Radiation therapy for Ewing sarcoma family of tumors

INTRODUCTION — Ewing sarcoma is a rare malignancy that most often presents as an undifferentiated primary bone tumor; less commonly, it arises in soft tissue (extraosseous Ewing sarcoma). Both are part of a spectrum of neoplastic diseases known as the Ewing sarcoma family of tumors (EFT), which also includes the more differentiated peripheral primitive neuroectodermal tumor (PNET, previously called peripheral neuroepithelioma, adult neuroblastoma, and Askin tumor of the chest wall) [1]. PNET can also present either in bone or soft tissue. Because these tumors share similar histological and immunohistochemical characteristics and unique nonrandom chromosomal translocations, they are considered to have a common origin. (See "Epidemiology, pathology, and molecular genetics of Ewing sarcoma".)

EFT also share important clinical features, including a peak incidence between the age of 10 and 20, a tendency toward rapid spread to the lungs, bone, and bone marrow, and a responsiveness to the same chemotherapeutic regimens and radiation therapy (RT). Because relapse rates are high in patients undergoing local therapy alone (80 to 90 percent), it is surmised that the majority have subclinical metastatic disease at the time of diagnosis, even in the absence of overt metastases. The routine administration of intensive multiagent chemotherapy, which can eradicate these deposits, has had a dramatic impact on outcomes. For Ewing sarcoma, from 1975 to 2010, the five-year survival rate increased from 59 to 78 percent for children younger than 15 years and from 20 to 60 percent for adolescents aged 15 to 19 years [2,3]. (See "Treatment of Ewing sarcoma".)

Local control at the primary tumor site can be accomplished by surgery, RT, or both. The choice of modality usually represents a tradeoff between functional result and treatment-related morbidity, particularly the risk of a secondary radiation-induced malignancy. Although modern treatment protocols emphasize surgery for optimal local control, patients who lack a function-preserving surgical option because of tumor location or extent, those who have clearly unresectable primary tumors following induction chemotherapy, and those whose tumors are resected with positive margins are appropriate candidates for RT.

Since more than 90 percent of patients with EFT have either detectable or subclinical metastases at diagnosis, local therapy, if delivered correctly, is probably not the critical event in determining survival. However, if local therapy is delivered poorly or given in such a way that it compromises the delivery of adequate chemotherapy, survival can be greatly compromised. Moreover, local failure is associated with a very poor survival outcome [4].

Here we will discuss the role of RT in the local management of the EFT. Epidemiology, pathology, molecular genetics, clinical presentation, and diagnosis of these tumors, as well as surgical principles and the use of chemotherapy are presented elsewhere. Central (supratentorial) PNET tumors are also discussed elsewhere. (See "Epidemiology, pathology, and molecular genetics of Ewing sarcoma" and "Clinical presentation, staging, and prognostic factors of Ewing sarcoma" and "Treatment of Ewing sarcoma" and "Bone sarcomas: Preoperative evaluation, histologic classification, and principles of surgical management" and "Uncommon brain tumors", section on 'CNS embryonal tumor NEC/NOS'.)

RADIATION THERAPY FOR LOCAL CONTROL OF THE PRIMARY

General principles — The radiosensitivity of Ewing sarcoma was first noted in the original description by James Ewing [5]. Historically, RT was preferred over surgery because of less acute morbidity, acceptable rates of local control (50 to 77 percent), and poor long-term prognosis, largely due to distant dissemination in the absence of effective chemotherapy [6].

The introduction of multiagent chemotherapy into the management of the Ewing sarcoma family of tumors (EFT) in the 1970s led to an increase in cure rates as well as higher local control rates after RT (72 to 90 percent) [7-12]. As more children survived their tumors, balancing long-term local disease control with long-term functional results assumed greater importance. The recognition of late effects of RT (eg, fracture, second malignant neoplasms) as well as improvements in surgical techniques resulted in a shift toward the use of surgery rather than RT.

Current protocols tailor local treatment to the individual patient with the goal of maximizing local tumor control while minimizing treatment-related morbidity. In general, patients are selected for local therapy in such a way that they are treated with surgery or RT but not both, since the combined use of surgery and RT places patients at risk for morbidity from both modalities [13]. However, there is a role for combined therapy in some circumstances, particularly for large tumors (particularly involving the pelvis), and for cases in which resection margins are positive or close. (See 'Adjuvant radiation therapy' below.)

Radiation therapy versus surgery — Although most clinical protocols for EFT emphasize surgery for treatment of the primary lesion, RT is an effective option for local control in patients who lack a function-preserving surgical option because of tumor location or extent, and those who have clearly unresectable primary tumors despite induction chemotherapy. If a tumor appears to be categorically unresectable following induction therapy, debulking surgery should be avoided, and the patient should be referred for definitive RT.

Both surgery and RT represent effective local treatment for Ewing sarcoma. There are no randomized trials which directly compare both modalities, and their relative roles continue to be debated. Contemporary treatment guidelines emphasize surgical resection as the local control modality of choice if it is believed that the lesion can be resected with negative margins, without excessive morbidity, and with the expectation of a reasonable functional result. Surgery is preferred for potentially resectable lesions, and for those arising in dispensable bones (eg, fibula, rib, small lesions of the hands or feet) for the following reasons:

●It avoids the risk of secondary radiation-induced sarcomas.

●An analysis of the degree of necrosis in the excised tumor can permit refinements in the estimate of prognosis. (See "Bone sarcomas: Preoperative evaluation, histologic classification, and principles of surgical management".)

●In the skeletally immature child, resection may be associated with less morbidity than radiation, which can retard bone growth and cause deformity.

Thus, in most cases, the decision to use RT for treatment of the primary site is made only after review of the potential surgical options for a given lesion. Patients who lack a function-preserving surgical option because of tumor location or extent, those resected with positive margins, and those who have clearly unresectable primary tumors following induction chemotherapy are appropriate candidates for RT.

Influence of tumor site — When considering the local control strategy for an individual patient, the need to attain complete tumor eradication must be weighed against the twin goals of maximizing function and minimizing long-term morbidity. Although treatment decisions must be individualized according to the expected deficits from surgical resection at any site, the following generalizations can be made regarding the influence of tumor site on the choice of local treatment (see "Bone sarcomas: Preoperative evaluation, histologic classification, and principles of surgical management"):

●It is generally agreed that surgery is preferred for lesions arising in dispensable bones (eg, fibula or rib).

●Tumors affecting the long bones of the leg, distal humerus, or ulna can usually be resected and reconstructed using intercalary techniques (allografts, autografts, or metallic prostheses) or joint replacement, depending on tumor location. (See "Bone sarcomas: Preoperative evaluation, histologic classification, and principles of surgical management", section on 'Allografts and endoprostheses'.)

On the other hand, primary tumors involving the proximal humerus and upper scapula may be best treated with RT, since limb reconstruction is difficult and shoulder morbidity may be substantial. (See "Bone sarcomas: Preoperative evaluation, histologic classification, and principles of surgical management", section on 'Extremity lesions: Limb-sparing procedures'.)

●Patients with lesions of the skull, facial bones, or vertebrae are often candidates for nonsurgical treatment because of the difficulty in achieving negative margins without substantial functional deficit [14].

●A number of series report better results with surgery over RT alone for pelvic bone tumors [15-19], but this is a controversial area. Surgical treatment for pelvic sarcomas, which are often bulky, is challenging. Lesions of the iliac wing, ischium, or pubis can usually be resected after a good response to chemotherapy without substantial functional morbidity. For tumors of the periacetabular region and those that cross the sacroiliac joint, there is less enthusiasm for surgical resection given the significant resulting functional deficit. (See "Bone sarcomas: Preoperative evaluation, histologic classification, and principles of surgical management", section on 'Pelvic tumors'.)

Outcomes — Several studies support the adequacy of RT alone for local control:

●One retrospective report included 39 patients who were treated between 1975 and 1991 at two Boston hospitals with multiagent chemotherapy and varying combinations of surgery and RT [20]. Twenty had RT alone, 16 had both surgery and RT, and 3 had surgery alone. Despite the use of combination chemotherapy, five-year disease-free survival (DFS) was only approximately 30 percent, emphasizing the high-risk nature of these mainly axial lesions. However, local control rates were similar among patients treated with RT alone or surgery (85 and 84 percent, respectively).

●Another case series included 60 patients treated for EFT at Memorial Sloan Kettering Cancer Center between 1990 and 2004; 31 underwent RT as the sole modality for local control, the remainder underwent RT in conjunction with surgery [21]. The relatively poor prognosis of this group is reflected by the fact that most were centrally located tumors (spine, pelvis, proximal extremities), 52 percent were ≥8 cm in size, and one-third were metastatic at diagnosis.

●Adult Ewing sarcoma patients who are treated with modern chemotherapy protocols and RT techniques have outcomes that are similar to those of younger patients, with anticipated local control rates of approximately 80 percent after definitive RT [22]. With a median follow-up of 41 months, DFS and overall survival rates for patients with nonmetastatic disease at presentation were 70 and 86 percent, respectively. There were only nine local recurrences, five of which developed in patients who presented with overt metastases. The three-year actuarial local control rates for the entire group and those undergoing definitive RT alone were 77 and 78 percent, respectively. The presence of overt metastases emerged as the only significant adverse prognostic factor for local control, a finding that has been confirmed by others [10,23]. Long-term treatment-related complications were minimal, and there were no secondary malignancies. (See 'Late effects' below.)

●The relationship between local control modality (surgery, RT, or both) and the subsequent risk for local failure was examined in a subgroup of 75 patients treated for nonmetastatic pelvic EFT in United States Intergroup study 0091 (Intergroup Ewing Sarcoma Study Group [IESS] III), a randomized comparison of VACA (vincristine, doxorubicin, cyclophosphamide, and dactinomycin) with or without alternating IE (ifosfamide plus etoposide) [24]. Choice of local control modality was left to the treating clinicians; 12 underwent surgery, 44 received RT, and 19 received both.

The rates of five-year EFS and cumulative incidence of local failure (as any component of first failure) in the entire group were 49 and 21 percent, respectively. There was no significant difference in either endpoint according to tumor size (<8 versus ≥8 cm) or choice of local control modality. However, there was a trend toward improved local control in patients receiving VACA/IE as compared with VACA alone (cumulative incidence of local failure 11 versus 30 percent), which was present regardless of the local control modality.

Although not a universal finding [7,21,25], many studies suggest that smaller tumors are more likely to be locally controlled after RT alone [10,16,26-31]. As an example, in a report of 100 patients with localized Ewing sarcoma of bone, local control rates were significantly higher after RT for the 14 lesions that were ≤100 cm³ in volume compared with larger lesions (93 versus 60 percent, respectively) [28]. Other series report higher local control rates with lesions <8 cm (94 versus 56 percent) [32].

Finally, certain sites within the pelvis, such as the sacrum, have been associated with a higher incidence of local recurrence after definitive RT in some series [33].

Radiation therapy versus surgery — Because no randomized trial has directly compared both modalities, only a relative comparison can be made from retrospective reports and the prospective trials that have mainly tested different multiagent chemotherapy regimens. (See "Treatment of Ewing sarcoma".)

In many retrospective series, rates of local control and survival are superior after surgery compared with RT alone [10,11,17,26,34-44]. However, larger cooperative group studies have failed to reflect this advantage, and selection bias likely accounts for at least some of these results. Smaller, more favorably situated peripheral tumors are more often resected while larger, axial lesions (which have a higher rate of local failure in most series and a poorer prognosis overall [10,34,35,45]) are referred for RT. (See "Clinical presentation, staging, and prognostic factors of Ewing sarcoma", section on 'Tumor site and size'.)

The following results are available to inform this issue:

●The Children's Oncology Group (COG) reported local control outcomes in patients treated with surgery, radiation, or both on three consecutive protocols using standard-dose, five-drug chemotherapy every three weeks [46]. Patients who underwent surgery were significantly younger and had more appendicular primaries. Compared with surgery, radiation was associated with a higher unadjusted risk of any event (hazard ratio [HR] 1.70; 95% CI, 1.18-2.44), death (HR 1.84; 95% CI, 1.18-2.85), and local failure (HR, 2.57; 95% CI, 1.37-4.83). On multivariate analysis, compared with surgery, use of RT was associated with a higher risk for local failure (HR 2.41; 95% CI, 1.24-4.68), although there were no significant differences in event-free survival (EFS; HR 1.42; 95% CI, 0.94-2.14), overall survival (HR 1.37; 95% CI, 0.83-2.26), or distant failure (HR 1.13; 95% CI, 0.70-1.84).

●In a later analysis, the COG analyzed the local failure rate for 956 patients treated with ifosfamide- and etoposide-based chemotherapy on three consecutive protocols [47]. The local treatment modalities were surgery, definitive RT, or surgery plus RT. The local failure rate for the entire cohort was 7.3 percent, and it was lower for surgically treated patients (3.9 percent for surgery alone and 6.6 percent for surgery plus RT) compared with those treated with RT alone (15.3 percent). The local failure incidence was higher for pelvis tumors (13.2 percent) than for other sites (5.4 percent for extremity tumors, 5.3 percent for axial non-spine tumors, 9.1 percent for extraskeletal tumors, and 3.6 percent for spine tumors). Among those treated with RT alone, the local failure rate was 22.4 percent for pelvic tumors compared with 15 percent for extremity tumors, but for surgically treated patients, local failure rates did not differ according to primary tumor site (3.7 percent for extremity tumors and 3.9 percent for pelvis tumors). Furthermore, there was no difference in local failure rate by local treatment modality for axial non-spine, spine, and extraskeletal tumors. The local failure incidence was higher in older individuals than in those less than 18 years (11.9 versus 6.7 percent). In multivariate analysis, age ≥18 years (HR 1.9) and treatment with RT alone were independent prognostic factors for a higher local failure rate. Tumor size (<8 versus ≥8 cm) was available in 40 percent of patients and did not correlate with local failure incidence. Hence, overall local tumor control was excellent and similar between surgery and RT for axial non-spine, spine, and extraskeletal tumors.

Taken together, these data support surgical resection when appropriate. RT remains a reasonable alternative in selected patients.

RT delivery techniques may have contributed to poorer outcomes in early series. As an example, in the Cooperative Ewing Sarcoma Study (CESS)-81 trial, relapse-free survival (RFS) for patients undergoing irradiation of the primary site in the initial phase of the trial was only 55 percent, significantly lower than for surgically treated patients [36]. On review, the high rate of local failure after RT alone (50 percent) was attributed to geographic "miss." After a quality assurance program was initiated, local control rates improved significantly, and RFS increased to 80 percent. In a subsequent study, CESS-86, which included central quality assurance, the five-year RFS rates following chemotherapy plus either RT or surgery were 67 and 65 percent, respectively [7]. The importance of RT dose and field planning to local control is discussed below. (See 'Radiation treatment planning' below.)

Timing of local control — The recommended timing for initiating local treatment, be it surgery, RT, or the combination, for patients with localized Ewing sarcoma in COG protocols has been immediately after completion of 12 weeks of induction chemotherapy. An analysis of data derived from the National Cancer Database noted that patients initiating local therapy by week 15 versus week ≥16 had improved five-year overall survival of 78.7 versus 70.4 percent and 10-year overall survival of 70.3 versus 57.1 percent, respectively (p<0.001). The difference in overall survival according to time to local therapy was particularly more important in patients receiving RT alone. (See "Treatment of Ewing sarcoma".)

Adjuvant radiation therapy — Indications for adjuvant RT in patients who have undergone resection of an EFT include the following:

●Bulky tumors in difficult sites (eg, the pelvis); in this setting, RT can be given either preoperatively or postoperatively, based on institutional protocols and experience. Our own preference, when positive margins are considered likely, is to use preoperative irradiation.

●If there is residual microscopic or gross disease after surgery, or there are inadequate surgical margins.

●Adjuvant hemithorax irradiation is indicated in patients with high-risk chest wall primary tumors (close or involved margins, initial pleural effusion, pleural infiltration, and intraoperative contamination of the pleural space).

RT is usually not recommended for patients who undergo a complete resection in order to avoid exposing them to the risk of a secondary malignancy. Although indications for postoperative RT are not firmly established, the following are common scenarios in which RT is recommended in conjunction with surgery for EFT [48].

Bulky tumors — For bulky tumors in difficult sites such as the pelvis, combined surgery and RT might allow for a more limited surgical procedure, better functional outcome, and enhanced local control as compared with single-modality therapy [10,16,21,49]. RT can be given either preoperatively or postoperatively, based on institutional protocols and experience.

Positive or inadequate margins — Patients who are left with macroscopic amounts of viable tumor in the resected specimen following neoadjuvant chemotherapy have worse survival than those with minimal or no residual tumor [11,27,31,34,50-53]. In addition, inadequate resection margins (ie, a marginal or intralesional resection (table 1)) are associated with a worse outcome as compared with radical or wide resection [41,54].

The impact of surgical margins on outcome was described in a series of 244 patients who were enrolled in CESS studies for localized EFT and underwent surgery as a component of local therapy [41]. Ninety-four had definitive surgery alone, 131 received postoperative RT, and 19 had preoperative RT. Surgical margins were radical, wide, marginal, and intralesional in 29, 148, 39, and 28 patients, respectively. The local and combined (local plus systemic) relapse rate was significantly lower after a complete resection (radical or wide) with or without RT than after an incomplete resection (marginal or intralesional) with or without RT (5 versus 12 percent). However, the relapse rate was not significantly lower in patients who received RT after incomplete surgery compared with those who did not (12 versus 14 percent). Ten-year survival rates for the cohorts undergoing radical, wide, marginal, and intralesional resection were 58, 65, 61, and 71 percent, respectively; the differences were not statistically significant.

However, the data to support the benefit of RT in this setting are surprisingly scant and conflicting:

●The efficacy of combined surgery and RT was evaluated in a retrospective series of 39 patients treated for localized EFT at the St. Jude Hospital between 1978 and 2001 [55]. With combined local therapy, local control rates were excellent, even for patients with positive surgical margins (eight-year local failure rates for patients with positive and negative surgical margins were 17 and 5 percent, respectively), and the corresponding overall survival rates were 71 and 94 percent, respectively.

●On the other hand, a lack of benefit for adjuvant RT was suggested in a large single-institutional retrospective series, in which the addition of moderate-dose adjuvant RT (45 Gy) to patients with inadequate margins did not seem to improve local control or overall DFS. The authors concluded that patients for whom inadequate margins were anticipated at the time of preoperative evaluation might be better treated with full-dose RT alone [56].

●The Euro-E.W.I.N.G group reported a large observational study of 599 patients who underwent surgery after chemotherapy for standard-risk localized disease [57]. Postoperative RT was recommended for patients with inadequate surgical margins, and it was also common practice for patients with a vertebral tumor or with a pleural effusion contiguous to a rib or thoracic wall primary tumor. After controlling for possible confounders, there was a significant reduction in local recurrence in patients treated with surgery plus postoperative RT compared with surgery alone (HR 0.43, 95% CI 0.21-0.88,). They also observed a nonsignificant trend for benefit associated with postoperative RT for DFS, event-free survival, and overall survival. Among the 132 patients for whom RT would have been recommended according to the policies of the Euro-E.W.I.N.G group, 50 did not receive it. Among these patients, local recurrence rates were 30 percent at eight years compared with 11 percent when RT was performed.

High-risk chest wall tumors — The role of adjuvant thoracic irradiation is unsettled. The IESS-I trial showed that bilateral whole lung irradiation was an effective adjuvant treatment for patients with localized EFT, but prolonged follow-up favored four-drug multiagent chemotherapy [9]. Prophylactic whole lung irradiation has not been studied in any subsequent trial, and no studies have used it in addition to modern combination chemotherapy regimens [58], except in patients with lung metastases at diagnosis. (See 'Pulmonary metastases' below.)

On the other hand, adjuvant hemithorax irradiation appears to improve outcomes in patients with high-risk chest wall primary tumors (close or involved margins, initial pleural effusion, pleural infiltration, and intraoperative contamination of the pleural space) [59,60].

RADIATION THERAPY FOR METASTATIC DISEASE — Patients with overt metastatic disease at presentation have a significantly less favorable outcome than do those with localized disease. However, aggressive multimodality therapy can relieve pain [61], prolong the progression-free interval, and provide long-term relapse-free survival (and possible cure) in some patients. As a result, aggressive treatment is warranted. (See "Clinical presentation, staging, and prognostic factors of Ewing sarcoma", section on 'Disease extent'.)

Management of the primary site — Local control can pose major issues, even for patients with overt metastatic disease. In such cases, it may be difficult to justify a large resection of the primary site because of the poor long-term prognosis [16,62]. However, resection may be reconsidered in selected patients if chemotherapy results in significant volume reduction, particularly if areas of small-volume metastatic disease are also amenable to surgical resection. (See "Treatment of Ewing sarcoma", section on 'Treatment for metastatic disease'.)

On the other hand, RT can provide adequate local control with acceptable morbidity [7,32,35,36]. If substantial amounts of bone marrow will need to be included in the field, RT may be delayed until the end of systemic therapy to avoid interfering with chemotherapy. (See "Treatment of Ewing sarcoma", section on 'Treatment for metastatic disease'.)

Pulmonary metastases — Patients with a limited number of lung metastases do not share the same dismal prognosis of metastatic disease at other sites (ie, bone or bone marrow), because of the ability to accomplish wide resections both easily and repeatedly. Chemotherapy is a necessary component of therapy. In highly selected patients, between 20 and 40 percent five-year survival can be achieved [63-66]. (See "Surgical resection of pulmonary metastases: Outcomes by histology".)

Retrospective reports from large cooperative groups and single institution series suggest that low-dose bilateral whole lung irradiation (15 to 20 Gy) benefits patients with the Ewing sarcoma family of tumors (EFT) presenting with pulmonary metastases, even if all lesions are resected or completely respond to chemotherapy [9,36,58,63,65-70]. In the Cooperative Ewing Sarcoma Study (CESS) and European Intergroup Cooperative Ewing Sarcoma Study (EICESS) trials, the rate of pulmonary relapse was reduced by 50 percent over that for patients who did not undergo lung RT, and this was accompanied by improvements in event-free survival (from 19 to 40 percent in one report [66]) as well [65,66]. However, most of the available data supporting benefit from RT in this setting are from nonrandomized series, and patient selection factors confound interpretation of the data. There are no randomized trials that confirm the benefit of this approach.

Despite the lack of controlled trials, low-dose bilateral lung irradiation (15 to 18 Gy, in daily 1.5 to 2.0 Gy fractions) with a focal boost dose to a total of 40 to 50 Gy to large deposits is commonly recommended for patients with pulmonary metastases who have had a good response to chemotherapy. Long-term toxicity appears to be acceptable. This was shown in a retrospective analysis of data from the prospectively performed EICESS, in which pulmonary function tests were available for 37 patients who were treated with whole lung irradiation (with or without a boost) [71]. At a median follow-up of 25 months, none, mild, moderate, or severe pulmonary complications were seen in 43, 29, 21, and 7 percent of patients treated with whole lung irradiation without further boost. Slightly higher complication rates were reported in patients who had an additional radiation boost or surgery to the thorax in addition to RT. The addition of consolidation whole lung RT appears to be well tolerated in adult patients with lung metastases, with minimal reported significant acute or late toxicity and a three-year rate of freedom from pulmonary relapse of 45 percent [70].

Bone and soft tissue metastases — Patients with solitary or circumscribed bone or soft tissue lesions fare less well in the long run than those with isolated pulmonary metastases. Nevertheless, with aggressive multimodality therapy, perhaps 10 percent will be long-term survivors. For patients with solitary or limited bone metastases, RT is delivered to metastatic lesions (doses equivalent to 40 to 50 Gy) in addition to irradiation of the primary tumor. One small series of 13 patients with metastatic Ewing sarcoma and rhabdomyosarcoma reported only one local failure at irradiated sites with minimal associated toxicity [72]. (See "Treatment of Ewing sarcoma", section on 'Treatment for metastatic disease'.)

Total body irradiation — Total body irradiation (TBI) or sequential hemibody irradiation as a component of systemic treatment for patients with high-risk features or metastatic disease at presentation has been investigated in a limited fashion at a few institutions [25,73-75]. Unfortunately, the results of these trials have been disappointing; TBI has not contributed significantly to the control of metastatic disease. This approach should not be considered at present outside of the context of a clinical trial.

The use of TBI as a conditioning regimen prior to high-dose chemotherapy and hematopoietic stem cell transplantation for patients with poor-risk disease is discussed elsewhere. (See "Treatment of Ewing sarcoma", section on 'Other strategies'.)

RADIATION TREATMENT PLANNING — Because of the relative rarity of the Ewing sarcoma family of tumors (EFT), the potential for cure, and the importance of minimizing late side effects from therapy (see below), it is important that RT be administered by radiation oncologists who are familiar with the optimal treatment techniques. RT is currently planned using computed tomography (CT) or magnetic resonance imaging (MRI) simulation. Biopsy sites and surgical scars should be marked with radio-opaque markers to allow inclusion into the target volume when deemed appropriate.

Treatment volume — Historically, Ewing sarcoma was thought to be a tumor of the bone marrow. Consequently, RT was administered to the entire marrow cavity of the involved bone, and the site of gross disease boosted to a higher dose. However, an analysis of the RT fields for the Intergroup Ewing Sarcoma Study Group (IESS) trial I suggested that most relapses were at the site of initially bulky tumor [8,9,76]. Subsequent efforts were geared toward reducing the irradiated field and targeting higher doses to the site of the initial primary tumor [7,26].

In 1983, the Pediatric Oncology Group attempted a randomized trial of whole bone versus tailored-field RT after 12 weeks of induction chemotherapy [35]. After preliminary analysis showed that tailored field was as effective as whole-bone RT, the randomized study was stopped and subsequent patients were treated with limited fields. At three years, the rates of local control and event-free survival were 76 and 54 percent, respectively.

IESS trial III, which opened in 1988, was the first cooperative group trial to include tailored RT ports, and the first to be carried out with modern MRI imaging of the primary site and CT-based treatment planning [77]. The main randomization was between standard versus more intensive chemotherapy (VAC with or without IE). The addition of IE significantly improved five-year survival (72 versus 61 percent) and event-free survival rates (69 versus 54 percent), particularly for patients with pelvic primary tumors. Moreover, the five-cumulative rate of local recurrence either alone or with systemic relapse was 20 percent in the VAC group compared with only 7 percent in the VAC/IE group. These data underscore the contribution of intensive chemotherapy to both local control and survival for patients with nonmetastatic EFT at diagnosis. (See "Treatment of Ewing sarcoma".)

Current recommendations for RT define an initial clinical target volume to include the original bone and soft tissue tumor extent with a 1.0 cm margin. As noted above, for patients with chest wall tumors and malignant effusions, the field includes the entire ipsilateral hemithorax to a dose of 15 Gy prior to cone-down to the chest wall tumor. For patients where the soft tissue tumor displaces but is not thought to violate an anatomic compartment (ie, pelvic Ewing sarcoma), the postchemotherapy soft tissue extent of disease abutting the anatomic compartment, rather than the original prechemotherapy extent of disease, is included in this initial target volume.

The initial clinical target volume is treated to 45 Gy in 25 fractions followed by a field reduction to encompass the post-chemotherapy gross soft tissue tumor as well as all of the originally involved bone with a margin of 1 cm; this volume is treated to another 10.8 Gy in six fractions in patients undergoing definitive RT and to 5.4 Gy in three fractions for patients undergoing adjuvant RT for microscopic residual disease after surgery (with boost up to 10.8 Gy for any sites of gross disease after surgery). Additional margin expansions on these volumes are used for the planning target volume expansion to account for any variability in daily setup; this margin should be individualized by the treating physician based on the reproducibility of immobilization and the kind of image guidance used but is generally a minimum of 5 mm and usually in the range of 0.5 to 1.0 cm.

Attention to potential RT effects on normal tissue is critical in radiation planning to minimize late effects, particularly in children:

●Even partial treatment of uninvolved epiphyseal growth plates is avoided in order to minimize treatment-induced limb shortening [78].

●Circumferential irradiation of a limb is avoided to reduce the risk of limb edema and fibrosis.

●Gonadal avoidance or additional shielding (for the testes) is important to retain fertility.

●Excellent results can be achieved from RT of lesions of the hand or foot [79]. However, walking places repetitive stress on the soft tissues of the sole of the feet, and dose sparing for the sole of the foot can improve the functional outcome of treatment.

●Irradiation of the Achilles tendon is usually avoided. Lesions of the calcaneus can be treated; although some atrophy of the heel pad has been reported, the functional impairment is usually minor [80].

●Nail beds should be excluded from the radiated field when possible.

●For pelvic tumors, distention of the bladder prior to each day's treatment can reduce the amount of small bowel in the radiated field, but care must be taken to avoid radiation to significant portions of the bladder in patients receiving ifosfamide to reduce the risk of hemorrhagic cystitis.

It is anticipated that improvements in RT technique over the last 20 years (ie, the adoption of tailored radiation ports as opposed to whole bone ports used in the period 1960 to 1980, three-dimensional conformal radiation therapy [3D-CRT], intensity-modulated radiation therapy [IMRT], and proton therapy [see below] and use of lower and risk-adapted RT doses [32,81]) will reduce the risk of late effects. Nevertheless, the relatively high complication rate seen with earlier treatment approaches and the delayed nature of many of these complications underscore the need for long-term posttreatment follow-up. (See 'Sequelae of treatment' below.)

Proton beam therapy — One way to reduce the volume of normal tissue irradiated is with charged particle irradiation using protons. (See "Radiation therapy techniques in cancer treatment", section on 'Particle therapy'.)

The advantage of protons over conventional photons is in their dose distribution. The physical characteristics of the proton beam result in the majority of the energy being deposited at the end of a linear track, called a Bragg peak, with the dose falling rapidly to zero beyond the Bragg peak. Compared with photon beam irradiation, proton beam therapy permits the delivery of high doses of RT to the target volume while reducing the radiation dose received by normal tissues distal to the target. Depending on the tumor target configuration and the portal selection, this can result in up to 60 percent reduction in the dose received by uninvolved normal tissue [82].

This approach may be particularly beneficial for Ewing sarcomas arising in the bones of the paranasal sinus (image 1A-B), the pelvis, and the spine:

●Because of the proximity of the spinal cord, the dose to vertebral body primaries using conventional photon irradiation has been often limited to 45 Gy. Proton beam therapy permits the delivery of higher doses while respecting spinal cord constraints (image 2) [83]. Of note, spinal cord constraints can also be respected with photons using stereotactic body radiation therapy (SBRT) [84,85].

●When used for pelvic lesions, proton beam therapy is associated with better sparing of the intestine, rectum, bladder, pelvic bone marrow, and femoral head as compared with photon irradiation [86].

In addition to less acute morbidity, one would also anticipate a reduction in late, RT-induced tumors in patients treated with protons because of the lower volume of normal tissue in the irradiated field (see below) [87]. Proton beam irradiation is approved for use in Children's Oncology Group (COG) protocols for EFT [88].

Our group reported experience with proton beam irradiation in 30 children undergoing chemoradiation with or without surgery for Ewing sarcoma [89]. At a median follow-up of 38.4 months, the three-year actuarial rates of event-free survival, local control, and overall survival were 60, 86, and 89 percent, respectively. RT was acutely well tolerated, with mostly mild-to-moderate skin reactions. The only serious late effects were four hematologic malignancies (three acute myelogenous leukemia [AML] and one myelodysplastic syndrome [MDS]), which are known risks of topoisomerase and anthracycline exposure (see 'Second malignancies' below). These children were treated with passively scattered protons. Increasingly, patients can be treated with spot-scanned protons which reduce the skin dose and contour the dose better than passively scattered protons. Additional studies from other centers confirm the efficacy and limited toxicity with protons for these patients [90,91].

Concern has been raised that neutron scatter radiation associated with passively scattered proton beam lines may also result in secondary malignancies [92]. The magnitude of this risk is uncertain and may be low in the beam lines in modern facilities [93]; it is also significantly reduced by the use of scanned proton beams [94]. Modeling studies comparing protons with IMRT suggest that proton beam irradiation can significantly reduce the risk for developing an in-field second malignancy [95]. A review from the National Cancer Database of pediatric and adult patients treated for nine tumor types suggested a reduction in radiation-associated second cancers with protons compared with photons (relative risk of 0.31, 95% CI 0.26-0.36) [96]. Further follow-up of pediatric patients treated with protons and photons using specific registry data are necessary to confirm these results. [97]

Intensity-modulated radiation therapy — Intensity-modulated radiation therapy (IMRT) requires the same careful three dimensional radiation treatment planning used for 3D-CRT. However, IMRT utilizes variable, computer-controlled intensities within each RT beam, in contrast to the uniform doses within each 3D-CRT beam. Compared with most other treatment techniques, IMRT can achieve a higher degree of accuracy in conforming the radiation to the planned target while sparing normal tissue and reducing the incidence of some late complications. The advantages of IMRT are particularly evident when the target volumes have complex shapes or concave regions. A planning comparison study demonstrated that, compared with 3D-CRT, IMRT achieved significantly better dose conformity and bowel sparing at dose levels above 30 Gy [98]. (See "Radiation therapy techniques in cancer treatment", section on 'Intensity-modulated radiation therapy'.)

Initial experience with IMRT for Ewing sarcoma has been promising [21]. However, critics argue that the risk of a second malignancy might be higher because the multiple portals used with IMRT to achieve conformality of the high-dose region around the target result in delivery of low-moderate doses of radiation to a larger volume of surrounding tissue than with 3D-CRT techniques. In addition, for IMRT delivery, the linear accelerator is delivering more monitor units and, therefore, a larger total-body dose because of leakage of radiation [92]. IMRT supporters point out that since the risk of some radiation-associated tumors is dose dependent, this risk may not in fact be increased. One observational study of adult and pediatric patients treated for nine tumor types reported that second malignancy rates for IMRT were comparable to those with 3D conformal photons but were reduced with protons (relative risk 0.31, 95% CI 0.26-0.36) [96]. These questions will be further clarified by additional follow-up of patients treated with IMRT.

Dose — RT dose is an important factor in local control, particularly for large tumors [23,32]. For patients treated with chemotherapy and RT, more recent cooperative studies in the United States have employed 45 Gy in 25 fractions to the initial clinical target volume as defined above, followed by a 10.8 Gy boost in six fractions to the site of original bony disease and any residual soft tissue disease that remains following chemotherapy. Because of the proximity of the spinal cord, the dose to vertebral body primaries has been often limited to 45 Gy, although techniques such as protons, IMRT, and SBRT can permit the safe delivery of higher doses [99]. (See 'Proton beam therapy' above and 'Intensity-modulated radiation therapy' above.)

For patients undergoing adjuvant postoperative RT, doses of 45 to 50.4 Gy are recommended for microscopic, and 55.8 Gy for gross residual disease. If preoperative RT is given for bulky tumors, doses in the range of 45 Gy are used.

Lower doses (eg, 30 to 36 Gy) have been studied at St. Jude and Memorial Sloan Kettering [32,100]. In one early report of patients undergoing definitive RT, local control rates were 90 percent using RT doses of 35 Gy for smaller lesions (≤8 cm) with a favorable response to chemotherapy, but they were inferior (52 percent) for larger lesions [32]. Others suggest that doses as low as 30 Gy may be sufficient in the adjuvant setting [100]. However, further studies are needed before lower doses of RT can be accepted as standard practice.

Dose-escalation RT to limited fields may be beneficial for high-risk tumors and may reduce local recurrence rates. A single-institutional phase 2 trial employed focal limited-margin RT using conformal or intensity-modulated techniques [101]. The treatment volume incorporated a 1 cm constrained margin around the gross tumor. In the definitive setting, the gross tumor volume included residual gross tumor, initially involved bone, and adjacent soft tissues, such as pleural or fascial surfaces thought to be initially infiltrated but would not regress back to the involved bone with the soft tissue mass. Unresected tumors <8 cm at diagnosis received a standard dose of 55.8 Gy, and tumors >8 cm received an escalated dose to 64.8 Gy. Patients with microscopic residual disease after resection received adjuvant RT to 50.4 Gy. Adjuvant brachytherapy was permitted in selected patients. Forty-five patients were enrolled, 26 with localized and 19 with metastatic disease. Seventeen patients received adjuvant RT, 16 received standard-dose and 12 received escalated-dose RT. Failures included 1 local, 10 distant, and 1 combined local/distant. The estimated 10-year cumulative incidence of local failure was 4.4 +/- 3.1 percent, with no statistical difference seen between RT treatment groups; there were no local failures in the escalated-dose RT treatment group.

This approach, if validated, has particular appeal because of the reduction in radiation target volume, which should reduce both acute and late toxicity, as well as the selective use of dose escalation for high-risk lesions >8 cm, reported in multiple series to be at higher risk for local failure with conventional RT doses. Another report also suggests that patients treated with definitive radiation doses ≥5600 centigray (cGy) had a lower incidence of local recurrence (17 versus 28 percent) [33].

Radiation schedule — Conventional RT schedules usually consist of once daily RT doses of 1.8 to 2.0 Gy per fraction. Accelerated fractionation RT does not seem to improve rates of local control or survival [7,21,26,36,102]. However, data from the University of Florida suggest that hyperfractionated RT (1.2 Gy twice daily with a six hour interfraction interval) may be associated with less long-term toxicity [26].

The series consisted of 75 patients with localized Ewing tumor of the extremity or pelvis (lower extremity, 30; upper extremity, 19; pelvis, 26) who were treated with definitive RT at the University of Florida between 1970 and 2006. RT was performed on a once-daily (40 percent, median dose was 55.2 Gy in 1.8-Gy daily fractions) or twice-daily (60 percent, median dose 55.0 Gy in 1.2 Gy twice-daily fractions) basis. Functional outcome was assessed using the Toronto Extremity Salvage Score [103].

Larger tumors had significantly worse cause-specific survival (81 versus 39 percent for tumors <8 cm versus ≥8 cm), but there were no patient characteristics or treatment variables that were predictive of local failure. No fractures occurred in patients treated with hyperfractionation or with tumors of the distal extremities. Severe late complications were more frequently associated with use of <8 MV photons and fields encompassing the entire bone or hemipelvis. In terms of late effects, a significantly better late effect Toronto Extremity Salvage Score was associated with biologically effective doses of <91.7 Gy [26]. The authors concluded that limited field sizes with hyperfractionated high-energy RT could minimize long-term complications and provide superior functional outcomes.

Intraoperative radiation therapy — A benefit for intraoperative radiation therapy (IORT) has been suggested in retrospective series involving a small number of patients [104,105]. However, peripheral nerves are dose-limiting tissue structures for IORT, so the risk of severe neuropathy and soft tissue necrosis must be considered if this approach is used.

SEQUELAE OF TREATMENT — Potentially curative treatment of the Ewing sarcoma family of tumors (EFT) is necessarily aggressive. The toxicity of RT given in conjunction with intensive chemotherapy can be separated into short-term and long-term complications. The severity of both can be limited by careful attention to technique.

Acute effects — Acute reactions are those that occur during or shortly after the completion of RT. The most prominent affect tissues within the radiated field that contain rapidly dividing cells, and include desquamation of the skin, myelosuppression, mucositis, diarrhea, nausea, and cystitis. All except myelosuppression are site-specific. Severity depends on the amount of normal tissue in the radiated field, the radiation fraction size, and the timing of chemotherapy in relation to the RT.

Many acute effects, particularly myelosuppression, are exacerbated or potentiated by the use of intensive systemic chemotherapy. As examples:

●Both cyclophosphamide and ifosfamide can cause hemorrhagic cystitis; the amount of bladder in the treatment field should be limited to avoid compounding this problem. (See "Chemotherapy and radiation-related hemorrhagic cystitis in cancer patients".)

●Dactinomycin and doxorubicin act as radiation sensitizers and can enhance acute radiation toxicity. If these drugs are used as a component of systemic chemotherapy, breaks should be instituted during the RT or administration of these agents should be delayed until acute RT reactions subside. Current practice in the Children's Oncology Group (COG) studies is to administer RT concurrently with the ifosfamide/etoposide cycles of chemotherapy. (See "Treatment of Ewing sarcoma".)

Acute reactions are usually self-limited and subside within 10 to 14 days of RT completion. Dry desquamation can be managed topically with emollient creams. The development of wet desquamation may require a temporary break in therapy in addition to thrice daily cleansing with warm water and mild soap, petrolatum-based ointments, nonadherent dressings, and, if patients are neutropenic, a topical antibiotic ointment or systemic antibiotics depending on the clinical scenario.

Mucositis and diarrhea are managed with supportive measures; if severe, a treatment break may be required. Patients receiving proton beam irradiation for treatment of vertebral lesions tend to have less nausea and diarrhea than those undergoing photon beam irradiation because of the lack of an exit dose to bowel anterior to the spine. Patients receiving whole lung irradiation are at risk for radiation pneumonitis. (See 'Proton beam therapy' above and "Radiation-induced lung injury".)

Late effects — Late reactions occur months to years after completing a course of RT. Severity is not always predicted by the severity of acute effects. Late changes in normal tissues caused by RT are related to field size, the volume of normal tissues in the field, and the dose received by normal tissue. Other important factors include patient age at the time of treatment, skeletal maturity, adherence to a rehabilitation program, and whether a pathologic fracture is present at diagnosis. As examples:

●Younger, prepubertal children are at greatest risk for radiation-induced arrest of bone growth [106]. Sparing of uninvolved epiphyseal plates minimizes limb shortening after RT of extremity lesions.

●RT doses above 60 Gy are associated with markedly increased rates of soft tissue induration and fibrosis [107,108].

●High-dose circumferential irradiation of an extremity is associated with edema, fibrosis, and compromised limb function [109]. This can be avoided by sparing of an adequate strip of tissue.

●Weight-bearing bones are at risk for pathologic fractures [108]. The highest risk is within the first 18 months of RT completion.

●Survivors of Ewing sarcoma treated with RT (and/or surgery) are also at risk for increased late musculoskeletal surgeries in the treated region. Such interventions include arthroplasty, amputation, or prosthetic revision due to infection, device failure, or associated fractures [110].

Refinements in diagnostic imaging, RT planning, and techniques over time (tailored field size, hyperfractionated treatment schedules, intensity-modulated radiation therapy [IMRT], proton beam irradiation) have resulted in better limb function among long-term survivors, and more recent series suggest that excellent functional results can be obtained in the majority of patients following RT for EFT [26,80,107]. A posttreatment rehabilitation program, including active range of motion of affected joints, is also important in improving and maintaining limb function.

Second malignancies — There is an increased risk of secondary neoplasia after treatment for EFT [81,111-114]. RT-induced osteosarcomas and therapy-related (alkylating agents, epidophyllotoxins) leukemias predominate. (See "Treatment of Ewing sarcoma", section on 'Complications in long-term survivors' and "Osteosarcoma: Epidemiology, pathology, clinical presentation, and diagnosis", section on 'Risk factors'.)

Most of the radiation-related bone tumors are osteosarcomas, although other tumors are reported. The magnitude of risk is variable [52,81,111-113,115-117]. Early data suggested that the cumulative risk of a secondary sarcoma after treatment for EFT was approximately 35 percent at 10 years [116]. The Late Effects Study Group, examining the late effects of treatment for childhood cancer, reported an estimated cumulative incidence of secondary sarcomas after treatment of EFT that approached 22 percent at 20 years [112]. These and other investigators identified RT dose as the major predisposing factor for development of a radiation-induced sarcoma, particularly above 60 Gy.

Subsequent experience with protocols utilizing lower doses of RT and tailored RT fields suggest that the magnitude of the risk is somewhat lower [52,81,113,117]. Data are as follows:

●One series included 266 survivors of EFT treated at St. Jude, the National Cancer Institute, and the University of Florida, and followed for a median 9.5 years (range 3 to 30) [81]. The cumulative RT doses were none, 21 to 48, 48 to 60, and >60 Gy in 8, 23, 45, and 24 percent, respectively. Overall, 16 children developed second malignancies, which included 10 sarcomas. The estimated cumulative risk at 20 years was 9.2 percent for any malignancy, and 6.5 percent for a secondary sarcoma. The median time to the diagnosis of the second malignancy after completion of therapy was 7.6 years. All the secondary sarcomas occurred near or at the primary site of the Ewing sarcoma and within the primary irradiated field. As has been noted by others, the cumulative incidence of secondary sarcoma was radiation dose-dependent. No secondary sarcomas developed among patients who had received less than 48 Gy, while the absolute risk of secondary sarcoma was 130 cases per 10,000 person-years of observation among patients who had received ≥60 Gy. (See "Radiation-associated sarcomas", section on 'Radiation dose and age of exposure'.)

●Another series from the Italian Sarcoma Group detailing their experience during the years 1983 to 2006 examined 543 Ewing sarcoma patients (of whom 276 received RT) and observed a cumulative 10-year and 20-year incidence of a second malignant neoplasm (+/-standard error) of 3.4 +/- 0.9 percent and 4.7 +/- 1.6 percent. Of the 15 patients who developed a second malignancy, 10 had received RT [118].

●Another study reported on second malignancies among 674 patients enrolled in the Cooperative Ewing Sarcoma Study Group (CESS)-81 and 86 studies [113]. At average follow-up of 5.1 years, only eight developed a second malignancy, four were treatment-related leukemias/myelodysplastic syndrome, and there were three sarcomas. The cumulative risk of a second malignancy at 10 and 15 years was 2.9 and 4.7 percent, respectively.

●In a systematic review of 52 reports of patients treated for Ewing sarcoma, cumulative incidence rates of second malignant neoplasm ranged from 0.9 to 8.4 percent and 10.1 to 20.5 percent at 5 and 30 years, respectively, after initial diagnosis [119]. Of the 327 reported second malignant neoplasms, 63.6 percent were solid tumors, although acute myeloid leukemia/myelodysplastic syndrome was the single most commonly diagnosed second malignant neoplasm, with generally poor outcomes.

Long-term follow-up guidelines after treatment of childhood malignancy have been published by the COG and are available online [120].

RT and other risk factors for secondary sarcomas after treatment of childhood cancer are discussed further elsewhere. (See "Pathogenetic factors in soft tissue and bone sarcomas", section on 'Radiation therapy and chemotherapy'.)

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Soft tissue sarcoma" and "Society guideline links: Bone sarcomas".)

INFORMATION FOR PATIENTS — UpToDate offers two types of patient education materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces are written in plain language, at the 5^th to 6^th grade reading level, and they answer the four or five key questions a patient might have about a given condition. These articles are best for patients who want a general overview and who prefer short, easy-to-read materials. Beyond the Basics patient education pieces are longer, more sophisticated, and more detailed. These articles are written at the 10^th to 12^th grade reading level and are best for patients who want in-depth information and are comfortable with some medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you to print or e-mail these topics to your patients. (You can also locate patient education articles on a variety of subjects by searching on "patient info" and the keyword(s) of interest.)

●Basics topics (see "Patient education: Ewing sarcoma (The Basics)" and "Patient education: Bone cancer (The Basics)")

SUMMARY AND RECOMMENDATIONS

●Ewing sarcoma is a rare malignancy that most often presents as an undifferentiated primary bone tumor; less commonly, it arises in soft tissue. Both are part of a spectrum of neoplastic diseases known as the Ewing sarcoma family of tumors (EFT), which also includes the more differentiated peripheral primitive neuroectodermal tumor (PNET). These entities share important clinical features, including a tendency toward rapid spread to the lungs, bone, and bone marrow, and a responsiveness to the same chemotherapeutic regimens and radiation therapy (RT). Because relapse rates are high in patients undergoing local therapy alone (80 to 90 percent), it is surmised that the majority have subclinical metastatic disease at the time of diagnosis, even in the absence of overt metastases. Intensive multiagent chemotherapy can eradicate these deposits, and it is a critical component of therapy for both localized and advanced disease. (See 'Introduction' above.)

●For patients with localized EFT, local and systemic therapy are both necessary to achieve cure. (See 'General principles' above and "Treatment of Ewing sarcoma".)

Either surgery or RT can provide effective local control. Contemporary treatment guidelines emphasize surgical resection as the modality of choice if it is believed that the lesion can be resected with negative margins, without excessive morbidity, and with the expectation of a reasonable functional result. A major advantage of surgery is the lack of association with treatment-related sarcomas. For patients who lack a function-preserving surgical option because of tumor location or extent, and those who have clearly unresectable primary tumors following induction chemotherapy, we recommend RT (Grade 1A). (See 'Radiation therapy versus surgery' above.)

●Although the indications for adjuvant RT are not firmly established in patients undergoing surgery for localized EFT, we suggest adjuvant RT in the following situations (Grade 2C) (see 'Adjuvant radiation therapy' above):

•Bulky tumors in difficult sites (eg, the pelvis); in this setting, RT can be given either preoperatively or postoperatively, based on institutional protocols and experience.

•If there is residual microscopic or gross disease after surgery, or there are inadequate surgical margins (ie, a marginal or intralesional resection as compared with a wide or radical excision).

•Adjuvant hemithorax irradiation is indicated in patients with high-risk chest wall primary tumors (close or involved margins, initial pleural effusion, pleural infiltration, and intraoperative contamination of the pleural space).

●Although prognosis is clearly worse than that for localized disease, a subset of patients who present with overt metastatic disease can be cured, particularly if they have limited pulmonary metastases. (See "Treatment of Ewing sarcoma", section on 'Treatment for metastatic disease'.)

●For patients with metastatic EFT, we suggest RT rather than surgery for treatment of the primary site in most patients (Grade 2B). Surgery could be reconsidered in selected patients if chemotherapy results in significant volume reduction, particularly if areas of small-volume metastatic disease are also amenable to surgical resection. (See 'Radiation therapy for metastatic disease' above.)

●For patients with pulmonary metastases that are resected or completely respond to chemotherapy, we suggest evaluation for bilateral low-dose lung irradiation (15 to 18 Gy) rather than observation (Grade 2C). Although the role of whole lung irradiation after resection of pulmonary metastases is unclear, patients may be offered this approach on an individual basis due to the reduced risk of pulmonary relapse, improved event-free survival, and the low rate of expected pulmonary toxicity. (See 'Pulmonary metastases' above.)

ACKNOWLEDGMENT — The UpToDate editorial staff acknowledges Thomas F DeLaney, MD, who contributed to an earlier version of this topic review.