuvant therapy trials. One difficulty in current neoadjuvant therapy trials is the differences in the patient inclusion criteria. Patients with a higher risk of local or distant recurrences should be included in these trials, but what constitutes a “high-risk patient”? Certainly, patients with higher-grade tumors are at increased risk [24]. However, there is no universal soft tissue grading system. The two most commonly used systems are the national cancer institute (NCI) grading system, and the Federation Nationale des Centres de Lutte Contre le Cancer (FNCLCC) grading system. Both of these systems divide patients into three groups, low grade or grade 1, intermediate grade or grade 2, and high grade or grade 3, and are roughly equivalent in predicting patient outcomes [40]. Patients with low-grade tumors are at the lowest risk of recurrence and need not be included in neoadjuvant therapy trials. Conversely, high-grade patients are at the highest risk of recurrence and are the obvious candidates for neoadjuvant therapy trials. Patients with intermediate grade tumors have a 70 % 5-year overall survival [40]. So, the majority of these patients may not derive benefit from neoadjuvant therapy, though there is no universal agreement on this point. Additionally, given differences in pathology interpretation, any neoadjuvant therapy trial in soft tissue sarcomas should include a centralized pathology review to ensure concordance among grade determination.

Tumor size is an important prognostic factor as well. Patients with a large tumor >5 cm are at significantly increased risk of recurrence [24]. In fact, increasing tumor size correlates with worsening patient survival [13]. However, what tumor size constitutes the highest risk of recurrence and thus inclusion in neoadjuvant therapy trials has not yet been determined. Size cutoffs previously used included ≥5, ≥8, or ≥10 cm. Patients with greater than 5 cm have a ~50 % [13] chance of tumor recurrence. So, while some of these patients may benefit from neoadjuvant therapy in hopes to improve their chances of cure, 50 % of patients will be unnecessarily exposed to the short-term and long-term consequences of neoadjuvant therapy. A tumor size cutoff of ≥8 or ≥10 cm may be more appropriate for neoadjuvant studies. Soft tissue sarcoma staging is simply a marker for tumor depth, size, and grade [41]. Thus, only patients with high grade and large tumors, which corresponds to stage II and III patients, should be included in these trials.

Other important patient selection criteria include tumor location, histologic subtype, surgical criteria, and patient characteristics. Adjuvant chemotherapy is potentially beneficial in extremity soft tissue sarcomas [10]. Many neoadjuvant therapy trials are limited to patients with extremity or trunk sarcomas. However, is there an intrinsic difference in tumor biology at one location or another, or is the difference in outcomes by tumor location a result of the differences in being able to obtain an adequate surgical resection? Sarcomas located in the abdomen, retroperitoneum, thorax or head and neck represent more surgically challenging cases than those located in the extremities or trunk. If the latter is the case, then perhaps patients with non-extremity/trunk soft tissue sarcomas may derive an increased benefit from neoadjuvant chemotherapy. The type of tumor locations to include in neoadjuvant therapy trials should be evaluated on a per-trial basis, depending on the proposed intervention.

There are many histologic subtypes of soft tissue sarcomas, some of which are intrinsically more sensitive to chemotherapy than other subtypes, such as synovial sarcoma and angiosarcoma. Any neoadjuvant therapy trial will likely include a heterogeneous population of soft tissue sarcomas, and differences in the chemoresponsiveness of different tumors must be accounted for in the pre-planned subgroup analysis.

The surgical selection criteria for neoadjuvant therapy trials vary greatly, and, unfortunately, there is no current consensus for the ideal soft tissue sarcoma patient from a surgical perspective. Most patients who are candidates for limb salvage operations would be candidates for neoadjuvant therapies. However, which patients are considered able to undergo limb salvage operations is a very surgeon-dependent, subjective decision.

Also, any treatment, surgical, radiation, or chemotherapy will have its complications and side effects. Patients will have to be healthy enough, with intact organ function, and intact functional status, to be able to tolerate such an intensive therapy. Given that differences in organ function and functional status can vary among patients of the same age, age should not be used to exclude patients from neoadjuvant trials. Rather a patient’s functional status and expected organ reserve should be used as markers to predict their ability to handle intensive chemotherapy and radiation, and thus their suitability for neoadjuvant therapy trials.

Overall patients with high-grade tumors, ≥5 cm tumors, stage II or III patients, who have adequate functional status and organ status, and are candidates for limb salvage, should be considered for neoadjuvant therapy trials. However, this is still a rough estimate of which patients are at “high risk.” A nomogram, including age, size, depth, grade, location, and histologic subtype, for postoperative 12-year sarcoma-specific death, has been developed by Memorial Sloan-Kettering Cancer Center [42]. This nomogram has been adapted to include a three-tier grading system [43]. It has also been prospectively validated in a separate cohort [44]. This nomogram provides an estimation of the postoperative 12-year risk of death from soft tissue sarcomas. Perhaps a similar methodology needs to be applied to neoadjuvant therapy trials, with pre-planned subgroup analysis for tumor size, grade, location, and histologic subtype. With an eventual plan for the development of preoperative nomogram that can predict the patients at high risk for recurrence, and thus, those that might be considered for neoadjuvant therapy. Unfortunately, the rarity of soft tissue sarcoma does make such an analysis difficult to accomplish.

Finally, even when appropriate inclusion criteria are defined, recruitment into neoadjuvant chemotherapy trials for soft tissue sarcomas can be challenging in such a rare disease. Patients may be treated by local oncologists with surgery, radiation or even chemotherapy, prior to presentation to a tertiary care center specialized in the treatment of soft tissue sarcomas, invalidating them for any neoadjuvant chemotherapy trial. Given the limited number of centers with expertise in soft tissue sarcomas, it is not uncommon for patients to present after initiation of therapy locally. Additionally, the financial cost of traveling to and remaining at a specialized sarcoma treatment center for purposes of undergoing a clinical trial can be limiting for many patients. Furthermore, patients may have preconceived notions regarding the effectiveness of chemotherapy for soft tissue sarcomas. Depending on whether patients believe that chemotherapy would provide a benefit, some patients may decide to spare themselves the toxicity associated with anthracycline-based regimens versus other patients may not want to be randomized to a potentially inferior treatment arm. In addition to patient derived reasons for not enrolling in sarcoma neoadjuvant therapy trials, many physicians may choose not to enroll patients. Any phase III trial in neoadjuvant chemotherapy is by necessity a multicenter study, as a single institution is unlikely to be able to accrue sufficient patients. This requires agreement among sarcoma experts with regard to the neoadjuvant therapy regimen, which can be difficult to achieve. Sarcoma physicians may have preconceived treatment philosophies and may not want to enroll patients on a neoadjuvant therapy trial which could randomize a patient to, in their view, a potentially inferior treatment. These reasons make it particularly difficult to enroll sufficient patients for neoadjuvant chemotherapy trials in soft tissue sarcomas.

Assessment of tumor response

A second difficulty in designing neoadjuvant sarcoma trials is how to assess tumor response and its correlation with improved survival. Different imaging modalities used in sarcomas are CT scans, MRI scans, and more recently PET scans. The standard response criteria employed in most clinical trials is RECIST or RECIST 1.1 [45], which defines tumor responses based solely on changes in tumor size, with partial response defined as a ≥30 % decrease from baseline and progressive disease defined as a ≥20 % increase from baseline Two retrospective series reported improvement in overall survival with tumor shrinkage [46, 47]. Conversely, increase in tumor size does not necessarily correlate with prognosis [48]. Reconciling these retrospective reports is possible, as soft tissue sarcomas may have varying responses to neoadjuvant therapy. Rarely, patients will have shrinkage of their tumors which usually does correspond to a histologic response. Patients with an increase in their tumor size may still have a pathologic response, with tumor swelling from cystic transformation, hemorrhage or necrosis [49]. Thus, RECIST or RECIST 1.1 may be inadequate for assessment of response in soft tissue sarcomas [50].

Alternative response criteria utilize density changes on CT scan or FDG changes on PET scan. Both of these modalities have been studied in retrospective reports in soft tissue sarcoma. Choi criteria are based on changes in tumor size and tumor density on CT scan [51]. Choi criteria are used in the assessment of response of gastrointestinal stromal tumors to tyrosine kinase inhibitor therapy [52]. Choi criteria have been shown to correlate better with pathologic response than RECIST in a retrospective series of 37 patients, though response did not correlate with progression-free survival or overall survival [53]. A follow-up retrospective analysis of 69 patients treated on a phase III trial of neoadjuvant chemotherapy with epirubicin/ifosfamide showed improved correlation of overall survival with Choi criteria versus RECIST [54]. In fact, Choi criteria may better differentiate between patients with stable disease versus progressive disease [55]. However, these reports are limited by their small sample size and retrospective nature. No prospective validation of Choi criteria versus RECIST in soft tissue sarcomas exists. At the present time, Choi criteria cannot replace RECIST in the radiographic assessment of tumor response in soft tissue sarcomas. An appropriate strategy is to include both RECIST 1.1 and Choi criteria on future prospective trials of neoadjuvant chemotherapy in soft tissue sarcomas.

Another alternative imaging modality to assess tumor response is PET scans. Early metabolic response to neoadjuvant chemotherapy as determined by PET may correlate with improved tumor survival [56, 57]. However, which PET criterion is the best predictor of response to therapy and overall survival in soft tissue sarcoma has yet to be determined. One retrospective report showed that a >35 % reduction in FDG uptake had a sensitivity of 100 % for a histopathologic response, defined as ≥95 % pathologic necrosis, and a specificity of 67 % [58]. A second retrospective report showed that patients with an SUV max of ≥6 and a <40 % decrease in FDG uptake were at a high risk of recurrence, ~90 % at 4 years, versus those with a ≥40 % decrease in FDG uptake [59]. These reports are intriguing, but limited by their small sample size and retrospective nature. Future studies are required to evaluate the role of PET imaging in neoadjuvant soft tissue sarcoma trials. The EORTC defines complete metabolic response (CMR) as complete resolution of 18F-FDG uptake, partial metabolic response (PMR) as a reduction of 15 ± 25 % in tumor 18F-FDG SUV after one cycle of chemotherapy and >25 % reduction after more than one cycle of chemotherapy, stable metabolic disease (SMD) as a <25 % increase or <15 % decrease in tumor 18F-FDG SUV, and progressive metabolic disease (PMD) as a >25 % increase in tumor 18F-FDG SUV [50]. PET response criteria in solid tumors (PERCIST) have, also, been defined. PERCIST defines CMR as complete resolution of 18F-FDG uptake as compared to mean SUV uptake in the liver or surrounding background blood-pool. PMR as a 30 % reduction in tumor 18F-FDG SUV, SMD as not meeting criteria for CMR, PMR, or PMD, and PMD as a >30 % increase in tumor 18F-FDG SUV [50]. No prospective validation of PERCIST or EORTC PET response criteria versus RECIST in soft tissue sarcomas currently exists, but these are standardized criteria for assessing response on PET scans, which can be applied to future soft tissue sarcoma neoadjuvant clinical trials. A significant downside to the use of PET scans is the added cost of PET imaging versus contrast enhanced CT scan, ~250.00 versus ~700.00 € [60]. The added financial cost of including PET scans in neoadjuvant therapy trials must be considered.

Unfortunately, most neoadjuvant therapy trials have shown the development of progressive disease in ~10 % of patients. These patients may lose the chance for a limb salvage operation or cure, and do not derive any benefit from neoadjuvant therapy due to non-responsiveness of their disease. Currently, RECIST is inadequate in determining which patients will be treatment failures, as an increase in size does not correlate with prognosis [48]. Furthermore, conventional radiographic findings with CT or MRI are not concordant with pathologic findings [61–63]. Choi criteria may have improved correlation with pathologic response than RECIST [53, 54]. As noted previously, a reduction in FDG uptake from PET scans may be a better predictor of pathologic response [58]. The utility of Choi criteria or response on PET imaging to predictive pathologic response has to be prospectively validated in future clinical trials.

An additional problem is that pathologic response has been inconsistent in predicting survival or recurrence. Some retrospective studies reported positive [64–69] and some reported negative results [70, 71]. Typically, response was defined as percent necrosis, with different cutoffs used to define response, >90 % necrosis, >95 % necrosis, or >98 % necrosis [70]. Complicating the matter is the fact that necrosis is a component of tumor grade and has intrinsic prognosis value irrespective of neoadjuvant treatment. Many soft tissue sarcomas have extensive preexisting necrosis. It is difficult to distinguish preexisting necrosis from post-treatment necrosis [49]. Thus, necrosis may not be the best histopathologic marker to assess response to treatment. There was no previously defined histopathologic assessment for treated soft tissue sarcomas. Recently, the EORTC proposed a soft tissue sarcoma response score, with five response grades depending on the amount of stainable tumor cells: grade A no stainable tumor cells; grade B single stainable tumor cells; grade C ≥ 1 to <10 % stainable tumor cells; grade D ≥ 10 to <50 % stainable tumor cells; and grade E ≥ 50 % stainable tumor cells [49]. This histopathologic criterion may be used in future neoadjuvant therapy trials to assess pathologic response to therapy. Overall, the imaging response criteria used in neoadjuvant studies and their correlation with pathologic response and improved survival must be further studied and improved.

Outcome measures in neoadjuvant therapy trials

Similar to patient selection criteria and tumor response assessment criteria, outcome measures have greatly varied in previous neoadjuvant therapy trials. The primary outcome for all neoadjuvant therapy trials should be overall survival. It is only by this criterion that an estimation of the long-term impact of neoadjuvant therapy can be obtained. Disease-specific survival, local recurrence-free survival, distant recurrence-free survival, and progression-free survival can be included as secondary outcomes. Several neoadjuvant therapy trials have included tumor response as a surrogate marker for overall survival; however, this has yet to be definitively proven. So, it will be essential to note the rates of complete response, partial response, stable disease, and progressive disease both by radiographic and by pathologic criteria to help identify whether there is an accurate surrogate outcome for overall survival. Also, the limb salvage rates and rates of amputation should be reported in any neoadjuvant therapy trial. If the therapy is effective, then the rates of limb salvage should increase, limb functionality after surgery should improve, and the rate of amputations should decrease.

Due to the intensive nature of chemotherapy regimens used in soft tissue sarcomas, the toxicities and side effects of chemotherapy will have to be carefully assessed. The short-term toxicities, such as wound complication rate, and long-term toxicities, such as cardiomyopathy and myelodysplasia, will need to be reported, as well as the fraction of patients completing the planned chemotherapy regimen, the fraction of patients that require chemotherapy dose reductions, and the death rate from chemotherapy. Effect of chemotherapy on surgical morbidity is a concern as well. There is, at least, one retrospective report that shows neoadjuvant chemotherapy does not affect the morbidity of surgical resections [72]. So, this concern may be unfounded and should not preclude a patient from undergoing neoadjuvant chemotherapy. The inclusion of these criteria as secondary outcomes will be important in evaluating any neoadjuvant therapy trial in soft tissue sarcomas.

Choice of regimen

The choice of the neoadjuvant chemotherapy regimen is a most important variable when designing a neoadjuvant therapy trial. Neoadjuvant radiation therapy has been shown to be as effective as adjuvant radiation therapy, with a local recurrence rate of only ~10 % though radiation therapy does not influence distant recurrence or overall survival [8]. Thus, radiation therapy alone is likely to be insufficient therapy to improve cure rates. Chemotherapy must be included to improve cure rates. In deciding a neoadjuvant therapy regimen with chemotherapy, choices include chemotherapy alone, concurrent chemoradiation, or more commonly used is sequential chemotherapy and then radiation. Various chemotherapy regimens have been tried, usually anthracycline-based, which has been the most effective agent for soft tissue sarcomas since its introduction in the 1970s [73]. The most active regimen in the adjuvant setting is the combination of an anthracycline and ifosfamide [9–11] [epirubicin/ifosfamide, doxorubicin/ifosfamide/mesna (AIM)]. Both infusional doxorubicin and bolus doxorubicin were studied in the neoadjuvant setting [74]. Infusional doxorubicin is associated with a lower cardiac side-effect profile [75].

A considerable “downside” to these regimens is their significant toxicity. Patients without favorable performance status or with significant co-morbidities cannot be included in neoadjuvant chemotherapy trials. Additionally, the long-term toxicities of these regimens include acute leukemia and heart failure, which must be weighed against an uncertain contribution of these regimens to cure. In patients with more than a 50 % chance of cure, the long-term risks of these regimens may outweigh the benefits. More recently, non-anthracycline-based regimens with improved side-effect profiles include trabectedin [76] and gemcitabine/docetaxel [77, 78] which have been studied in some neoadjuvant/adjuvant sarcoma trials particularly in myxoid liposarcoma and uterine leiomyosarcoma. Eribulin was recently FDA-approved for liposarcomas, and represents another chemotherapeutic which could be moved from the metastatic to neoadjuvant setting [79]. Pazopanib was studied in the neoadjuvant setting as well as in combination with gemcitabine and docetaxel [80]. Chemotherapy used concurrently with radiation has included doxorubicin [81–85], ifosfamide [86, 87], ifosfamide and doxorubicin [88] razoxane [89], iododeoxyuridine [90, 91], temozolomide [92], bevacizumab [93], sorafenib [94], and pazopanib [95]. Additionally, intra-arterial chemotherapy and limb perfusion have been used in neoadjuvant sarcoma trials [96]. There is currently no defined superior neoadjuvant chemotherapy regimen for soft tissue sarcomas. The status of current neoadjuvant therapy trials in soft tissue sarcomas is reviewed below.

Current neoadjuvant therapy trials in soft tissue sarcomas

Neoadjuvant chemotherapy

Neoadjuvant chemotherapy has been studied in the treatment of soft tissue sarcomas since the 1980s [97]. Initial small, retrospective or single-institution experiences suggested responses to chemotherapy and a benefit to neoadjuvant chemotherapy with an improved survival in those patients who obtained a response [97–100]. However, a retrospective review of 76 patients with stage IIIb soft tissue sarcoma treated at MD Anderson Cancer Center with doxorubicin-based neoadjuvant chemotherapy, without ifosfamide, showed a 5-year disease-free survival of 52 % and a 5-year overall survival of 46 % [101]. This result was similar to results with surgery alone or postoperative chemotherapy. A larger retrospective review of 674 patients from Memorial Sloan-Kettering Cancer Center and MD Anderson Cancer Center identified 214 patients treated with neoadjuvant chemotherapy and 112 patients treated with adjuvant chemotherapy. The treatment regimens included AIM, MAID (MESNA/doxorubicin/ifosfamide/dacarbazine), doxorubicin alone, ADIC (doxorubicin/dacarbazine), and CyADIC (cyclophosphamide/doxorubicin/dacarbazine). There was an initial response to chemotherapy with improved survival in the first year, but this trend was not sustained over time. The 5-year overall disease-free survival was 47 % for patients treated with chemotherapy and 49 % for patients who received no chemotherapy [102]. These results were disappointing. The chemotherapy regimen associated with the highest response rates in the adjuvant setting is doxorubicin and ifosfamide [9]. This regimen has the highest tumor response rates of ~45–47 %, in the neoadjuvant setting, as well [103]. Further retrospective neoadjuvant chemotherapy trials have included a doxorubicin/ifosfamide-based regimen; Table 1.

Table 1 Neoadjuvant chemotherapy studies
Full size table

A retrospective experience with the use of doxorubicin, intra-arterial cisplatin, and ifosfamide was reported by Henshaw et al. [66]. There were increased complications with arterial thrombosis and myocutaneous necrosis, but the 5-year overall survival was 88 %, a significant improvement from historical controls [66]. A more extensive retrospective experience was reported by Grobmyer et al. in 2004, with 74 patients who received an average of three cycles of AIM chemotherapy compared to 282 patients treated with surgery alone. The greatest effect of chemotherapy was seen in patients with a >10 cm tumor. The disease-specific mortality hazard ratio was 0.45 (95 % CI 0.25–0.83) for patients receiving chemotherapy, compared to those who received surgery alone. The 3-year disease-specific survival was 83 % for those receiving chemotherapy and 62 % for those with surgery alone [104]. These results support the hypothesis that patients with larger tumors do benefit, at least, upfront, from neoadjuvant therapy with a doxorubicin/ifosfamide-based regimen, though a follow-up of only 3 years is too short to determine whether the effect of chemotherapy will be long term. A further retrospective report from the French sarcoma group showed no benefit to adjuvant or neoadjuvant chemotherapy in patients with synovial sarcoma, though only 19 % of patients were treated with neoadjuvant chemotherapy, and only 59 % of the chemotherapy regimens were doxorubicin/ifosfamide-containing [105]. A retrospective study from Japan in 2013, using MAID neoadjuvant chemotherapy, showed a 5-year overall survival of 86 % and 5-year disease-free survival of 77 % for patients with grade 2 or 3 extremity or trunk STS [106]. This result compares favorably with historical controls. Overall, retrospective reviews and single-institutional experience suggest a benefit of neoadjuvant doxorubicin/ifosfamide-based chemotherapy for large, extremity STS. Unfortunately, prospective studies to confirm this have been limited, Table 1.

A phase II study reported 2001 by Gortzak et al. [107] showed a 5-year overall survival of 64 % in the surgery-alone arm compared to 65 % in the neoadjuvant chemotherapy arm, p = 0.22. There were only 67 patients in each arm, so this study may not have been appropriately powdered to detect a difference. Due to slow accrual, a planned phase III trial was never completed. In addition, lower-risk patients with <8 cm tumors or grade II tumors were included, and these patients may not have benefited from neoadjuvant chemotherapy as much as high grade or large tumors. Finally, the chemotherapy regimen used was doxorubicin 50 mg/m2 day 1 and ifosfamide 5 g/m2 as a 24 h continuous infusion day 1, a dosing schedule that was much lower when compared to the regimens used in the retrospective studies that suggested a benefit to neoadjuvant chemotherapy. A second phase II study was conducted in the pediatric population with advanced non-rhabdomyosarcoma soft tissue sarcomas using vincristine (1.5 mg/m2 weekly for 13 doses), ifosfamide (3 g/m2 daily for 3 days every 3 weeks for seven cycles), and doxorubicin (30 mg/m2 daily for 2 days for six cycles) [108]. There was no control group though the patients with clinical group III disease did have a 3-year overall survival of 80 %. The majority of patients who responded to chemotherapy had synovial sarcoma. A third phase II study conducted in Japan with adriamycin 30 mg/m2 day 1, 2 and ifosfamide 2 gm/m2 day 1–5 given for three cycles preoperatively and two cycles postoperatively showed a 5-year overall survival of 82.6 %, a promising result [109, 110]. However, a phase III randomized controlled trial of a high-dose doxorubicin/ifosfamide (AIM)-based neoadjuvant chemotherapy regimen still needs to be conducted in patients with extremity, high grade, large soft tissue sarcomas to confirm the ~80 % 5-year overall survival and if this benefit is durable.

High-dose doxorubicin and ifosfamide have considerable toxicities with risks for heart failure, acute myeloid leukemia, neutropenic fever, infection, and even death. Unfortunately, these are the most active chemotherapeutic agents in soft tissue sarcomas. A recent neoadjuvant chemotherapy phase II study of a new agent, trabectedin, in myxoid liposarcoma showed a 13 % complete pathologic response and a 24 % partial response by RECIST [76]. This agent was well tolerated. These results and the tolerability of this agent suggest that this drug should be studied further in the treatment of myxoid liposarcomas particularly in the neoadjuvant setting. Another alternative which can be applied to soft tissue sarcomas is gemcitabine/docetaxel, which has shown improved progression-free and overall survival in uterine leiomyosarcoma compared to historical controls. However, follow-up in these phase II studies was only 2–3 years [77, 78]. A phase Ib/II study of gemcitabine/docetaxel and pazopanib was conducted. However, the study was terminated early due to significant hepatic and myelotoxicity. Only five patients were enrolled, and no objective responses occurred [80].

Neoadjuvant chemotherapy with radiation

One of the important goals of neoadjuvant chemotherapy is producing a local response to decrease the extent of resection and improve the limb salvage rate. Not all patients will respond to chemotherapy, so the addition of radiation therapy to the primary tumor may rescue poor chemotherapy responders in terms of local response. Additionally, chemotherapy may potentiate the effect of radiation. Thus, the combination of chemotherapy with radiation in the neoadjuvant setting may improve local response rates and still increase the chances of a long-term cure. Chemotherapy can be given alternating with radiation therapy or concurrently. Chemotherapy regimens given sequentially with radiation include MAID, epirubicin/ifosfamide, and ifosfamide/mitomycin/doxorubicin/cisplatin (IMAP). Each of the MAID chemotherapy drugs as a single agent has a ~20 % response rate in sarcomas, so the combination is reasonable [73, 74, 111, 112]. This regimen has been studied and is active in soft tissue sarcomas. A phase II study showed an overall response rate of 47 % in advanced soft tissue sarcomas [113]. A second phase II study of MAID chemotherapy showed a 49 % response rate in adult osteosarcoma, Ewing’s sarcoma and rhabdomyosarcoma [114]. Later, a phase III study was conducted of doxorubicin and dacarbazine versus doxorubicin, ifosfamide, and dacarbazine [115]. This study showed an overall response rate of 32 % in 340 patients with significantly longer time to progression, though no difference in overall survival. Given good responses in the advanced and metastatic setting, it is reasonable to consider the use of MAID chemotherapy in the neoadjuvant setting, particularly because of the higher response rates which are important in the neoadjuvant setting before surgical resection. MAID has been studied in this setting, particularly in the Delaney experience [116–118], the RTOG 9514 trial [119, 120], and in the Ogura et al. [106] experience. These results, shown in Table 2, have been encouraging, but caution must still be advised [121].

Table 2 Neoadjuvant chemotherapy and radiation therapy studies
Full size table

In the initial Delaney retrospective experience, reported in 2003, 48 patients were treated with three cycles of preoperative MAID chemotherapy alternating with 44 Gy of radiation, and then three cycles of postoperative MAID chemotherapy with an additional 16 Gy of radiation for positive surgical margins. These results showed a significant improvement at 5 years in freedom from distant metastasis 75 versus 44 %, p = 0.0016, disease-free survival 70 versus 42 %, p = 0.0002, and overall survival 87 versus 58 %, p = 0.0003 compared to historical controls [116]. Only one patient developed a fatal myelodysplastic syndrome. The long-term results of this retrospective experience, reported in 2012, showed a 10-year disease-free survival of 65 % compared to 30 % in the control group p = 0.0002, and a 10-year overall survival of 66 versus 38 %, p = 0.003 [117]. While there were relapses and death from sarcoma after 5 years, the benefit of chemotherapy was still seen at 10 years. This result contradicts the Cormier et al. [102] retrospective experience which suggested that the benefit from adjuvant chemotherapy decreases with time. In an additional report from the Delaney group, a retrospective experience over 10 years of 66 patients showed a similar disease-specific survival of 89 % and overall survival of 86 % at 5 years with local recurrence of only ~10 % [118]. It should be noted that this is similar to the Ogura et al. [106] retrospective experience which had an 86 % 5-year overall survival. There are differences between these two experiences, in particular only six patients received preoperative radiation therapy in addition to chemotherapy, and nine patients received postoperative chemotherapy, out of a total of 40 patients. Also, while the total dose of doxorubicin per cycle was 60 mg/m2, in both reports the amount of ifosfamide per cycle 7.5 versus 6 gm/m2 and dacarbazine per cycle 900 versus 1000 mg/m2 is different between the Ogura and Delany retrospective experiences. Despite these differences, outcomes were similar.

An additional phase II study with similar “sandwiching” of chemotherapy and radiation was conducted by the radiation therapy oncology group, RTOG 9514. Short-term results at 3 years showed a disease-free survival of 56.6 %, distant disease-free survival of 64.5 %, and overall survival of 75.1 % [119]. Longer-term results at 5 years showed a disease-free survival of 56.1 %, distant disease-free survival of 64.1 %, and overall survival of 71.2 % [120]. These results are slightly inferior to the Delany or Ogura retrospective experiences. Two possible explanations are the chemotherapy completion rate, which was 83 % in the Delany retrospective experience and only 59 % in the RTOG study, or the chemotherapy doses. There was significantly more toxicity in the RTOG trial with three treatment-related deaths, one from infection, and two from acute leukemia. The RTOG trial used 7.5 gm/m2 of Ifosfamide per cycle compared to the Delany retrospective experience of 6 gm/m2, and only 675 mg/m2 of dacarbazine compared to 1000 mg/m2 in the Delany retrospective experience. In addition, these are small phase II studies and difference in outcome could be due to patient selection or random chance.

Overall, the Delany retrospective experience, the Ogura retrospective experience, and the RTOG Phase II trial suggest that in appropriately selected patients, the neoadjuvant use of MAID chemotherapy with radiation can improve local control with only an ~10 % failure rate, and produce a durable overall survival benefit, ~70–80 % at 5 years which is an improvement compared to historical controls. Historically, patients with large ≥8 cm high-grade tumors are predicted to have only a ~50 % 5 year overall survival. Despite these encouraging results, problems do still exist. These trials were small or retrospective, so prone to error from selection bias. Particularly in the Delany retrospective experience, the control group differed from the treatment group in terms of the portion of grade 2 and 3 tumors and by tumor histologic subtype [118]. A phase III trial that would eliminate such difficulties has not yet been conducted, but would be helpful in proving the benefit of this neoadjuvant approach.

Questions remain in the design of such a trial. With respect to patient selection, patients with large primary tumors should be included, but whether this should be limited to grade 3 tumors, which are at highest risk of recurrence or grade 2 and 3 tumors, is problematic. Furthermore, should large tumors be defined as ≥5, ≥8, or ≥10 cm. Additionally, only 15 patients in the Ogura retrospective experience received radiation, but 5-year overall survival was similar to the Delany retrospective experience. So, is radiation necessary in all patients or just those with positive margins, and when is the best timing of radiation? Furthermore, is dacarbazine necessary, as a neoadjuvant chemotherapy alone trial with only doxorubicin and ifosfamide showed a similar ~80 % 5 year overall survival [110]. Excluding dacarbazine would allow optimization of the dose of the two most active agents, doxorubicin and ifosfamide. Additionally, the dosing of dacarbazine and ifosfamide differed between these small studies. The best balance between chemotherapy dose and toxicity is unknown. Certainly, high doses of doxorubicin, ifosfamide and dacarbazine are required to produce a higher response rate in soft tissue sarcomas, but higher doses predictably lead to more toxicity and a lower therapy completion rate which may produce inferior outcomes. All of these questions remain unanswered, and given different preferences at different sarcoma centers and the need for a phase III trial in sarcomas to be a multicenter affair, a phase III trial with MAID or AIM chemotherapy is unlikely to be undertaken.

Further circumstantial evidence for the utility of a combined neoadjuvant radiation and chemotherapy approach comes from two phase II trials and one phase III trial with IMAP and epirubicin and ifosfamide. In 2001, Edmonson et al. [122] reported the results of a phase II trial of preoperative ifosfamide, mitomycin, doxorubicin and cisplatin for two cycles and 45 Gy of radiation, with three cycles of postoperative mitomycin, doxorubicin, and ifosfamide with a postoperative radiation boost to a total of 65 Gy. Local recurrence-free survival at 5 years was 92 %, and 5-year overall survival was 80 %, comparable to the Delany and Ogura retrospective experiences. In 2008, Ryan et al. [123] reported a phase II trial of epirubicin and ifosfamide for 25 patients with intermediate or high-grade >5 cm soft tissue sarcomas. Two-year overall survival was 84 %, and the toxicity of the regimen was similar to other neoadjuvant therapy trials. Later, a phase III trial of epirubicin and ifosfamide was conducted by Gronchi et al. [124, 125]. In this trial, patients received three cycles of preoperative chemotherapy with epirubicin 120 mg/m2 per cycle and Ifosfamide 9 gm/m2 per cycle and were then randomized to receive two further postoperative cycles of chemotherapy. Radiation was delivered either preoperatively or postoperatively. Five-year overall survival was 73 %, and the addition of two cycles of postoperative chemotherapy did not influence outcomes. These trials suggest alternative chemotherapy regimens that can have similar efficacy as MAID or AIM chemotherapy though still do not represent conclusive phase III randomized controlled evidence of the benefit of the neoadjuvant radiation and chemotherapy as compared to adjuvant chemotherapy or just radiation therapy alone. None of these results can be considered practice changing, and there remains no standard of care for neoadjuvant therapy in soft tissue sarcomas. Despite this, it is reasonable to consider treating a healthy high-risk (≥5–10 cm, extremity, high grade) sarcoma patients per these protocols, after discussing the risks and potential benefits with the patient, in an attempt to improve their overall survival.

Neoadjuvant concurrent chemotherapy

Undoubtedly, there will be those patients that choose to forgo neoadjuvant chemotherapy due to patient preference or concerns related to toxicity and those patients with too many co-morbid medical conditions. Most of those patients will receive radiation therapy alone. However, this may not be sufficient to reduce local recurrence, improve surgical outcomes, and limb salvage rates. An alternative strategy is to use radiation concurrently with chemotherapy. Though this may help local response, the doses of chemotherapy used are unlikely to affect distance recurrences. Trials using concurrent chemoradiation in soft tissue sarcomas are listed in Table 3. The first agents studied concurrently with radiation include the chemotherapeutic agents with the highest response rates in sarcoma, doxorubicin, and ifosfamide. Using low-dose doxorubicin 4 mg/m2 bolus and then daily continuous infusion for 4 days concurrently with radiation was shown to be feasible with acceptable toxicities, though with few complete responses [81, 82]. Increasing the dose of doxorubicin to 12 mg/m2 per day does increase the response rate to a complete response rate of 11 % and partial response rate of 56 % [83]. This result, however, has not yet been shown to improve clinical outcomes. In fact, a phase II study of concurrent doxorubicin with radiation had a 5-year overall survival of only 63 % [84]. Ifosfamide has been tried currently with radiation as well. A retrospective study showed a pathologic complete response rate of 14 % and >90 % necrosis in another 43 % with ifosfamide concurrently with radiation [86]. A prospective phase I/II study, however, only showed a radiographic response rate of 8 % and 5-year overall survival of 59 % [87]. In a combination approach, a retrospective series of concurrent doxorubicin with radiation followed by adjuvant doxorubicin and ifosfamide had only a 58.3 % 5-year overall survival [85]. These studies do not indicate improved outcomes with the use of concurrent doxorubicin or ifosfamide. A German phase I/II study 60–64 Gy of radiation with doxorubicin 50 mg/m2 on day 2 and day 30 and ifosfamide 1.5 gm/m2 on days 1–5 and 29–33 showed an 83 % 3-year overall survival [88]. These doses of doxorubicin and ifosfamide are higher than those used in the other concurrent chemoradiation trials, though lower than the neoadjuvant MAID regimens discussed above. These phase II and retrospective reports seem to indicate that it is the dosing of doxorubicin and ifosfamide that matter, rather than the use concurrently with radiation. This outcome supports the hypothesis that the best use of doxorubicin and ifosfamide is alternating with radiation, as a higher dosing of these medications can be obtained with acceptable toxicity when dosed separately from radiation.

Table 3 Neoadjuvant concurrent chemotherapy and radiation therapy studies
Full size table

With a lack of significant improvement in local response with doxorubicin and ifosfamide concurrently with radiation, newer chemotherapeutic agents have been studied concurrently with radiation. Unfortunately, these agents have failed to show a benefit over radiation alone, Table 3. Razoxane is a radio-sensitizer. In a phase III trial, razoxane 150 mg/m2 was given orally during radiation compared to radiation alone. There was no difference in local control or overall survival of 65 % in the radiation alone group and 55 % in the radiation plus razoxane group [89]. Another radio-sensitizer iododeoxyuridine was tested in two early-phase clinical trials, but no significant improvement in local responses was seen, and overall survival was only 39 % [90] and 16.6 % [91]. Additionally, temozolomide [92], bevacizumab [93], sorafenib [94] and pazopanib [95] were studied in combination with radiation therapy in small phase I or phase II trials. In particular sorafenib combined with radiation showed near-complete pathologic response (pCR) in 38 % of patients and pazopanib combined with radiation showed near pCR in 40 % of patients, with near pCR defined as necrosis greater than 95 % [94, 95]. These regimens were well tolerated, and improvements in local responses were suggested but have yet to be tested in a randomized controlled setting. An alternative approach is the inclusion of immunotherapy. A phase I/II trial tested the feasibility of intratumoral injection of dendritic cells during radiation therapy [126]. An immune response was induced in ~50 % of patients, but overall survival at 3 years was only 23.5 %. However, the approach of combining immunotherapy with radiation therapy does warrant further study in future clinical trials. Overall, the experience with concurrent chemoradiation using doxorubicin, ifosfamide, radio-sensitizers (razoxane, iododeoxyuridine) or newer agents (temozolomide, bevacizumab, sorafenib, pazopanib) has shown no definitive benefit over radiation alone, or chemotherapy alternating with radiation. It is still worthwhile conducting studies of concurrent chemoradiation, particularly with sorafenib or pazopanib, or concurrent immunotherapy and radiation in hopes to improve local response rates, but these regimens can only be considered investigational and not standard therapies at this time.

Neoadjuvant chemotherapy with radiation and regional hyperthermia

A further therapy under investigation is the addition of regional hyperthermia to neoadjuvant radiation and chemotherapy, Table 3. Radiation with neoadjuvant chemotherapy, etoposide, doxorubicin, and ifosfamide has been combined with regional hyperthermia in two phase II trials. The first included high-grade soft tissue sarcomas >8 cm and showed a response rate of 17 % with a 5-year overall survival of 49 % [127]. The second included high-grade retroperitoneal and visceral sarcomas and showed a response rate of 13 % with a 5-year overall survival of 32 % [128]. Neither of these studies showed a dramatic improvement in local response rates or long-term survival with the addition of regional hyperthermia. A third retrospective review of 28 patients who received radiation ± concurrent ifosfamide ± regional hyperthermia showed a 3-year overall survival of 72 % and no difference in local control rates with the addition of regional hyperthermia [129]. Thus, regional hyperthermia by itself does not add significantly to local control of soft tissue sarcomas.

Neoadjuvant intra-arterial chemotherapy and isolated limb perfusion therapy

Another method to improve local response rates in soft tissue sarcomas is the use of intra-arterial chemotherapy. This technique has been applied since the 1970s [96], with the use of agents such as doxorubicin, cisplatin, and melphalan [130]. The use of intra-arterial doxorubicin chemotherapy has shown a reasonable local response rate, but this comes at a price, with significant local toxicities. The southeastern cancer study group showed a 95 % limb salvage rate and 98.5 % local control rate with 90 mg of intra-arterial doxorubicin and 30–45 Gy of radiation [131–133]. However, the postoperative wound complication occurred in 41 % of patients [132]. Additionally, 54 % of patients developed distant disease, and the 5-year overall survival was 59 % [132]. These results are not clearly superior to historical controls. The addition of hyperthermic limb perfusion with intra-arterial doxorubicin or cisplatin and radiation has shown improved local responses 94–98 % and limb preservation rates of 91–97 % [134–136]. Again, overall survival is not clearly improved, with a median overall survival of 12 months in one phase II study [135]. However, there have been no randomized controlled trials that have compared intra-arterial chemotherapy and radiation or intra-arterial limb hyperthermic perfusion and radiation to neoadjuvant or adjuvant sequential high-dose chemotherapy and radiation.

Further advances in loco-regional therapy lead to the development of isolated limb perfusion (ILP) [137]. Such a technique allows for higher concentrations of chemotherapy, as long as exposure to the systemic circulation is limited. The inclusion of tumor necrosis alpha in regional limb perfusion increased local response rates in soft tissue sarcomas to 82 % [138, 139]. Standard dosing of TNFα is 1 mg, as higher doses increase toxicity, but not response [140–142]. This technique is limited by significant toxicities including delayed wound healing, wound infections, tissue fibrosis, sensor and motor neuropathies, lymphoceles, and arterial thrombosis [138, 143]. These complications can be severe enough to require limb amputation [144]. The Wieberdink scale grades toxicity from grade I no reaction, grade II mild edema and hyperemia, grade III severe edema and blistering, grade IV epidermolysis and compartment syndrome, to grade V tissue necrosis and amputation. Local toxicity is lower with melphalan-based versus doxorubicin-based TNFα ILP [145]. Grade IV and grade V reactions occur in 2–7.5 % of patients treated with melphalan-based TNFα ILP versus ~10 % in doxorubicin-based TNFα ILP with a TNFα dosing of 1 mg [144, 146–148]. Additional toxicity results from systemic leakage. Reports of systemic leakage vary from 2 to 13 % [145, 149]. Systemic exposure to chemotherapy can lead to cytopenias, and systemic exposure to TNFα can lead to hypotension and shock.

Counterbalancing these severe toxicities is improved local response rates, limb salvage rates, and decrease in local recurrence. In retrospective series of 186 patients with grade II or III soft tissue sarcomas, with median tumor size of 16 cm, treatment with hyperthermic isolated limb perfusion with melphalan and TNFα leads to a radiograph complete response of 18 %, partial response of 57 %, and stable disease of 22 % [146]. The limb salvage rate was 82 % [146]. Another retrospective series of 106 patients treated with doxorubicin and TNFα ILP showed a 22 % complete response, 77 % partial response, and 77 % limb salvage rate [148]. Many of these patients would have required amputation prior to treatment with ILP. The addition of radiation after ILP significantly improves local control rates, 96 versus 52 % in one retrospective series of 77 patients [150]. These retrospective studies have led to the approval of ILP with TNFα in Europe for soft tissue sarcomas.

However, there have been no randomized controlled trials that have compared isolated limb perfusion therapy with doxorubicin or melphalan ± radiation and TNFα to neoadjuvant or adjuvant sequential high-dose chemotherapy and radiation. The majority of patients with soft tissue sarcomas die from metastatic disease, not local recurrence. Loco-regional therapies, such as ILP with melphalan and TNFα, may be effective in increasing local response rates (63–91 %) leading to limb salvage operations (58–97 %) particularly in the setting of large tumors [145]. This limb salvage rate is not definitely improved compared to sequential radiation and chemotherapy, especially in the absence of a randomized trial. Furthermore, the goal of neoadjuvant therapy in soft tissue sarcomas is not only tumor response and limb salvage, but also to cure the patient. Loco-regional therapies such as ILP do not have improved overall survival (40–60 %) compared to historical controls [147, 150]. Retrospective neoadjuvant chemotherapy trials with an anthracycline and ifosfamide do suggest an improved survival compared to historical controls (60–80 %). Without an improvement in overall survival, the increased toxicities associated with ILP are hard to justify in most patients. Additionally, TNFα, arguable the most important agent in ILP, is not approved in the USA. Thus, ILP is rarely performed in the USA. TNFα ILP was approved in Europe for the treatment of advanced soft tissue sarcomas in 1988 [145]. ILP with TNFα is used in many specialized European centers experienced with its delivery for soft tissue sarcoma patients with large, extremity, non-operable tumors. ILP with TNFα has been reported not to impact the ability to give high-dose systemic chemotherapy [151, 152]. The role of ILP with TNFα may be in soft tissue sarcoma patients with large, extremity, tumors that are still in-operable after high-dose chemotherapy and radiation. However, this needs to be studied in a randomized fashion.

Conclusions

The long-term mortality from soft tissue sarcomas despite surgical resection and adjuvant or neoadjuvant radiation remains unacceptable high. Better strategies are required in high-risk patients, those with large and high-grade tumors who have a <50 % 5-year overall survival. Neoadjuvant chemotherapy offers the potential benefit of improved long-term survival [153–155]. There is no universally accepted neoadjuvant therapy standard of care. Based on retrospective series and phase II trials, neoadjuvant therapy with doxorubicin and ifosfamide (AIM), epirubicin and ifosfamide, or doxorubicin, ifosfamide and dacarbazine (MAID) with or without radiation therapy are possible treatment options. These small studies suggest improved survival in high-risk patients compared to historical controls to 70–80 % at 5 years. However, there is a lack of randomized phase III trials comparing these regimens to each other and to observation alone, which would prove a benefit. Additionally, unanswered questions which remain are the benefit of dacarbazine when added to ifosfamide and an anthracycline, and the timing of radiation therapy in regards to neoadjuvant chemotherapy. Another complication is the toxicity of these regimens and inability of older patients or patients with multiple medical co-morbidities to tolerate these regimens. Further, investigation of regimens with lower side-effect profiles, such as gemcitabine/docetaxel, trabectedin, eribulin, pazopanib or sorafenib in the neoadjuvant setting, is required. Overall treatment decisions must be individualized between the patient and provider. There is a role for neoadjuvant chemotherapy in select patients. Neoadjuvant chemotherapy with or without radiation should, at least, be considered in patients with high-risk disease, large (>5 or >8 cm), high-grade, extremity tumors, who are young and healthy enough to be able to tolerate the neoadjuvant chemotherapy. However, without randomized controlled trial evidence, the routine use of neoadjuvant chemotherapy cannot be recommended in all patients with soft tissue sarcomas.

The use of radiation should be considered in all high-risk patients to decrease the risk of local recurrence, but this is unlikely to affect overall survival, and radiation can be delivered in a neoadjuvant or adjuvant fashion. For older patients with more medical co-morbidities, who would not be able to tolerate AIM or MAID chemotherapy, current evidence does not support a definitive benefit for lower-toxicity treatments, such has concurrent chemoradiation and cannot be recommended at this time. Additionally, limb perfusion in soft tissue sarcomas has not been studied in randomized clinical trials and has the cost of significant toxicity. These treatment strategies cannot be recommended in the USA outside of the context of a clinical trial.

Future directions to improve our treatment of soft tissue sarcomas would be to conduct phase III randomized controlled trials to definitively answer the question of whether neoadjuvant chemotherapy with the AIM or MAID regimens results in an overall survival benefit as well as directly compare these regimens to determine the added benefit of dacarbazine. Though, conducting such trials in a rare very heterogeneous disease, without an agreement on what constitutes high-risk patients, can be quite challenging. Furthermore, as new techniques and strategies are being applied to the metastatic setting, such as recently FDA-approved trabectedin, eribulin, pazopanib, or immunotherapy, those that show a significant benefit will need to be applied to the adjuvant and neoadjuvant setting. Currently, the Italian, Spanish and French Sarcoma groups are enrolling a histology tailored phase III neoadjuvant chemotherapy trial comparing epirubicin (120 mg/m2) and ifosfamide (9 gm/m2) to gemcitabine/docetaxel (900, 75 mg/m2) in undifferentiated pleomorphic sarcoma, trabectedin (1.3 mg/m2) in myxoid liposarcomas, high-dose ifosfamide (14 mg/m2) in synovial sarcoma, ifosfamide/etoposide (9 gm/m2, 450 mg/m2) for malignant peripheral nerve sheath tumor, and gemcitabine/dacarbazine (1800, 500 mg/m2) for leiomyosarcoma, NCT01710176. Immunotherapy may permit improved overall survival with limited toxicity. The effectiveness of immunotherapy in the metastatic setting of soft tissue sarcomas is currently being studied. We await the results of these trials with great anticipation.