Sites of origin

Sarcomas of nonosseous tissues have been noted to arise at virtually all anatomic sites. The anatomic sites and site-specific histologic subtypes of more than 5113 sarcomas treated at a single referral institution are outlined in Figure 1. Approximately one-third to one-half of all sarcomas of nonosseous tissues occur in the lower extremities, where the most common histopathologic subtypes have traditionally been noted to include liposarcomas and the entity “UPS”, formerly termed malignant fibrous histiocytoma (MFH). With improved pathologic tools to categorize sarcomas (e.g., immunohistochemistry, DNA, and RNA analyses), it is increasingly recognized that UPS may have some features in common with poorly differentiated liposarcomas or leiomyosarcomas, as well as other histologic subtypes.²⁹ RPSs comprise 15–20% of all STSs, with well-differentiated/dedifferentiated liposarcoma and leiomyosarcoma being the predominant histologic subtypes. Visceral sarcomas make up an additional 24%, and the head and neck sarcomas approximately 4% of sarcomas.

Clinical presentation

The majority of patients with nonosseous sarcomas present with a painless mass, although pain is noted at presentation in up to one-third of cases.³⁰ Delay in diagnosis of sarcomas is common, with the most common incorrect diagnosis for extremity and trunk lesions being hematoma or “lipoma.” Late diagnosis of RPS is extremely common, as tumors in this area can grow to massive size before causing any symptoms (such as abdominal distention or psoas irritation with back or groin discomfort) or functional compromise such as hydronephrosis from ureteric obstruction.

Physical examination should include an assessment of the size and mobility of the mass. Its relationship to the fascia (superficial vs deep) and nearby neurovascular and bony structures should be noted. A site-specific neurovascular examination and assessment of regional lymph nodes should also be performed. Sarcomas rarely metastasize to lymph nodes, with those that do being limited to a few specific histopathologic subtypes. Presence of true nodal metastases should prompt the clinician to investigate whether the diagnosis of sarcoma is accurate.

Histopathologic classification

Histologic grade

Biologic aggressiveness can often be predicted based on histologic grade.³⁴ The spectrum of grades varies among specific histologic subtypes (Figure 2). In careful comparative multivariate analyses, histologic grade has been the most important prognostic factor in assessing the risk of distant metastasis and tumor-related death.^{34, 36} Several grading systems have been proposed, but there is no consensus regarding the specific morphologic criteria that should be employed in the grading of STS.

Two of the most commonly employed grading systems, both first published in 1984, are the US National Cancer Institute (NCI) system developed by Costa and colleagues and the system developed by the Federation Nationale des Centres de Lutte Contre le Cancer (FNCLCC) Sarcoma Group.^{29, 30, 37} The NCI system is based on the tumor’s histologic subtype, location, and amount of tumor necrosis, but cellularity, nuclear pleomorphism, and mitosis count are also to be considered in certain situations. The FNCLCC system employs a score generated by the evaluation of three parameters: tumor differentiation, mitotic rate, and amount of tumor necrosis. In a retrospective comparison of these two grading systems, in a population of 410 adult patients with nonmetastatic STS, univariate and multivariate analyses suggested that the FNCLCC system has a slightly better ability to predict distant metastasis and tumor-related death.³⁸ Significant discrepancies in assigned grade were observed in one-third of cases. An increased number of grade 3 tumors, reduced number of grade 2 tumors, and better correlation with overall and metastasis-free survival were observed in favor of the FNCLCC system. The FNCLCC system is the best presently available grading system and is employed as part of the AJCC/UICC STS staging system, with the caveat that several new diagnostic categories have been identified since 1984 whose histological grades are undefined by FNCLCC criteria.

In discussing grade, it is important to note well-described characteristics of sarcomas. First, there is often substantial intratumoral heterogeneity within individual sarcomas. Therefore, diagnoses based on very limited amounts of tumor may be inaccurate [e.g., diagnoses based only on fine-needle aspiration (FNA) biopsy specimens]. This is particularly true for such histopathologic subtypes as dedifferentiated liposarcomas, where one area of the tumor might have a relatively low-to-intermediate-grade appearance and another area within the same tumor might have high-grade components more evident. Any discussion of the clinical relevance of grading must take into account this variability inherent in the diagnostic process, which will add to the clinical variability in outcomes among patients with any given grade of sarcomas.

Second, the grade of tumors may evolve over time. This process is best described in the evolution of dedifferentiated liposarcoma arising in conjunction with well-differentiated liposarcoma in the same patient. Additional examples include the round-cell liposarcoma growing from what was previously myxoid liposarcoma and fibrosarcomatous degeneration that will occasionally accompany multiply recurrent DFSPs.

Imaging

Optimal imaging of the primary tumor is dependent on the anatomic site. For soft tissue masses of the extremities, trunk, and occasionally head and neck, magnetic resonance imaging (MRI) generally has been regarded as the imaging modality of choice (Figures 3 and 4) because MRI enhances the contrast between tumor and muscle and between tumor and adjacent blood vessels and provides multiplanar definition of the lesion.³⁹ However, a study by the Radiation Diagnostic Oncology Group that compared MRI and computed tomography (CT) in patients with malignant bone (n = 183) and soft tissue (n = 133) tumors demonstrated no specific advantage of MRI over CT.⁴⁰ That said, although it may be true that the diagnostic evaluation is equally served by both modalities, surgery and RT planning may require additional information provided by the multiplanar capability of MRI and the ability to perform MRI/CT image fusion.^{41, 42} For pelvic lesions or evaluation of specific fixed organs, such as the rectum or the liver, the multiplanar capability of MRI may provide superior single-modality imaging (Figure 4), whereas in the retroperitoneum and abdomen, CT usually provides satisfactory anatomic definition of the lesion. Occasionally, MRI with gradient sequence imaging can better delineate the relationship of a tumor to midline vascular structures, particularly the inferior vena cava and aorta (Figure 5). More invasive studies such as angiography or cavography are rarely used in evaluation of STS.

Cost-effective imaging to exclude the possibility of distant metastatic disease is dependent on the size, grade, and anatomic location of the primary tumor. In general, patients with low- and intermediate-grade tumors or high-grade tumors 5 cm or less in diameter require only a chest radiograph for satisfactory staging of the chest. This directly reflects the comparatively low risk of presentation with pulmonary metastases in these patients.^{43, 44} However, patients with high-grade tumors larger than 5 cm (T2) should undergo more thorough staging of the chest by CT owing to the increased risk of presentation with established metastatic disease in this group.^{44, 45} Patients with RPS and intra-abdominal visceral sarcomas should undergo imaging of the liver to exclude the possibility of synchronous hepatic metastases; the liver is a more common site of first metastasis from these lesions. CT is usually adequate in these patients to assess the liver, although the increased sensitivity of MRI of the liver may be valuable if any questionable findings are noted on initial CT.

Positron emission tomography (PET) scans may be used selectively to look for extent of disease, particularly when evaluating an ambiguous lesion that could represent a potential metastasis noted on other imaging. However, PET scans are not routinely utilized in staging work-up of STS.

Biopsy

Biopsy of the primary tumor is essential for most patients presenting with soft tissue masses. In general, any soft tissue mass in an adult that is enlarging (even if asymptomatic), is larger than 5 cm, or persists beyond 4–6 weeks should be biopsied. The preferred biopsy approach is generally the least invasive technique required to allow a definitive histologic diagnosis, assessment of grade. In most centers, core-needle biopsy provides sufficient tissue for diagnosis and results in substantial cost savings compared with open surgical biopsy.^{46, 47} When core-needle biopsy yields insufficient tissue for diagnosis, incisional biopsy is considered to yield optimal amounts of tissue to assess histopathology over a larger area of tumor volume, given the known heterogeneity of sarcomas, as well as to provide sufficient material for detailed molecular and cytogenetic assays. Direct palpation can be used to guide needle biopsy of most superficial lesions, but less accessible sarcomas often require imaging-guided biopsy for safe percutaneous sampling of the most radiographically suspicious area(s) of the mass. Tumor recurrences within the needle track after percutaneous biopsy are exceedingly rare but have been reported, leading some physicians to advocate tattooing the biopsy site for subsequent excision. FNA generally does not provide sufficient material for initial diagnosis, but can be used to confirm recurrence or metastatic disease. Exceptions to this idea exist; endoscopic ultrasound-guided FNA for visceral sarcomas such as GISTs may provide enough tissue for diagnosis while minimizing risk of tumor rupture; in this scenario, it is not feasible to assess mitotic rate. The need for sufficient tissue to conduct more specific molecular testing is a final major rationale for use of core-needle biopsy over FNA. Another major limitation of FNA (compared to core-needle biopsy) is that there is no semblance of preserved tissue architecture to evaluate characteristics such as degree of tissue necrosis.

A practical approach for biopsy and staging of the patient who presents with a primary extremity soft tissue mass is outlined in Figure 6. Small (<5 cm) superficial lesions on an extremity where the morbidity of excisional biopsy is minimal (i.e., remote from joints, tendons, and neurovascular structures that would compromise the surgical margin) are easily biopsied by excisional biopsy with microscopic assessment of surgical margins. For extremity lesions, incisions used for excisional biopsies should be oriented longitudinally along the length of the limb. T2 lesions, T1 lesions located beneath the investing fascia of the extremity, or superficial T1 lesions situated in proximity to joints, tendons, or neurovascular structures are best biopsied by percutaneous core-needle biopsy.

Staging

The relative rarity of STS, the anatomic heterogeneity of these lesions, and the presence of more than 50 recognized histologic subtypes of variable grades have made it difficult to establish a functional system that can accurately stage all forms of this disease. The staging system (7th edition) of the American Joint Committee on Cancer (AJCC) and the Union for International Cancer Control is the most widely employed staging system for STS ( Table 2).⁴⁹ The system is designed to optimally stage extremity tumors but is also applicable to torso, head and neck, and retroperitoneal lesions; a separate staging system is provided for GISTs.

Source: Edge et al. 2010.⁴⁹ Reproduced with permission of Springer.

A major limitation of the present staging system is that it does not take into account the anatomic site of STS. Anatomic site, however, has been recognized as an important determinant of outcome.^{50, 51} Therefore, although site is not a specific component of any present staging system, outcome data should be reported on a site-specific basis, when feasible. Furthermore, the staging system also fails to include histology, a critical prognostic factor.

Conventional prognostic factors

A thorough understanding of the clinicopathologic factors known to impact outcome is essential in formulating a treatment plan for the patient with STS. Several multivariate analyses of prognostic factors for patients with localized sarcoma have been reported.^52–54 However, with few exceptions, most studies have analyzed fewer than 300 patients.

The largest studies established the clinical profile of what is now accepted as the high-risk patient with extremity STS: the patient with a large (≥5 cm), high-grade, deep lesion. In addition, unappreciated prognostic significance includes specific histologic subtypes, for example, MPNST, and the increased risk of adverse outcome associated with a microscopically positive surgical margin or presentation with locally recurrent disease. The type of microscopically positive surgical margins also appears important. Patients with low-grade liposarcomas have a relatively low risk of local recurrence (LR), as do those patients in whom the positive margin is planned before surgery to preserve critical structures and RT can sterilize the small amount of residual disease. However, patients with two categories of positive margin remain at relatively higher risk of LR. These include patients who underwent “unplanned” excision and still have positive margins on re-excision and those with unanticipated positive margins after primary resection.⁵⁵ An “unplanned excision” is defined as an excisional biopsy or resection carried out without adequate preoperative staging or consideration of the need to remove normal tissue around the tumor.

Unlike for other solid tumors, the adverse prognostic factors for LR of an STS are distinct from those that predict distant metastasis and tumor-related death.⁵² In other words, patients with a constellation of adverse prognostic factors for LR are not necessarily at increased risk of distant metastasis or tumor-related death. Therefore, staging systems that are designed to stratify patients for risk of distant metastasis and tumor-related death will not necessarily stratify patients for risk of LR.

Kattan and colleagues from the Memorial Sloan-Kettering Cancer Center (MSKCC) have utilized a database of over 2000 prospectively followed adult patients with STS to predict the probability of sarcoma-specific death by 12 years.⁵¹ The results have been used to construct and internally validate a nomogram to predict sarcoma-specific death (Figure 6); this and similar nomograms have been validated in a variety of clinical situations, for example, RPS, or for disease-specific contexts, for example, liposarcoma, for individual patients.^56–59 These tools may be used for patient counseling, follow-up scheduling, and clinical trial eligibility determination.

Potential molecular prognostic factors

Specific molecular parameters evaluated for prognostic significance in STS have included TP53 mutation, MDM2 amplification, Ki-67 status, altered expression of the RB gene product in high-grade sarcomas, and histologic grade, but not SS18-SSX fusion type, which appears to be an important prognostic factor in patients with synovial sarcoma.⁶⁰ Complete discussion of the extensive literature on molecular prognostic factors in sarcoma is beyond the scope of this chapter. Readers are referred to more detailed reviews.^{61, 62}

As an example of the difficulty of using even the most commonly recognized markers as prognostic factors, one needs to look no further than Ki-67, an antigen expressed throughout the majority of the cell cycle. Ki-67 is used as a measure of the fraction of cells undergoing division. Preliminary reports of series of heterogeneous sarcomas in adults suggested that Ki-67 nuclear staining correlated with histologic grade, but was not an independent prognostic factor when histologic grade was taken into account.⁶³ Conversely, additional studies in larger numbers of patients indicated that Ki-67 status was an independent prognostic factor for clinical outcomes.⁶⁴ It is only with the development of consensus guidelines regarding the nature of Ki-67 immunohistochemistry and its interpretation that we can expect to see more careful accurate assessment of this biomarker in sarcoma outcomes.⁶⁵ It is also with the inconsistencies in the Ki-67 data that one can extrapolate the difficulty of using increasingly available genetic markers for outcome determination. Although specific protein, DNA, and RNA parameters have been identified as having independent prognostic significance, there is presently no consensus on how these prognostic factors should be used in clinical practice.

Surgery

Radiotherapy

Radiation dose and target volumes

Guidelines have recently been published on how to address the technical design of the radiation volumes and should be discussed for additional detail regarding this topic.⁸⁹

Many STS respect barriers to tumor spread in the axial plane of the extremity, such as bones, interosseous membranes, or major fascial planes. Consequently, extremity STS tend to spread longitudinally within the specific muscle groups of the extremity. Therefore, the margins of the RT volume must be wide in the cephalocaudal direction. In the cross-section, there may be much greater security in defining nontarget structures, especially those delimited by an intact barrier to tumor spread. Bone, interosseous membranes, and fascial planes are considered barriers to tumor spread in the axial direction, and, therefore, descriptions of radiation margins employed are principally in the cephalocaudal direction. For nonextremity lesions, the preferred direction of spread is also along the direction of the involved musculature, but care must be taken to ensure that the fascial planes are appropriately recognized and encompassed in the radiation target volume.

Earlier, this chapter summarized principles concerning anatomic planes and the preferential pathways for sarcomas to spread within tissues. This information facilitates the design of target areas for RT. The basic elements in RT planning are to first define a gross tumor volume (GTV) and then place a margin around it to encompass tissues at risk of harboring microscopic residual disease [clinical target volume (CTV)] (Figures 7a–c and 8).⁹⁰ Generally, RT is phased so that an initial volume (phase 1) around the risk zone is treated to doses that are capable of sterilizing microscopic amounts of tumor cells (e.g., 45–50.4 Gy in 1.8 or 2.0 Gy fractions). When delivering RT postoperatively, it is customary to have at least one field reduction to permit an augmented dose to a smaller volume surrounding the highest risk zone (phase 2). This dose is usually 15–16 Gy but can be higher if there is gross residual disease. For the phase 1 volume, the surgical bed is expanded with a 1.5-cm radial margin and a 4-cm craniocaudal margin to encompass microscopic disease in the surrounding tissues; the boost is applied to the original sarcoma localization with a 1.5-cm radial margin and a 2-cm craniocaudal margin. In preoperative RT, historically more recent but potentially the most prevalent approach used today, the GTV is treated to a dose of 50 Gy in 25 fractions over 5 weeks with surgery following 4–6 weeks later. In general, the CTV encompasses the GTV with a 4-cm craniocaudal and a 1.5–2.0 cm radial margin for microscopic disease coverage. CTV should also include peritumoral edema as it may harbor tumor cells at some distance from the GTV.⁹¹ Following preoperative RT, a postoperative “boost” has traditionally been used but is generally restricted to patients who received preoperative RT and have margin-positive disease at surgery. This is because the local control rate for margin-negative cases is in excess of 90% even when a boost is not provided.^{81, 92, 93} A positive margin is declared when the tumor reaches the inked surface of the specimen, and clear margins can be declared if the tumor does not reach the ink irrespective of how close it is.⁸¹

The need for a radiation therapy boost following preoperative RT and surgery with positive resection margins has been questioned in a retrospective review of 216 ESTS (extremity soft tissue sarcomas) patients; 52 received preoperative RT (50 Gy) alone and were compared to 41 who received preoperative RT with a postoperative boost (generally 16 Gy). A portion of the population did not receive RT at all, or received postoperative RT and were excluded (123 of 216). The postoperative boost cohort had lower 5-year LR-free rates (74% vs 90% for preoperative RT only) indicating that a postoperative boost provides no obvious advantage.⁹⁴ A similar study (n = 67) yielded almost identical results.⁹⁵ These results suggest that a benefit from a delayed postoperative boost following preoperative RT and surgery is at best debatable, and the increased risk and challenges of managing later RT morbidity (e.g., radiation-induced fractures) resulting from the higher radiation doses involved should be considered when treating STS.

The defined external beam volumes for extremity STS reflect those of the prospective Canadian Sarcoma Group randomized clinical trial discussed later.⁹⁶ However, the recently completed UK postoperative phase III randomized “VORTEX” trial (NCT NCT00423618) compared a 5-cm longitudinal margin from GTV in the standard arm to a 2-cm margin and results are awaited.⁹⁷ A completed RTOG-0630 trial (NCT NCT00589121) addressing the role of IMRT defined slightly smaller preoperative RT volumes (longitudinal 3 cm), although its results may be less easily interpreted owing to its nonrandomized nature.⁹⁸ In any event, the size of the RT volume margins is not appreciably smaller than the 4-cm longitudinal margin noted above and further outcome reports will be needed in the future to guide practice if target volumes are being reduced in the interest of normal tissue protection.

Despite the variations noted in target volume coverage, the local control rates reported for extremity STS using combinations or surgery and RT are approximately 90% and may suggest that the zone of microscopic involvement may be less than that was previously realized. Recent improvements in surgical technique may lessen the degree of intraoperative tumor dissemination, and irradiation of all surgically handled tissues, scars, and drain sites may be unnecessary. This seems particularly relevant for major centers where surgery is performed by teams with extensive experience in sarcoma management. One must also consider the possibility that case selection factors may explain apparent variations in practice between BRT and external beam RT approaches.

Sequencing of radiotherapy and surgery

The two most common methods of EBRT delivery are preoperative and postoperative RT. Preoperative RT is delivered to an undisturbed and potentially better oxygenated tumor site, which may be one reason why lower preoperative radiation doses do not appear to compromise local control.⁹⁹ Nielsen et al.¹⁰⁰ repeated RT planning in patients who had undergone preoperative RT and surgery and observed that the field size and number of joints irradiated in preoperative RT were significantly less than if the treatment had been administered postoperatively. Another advantage of preoperative RT is that it promotes collaboration between the surgical and radiation oncologists and facilitates the formulation of a coordinated management plan before any treatment.

The Canadian Sarcoma Group SR2 clinical trial represents the only prospective randomized comparison of preoperative versus postoperative RT.⁹⁶ As was anticipated, the trial showed that preoperative RT results in an increased rate of acute wound complications (WCs). On the other hand, as also anticipated, the trial also showed that postoperative delivery is associated with increased limb fibrosis, edema, joint stiffness, and bone fractures.

Long-term follow-up of patients treated in the Canadian Sarcoma Group NCIC trial (SR2) showed that, of 129 patients evaluable for late toxicity, 48% in the postoperative group compared to 32% in the preoperative group had grade 2 or greater fibrosis (p = 0.07).¹⁰¹ Edema was more frequently seen in the postoperative group (23% vs 16%), as was joint stiffness (23% vs 18%). Patients with these complications had lower function scores (all p values <0.01) on the Toronto Extremity Salvage Score and the Musculoskeletal Tumor Society Rating Scale. Field size predicted greater rates of fibrosis (p = 0.002) and joint stiffness (p = 0.006), and marginally predicted edema (p = 0.06). Acute wound-healing complications were twice as common with preoperative compared to postoperative RT. The increased risk was almost entirely confined to the lower extremity (43% associated with preoperative vs 21% with postoperative timing; p = 0.01). Of interest, additional reports, including one from the University of Texas M.D. Anderson Cancer Center, using the same criteria for classifying WCs as were used in the Canadian NCI trial, found almost identical results.⁶⁶

The influence of time interval between preoperative EBRT and surgery on the development of WC in extremity sarcoma has also been studied. While the interval had little influence, the data still suggested that the optimal interval to reduce potential WC was 4 or 5 weeks between RT and surgery.⁶⁷

In the initial report of the SR2 trial with 3.3 years median follow-up, an improvement in overall survival (OS) (p = 0.048) in the preoperative RT arm was noted and partially explained by increased deaths in the postoperative RT unrelated to sarcoma.⁹⁶ The local failure rate was identical in the 2 arms (7%) (Figure 9). However, updated results were recently presented and the preliminary survival difference had dissipated.¹⁰¹ The 5-year results for preoperative versus postoperative, respectively, were local control, 93% versus 92%; metastatic relapse-free, 67% versus 69%; recurrence-free survival, 58% versus 59%; OS, 73% versus 67% (p = 0.48); and cause-specific survival, 78% versus 73% (p = 0.64). Cox modeling showed only resection margins as significant for local control. Tumor size and grade were the only significant factors for metastatic relapse-free, OS, and cause-specific survival. Grade was the only consistent predictor of recurrence-free survival.

For the present, decisions about preoperative versus postoperative RT for extremity soft tissue sarcoma should be individualized, taking into account tumor location, tumor size, RT volumes needed, comorbidities, and risks. In general, preoperative RT provides some advantages over postoperative RT, but exposes the patient to significantly increased risks of serious postoperative WCs. A summary of the relative indications that can be used to select patients for preoperative RT is provided in Table 4. In addition, although much of the discussion about preoperative RT is focused on extremity lesions, patients with RPS tolerate preoperative RT substantially better than postoperative RT. This is because the tumor acts as a tissue expander to exclude the bowel from the RT volume (Figure 7a–c). This is discussed in detail later in the section titled “Retroperitoneal Sarcomas.”

Treatment context/sarcoma site	Issues of concern	Comments
Head and neck
Paranasal sinus	Proximity to optic apparatus (eye, orbit, chiasma)	Major visual functional deficit can be minimized
Skull base	Proximity to spinal cord, brain stem	Other “lesser” morbidities (dental, xerostomia) may also be less due to reduced doses and volumes
Cheek and face	Xerostomia	Early caries or loss of teeth, loss of sense of taste
Split thickness skin graft reconstruction (especially lower limb)	Skin graft breakdown and consequent infection	Many months to years of recreational and/or vocational disability may occur during healing (rare)
Large volume GTV or CTV occupying coelomic cavities
Retroperitoneum	Proximity to the bowel, the liver, and the kidney	Critical organs may be displaced by tumor or not fixed or adherent as is likely in postoperative setting
		Entire tumor treated before possible contamination of cavity
Some small bowel lesions	Proximity to critical anatomy, especially intestine with side wall adherence	Contamination of abdominal cavity renders postoperative RT unsuitable
Thoracic wall/pleura	Proximity to the lung or the cardiac structures	The lung may be displaced by the chest wall or pleural tumor and can be avoided with preoperative RT, or permits GTV to be treated before operative contamination
Abdominal trunk walls, pelvic side wall	Proximity to the kidney, the bowel, the liver, and the ovaries	Avoid CTV encroachment on vulnerable anatomy
GTV adjacent to dose limiting critical anatomy
Thoracic inlet/upper chest	Proximity to brachial plexus	Dose limitation of critical anatomy lends itself to preoperative wall low neck RT.
Additional volume considerations
Medial thigh (young male)	Proximity to testes	Permanent infertility may be avoided
Central limb tumor	Proximity to other compartments	Permits partial circumferential sparing, which would not be feasible in postoperative setting

Abbreviations: CTV, clinical target volume; GTV, gross tumor volume; RT, radiotherapy.

Source: O’Sullivan et al. 1999.⁷⁷ Reproduced with permission of Elsevier.

Intensity-modulated radiotherapy

Target coverage and protection of normal tissues from high-dose areas appear to be superior for IMRT compared to traditional techniques in ESTS.¹⁰² A recent retrospective review spanning noncoincident treatment time periods compared surgery combined with either IMRT (n = 165) or conventional EBRT (n = 154). Allowing for known limitations associated with studies involving treatments deployed over different eras, IMRT showed significantly reduced LR for primary ESTS (7.6% LR for IMRT vs 15.1% LR for conventional RT; p = 0.02).¹⁰³

Two prospective phase II trials [from Princess Margaret (NCT00188175) and the Radiation Therapy Oncology Group (RTOG 0630: NCT00589121)] investigated if preoperative image-guided radiotherapy (IGRT) using conformal RT/IMRT could reduce RT-related morbidities.^{98, 104} The characteristics of the Princess Margaret (PMH) and RTOG 0630 trials differed in several ways, particular relating to the exclusion of upper extremity lesions in the PMH trial, the use of a boost following preoperative RT in RTOG 0630 as well as the potential to use chemotherapy in the RTOG trial, and some aspects of the choice of target coverage mentioned below. The two trials also differed regarding their primary endpoints.

The PMH trial showed reduced wound-healing complication (WC) rates (31%) in lower extremity compared to the 43% risk in the previous Canadian Sarcoma Group NCIC SR2 trial that only used 2D and 3D RT.⁹⁶ The need for tissue transfer, RT chronic morbidities, and subsequent secondary operations for WCs were reduced while maintaining good limb function and local control (93%). The RTOG-0630 trial reported a significant reduction of late toxicities in comparison to the NCIC-SR2 trial (11% vs 37% in SR2), which is very similar to the IGRT PMH trial. Importantly, both the PMH and RTOG-0630 trials defined the CTV differently (longitudinal margin of 3 cm from the gross tumor for high-grade lesions and 2 cm for low-grade lesions versus 4 cm longitudinal margins in the PMH trial). Potentially, the reduction in CTV margins of this degree could explain the improvement in limb function with comparable local control, although an alternative possibility is the reduction in normal tissues receiving the target dose in all dimensions, which is shared by both studies. In the end, it also seems that IMRT is capable of conforming the dose more suitably to the desirable target volume compared to traditional conformal techniques.

Perhaps the greatest advantage to this approach is not only the possibility of improving local control but also the ability to spare bone toxicity and late fractures by achieving bone avoidance, which are often overlooked in the discussion of combined-modality treatments of extremity sarcoma. A recent study addressed evidence-based dose volume bone avoidance objectives for IMRT planning in 230 patients (176 lower and 54 upper extremity) with a median follow-up of 41.2 months.¹⁰⁵ The overall risk of fracture was 2% (4/230 patients), which compares favorably to a previous reported incidence of 6%, and suggests that efforts to achieve bone avoidance are appropriate.

Brachytherapy

BRT has some putative advantages over external beam, including a shorter overall treatment time (4–6 days vs 5–6.5 weeks) and quicker initiation of RT after surgery while clonogenic numbers are at a minimum. Because of its brevity, BRT is also more easily integrated into protocols that include systemic chemotherapy than is external beam RT, with its protracted courses. The irradiated volume is also smaller with BRT, which may confer functional advantages. BRT may also have an advantage in situations in which normal tissue tolerance to RT is compromised. One such scenario would be when a postoperative RT boost to the operative bed is desired in patients who received preoperative RT. The use of BRT with surgery in previously irradiated tissues is another situation to achieve limb salvage.^{106, 107} As noted earlier, no apparent benefit for BRT over surgical excision alone is evident with low-grade lesions, and external beam appears more effective for these tumors (Table 3).^{69, 70, 108} BRT also permits radiation volumes to be mapped according to intraoperative findings. The American Brachytherapy Society Guidelines differ from those for EBRT and also advise that BRT as a sole treatment modality is contraindicated in the following situations: (1) the CTV cannot be adequately encompassed by the implant geometry, (2) the proximity of critical anatomy, such as neurovascular structures, is anticipated to interfere with meaningful dose administration, (3) the surgical resection margins are positive, and (4) there is skin involved by tumor.^{109, 110}

BRT seems less useful where implant geometry is not optimal, such as in the upper extremity or more proximal limb regions.¹¹¹ The results of the BRT randomized trial were discussed earlier. In addition, BRT was compared retrospectively to IMRT in 134 high-grade ESTS with similar adverse features.¹¹² The 5-year local control rate was 92% for IMRT compared to 81% for BRT (p = 0.04). Unfortunately, while the results of BRT, including its more restrictive criteria for use, suggest lower efficacy compared to EBRT, there is no randomized controlled trial comparing these seemingly effective local adjuvants.

Traditionally BRT studies, including those mentioned above, used low-dose rate techniques. High-dose rate (HDR) BRT has potential logistic advantages including lower radiation staff exposure, outpatient delivery, and optimized dose distributions by varying dwell times. However, wound-healing complications may occur in sarcoma management and caution is also recommended when placing catheters adjacent to neurovascular structures. As yet, no large series evaluating HDR BRT for STS is available nor has it been directly compared to LDR, partly because of technical differences.

Additional approaches to RT delivery

In addition to external beam RT, BRT, and IMRT, several other approaches for RT delivery exist. These include particle beam RT (electrons, protons, pions, or neutrons), intraoperative radiotherapy (IORT) using external beam or BRT approaches, and combinations of other techniques (e.g., hyperthermia) with RT. IORT has been used most often in the management of RPS and will be discussed later. Some reports also describe IORT for extremity sarcomas.^{113, 114} Formal clinical trials have not compared the relative merits of these approaches, and their use may be governed as much by the availability of an approach at a given center as by any special advantage that it may confer. In the case of proton beam RT, its ability to achieve accurate targeting provides an advantage when tumors lie in proximity to critical structures.^{115, 116} In general, however, although reports on the use of many of these approaches exist, the problems of selection bias need to be considered in interpreting these small series in which treatments were not randomly assigned.^{117, 118}

Systemic therapy

Systemic agents, including both traditional cytotoxic chemotherapy drugs and newer small molecule oral kinase inhibitors (SMOKIs), are used widely in the metastatic setting for patients with STSs. The use of chemotherapy in the adjuvant setting remains somewhat controversial. However, if chemotherapy is going to have the same impact as radiation and surgery in the management of sarcomas, more effective drugs must be identified to help improve the cure rate for patients with primary tumors and unseen microscopic metastatic disease. This section will review the use of chemotherapy in the adjuvant and metastatic settings. A brief discussion of chemotherapy combined with radiation therapy is also included in this section.

Adjuvant systemic therapy following primary surgical resection

Although local or local–regional recurrence is a problem for a small subset of patients following primary therapy, the major risk to life in sarcoma patients is uncontrolled systemic disease. The availability of systemic therapy with proven, albeit often limited, ability to induce shrinkage of advanced sarcomas has raised the question of whether the early use of systemic treatment might affect microscopic metastatic disease and yield improvements in OS and disease-free survival (DFS).

Certainly for Ewing sarcoma/PNET, rhabdomyosarcoma, and osteogenic sarcoma, adjuvant or neoadjuvant chemotherapy is an appropriate standard of care.^119–122 However, for more common STSs such as leiomyosarcoma, liposarcoma, and high-grade UPS (formerly known as MFH), the benefit of chemotherapy, if there is one, is small.¹²³ As adjuvant therapy is utilized by many practitioners for more common diseases where the benefit is a relatively small one, such as stage I breast cancer and stage II colon cancer, this small potential benefit is an issue that needs to be discussed on an individualized basis. Certainly, the lack of available effective agents for metastatic sarcoma has impeded progress in this area, but the utility of imatinib in both the metastatic and adjuvant setting in GIST gives hope that new agents will contribute to the ultimate goal of any type of systemic therapy specifically increasing the cure rate for people with new diagnoses.

There have been over a dozen studies of anthracycline-based adjuvant chemotherapy for STSs that date back nearly as long as the initial development of doxorubicin.^{124, 125} These will not be reviewed here, as anthracycline/ifosfamide-based therapy constitutes a better standard of care in patients offered adjuvant chemotherapy, and only one of the studies completed by 1992 had used ifosfamide.

In one of the largest combination chemotherapy studies, the Italian Sarcoma Study Group (ISSG) examined patients with primary or recurrent resected STS of the extremity or limb girdle treated or not treated with radiation.^{126, 127} A total of 104 patients were randomized to receive no chemotherapy or to receive ifosfamide (1800 mg/m²/day for 5 consecutive days with mesna) and epirubicin (60 mg/m² on 2 consecutive days), with filgrastim support. Interim analysis in 1996 led to early conclusion of the trial because the study had reached its primary endpoint, specifically improved DFS. At a median follow-up of 36 months, OS in the chemotherapy arm was 72%, compared with 55% for the control arm (p = 0.002). However, with long-term follow-up, the OS difference showed only a trend to statistical significance on an intention-to-treat analysis.¹²⁶ This study is the strongest single study in the literature supporting the use of adjuvant chemotherapy for STS. Interpretation of the study is made more difficult with the observation of equivalent distant ± LR rates at 4 years.

Conversely, no survival benefit was observed in an EORTC phase III study of adjuvant chemotherapy versus observation (doxorubicin 75 mg/m², ifosfamide 5 g/m² per cycle for 5 cycles, with filgrastim support).¹²⁸ A total of 351 patients were accrued between 1995 and 2003, and 130 (80%) of the 163 patients receiving chemotherapy completed all 5 cycles. OS was not statistically different between arms (HR 0.94, 95% confidence interval 0.68–1.31 for the treatment arm, p = 0.72) and relapse-free survival was also not statistically different (HR 0.91, CI 0.67–1.22, p = 0.51). The 5-year OS rate was 67% for the treatment arm and 68% for the control group. The major differences between this study and the ISG study were a lower dose of ifosfamide and use of epirubicin in the ISSG trial, but it is unclear based on data from metastatic disease of the relevance of these differences.

The most recent and comprehensive overview of adjuvant chemotherapy for extremity sarcomas to date was the 2008 meta-analysis of 18 studies encompassing sarcomas of all anatomic sites.¹²⁹ In this analysis, 93 potential studies were considered and 18 ultimately selected, constituting 1953 patients with STSs of extremity and nonextremity. Pathology review was not centralized. The results of the meta-analysis, including the actuarial outcome probabilities and the hazard ratios, are summarized in Figure 10 and Table 5.

	Doxorubicin	Doxorubicin–ifosfamide	Combined
Local RFI	0.75 (0.56–1.01)	0.66 (0.39–1.12)	0.73 (0.56–0.94)
Distant RFI	0.69 (0.56–0.86)	0.61 (0.41–0.92)	0.67 (0.56–0.82)
Overall RFS	0.69 (0.56–0.86)	0.61 (0.41–0.92)	0.67 (0.56–0.82)
Overall survival	0.84 (0.68–1.03)	0.56 (0.36–0.85)	0.77 (0.64–0.93)

Source: Pervaiz et al. 2008.¹²⁹ Reproduced with permission of Wiley.

Study data from the 18 trials were combined without examining individual patient data. Combining all data, LR risk, distant recurrence risk, overall recurrence risk, and OS were superior with chemotherapy. For OS, the relative risk of death was 0.77 with chemotherapy (95% CI, 0.64–0.93), p = 0.01, and relative recurrence risk was 0.67 (95% CI, 0.56–0.82), p = 0.0001. The absolute risk reduction of any recurrence was 10% and absolute risk reduction for OS was 6%. However, as pointed out in a commentary on a prior meta-analyses,¹³¹ these data have to be interpreted with caution. For example, (1) individual patient data were not reviewed and interrogated, (2) in an older analysis, 18% of patients did not have histology available for review, recognizing the error rate in pathology review at expert centers, (3) ineligibility rates were high, and (4) the largest individual trial was not included (published after the meta-analysis publication). Although the meta-analysis cannot replace a well-designed randomized study, it reinforces many of the findings from smaller studies that local and distant recurrence-free survival is definitely improved, and OS modestly so, at least in unselected patients.

Combined preoperative chemotherapy and RT

Data regarding combination chemotherapy and radiation are more developed than hyperthermia and chemotherapy, largely from the commonality of the former being available at various institutions. As with combinations with hyperthermia, the primary putative advantage of preoperative chemotherapy with radiation is the potential to shrink selected lesions resectable only by amputation sufficiently that they become amenable to a limb-sparing approach.

Concurrent chemotherapy and radiation has been employed extensively by Eilber and colleagues at UCLA and has been modified and examined by other groups.^{139, 140} The first chemotherapy–radiation treatment protocol typically involved intra-arterial doxorubicin with high dose per fraction RT (35 Gy of external beam radiation delivered in 10 daily fractions, which was reduced to 17.5 Gy in 5 daily fractions to minimize local toxicity). Although the intra-arterial route delivers chemotherapy more directly to the tumor, it is more complex, expensive, and prone to complications than intravenous chemotherapy.¹⁴¹ Indeed, a prospective randomized trial comparing preoperative intra-arterial doxorubicin with intravenous doxorubicin, both followed by 28 Gy of radiation delivered over 8 days and then surgical resection, showed no differences in LR or survival.¹⁴⁰

The largest study to date directly studying systemic chemotherapy combined with RT examined razoxane as the radiation-sensitizing agent in a randomized study of drug versus no drug in combination with radiation therapy for resectable or unresectable STSs.¹⁴² Acute skin reactions were enhanced in the razoxane arm, but late toxicity was not greater than in the control arm. Although there are imbalances in the arms of the study, for the 82 of 130 evaluable cases examined with gross disease RT (median dose 56–58 Gy) with razoxane (daily oral doses of 150 mg/m² throughout RT) showed an increased response rate (74% vs 49%) and improved local control rate (64% vs 30%; p < 0.05) compared with external beam radiation alone.

Either ifosfamide or cyclophosphamide has been routinely combined with radiation therapy as part of the definitive therapy for Ewing sarcoma and rhabdomyosarcoma in an attempt to continue systemic therapy at the same time as maximizing local control.¹¹⁹ In general, toxicity does not appear to be greater than that seen for radiation alone. However, skin toxicity from the combination was greater in one study than that seen with radiation alone.¹⁴³

An alternative sequential chemotherapy and radiation strategy in patients with localized, high-grade, large (>8 cm) extremity STSs has been examined.^144–146 This treatment protocol involved 3 courses of doxorubicin, ifosfamide, mesna, and dacarbazine (MAID) with two 22-Gy courses of radiation (11 fractions each) for a total preoperative radiation dose of 44 Gy. This was followed by surgical resection with careful microscopic assessment of surgical margins. An additional 16-Gy (8 fraction) boost was delivered for microscopically positive surgical margins. The outcomes of 48 patients treated with this regimen were compared with those of matched historic controls and was superior to that of the historical control patients.¹⁴⁴ The 5-year actuarial local control, freedom from distant metastasis, DFS, and OS rate were 92% versus 86% (p = 0.1155), 75% versus 44% (p = 0.0016), 70% versus 42% (p = 0.0002), and 87% versus 58% (p = 0.0003) for the MAID and control patient groups, respectively. Febrile neutropenia was a complication in 25% of patients. Wound-healing complications were substantial and occurred in 14 (29%) patients receiving the chemotherapy/radiation sequential therapy. One patient who received chemotherapy developed late fatal myelodysplasia. Given the favorable results of this study in comparison to historical controls for high-risk extremity STS, the Radiation Therapy Oncology Group conducted a multi-institutional trial, modifying the chemotherapy in an attempt to address the local toxicity issue. The report of the trial suggested combined-modality treatment can be delivered successfully in a multi-institutional setting albeit with some toxicity. Efficacy results are consistent with previous single-institution results.^{145, 146} The question remains whether this approach in high-risk, extremity STS confers significant survival benefits following and intense regimen of neoadjuvant chemoradiotherapy and surgery, which seems sustained even with long-term follow-up.

Although significant toxicity was observed, local control was improved in the prior studies compared to historical controls, raising the possibility that combined chemotherapy and radiation could be combined safely to decrease local control risk. Pisters et al.¹³² examined concurrent doxorubicin and irradiation in the neoadjuvant setting in 27 patients with extremity STS. Preoperative external beam radiation was administered in 25 fractions of 2 Gy each. Doxorubicin was administered in escalating doses with a bolus followed by 4-day continuous infusion weekly. Radiographic restaging was performed 4–7 weeks after chemoradiation. Patients with localized disease underwent surgical resection. The maximum tolerated dose of continuous infusion doxorubicin combined with standard preoperative radiation was 17.5 mg/m²/week; 7 of 23 (30%) patients had grade 3 dermatologic toxicity at this dose level. Macroscopically complete resection (R0 or R1) was performed in all 26 patients who underwent surgery. In 22 patients who were treated with doxorubicin at the maximum tolerated dose and subsequent surgery, an encouraging 11 patients (50%) had 90% or greater tumor necrosis, including 2 patients who had complete pathologic responses. This approach appears valid with other radiation-sensitizing agents as well, such as gemcitabine.¹⁴⁷ Further studies of combination therapy are also discussed later in the section titled “Retroperitoneal Sarcomas.”

Other multicenter trials of adjuvant therapy for STS

It is well recognized that different sarcoma subtypes have different chemotherapy resistance/sensitivity patterns, details of which are discussed in greater detail elsewhere.⁶¹ For example, synovial sarcoma is typically resistant to gemcitabine–docetaxel, and leiomyosarcomas are typically less sensitive to ifosfamide than other forms of sarcoma. Synovial sarcoma and myxoid/round-cell liposarcoma appear to be more sensitive to chemotherapy in the metastatic setting than other subtypes of sarcoma and may well be two subtypes that respond to both anthracyclines and ifosfamide. These data argue that adjuvant chemotherapy should be examined on a subtype-specific basis. Combined nonrandomized data from UCLA and MSKCC showed that adjuvant chemotherapy may indeed be useful in the setting of synovial sarcomas and myxoid/round-cell liposarcoma and argued for the use of chemotherapy in the neoadjuvant setting for all types of STS, based on institutional databases.^148–150 Notably, data of all patients treated or not treated with chemotherapy from MD Anderson and MSKCC indicated in the adjuvant or neoadjuvant setting that there was no statistical difference in OS in the group of patients who received chemotherapy versus those who did not.¹⁵¹ However, these data are inherently biased in that it is likely that younger, healthier patients with higher risk tumors were those selected to receive chemotherapy. Even though there was no statistically significant difference between the group of patients who received chemotherapy and those who did not, there were still shifts in the frequency of patients with larger sarcomas with a predominance of liposarcoma toward receiving chemotherapy, perhaps representing the selection bias that allowed a group of patients with an inherently poorer outcome to do as well as those with a better outcome.

In summary, for AJCC stage III STS, if there is a benefit to chemotherapy in the adjuvant setting, it appears to be a modest one. There is variation in practice between centers and between practitioners as to who is an appropriate candidate for chemotherapy. Given this situation, it is the authors’ practice to attempt to compare benefits and risks of systemic therapy and individualize the plan of treatment for a given clinical setting. Patients below age 50 may be those who benefit most among this very heterogeneous patient population. Certainly, the finding of 35% or more of STS with specific genetic translocations or mutations brings hope that the benefit seen with GIST and adjuvant imatinib therapy will carry over to the adjuvant setting for a subset of patients with STS of the extremities and trunk when traditional adjuvant cytotoxic chemotherapy is combined with novel therapeutics.

“Pediatric” sarcomas and adjuvant therapy

The standard of care for nearly all sarcomas specific to the pediatric population involves chemotherapy. It is of proved benefit in patients with osteogenic sarcoma, Ewing sarcoma, and rhabdomyosarcoma. A few brief comments on these chemotherapy-responsive tumors follow.

Neoadjuvant chemotherapy is the standard of care for the initial treatment of osteogenic sarcoma.¹²¹ In the era before systemic therapy, cure rates for osteogenic sarcoma, even in the setting of amputation for primary disease, was only on the order of 15%. With chemotherapy, the survival rate is 65–70%. Yet the 30–35% of patients who relapse remain a frustrating problem because the addition of new agents has changed the chance for cure comparatively modestly, and the one agent demonstrating benefit in a randomized trial was not approved in the United States.¹²⁰ The outcome of adjuvant chemotherapy (degree of tumor necrosis after chemotherapy) is directly associated with improved clinical outcome and provides the opportunity for changing chemotherapy for a poor initial response. The standard of care for adjuvant chemotherapy is a backbone of cisplatin and doxorubicin, with methotrexate employed in most pediatric and some adult patients. The benefit for methotrexate as part of adjuvant treatment was called into question in a 1997 paper in which the combination of doxorubicin and cisplatin alone was shown equivalent to a more complicated and toxic regimen containing high-dose methotrexate.¹⁵² However, the methotrexate component of neoadjuvant therapy has been observed to be important in a variety of studies and remains an integral part of combination therapy for osteosarcoma.¹⁵³ Less questionable is the benefit from nonspecific immune system stimulator muramyl tripeptide (MTP), which was shown in a large pediatric clinical trial to improve OS when used in the adjuvant setting, while there was no benefit in survival seen with the addition of ifosfamide to the standard methotrexate–doxorubicin–cisplatin (MAP) backbone.¹²⁰

Through a series of international clinical studies, the adjuvant program for rhabdomyosarcoma typically involves an induction course of chemotherapy, followed by combination chemotherapy and radiation, followed by the completion of chemotherapy, which will last approximately 48 weeks in pediatric population. The standard of care is the combination of vincristine, dactinomycin, and cyclophosphamide, as this combination was shown as effective as VAI and VIE and less toxic.¹¹⁹ The addition of doxorubicin to the vincristine, dactinomycin, and cyclophosphamide regimen did not appear to improve OS and is omitted in the treatment of pediatric rhabdomyosarcoma.¹⁵⁴ The frequent dosing of vincristine is extremely difficult to complete for adults and requires dose adjustment or shorter courses of therapy for adults than for children with this diagnosis. Children with this diagnosis appear to fare better than adults with the same diagnosis stage in most studies, as well as in everyday practice.^{155, 156}

For patients with Ewing sarcoma, in distinction from rhabdomyosarcoma, the addition of additional agents (ifosfamide and etoposide) to an existing backbone of vincristine, doxorubicin, and cyclophosphamide chemotherapy improved outcome for localized disease, with a 2-week schedule superior to a 3-week schedule of treatment;¹²² however, not published were data that the 2-week schedule was not beneficial for patients over age 18. This 5-drug regimen is a good standard of care for patients with a new diagnosis of Ewing sarcoma. It is often difficult to administer the 14 cycles of chemotherapy to adults, and abbreviation to the adjuvant therapy program is often necessary. It is also not clear if all 14 cycles are necessary for best outcomes. As with rhabdomyosarcoma, children with Ewing sarcoma fare better than adults with the same diagnosis, all else being equal.^157–159

Hyperthermic isolated limb perfusion, isolated limb infusion, and regional hyperthermia

Hyperthermic isolated limb perfusion (ILP), an investigational technique in the United States (although recently approved by regulatory agencies in other parts of the world), has received considerable attention in the treatment of locally advanced, unresectable sarcomas of nonosseous tissues. ILP involves local perfusion of high-dose chemotherapy (most commonly melphalan) and, when available, tumor necrosis factor alpha (TNFα) under hyperthermic conditions. An oxygenated circuit is established by local arterial and venous cannulation on a bypass pump. Systemic circulation is minimized by placement of a tourniquet proximally.

ILP has been evaluated in two settings: (1) attempted limb preservation in cases of locally advanced extremity lesions surgically amenable only to amputation and (2) function extremity preservation for the short survival duration anticipated in cases of locally advanced extremity lesions with synchronous pulmonary metastases (stage IV disease).

A multicenter phase 2 trial evaluated a series of 55 patients with radiologically unresectable extremity STS using HILP with high-dose TNFα and melphalan, and interferon-α in some patients. ¹⁶⁰ A major tumor response was seen and limb salvaged in over 80% of patients. Regional toxicity was limited, and systemic toxicity was minimal to moderate. There were no treatment-related deaths. Despite the high rate of complete responses (15–30%) and limb sparing (>80%) achieved by ILP, no randomized trials have compared ILP to aggressive limb-sparing resection with RT for STS. Therefore, ILP should be considered as a potential treatment, when other options are limited or not available, in appropriately selected patients. Eligible patients should be referred to centers where this therapy is available.

Isolated limb infusion (ILI) has been evaluated in extremity STS in a more limited manner, and parallels work done with intra-arterial chemotherapy. Like IPL, ILI relies on circulating high-dose chemotherapy in an isolated extremity. Unlike ILP, ILI is conducted through percutaneously placed catheters and is performed under hypoxic conditions. However, there is limited experience with ILI in extremity STS. Similar to ILP, ILI has not been directly compared to aggressive limb-sparing resection with EBRT in a randomized trial. However, patients under consideration for ILP and ILI are often no candidates for surgery with EBRT at first evaluation, and therefore ILP and ILI may be considered as potential therapies in appropriately selected patients.

Definitive radiation for local control

Apart from patients with some very radiosensitive subtypes of sarcomas, most patients who undergo RT as the sole treatment modality for sarcoma have been deemed to have locally advanced unresectable disease. RT alone is a rare treatment choice that should be done only at centers skilled in the management of sarcomas; medically fit patients with grossly “unresectable” but nonmetastatic disease should always be referred to a specialty center for multidisciplinary management, which may combine surgery, RT, and possibly chemotherapy. For example, proximal inguinal or axillary tumors that encircle major vascular structures in the proximal leg or arm may be resected along with the involved vasculature and the vessels reconstructed. Adjuvant RT is also generally used. Rarely, a patient with truly inoperable locally advanced disease may require RT alone, with either photon or particle (proton, neutron, or pion) beams.^161–163 No formal clinical trials have been performed to compare these strategies with each other, and they are generally administered in an adverse clinical setting. Local control has been reported in 40–70% range.

Clinical problem of metastases

The diagnosis of recurrent or metastatic disease in patients with STSs is often heartbreaking. Patients and physicians are aware that, in general, such a diagnosis is typically fatal. The role of the multidisciplinary sarcoma team in the management of patients with metastatic sarcoma is to recognize opportunities in which multimodality care might still improve important outcomes such as survival or quality of life. Both surgery and systemic chemotherapy can play an important role in improving these outcomes in selected patients. Overall, it is important to recognize that chemotherapy is usually given with the palliative aim of prolonging life and improving quality of life.

The most common site of metastasis from STS of the extremities is the lungs. Indeed, the lungs are the only site of metastasis in approximately 80% of patients with metastases from primary extremity and trunk STS.⁶¹ Primary visceral and gastrointestinal sarcomas such as GIST commonly metastasize to the liver, while other visceral sarcomas may metastasize to lungs as well. Extrapulmonary metastases are uncommon forms of first metastasis from extremity sarcomas and usually occur as a late manifestation of widely disseminated disease. The median survival after development of distant metastases is approximately 12 months (Figure 11),¹⁶⁵ although more contemporary data suggest that median survival may presently be 15–18 months for unselected patients who receive cytotoxic chemotherapy; the optimal treatment of patients with metastatic STS requires an understanding of the natural history of the disease and individualized selection of treatment options based on patient factors, disease factors, and limitations imposed by prior treatment.

The approach to patients with advanced or metastatic sarcomas is changing over time. We increasingly realize clinical trials must be stratified rationally for data of value to be derived. Studies of “sarcomas” without stratification by histology or molecular features will soon seem as naïve as studies of “cancer” without further qualification. These mesenchymally derived diseases lumped under the heading of “sarcomas” can be quite different, and studies need to take that into account. In a similar manner, the consideration of a specific sarcoma histology may be supplemented or even superseded by the genetic profile of the tumor. To generate studies of sufficient size and power, large-scale collaborations on a national and international level will be required. Such collaborations are already in place among the nations of Europe individually and collectively (e.g., the Soft Tissue and Bone Sarcoma Group of the EORTC), Canadian centers, and cooperative trials groups and other clinical trials collaborations such as SARC (Sarcoma Alliance for Research through Collaboration) in the United States. With these collaborations, it is hoped that further research will rapidly translate research findings into the novel therapeutics that are so desperately required by patients with sarcomas.

Resection of metastatic disease

Many investigators have reported their experience with pulmonary metastasectomy for metastatic STS in adults.^{166, 167} Three-year survival rates following thoracotomy for pulmonary metastasectomy range from 23% to 54%. As a result, some patients can be resected with curative intent, similar to the situation for osteogenic sarcoma. As the ability to completely resect all metastatic disease is an important determinant of outcome, the reported interinstitutional variability in postmetastasectomy survival rates is partially a function of whether survival was reported for all patients who underwent thoracotomy or only for the subset who underwent complete resection.

It remains difficult to predict which patients with pulmonary metastases will benefit from pulmonary resection. A number of clinical criteria have been evaluated by univariate analysis in this regard, including the disease-free interval, number of metastatic nodules, and tumor doubling time. Multivariate analyses from both the NCI and Roswell Park Cancer Institute confirm that a short disease-free interval and incomplete pulmonary resection are adverse prognostic factors for survival for patients with pulmonary metastases.^166–169 A multivariate analysis from MD Anderson suggested that, in addition, the presence of more than three metastatic pulmonary nodules on preoperative chest CT is an adverse prognostic sign. The most important prognostic factor impacting survival appears to be the ability to completely resect all disease. In the review of postmetastasectomy outcomes in one series, the median survival among patients who were able to undergo complete resection of metastases was 20 months as compared with 10 months among patients who did not have complete resection (Figure 12).¹⁶⁵ In summary, the ability to achieve complete resection and the number of pulmonary nodules present appear to best define the postoperative prognosis for these patients.

Unfortunately, metastasectomy benefits only a fraction of patients who develop pulmonary metastases. This is best illustrated by data from MSKCC, where a population of 716 patients who presented with primary extremity sarcoma were followed for the subsequent development and treatment of pulmonary metastases (Figure 13).¹⁷⁰ Of an initial group of 716 patients, 148 patients (21%) developed pulmonary metastases. Isolated pulmonary metastases occurred in 135 (91%) of these 148 patients. Of the 135 patients with pulmonary-only metastases, 78 (58%) were considered to have operable disease, and 65 (83%) of those taken to thoracotomy were able to undergo complete resection of all of their pulmonary metastatic disease. Thus, 44% of all patients with pulmonary metastases were able to undergo complete metastasectomy. The median survival from the time of complete resection was 19 months, and the 3-year survival rate was 23%. All patients who did not undergo thoracotomy died within 3 years. For the entire cohort of 135 patients developing pulmonary-only metastases, the 3-year survival rate was only 11%.

Several series of repeat pulmonary metastasectomy also have been published. In a series of 43 patients thus treated at the NCI, 72% of patients could be rendered free of disease at the second thoracotomy, with a median survival duration from the time of second thoracotomy of 25 months.¹⁷¹ In a report from MD Anderson of a series of 39 patients undergoing reoperation for a second pulmonary metastasis after successful initial metastasectomy, factors predicting long-term survival included the presence of a solitary metastasis and the ability to perform a complete resection.¹⁷² Patients with isolated second metastatic sites fared better than those with multiple resected lesions.

The disappointing overall results of treatment of metastatic disease underscore the importance of careful patient selection for resection of pulmonary metastases. The following criteria are generally agreed upon: (1) the primary tumor is controlled or is controllable, (2) there is no extrathoracic disease, (3) the patient is a medical candidate for thoracotomy and pulmonary resection, and (4) complete resection of all disease appears possible. With careful patient selection, the morbidity of thoracotomy (or repeated thoracotomies) can be limited to the subset of patients who are most likely to benefit from this aggressive treatment approach. The potential role of systemic adjuvant chemotherapy following complete metastasectomy is discussed below in the section “Individualized Therapy” on individualizing chemotherapy for metastases.

Chemotherapy for metastatic disease

Strategies to improve the therapeutic index of chemotherapy

Pre- and postoperative radiotherapy approaches

A variety of adjuvant RT approaches have been used for RPS. One approach, evaluated in a prospective randomized trial, used an IORT boost (20 Gy) to the tumor bed followed by postoperative external beam (35–40 Gy); this approach was compared with conventional postoperative RT (50–55 Gy). In this study of 35 patients, the incidence of local–regional recurrence was lower in the experimental treatment arm, but no improvement in survival was demonstrated.²³⁶ IORT was associated with a high rate of peripheral neuropathy when large, sometimes overlapping, RT portals were used to cover the sacral plexus region. However, gastrointestinal complications were more common in the control group, in whom higher doses were delivered to the bowel.

Other researchers have investigated strategies employing preoperative RT. Gieschen et al.²³⁷ from the Massachusetts General Hospital reported on 37 patients who underwent preoperative RT, resection, and then, when feasible, electron beam IORT. The grossly complete resection rate was 83%, and the 5-year actuarial OS, DFS, local control, and freedom from distant disease rates were 50%, 38%, 59%, and 54%, respectively. Earlier reports from the same group described a 70% complete resection rate and 81% 4-year local control rate. The more recent report also indicates that complete resection and IORT improved OS and local control rates (74% and 83%, respectively) compared with no IORT (30% and 61%, respectively). However, the study was not randomized, and potentially more favorable and accessible lesions may have been chosen for IORT. Petersen and colleagues at the Mayo Clinic found improved local control, at least of primary tumors, using a similar treatment approach.²³⁸

The University of Toronto Sarcoma Group described an unusually favorable outcome, especially for primary lesions, in a single-arm prospective trial of preoperative RT (median dose of 45 Gy in 25 fractions) plus postoperative BRT in selected cases.²³⁹ Of interest, acute toxicity resulting from preoperative RT was differentiated prospectively from the effects of other treatments. Although the median radiation volume exceeded 7 L, preoperative external beam RT was associated with extremely low gastrointestinal toxicity scores. Furthermore, no patient was hospitalized for acute toxicity, and there were no treatment interruptions or cessations of treatment because of acute toxicity. The remarkably low toxicity of the preoperative RT with enormous volumes has been attributed to the displacement of the bowel outside the target volume. In contrast, the selective use of postoperative BRT was associated with toxicity, and there was little evidence that BRT contributed to the enhanced tumor outcome reported.

The University of Texas MD Anderson Cancer Center and Princess Margaret Hospital groups combined the data from their phase 2 and pilot studies. This pooled analysis demonstrated very favorable OS rates for patients treated with preoperative radiation combined with surgical resection.²⁴⁰ However, we again caution against overinterpretation of results that could be explained by surgical technique at major referral centers or by case selection. Similarly, it remains unclear what contribution was provided by the BRT or IORT, which should probably remain protocol based or be reserved for nonstandard use in selected cases. Notably, however, the preoperative RT approach seems to continue to demonstrate favorable local control and OS in very long-term follow-up of the PMH study.

The results of these three reports add credence to the advantages posited for preoperative RT in RPS: (1) the tumor bulk often displaces the small bowel from the high-dose RT region, resulting in a safer and less toxic treatment; (2) the bowel is unlikely to be fixed by surgical adhesions as when RT is given postoperatively, enabling safe delivery of a higher dose to the true area at risk; (3) optimum knowledge of the gross tumor location is possible, permitting better radiation targeting; (4) the tumor is contained by an intact peritoneal covering, providing a physical barrier to immediate tumor dissemination; (5) the risk of intraperitoneal tumor dissemination at the time of surgery may be reduced by the biologic impact of preoperative RT; and (6) using traditional principles of sarcoma RT, the radiation dose believed to be biologically effective is lower in the preoperative setting. Although RT planning for RPS can be complex, the use of conformal techniques or intensity modulation usually permits the RT dosage to be administered safely to the critical organ preoperatively. It is imperative that members of both the surgical and radiation oncology teams be present in the operating room when a significant amount of the liver has been irradiated preoperatively to evaluate the planned residual liver volume. We cannot overemphasize the detailed evaluation of dosimetry and treatment planning films are needed to be certain that a sufficient volume of unirradiated liver is left unresected at the time of surgery.²⁴¹

The only standard that exists at present for RPS treatment is that complete surgery should be performed wherever possible. The adjuvant RT approaches described above should be considered experimental and be the subject of further assessment. Recently, it was also suggested that a more scientifically sound approach to study design should be undertaken in RPS, particularly exploring the role of external beam RT through a randomized controlled trial.¹⁷⁴ The European Organisation for Research and Treatment of Cancer (EORTC) is currently conducting a phase 3 randomized study of preoperative RT plus surgery versus surgery alone (NCT01344018) that is readily on target for completion.²⁴²

Chemoradiation approaches

Although chemotherapy alone has not been associated with obvious improvements in outcome of RPS, chemoradiation, as in other solid tumors, is a subject of interest as a treatment strategy for RPS. This is especially relevant in high-grade RPS, which have a more adverse prognosis than the more common low-grade lesions.

Pilot studies using concurrent preoperative chemotherapy and external beam radiation have been reported, often in patients with extremely large tumors, with acceptable toxicity and with achievement of local control in patients in whom a negative-margin resection was possible.^{141, 243} These reports demonstrate that chemoradiation approaches are feasible. Additional phase 2 studies will be necessary to clarify response rates and toxicity profiles and determine whether chemoradiation should be tested in phase 3 trials. As discussed earlier, the Radiation Therapy Oncology Group completed a phase 2 study of preoperative doxorubicin and ifosfamide followed by preoperative RT and then by surgical resection with an intra- or postoperative radiation boost in patients with intermediate- and high-grade RPS. This study demonstrated significant toxicities and event-free outcomes that were considered modest.^{144, 146}

Functional outcome and morbidity of treatment

The functional result of extremity sarcoma management has become an important component of outcome assessment. Assessing the functional result is difficult, as it requires methods that are valid and reproducible. Many centers have yet to become experienced in the development and use of these methods, and much of the literature contains significant heterogeneity in patient samples.

Thus far, the variables associated with poorer functional outcome include large tumor size, higher doses and larger volumes of radiation, nerve sacrifice, postoperative fracture, and wound-healing complications.^{244, 245} To evaluate and compare functional outcome, it is imperative that functional data be reported consistently. Three disease-specific scoring scales have been reported as useful in assessing functional outcome.²⁴⁶ This area has been discussed in detail by Davis, who observed that “function” has many meanings in the literature.²⁴⁶ The concepts of impairment, disability, and handicap following extremity STS are likely misunderstood and certainly not used consistently. Davis noted that impairment is a disorder of structure or function whereas disability is a restriction or lack of ability to perform an activity. Handicap results from impairment and disability and prevents or restricts an individual from performing in a role that is normal for the individual. For sarcoma patients, impairments can be manifested as soft tissue fibrosis, loss of motion at a joint, and decreased muscle strength; disability can be manifested as limited mobility and difficulty performing routine self-care and activities of daily living; and handicap can be evident in limitation in family roles, social functioning, and the capacity for employment.

Impairments are the most frequently reported deficits following limb-preserving therapy for extremity STS, and up to 50% of patients appear to experience significant impairments.²⁴⁶ Disability occurs less frequently, although reports are contradictory. It seems likely that many sarcoma patients learn to accommodate their impairments. Handicap has received little attention in the literature. However, the limited data suggest that up to 50% of patients may experience changes in their employment and vocation status after treatment of extremity STS. The continuing challenge in treating sarcomas is to define the therapeutic ratio for the patient with sarcoma of the extremity. Specifically, the aim of the multidisciplinary team will be to minimize the amount of treatment while maintaining or improving current standards of disease control to reduce treatment morbidity and enhance patient outcome.

Molecularly and pathobiologically based sarcoma management

Management of sarcomas is increasingly being driven by the specific nature of the disease entity, most importantly the pathophysiologic subtype. The work of Pasteur and Koch was fundamental for the recognition and definition of pathogenic microbes; similarly, many laboratories today are identifying molecular lesions that will redefine the field of sarcoma research. An example of this work is in the recognition of soft tissue Ewing sarcoma/PNET family of tumors as relatives of their bony primary counterparts by virtue of the same spectrum of translocations in each clinical setting. These tumors should be treated with curative intent using an aggressive multimodality approach that begins with multiagent chemotherapy. If primary surgery has removed measurable disease, adjuvant chemotherapy is definitely indicated, with consideration of adjuvant RT. By adopting similar strategies for PNETs and Askin tumors as for conventional Ewing sarcomas, outcomes improved. The molecular similarities between tumors of this family have led to the current convention of considering them morphologic and clinicopathologic variants of the same underlying molecular disease process.^{254, 255}

Greater understanding of the biology of translocation gene products points to epigenetic events being important in sarcomagenesis and should impact on the efficacy of chemotherapy. Ewing sarcoma and synovial sarcoma are chemotherapy-responsive STS subtypes. While TP53 could be posited as a reason that certain sarcomas are chemotherapy sensitive, TP53 status is not the only factor dictating chemotherapy sensitivity as there are other TP53 wild-type sarcomas that are far less sensitive to chemotherapy, for example, extraskeletal myxoid chondrosarcoma and alveolar soft part sarcoma. Striking research on the epigenetic mSWI/SNF (BAF) complexes and their role in synovial sarcoma hopefully will shed light on some of the basic mechanistic aspects permitting synovial sarcoma survival, in order to better engender tumor cell death.^{256, 257}

As a final example, the multiple histopathologic subtypes of liposarcomas are becoming increasingly well researched and well understood, and mechanisms by which some forms acquire their characteristic DNA signature are beginning to become understood.²⁵⁸ Myxoid and round-cell liposarcomas usually exhibit a characteristic chromosomal rearrangement t(12;16) (q13;p11) FUS-DDIT3. These liposarcomas tend to be relatively sensitive to chemotherapy, with trabectedin standing out as most active in this specific diagnosis. Trabectedin appears to inhibit the binding of the FUS-DDIT3 fusion protein to DNA, which then causes liposarcoma cell death or differentiation.^{259, 260} The rarer pleomorphic liposarcoma subtype has more in common with undifferentiated pleomorphic sarcoma than other liposarcomas. Finally, well-differentiated liposarcomas exhibit ring and giant marker chromosomes on cytogenetic analysis and these karyotypic abnormalities, often involving massive amplification of chromosome 12q, carry through in dedifferentiated liposarcomas. Unraveling the ability of the CDK4 and HDM2 loci on 12q to amplify hundreds of times in well-differentiated–dedifferentiated liposarcoma may provide insight on how better to attack these relatively chemotherapy-insensitive sarcomas.^261–263