RMS is one of the “small, round, blue-cell tumors” of childhood. It is the third most common extracranial solid tumor occurring in children. Almost two-thirds of RMS cases develop in children younger than the age of 10 years.^2–5 Since the inception of the pediatric cooperative group process, more than 4,000 patients with RMS have been enrolled in therapeutic trials. The first three decades of investigations were conducted by the Intergroup Rhabdomyosarcoma Study Group (IRSG), an independent National Cancer Institute (NCI)-funded cooperative group that has become a part of the Children's Oncology Group. These multidisciplinary studies led to successive improvements in outcome for children with RMS. The proportion of patients alive at 5 years after start of therapy increased from 55% on IRS-I² to 63% on IRS-II³ to 69% on the IRS-III and 73% on IRS-IV protocols.^4,6

RMS can occur in almost any soft-tissue site in the body; the most common primary sites are genitourinary (24%), parameningeal (16%), extremity (19%), orbit (9%), other head and neck (10%), and miscellaneous other sites (22%).^5,7 Most often RMS presents as a mass; however, the presenting signs and symptoms often correspond with the site of primary tumor. Genitourinary tumors may present with hematuria, difficulty voiding or urinary obstruction, and/or a scrotal or vaginal mass, occasionally associated with vaginal bleeding. Parameningeal tumors may present with cranial nerve dysfunction, signs suggestive of sinusitis, ear discharge, headaches and facial pain. Proptosis or periorbital swelling are common manifestations for orbital tumors. Extremity tumors and those occurring at other sites most often present as a rapidly enlarging, firm mass lesion.

In most cases, there are no clear predisposing risk factors for the development of RMS. Although RMS has been reported in association with neurofibromatosis type 1,⁸ Beckwith-Wiedemann syndrome,^9,10 Li-Fraumeni syndrome,^11,12 cardio-facio-cutaneous syndrome,¹³ Costello syndrome,¹⁴ a variety of congenital anomalies,¹⁵ and parental use of cocaine and marijuana,¹⁶ the vast majority of cases appear to be sporadic.

Accurate and specific diagnosis of RMS is critical for appropriate treatment assignment and analysis of outcome data. RMS is a small, round, blue-cell neoplasm and must be distinguished from neuroblastoma, Ewing tumor, and lymphoma. The hallmark for diagnosis of RMS is the demonstration of malignant skeletal muscle differentiation. On light microscopy, cross-striations in tumor cells, characteristic of skeletal muscle or plump rhabdomyoblasts, are occasionally noted. Often immunohistochemical staining (eg, myogenin, MyoD, muscle-specific actin, myoglobin, and/or desmin) is required for accurate diagnosis. Tumors are classified using the modified International Classification of Rhabdomyosarcoma.¹⁷ This system identifies a favorable histologic group comprising the embryonal histology tumors, including the spindle cell and botryoid variants, and an unfavorable group comprising tumors with alveolar histology and undifferentiated sarcoma.

Embryonal RMS (ERMS) is most common, comprising more than half of all RMS cases. ERMS tends to occur in younger children; these tumors typically occur in the orbit, head and neck, and genitourinary system. Histologically, ERMS is characterized by primitive spindle cells, often with more myxoid areas. The botryoid variant often arises in mucosa-lined hollow organs (nasopharynx, vagina, and bladder), presenting as a grape-like protruding mass and characterized histologically by a subepithelial condensation of tumor cells (“cambium layer”) (Figure 142d-2). Analysis of histologic subtypes of RMS may have additional prognostic significance.¹⁸ For example, spindle-cell RMS histology, accounting for 3% of all RMS, is associated with a superior prognosis (95% 5-year survival).¹⁷ For paratesticular spindle-cell RMS, the prognosis is significantly impacted by extent of T-cell-rich lymphocytes infiltrating in tumor.¹⁹ The prognosis of conventional embryonal RMS is intermediate; this group may include tumors with either favorable or unfavorable features. Although a grading system often used for other soft-tissue sarcomas has not been routinely used for RMS, grading of ERMS by degree of maturation, proliferation, and apoptosis may serve, in lieu of known biologic markers, to further subdivide this intermediate prognostic category of RMS into better defined prognostic subgroups. RMS with diffuse anaplasia (2% of all RMS) has a poor prognosis (45% 5-year survival).¹⁷ The fate of RMS with focal anaplasia (2% of RMS) is less clear. ERMS is characterized by loss of heterozygosity on the short arm of chromosome 11 (11p15.5), suggesting inactivation of a tumor-suppressor gene.^20,21 This is the region of the insulin-like growth factor (IGF)-2 gene that is overexpressed in rhabdomyosarcoma.²²

Figure 142d-2

High-power photomicrograph of embryonal rhabdomyosarcoma, botryoid variant, with primitive spindle-shaped cells and the characteristic “cambium” layer of ovoid cells in the submucosal zone. (Four-color version of figure on CD-ROM)

Alveolar histology tumors comprise approximately 25% of RMS. This histology is more common in tumors arising in adolescents and in extremity primary tumors. Histologically, alveolar rhabdomyosarcoma (ARMS) is composed of ill-defined round- or oval-cell tumor aggregates, often with loss of cellular cohesion, forming “alveolar” spaces surrounded by a framework of fibrous septae (Figure 142d-3). A solid variant of ARMS with cellular nests separated by fibrovascular septa is less common. Numerous studies demonstrate that the PAX3-FKHR and PAX7-FKHR gene fusions that are generated by t(2;13) and t(1;13) chromosomal translocations are specific and consistent features of the alveolar subtype of RMS (Figure 142d-4).²³ The fusion transcripts expressed from these gene fusions can be detected by highly sensitive, efficient, and objective polymerase chain reaction (PCR)-based and fluorescence in situ hybridization (FISH) assays.^24,25 In conjunction with standard histopathologic analysis, this molecular testing facilitates accurate identification and subclassification of RMS tumors and subsequent assignment to risk-based therapeutic protocols. In particular, this molecular testing provides valuable confirmation of the “higher risk” alveolar subtype in cases where histopathologic evidence of an alveolar pattern is subtle or equivocal. The PAX3-FKHR fusion of t(2;13) is associated with classic cystic ARMS and giant cells; PAX7-FKHR t(1;13) tumors may have a more subtle ARMS histology and lower apoptotic/mitotic activity.²⁶

Figure 142d-3

High-power photomicrograph of alveolar rhabdomyosarcoma, with fibrovascular septae, loss of cellular cohesion and occasional giant cells. (Four-color version of figure on CD-ROM)

Figure 142d-4

Gene fusions in alveolar rhabdomyosarcoma. PAX3/FKHR [t(2;13)] is created by fusion of PAX3, localized to 2q25 (in blue and gray), and FKHR, localized to 13q14 (tan and yellow); PAX7/FKHR [t(1;13)] is created by fusion of PAX7, localized to 1q36 (in green (more...)

Open biopsy is usually required to obtain sufficient tissue for accurate diagnosis and also provides invaluable tissue for molecular characterization and other investigations. Complete clinical and radiologic evaluation should be done prior to consideration of primary tumor resection. The primary tumor should be assessed by either magnetic resonance imaging (MRI) or computerized tomography (CT) imaging to determine location, size, invasiveness, and anatomic boundaries that will determine the necessary local therapy. Appropriate metastatic evaluation includes CT scanning of lungs, bone scan, bilateral bone marrow aspirates/biopsies, and, for extremity and paratesticular tumors, imaging assessment of regional lymph nodes. Regional lymphatic spread of tumor often occurs in extremity tumors and in paratesticular tumors, particularly in those older than 10 years of age. Only approximately 15% to 20% of patients will have clinically detectable metastatic disease at diagnosis; however, all patients are considered to have micrometastatic disease, providing the rationale for universal chemotherapy. The most common sites for metastatic disease are lungs, bone marrow, and bones.

Extent of disease is among the strongest predictors for long-term outcome. Appropriate treatment planning is based primarily on the IRSG Clinical Group (Table 142d-1) and Stage (Table 142d-2). Clinical Group assignment has been used for all IRSG studies, and primarily is determined by the surgical resectability of the primary tumor and presence of metastatic disease.² In 1991, a site modification of the tumor, node, metastases (TNM) system was adopted that assigns stage based on pretreatment assessment of tumor site, size, and regional and systemic tumor spread.^27,28 Patients with completely excised tumors (group I), tumors arising in favorable sites (stage 1), and those without regional node involvement (N₀), consistently have the most favorable outcomes, whereas patients with alveolar histology tumors and metastatic disease have significantly poorer outcomes. The present Children's Oncology Group protocols for RMS assign patients to one of three risk groups (Table 142d-3). Multidisciplinary evaluation including assessment by the surgical subspecialist and radiation oncologist is critical for appropriate treatment planning.

Table 142d-1

Intergroup Rhabdomyosarcoma Study Group Clinical Grouping for Rhabdomyosarcoma.

Table 142d-2

Intergroup Rhabdomyosarcoma Study Group Site-Modified TNM (Tumor, Node, Metastases) Staging.

Table 142d-3

Predicted Outcome for Risk Groups of Rhabdomyosarcoma.

Complete surgical excision of tumor is correlated with better outcome and, for embryonal tumors, obviates the need for local irradiation. However, rhabdomyosarcoma arises in many sites that do not permit primary surgical excision without significant morbidity or loss of function. Fewer than 20% of patients have tumors that are completely excised with negative tumor margins (clinical group I). Data from IRS II and III demonstrate a benefit for pretreatment reexcision of those primary tumors occurring in the trunk and extremities to assure adequate resection margins²⁹; therefore such reexcision of the tumor bed is recommended for these sites when necessary. However, better methods are needed to assess adequacy of tumor margins and are proposed for future studies. For some tumors, delayed tumor resection after a period of chemotherapy may minimize surgical morbidity and decrease doses of radiotherapy required. Approximately half of patients older than age 10 years with paratesticular RMS,^30,31 and a similar proportion of patients with extremity and trunk primaries (regardless of age), have regional node involvement.^32,33 For these two groups, regional lymph node sampling is required. Regional node involvement in extremity RMS is associated with significantly poorer event-free survival (EFS) (p < .001).^32,33 The recent ability to assess sentinel node involvement³⁴ potentially provides a refined method for risk assessment in RMS.

Multiagent chemotherapy is indicated for all patients with rhabdomyosarcoma. Combination therapy including vincristine, dactinomycin, and cyclophosphamide (VAC) continues to be the mainstay of effective, curative therapy. Several other antineoplastic agents are active against RMS but have not improved outcome. The overall 5-year survival for the most favorable subset of patients with low risk RMS is 95% when treated with vincristine and dactinomycin chemotherapy administered for 48 weeks. European studies suggest that excellent outcome may be achieved with shorter therapy in this group of patients.³⁵ Future studies will test whether further substantial decreases in length of therapy will maintain excellent outcomes. For the less favorable low-risk patients three-drug therapy is necessary to achieve similar excellent outcome. Patients at intermediate risk of treatment failure also require three-drug chemotherapy. Based on activity of ifosfamide³⁶ and ifosfamide/etoposide in IRSG and other window trials and in relapse Phase II trials,^37,38 the IRS-IV study prospectively randomized patients with intermediate-risk RMS to VAC, or to vincristine, dactinomycin, ifosfamide (VAI), or to vincristine, ifosfamide, etoposide (VIE). The 3-year failure-free survivals were 75%, 77%, and 77%, respectively, indicating no advantage for ifosfamide-containing regimens.⁶ Recently the topoisomerase I poisons demonstrated clinical activity as predicted from human tumor xenograft models of RMS.^39–41 Both topotecan alone and topotecan/cyclophosphamide caused tumor regression in approximately 50% of patients enrolled on Phase II window trials^42,43; the two-drug combination also is active in patients with recurrent RMS.⁴⁴ The present cooperative group study is comparing VAC alone to VAC plus vincristine, topotecan, and cyclophosphamide in a prospective, randomized trial for patients with intermediate-risk rhabdomyosarcoma.

The highest-risk population comprises patients with alveolar tumors and those who are older than 10 years of age with embryonal tumors who have macrometastatic disease at diagnosis. Fewer than 25% of these children will survive. The IRSG has used the Phase II window strategy^45,46 to assess the activity of several new agent combinations, prioritizing for clinical evaluation those agents with high levels of activity in the preclinical human rhabdomyosarcoma xenograft model. Using this strategy, several single agents and two-drug combinations have shown substantial clinical activity, including vincristine/melphalan, ifosfamide/etoposide,⁴⁷ ifosfamide/doxorubicin,⁴⁸ topotecan,⁴² and topotecan/cyclopho- sphamide.⁴³ The present high-risk study tests the combination of irinotecan/vincristine, based on the preclinical⁴⁰ and clinical activity for prolonged exposures of irinotecan,^49–51 the activity of single-agent vincristine and the predicted synergism of this combination.⁵²

For patients with embryonal RMS and microscopic or gross residual disease after surgical excision (clinical groups II to IV) and for all patients with alveolar RMS, radiation therapy plays a critical role for successful local tumor control. The dose of radiotherapy required varies dependent on tumor site, group, and response to chemotherapy, with total doses ranging from 36 to 50.4 Gy. Higher doses are indicated for patients with gross residual disease after initial surgery and chemotherapy. The IRS-IV study was the first, prospective evaluation of radiation therapy technique in children, comparing once daily conventional fractionation to a twice daily hyperfractionation schedule.⁵³ To date, this study has shown no benefit for the hyperfractionation schedule. For patients with tumors that erode the skull base or extend into the intracranial space or cause spinal cord compression, immediate initiation of radiotherapy is indicated.^54,55 For most other patients radiotherapy is delayed until either week 3 or 12 of therapy. Patients with tumors arising in the vagina may have local control delayed even longer so that late effects of therapy may be minimized.^56,57 To minimize the late effects of radiotherapy and decrease toxicity to surrounding normal tissues, new innovative techniques may be indicated in selected patients, including brachytherapy, three-dimensional conformal treatment planning, and intensitymodulated radiation therapy.

Most children and adolescents with favorable and intermediate risk RMS will be cured with presently available multidisciplinary therapeutic approaches. Strategies in these groups to minimize the late effects of therapy are warranted. For the lower-risk patients, efforts are underway to decrease treatment length and minimize exposure to alkylating agents that will decrease the risk of infertility and second malignancies. Newer radiotherapy guidelines and treatment techniques are aimed at decreasing late toxicities, including neuroendocrine dysfunction, soft tissue and skeletal growth effects, and development of secondary sarcomas.

Of the children who relapse, a small subset with group I embryonal histology tumors at diagnosis and only locally recurrent tumor is potentially curable. However, more than 80% of children with tumor recurrence have a dismal outcome.⁵⁸ In these very-poor-risk patients, new and innovative therapy approaches must be attempted. Combinations of irinotecan and vincristine, shown to be synergistic in the xenograft model, and the hypoxic agent tirapazamine^59,60 now are being evaluated in the ongoing cooperative group relapse study. In addition, other potential modalities and targets for therapeutic interventions that are being evaluated in early clinical trials or preclinical models include vaccines generated against small peptide fragments spanning the PAX3-FKHR fusion,⁶¹ epidermal growth factor receptor tyrosine kinase inhibitors,^62,63 tumor necrosis factor-related apoptosis-inducing ligand (TRAIL),⁶⁴ and rapamycin analogs.