Pediatric Sarcomas of Bone

Etiology and Epidemiology

In the United States, 650 to 700 children and adolescents younger than 20 years of age are diagnosed with bone tumors every year. Ewing sarcoma family of tumors (ESFT) is the rarer of the two with approximately 200 cases diagnosed each year. The incidence of Ewing sarcoma is approximately 2.8 cases per million in children younger than 15 years of age. Generally, the disease occurs in the teenage years during the adolescent growth spurt (ages 10 years to 15 years). However, approximately 30% of cases occur in the first decade of life and 30% occur in the third decade. There is a slight male predominance (1.6:1). Genome-wide analysis suggests ethnic variations in susceptibility genes associated with Ewing sarcoma, but the strong white predominance is not completely understood.

The cause of Ewing sarcoma is unknown, and it does not appear to be induced by any known agents.

Osteosarcoma also is primarily a disease of adolescents and young adults; a different type of osteosarcoma linked to Paget disease occurs in older adults. This section focuses on osteosarcoma in the younger age group. Similar to Ewing sarcoma, the peak incidence of osteosarcoma coincides with a period of rapid bone growth. Osteosarcoma is known to be associated with retinoblastoma gene mutations as well as prior radiation therapy, particularly in children with retinoblastoma or other genetic abnormalities.

Prevention and Early Detection

The rate of secondary osteosarcomas has decreased with the declining use of radiotherapy (RT) in patients with retinoblastoma. Some evidence suggests that limiting dose to less than 60 Gy can reduce the risk of secondary osteosarcoma in patients receiving radiotherapy for Ewing sarcoma. Early detection is possible with careful evaluation of persistent swelling or pain.

Biologic Characteristics and Molecular Biology

The histogenesis of Ewing sarcoma is controversial. The tumor was first described as an endothelioma of bone. Another hypothesis suggests Ewing is a primitive cell of neural origin, specifically from postganglionic, parasympathetic, and primordial cells. Most recently, an alternative hypothesis suggests Ewing cells arise from mesenchymal progenitor or mesenchymal stem cells, which are found in bone marrow. Previously, extraosseous Ewing sarcoma and malignant peripheral neuroectodermal tumor (PNET) were considered separate entities from Ewing sarcoma of bone and were treated differently; genetic studies now confirm they are from the same family of tumors, and together are termed Ewing sarcoma family of tumors . The most frequent chromosomal translocation, t11:22 (q24;q12), results in balanced translocations between the EWSR1 gene and the FLI1 gene, a member of the ETS family of transcription factors; it is found in approximately 85% of Ewing sarcoma cases. Other less-common fusions with EWSR1 and ETS family members, including ERG , ETV1 , ETV4 , or FEV fusions, account for the remaining cases. More than 85% of patients with Ewing sarcoma share expression of a common surface antigen, CD99. Ewing sarcoma, atypical Ewing sarcoma, and PNET of bone exist in the spectrum within this family, from the most undifferentiated tumors to those with neural differentiation. PNET can be differentiated from Ewing sarcoma by the presence of a globular growth pattern, neuron-specific enolase positivity, and Homer Wright rosettes. EWS translocations are also seen in tumors increasingly recognized as having distinct natural history and treatment response. Desmoplastic small round-cell tumor has a characteristic chromosomal translocation involving the fusion of the Wilms’ tumor gene WT1 and the Ewing sarcoma gene EWSR1 , t(11;22)(p13q;q12, which confirms the diagnosis. Other EWS-ETS family of transcription factors fusion-negative Ewing-like tumors recently identified include round cell sarcoma with CIC-DUX4 rearrangement and round cell sarcoma with BCOR-CCNB3 fusions or rearrangements, which appear to behave distinctly from Ewing sarcoma despite histologic similarities.

Osteosarcoma is associated with inactivation of the retinoblastoma tumor-suppressor gene (13q14), which occurs in approximately one-third of cases. Other genetic abnormalities include translocations, gene amplification, and abnormal TP53 function.

Pathology and Pathway of Spread

Ewing sarcoma is an undifferentiated blue round-cell tumor usually of the bone. Its pathologic appearance is a monomorphic pattern of densely packed, small, round, malignant cells with hyperchromatic nuclei and varying amounts of cytoplasm. Immunohistochemical studies reveal cell-surface glycoprotein CD 99 and vimentin, HBA-71, and β ₂ -microglobulin positivity. Occasionally, cytokeratin and neuron-specific enolase are positive. These studies can help differentiate ESFT from other small round-cell malignant tumors of childhood. Approximately 87% of cases within the ESFT are Ewing sarcoma of bone. The remainder are primitive neurectodermal tumors (PNETs) or extraosseous Ewing sarcoma.

About 75% of patients with Ewing sarcoma present with localized disease at diagnosis ( Fig. 80.1 ). Approximately 80% of children experience distant metastases if treated with only local therapy. This suggests that in the majority of cases, unidentifiable micrometastases are present at diagnosis. The most common site of metastasis is the lung, followed by bone. Other distant sites include bone marrow, soft tissues, and rarely, the liver or central nervous system (see Fig. 80.1 ).

Osteosarcoma is derived from bone-forming mesenchyme and is described as a malignant sarcomatous stroma associated with the production of osteoid bone, its defining histopathologic feature. The most common types in the pediatric population are the conventional osteosarcomas, including osteoblastic, chondroblastic, and fibroblastic types. Each type has varying amounts of osteoid formation and a different predominant component. There is no difference in outcome or treatment recommendations among these different types. Other types of less-common osteosarcomas include telangiectatic, small-cell, juxtacortical, periosteal, paraosteal, and high-grade surface sarcomas.

Approximately 90% of children with osteosarcoma present with localized disease at diagnosis ( Fig. 80.2 ). However, if only the primary tumor is treated, about 90% will experience metastatic disease.

Clinical Manifestations, Patient Evaluation, and Staging

Patients with Ewing sarcoma present in general with localized pain, swelling, and a palpable mass. The most common primary tumor sites are illustrated in Fig. 80.1 . In Ewing sarcoma, plain radiographs show a lytic, destructive lesion, most typically of the diaphysis, with or without a soft-tissue mass. Codman's triangle may form from the elevated periosteal reaction. An “onion skin” effect, derived from the development of parallel, multilaminar, periosteal reactions, is typically seen.

Patients with osteosarcoma present with similar signs and symptoms; in general, localized pain, swelling, and a palpable mass. The frequency with which osteosarcoma occurs within the different regions of the body is illustrated in Fig. 80.2 . Plain radiographs of patients with osteosarcoma typically show sclerotic or lytic lesions of the metaphysis. The elevated periosteal reaction may cause a Codman triangle to form. In osteosarcoma, periosteal new bone formation may be present, with the blastic component showing a bony sunburst pattern.

The evaluation for patients with Ewing sarcoma and osteosarcoma is similar. A complete history and physical examination are performed, with particular attention to the duration of symptoms, the presence of pain, difficulty of function, neurologic symptoms, and the location and size of the mass. Studies commonly obtained to evaluate the extent of disease include routine blood work, and plain radiographs; a computed tomography (CT) scan to assess bone involvement and integrity and magnetic resonance imaging (MRI) to assess soft-tissue extent are both required of the primary region. For assessment of possible metastatic disease, ¹⁸ fluorodeoxyglucose-positron emission tomography ( ¹⁸ FDG-PET) has improved sensitivity to detect bone and lymph node metastases compared with bone scan and CT. Tumors that are not particularly ¹⁸ FDG-avid should still undergo bone scanning and, if high-resolution CT is not performed with PET, a diagnostic chest CT scan should be obtained to rule out lung metastases. Because of the high frequency of lung metastases in this population, Children's Oncology Group (COG) protocols consider a solitary solid nodule more than 5 mm or multiple solid nodules less than 5 mm to be metastatic disease unless proven otherwise. This has significant implications for subsequent treatment (i.e., whole-lung radiation) so indeterminate and suspicious lung nodules should be confirmed pathologically. An electrocardiogram and echocardiogram are included in the evaluation before chemotherapy is initiated. In the case of Ewing sarcoma, a bone marrow biopsy is obtained. A summary of staging and follow-up investigations is given in Table 80.1 .

A biopsy of the primary lesion should be obtained after complete evaluation, ideally by the surgeon who will perform the definitive resection. The biopsy should be placed carefully to avoid contamination of uninvolved areas, vital structures, and hematoma development. It must not increase the extent of surgery or preclude a limb-sparing procedure or sparing of a strip of skin outside the radiation port.

Currently, there is no staging system for Ewing sarcoma. Patients are classified as having either localized disease or metastatic disease and are treated accordingly.

In osteosarcoma, available staging systems are those of the Musculoskeletal Tumor Society and, lately, of the International Union against Cancer and American Joint Committee on Cancer (UICC/AJCC).

The prognosis for patients with Ewing sarcoma is dependent on a number of factors, the most important of which is the presence or absence of metastatic disease. Data from European studies suggest histologic response to chemotherapy is the primary predictor of outcome in patients undergoing surgery. This has yet to be confirmed in North American regimens, although single-institution reports suggest response correlates with improved survival. Historically, other prognostic factors include the site of the lesion and its size, and age and gender of the patient. In more recent studies, however, data are conflicting with no association of outcome with tumor site or size reported in INT-154, a COG study evaluating dose-intense multiagent chemotherapy. However, the COG AEWS0031 trial evaluating interval-compressed multiagent chemotherapy, pelvic site was associated with inferior event-free survival (EFS) and OS. The French study EW93 reported poorer outcomes in axial and pelvic tumor, as well as large tumor volumes; however, prognostic factors varied by type of local therapy. For surgically treated patients, location and histologic response correlated with survival. For patients treated with radiotherapy, tumor volume remained significant, but site did not. Historically, older patients and male gender were also associated with poor survival. Yet, when accounting for histologic response in more recent studies, age and gender have not remained significant. Age older than 18 was associated with poorer survival in AEWS0031, but histologic response was not assessed in this study. Multiple retrospective studies suggest molecular biomarkers, such as p53 and p16, may correlate with outcomes; however, prospective validation has not confirmed these findings.

The prognostic features of osteosarcoma are similar, namely, the presence or absence of metastatic disease, the tumor location and size, whether complete resection of the primary tumor was possible, and the response to adjuvant chemotherapy. No prognostic molecular markers have been validated in osteosarcoma.

Treatment of Ewing Sarcoma

Primary Therapy

Historically, cure rates of Ewing sarcoma were less than 20% when treatment was given to the primary lesion only. In the early 1960s, single-institution studies began to show an improved outcome with the addition of adjuvant chemotherapy. Today, the standard treatment consists of local therapy and neoadjuvant and adjuvant multiagent chemotherapy. Local therapy consists of surgery or RT or a combination of both. Concerning systemic therapy, a number of randomized trials have been performed to assess the value of different drug combinations.

The first Intergroup Ewing Sarcoma Study (IESS) investigated the use of multiagent chemotherapy from 1973 until 1978. Patients received RT to the primary lesion and were randomized among three adjuvant chemotherapy treatment arms: vincristine, actinomycin D, and cyclophosphamide (VAC); VAC plus doxorubicin (trade name Adriamycin, so the regimen is known as VACA); or VAC plus bilateral pulmonary radiotherapy. This study showed a significant improvement in all parameters with the addition of doxorubicin. Furthermore, VAC plus bilateral lung irradiation, although less effective than VACA, showed superior results to VAC alone. As a consequence of these results, doxorubicin was considered to be an essential drug in further trials.

The third large intergroup study, INT-0091, investigated the addition of etoposide and ifosfamide to VACA; at 5 years, data showed a statistically significant benefit in EFS and OS from the addition of these two drugs in patients with localized disease at diagnosis. In addition, local control was also significantly improved in patients on the experimental arm. In fact, the beneficial effect of the experimental therapy was associated with a greater reduction in the rate of local recurrence than in the rate of systemic recurrence. In the European Cooperative Intergroup Ewing Sarcoma Study (EICESS-92), the value of single-agent modifications were evaluated. In standard-risk patients (localized disease and tumor volume of < 100 mL), VACA was randomized against VAIA (ifosfamide instead of cyclophosphamide). There was no significant difference in EFS rates between the groups. In patients at high risk (larger tumors or metastatic disease), VAIA was randomized against VAIA plus etoposide; again, no significant difference in EFS rates was observed, but a marginal benefit was demonstrated for patients with large localized disease with the use of etoposide. The COG trial, AEWS0031, compared vincristine, doxorubicin, cyclophosphamide, ifosfamide-etoposide (VDC-IE) dosed every 3 weeks versus every 2 weeks in patients with localized Ewing sarcoma. This trial showed an 8% 5-year EFS benefit for interval compressed chemotherapy. This regimen remains the current standard in the United States for patients with either localized or metastatic disease (see Advanced Disease and Palliation ). Fig. 80.3 depicts the most common treatment schema for Ewing sarcoma in North America. An alternative for Ewing sarcoma treatment intensification used more commonly in European studies has been high-dose chemotherapy with stem cell rescue. Because of its toxicity, this treatment is mainly used for patients at very high risk. In the Euro-EWING 99 trial, patients with localized disease with poor pathologic response after six cycles of induction chemotherapy of vincristine, ifosfamide, doxorubicin, and etoposide (VIDE) were randomized to either a single cycle of high-dose chemotherapy with busulfan and melphalan (BuMel) followed by autologous stem cell transplant or conventional therapy with vincristine, dactinomycin, and ifosfamide (VAI) for seven cycles. Nearly 80% of patients at high risk were defined by poor histologic response (> 10% viable cells) or had a tumor volume at diagnosis greater than 200 mL It is important to note that owing to concerns of myelopathy and bowel injury, patients with central axis tumors were not eligible for randomization and patients with nonresectable pelvic tumors could not receive RT before transplant. The results of this trial showed high-dose chemotherapy was associated with significantly improved 8-year EFS (60.7% vs. 47.1%), which translated to improved 3- and 8-year OS (78.0% vs. 72.2%, and 64.5 vs. 55.6, respectively), suggesting that high-dose chemotherapy improves survival in patients with localized disease at high risk receiving Euro-EWING 99 induction regimen. North American interval compressed regimens include significantly more chemotherapy exposure by the time of local control so it is difficult to compare these results across the Atlantic. In addition, current outcomes are similar between both approaches, with greater than 70% 5-year overall survival for patients with localized disease. Efforts are underway to determine if a subgroup of patients treated with North American regimens may warrant further investigation with high-dose chemotherapy.

A summary of results of Phase III trials is given in Table 80.2 .

TABLE 80.2

Treatment Results in Selected Clinical Studies of Localized Ewing Sarcoma Family Tumors

Study	Reference	Schedule	No. Patients	5-Year Event-Free Survival
COG
IESS-I (1973–1978)	Nesbit, J Clin Oncol 1990;8:1664	VAC	342	24%
		VAC + WLI		44%
		VACD		60%
IESS-II (1978–1982)	Burgert, J Clin Oncol 1990;8:1514	VACD-HD	214	68%
		VACD-MD		48%
First POG-CCG (INT-0091) (1988–1993)	Grier, N Engl J Med 2003;348:694	VACD	200	54%
		VACD + IE	198	69% ( p = 0.005)
		VACD ± IE (metastatic)	120	22% ( p = 0.81)
Second POG-CCG (1995–1998)	Granowetter, J Clin Oncol 2009;27:2536	VCD + IE 48 wk	247	70%
		VCD + IE 30 wk	231	72% ( p = 0.57)
AEWS 0031	Womer, J Clin Oncol 2012;30:4148	VDC + IE 3/wk	284	65%
		VDC + IE 3/wk	284	73% ( p = 0.048)
MEMORIAL SLOAN-KETTERING CANCER CENTER
T2 (1970–1978)	Rosen, Cancer 1978;41:888	VACD (adjuvant)	20	75%
P6 (1990–1995)	Kushner, J Clin Oncol 1995;13:2796	HD-CVD + IE	36	77% (2 year)
P6 (1991–2001)	Kolb, J Clin Oncol 2003;21:3423	HD-CVD + IE	68	Localized: 81% (4 year)
				Metastatic: 12% (4 year)
ST. JUDE CHILDREN'S RESEARCH HOSPITAL
ES-79 (1978–1986)	Hayes, J Clin Oncol 1989;7:208	VACD	52	82% < 8 cm (3 year)
				64% ≥ 8 cm (3 year)
ES-87 (1987–1991)	Meyer, J Clin Oncol 1992;10:1737	Therapeutic window with IE	26	Clinical responses in 96%
EW-92 (1992–1996)	Marina, J Clin Oncol 1999;17:180	VCD-IE × 3	34	78% (3 year)
UKCCSG/MRC
ET-1 (1978–1986)	Craft, Eur J Cancer 1997;33:1061	VACD	120	41%
ET-2 (1987–1993)	Craft, J Clin Oncol 1998;16:3628	VAID	201	62%
CESS
CESS-81 (1981–1985)	Jürgens, Cancer 1988;61:23	VACD	93	< 100 mL 80%
				≥ 100 mL 31% (both 3 year)
				Viable tumor < 10%: 79%
				>10%: 31% (both 3 year)
CESS-86 (1986–1991)	Paulussen, J Clin Oncol 2001;19:1818	< 100 mL (SR): VACD	301	52% (10 year)
		≥ 100 mL (HR): VAID		51% (10 year)
EICESS (CESS PLUS UKCCSG)
EICESS-92 (1992–1999)	Paulussen, J Clin Oncol 2008;6:4385	SR: VAID/VACD	155	68%/67% ( p = 0.72)
		HR: VAID/EVAID	492	44%/52% ( p = 0.12)
ROI/BOLOGNA, ITALY
REN-3 (1991–1997)	Bacci, Eur J Cancer 2002;38:2243	VDC + VIA + IE	157	71%
SFOP/FRANCE
EW-88 (1988–1991)	Oberlin, Br J Cancer 2001;85:1646	VD + VD/VA	141	58%
EW-93 (1993–1999)	Gaspar, Eur J Cancer 2012;48:1376	< 100 mL >95% response (SR): VD + VD/VA	116	70%
		>100 mL, 70% to 95% response, < 50% size response (IR): VD + VD/VA + IE	46	54%
		> 100 mL, <70% response, < 50% size response (HR): VD + VD/VA + IE + HD	48	48%
SSG/SCANDINAVIA
SSG IX (1990–1999)	Elomaa, Eur J Cancer 2000;36:875	VID + PID	88	58% (metastases-free survival)
EURO-EWING (EICESS + SFOP)
Euro-EWING 99 (1999–2005)	Ladenstein, J Clin Oncol 2010;28:3284	R3 (multiple metastases): VIDE + VAI + HD	281	27% (3 year)
Euro-EWING 99 (2000–2014)	Dirkson, J Clin Oncol 2016;34:11001suppl	R2Pulm (lung only metastasis) VIDE+ VAI+ WLI vs. VIDE+ HD	265	53% (3 year)
Euro-EWING 99 and Ewing-2008	Whelan, J Clin Oncol 2018;36:3110	R2Loc (high-risk localized) VIDE + VAI vs. VIDE +HD	240	69% (3 year)

A, Actinomycin D; C, cyclophosphamide; D, doxorubicin; E, etoposide; HD, high dose; HR, high risk; I, ifosfamide; MD, moderate dose; P, cisplatin; SR, standard risk; V, vincristine; WLI, whole lung irradiation.