Rhabdomyosarcomas are highly malignant neoplasms that originate from the same embryonic mesenchyme that gives rise to striated skeletal muscle. These tumors can arise virtually anywhere in the body, and they tend to disseminate early in the course of the disease.
Survival rates of children with this tumor have increased remarkably over the past 15 years. This achievement was possible largely because of rigid application of established principles of cancer treatment. The clinical approach consists of comprehensive diagnostic evaluation to ascertain the extent of the disease, control of the primary tumor with modern surgical and radiotherapeutic techniques, eradication of micrometastasis through the use of concomitant adjuvant combination chemotherapy, and use of vigorous supportive therapy. A better understanding of the biological characteristics of this tumor and early identification of groups of patients at high risk of relapse also contributed to the improved outcome. The overall survival rate has risen from less than 20% to approximately 70%, and new techniques and therapeutic approaches promise to further improve the outlook of children with rhabdomyosarcomas in the future.
The earliest description of a rhabdomyosarcoma was published by Weber in 1854. ' The embryonal type, the most common form of childhood rhabdomyosarcoma, was identified by Stobbe and Dargeon in 1950. Previous to that, the pleomorphic form, which is actually rare in children, was thought to be the only histologic type.2
In 1958, Horn and Enterline established the presence of four subtypes: pleomorphic, embryonal, alveolar, and botryoid. The botryoid form was considered to be a morphological variant of the embryonal type. i Recently, three additional sarcomas of primitive cell types, occurring in soft tissues, have been recognized. These lesions have been termed "extraosseous Ewing's sarcoma" because of their morphological similarities to Ewing's sarcoma (a tumor of bone origin). Extraosseous Ewing's sarcomas have been further divided into special undifferentiated cell type I (small cell) and type II (large cell), plus a small number of lesions classified as small mesenchymal sarcoma, type indeterminate.4
Surgical removal of the primary tumor was the first successful method of treatment, but radical surgery was often required for long-term disease control. Radiation therapy was first used in 1950 when Stobbe and Dargeon showed that at least some rhabdomyosarcomas were relatively rathosensitive.2 In 1959, D'Angio noted the favorable synergistic effect of dactinomycin and radiation in the treatment of several tumors, including rhabdomyosarcoma.5
Diagnostic Imaging Methods: Effective Uses and Limitations
The initial use of chemotherapy for treatment of rhabdomyosarcomas consisted of giving single drugs to patients with metastases. Dactinomycin, vincristine, and cyclophosphamide produced either complete or partial regression of the tumor in some patients. However, improvement was of short duration. Subsequently, combined drug therapy using two or three agents was tried for treatment of metastatic disease. Pinkel and Pickren suggested a coordinated program of surgery, radiation therapy, and chemotherapy for the treatment of rhabdomyosarcoma. They also proposed the concept of prophylactic chemotherapy to be given following complete excision of the tumor.6
A study reported by Heyn7 for the Children's Cancer Study Group (CCSG) confirmed the validity of the concept of prophylactic therapy. A total of 28 children with localized tumors were made grossly tumor-free by surgery. These children were then given postoperative radiation therapy together with a regimen of dactinomycin and vincristine lasting one year. In this group, 24 (85. 7%) were relapse-free at two years following initial treatment. In the control group who received no chemotherapy, only 7 out of 15 (47%) were relapse-free at two years. Wilbur advocated giving more intensive chemotherapy prior to surgery for treatment of more extensive lesions. The goal was to make these large and invasive tumors more amenable to surgical removal or radiation therapy.8
In 1972, following the organizational model of the National Wilm's Tumor Study Group, the three national cancer treatment study groups pooled their resources. The Cancer and Leukemia Group B (CALGB), the Children's Cancer Study Group, and the Southwest Oncology Group (SWOG) combined forces and established the Intergroup Rhabdomyosarcoma Study (IRS). The members of this new group implemented a prospective and multidisciplinary study aimed at improving understanding of the natural history of rhabdomyosarcoma and finding more effective therapy for this malignancy.
INCIDENCE AND ASSOCIATED CONDITIONS
Rhabdomyosarcoma is the most common soft tissue sarcoma in children. It represents between 5% and 15% of all malignant solid tumors and 4% to 8% of all cases of malignant disease in children under 15 years of age.9 This tumor ranks seventh among the common malignant neoplastic diseases of childhood, preceded by leukemia, tumors of the central nervous system, lymphoma, neuroblastoma, Wilms' tumor, and bone cancers.9 The annual incidence rate of rhabdomyosarcomas in the United States is estimated to be 4-4 per million in white children and 1.3 per million in black children under age 15. Males are predominantly affected, with a male to female ratio of 1.4:14
Relatives of children with rhabdomyosarcoma have a high frequency of carcinoma of the breast and other sites. 10 Rhabdomyosarcomas occur with increased frequency in persons with neurofibromatosis. " Bone sarcomas occur as second malignant neoplasms in patients with rhabdomyosarcomas.12 This association raises the question of whether there is a genetic predisposition to both cancers, whether development of a second cancer is a side effect of therapy to the original rhabdomyosarcoma, or a combination of both.
Autopsies of 43 children in the IRS showed that seven had associated congenital anomalies. 0( these seven cases, four involved malformations of the central nervous system.13
As many as 70% of patients with rhabdomyosarcoma are under 10 years of age at the time of presentation; the peak incidence is in children between 2 and 5 years old. A second peak of incidence occurs between 15 and 19 years of age. In these older patients, tumors predominate in the head and neck, extremities, and genitourinary tract. In the latter site, paratesticular tumors predominate. Younger children, those under 2 years of age, tend to present with more localized tumors, whereas those between 11 and 15 years tend to present with more extensive or disseminated disease.
The common sites of origin in order of frequency are the head and neck including the orbit (38%), genitourinary tract (21%), extremities (18%), trunk (7%), retroperitoneum (7%), the intrathoracic region (3%), the gastrointestinal-hepatobiliary tract (3%), the perineum and anus (2%), and other sites (1%). I4
Signs and symptoms relate either to the location of the primary tumor or to metastases. The most common presenting sign is an asymptomatic mass that is usually first detected by the patient or parent. Orbital tumors produce proptosis. Nasopharyngeal tumors cause nasal voice, obstruction of the airways, epistaxis, sinusitis, pain, and dysphagia. Middle ear involvement results in pain and chronic otitis media. Parameningeal lesions, such as those in the nasopharynx, nasal cavity, paranasal sinuses and middle ear, may extend directly into the central nervous system in as many as 35% of patients. This parameningeal spread of the tumor may cause cranial nerve palsies, meningeal symptoms and respiratory paralysis as a result of infiltration to the brain stem.15
Rhabdomyosarcomas of the trunk, extremities, and the paratesticular area usually appear as enlarging soft tissue masses. Approximately 40% of patients with paratesticular tumors have involvement of retroperitoneal lymph nodes. I6 The tumor can arise in the prostate and bladder, causing obstruction of the urinary tract and other symptoms. In girls, the well known sarcoma, botryoides, may originate in the vagina or uterus, appearing as a polypoid lesion with a hemorrhagic discharge.
Rhabdomyosarcomas extend locally, they infiltrate along fascial planes and into surrounding tissues. The margins of the tumor are usually indistinct. Metastases result from lymphatic and hematogenous spread, with the most common sites of metastases being the lungs, bones, bone marrow, lymph nodes, brain, spinal cord, and heart.
The purposes of diagnostic evaluation are to: 1) define the extent of disease for staging and treatment planning, and 2) provide objective criteria (baseline data) for measuring response to therapy. Definite diagnosis can only be established by histopathological examination of the tumor.
The workup starts with a complete history and a thorough physical examination. A complete blood count, urinalysis, and a blood chemistry profile should be obtained as part of the initial evaluation. The blood chemistry studies should include creatinine, bilirubin, uric acid, alkaline phosphatase, serum glutamic oxalacetic transaminase (SGOT or AST) and lactic dehydrogenase (LDH). Bone marrow aspiration and biopsy should be obtained as this tumor can metastasize to the marrow.
The recommended use of the various modalities of diagnostic imaging will be discussed according to location of the tumor (Table I).17
Head and Neck
Meningeal extension from primary tumors in the parameningeal areas has been recognized with increased frequency. This form of spread has considerable prognostic significance. Computed tomography (CT scan)18 and magnetic resonance imaging (MRI) are the primary imaging modalities for patients with parameningeal tumors. CT or MRI examination can determine the presence or absence of bone erosions at the base of the skull (Figure 1) and detect intracranial extension of the tumor and the presence of metastases in the neck.
Figure 1. An 8-year-old boy with a rhabdomyosarcoma of the nasopharynx extending into the right maxillary sinus and base of the skull. MRI shows that the lateral wall of the maxillary sinus as well as the pterygoid plates are destroyed. A. Axial plane. B. Coronal plate.
Radionuclide bone scans are not reliable in detecting tumor involvement of the base of the skull. Gallium scans help detect metastatic involvement of lymph nodes in the cervical region. Brain scans are sensitive and specific in detecting intracranial metastases, especially those involving the meninges.19
The cerebrospinal fluid should also be examined. This study may reveal meningeal seeding by the presence of tumor cells or pleocytosis and by elevation of protein and reduction of glucose in the spinal fluid. CT scan or MRI are the best imaging modalities for detecting a change in the size of the primary tumor and for assessing the degree of osseous involvement in the follow-up of children with rhabdomyosarcomas of the head and neck.
The favorable impact of adjuvant chemotherapy on survival of children with pelvic rhabdomyosarcomas has resulted in a modified treatment approach. Chemotherapy is given before surgery and radiation therapy, instead of after these therapeutic modalities. This approach requires close and frequent evaluation of the extent of disease so that surgery can be done at the appropriate time with regard to the tumor's response to chemotherapy (Figure 2).
Ultrasound should be the initial imaging modality used to evaluate these patients. The presence of a soft tissue mass displacing or deforming the urinary bladder and the uterus may be the first evidence of a pelvic rhabdomyosarcoma.20 Computed tomography with oral and intravenous contrast should be performed. The regional and retroperitoneal lymph nodes should be assessed for evidence of tumor involvement by ultrasound or CT scan, thus obviating the need for lymphangiography.
Skeletal scintigraphy can be used to identify osseous involvement by direct extension of the tumor. Gallium scans of the pelvis are difficult due to the excretion of this radionuclide in urine and feces, resulting in a false negative examination. However, this technique can be helpful in demonstrating extrapelvic retroperitoneal nodal metastases.19
Follow-up of patients with pelvic rhabdomyosarcomas can be done using either ultrasound or CT scan, except in those patients who have had extensive surgery, which may have included total resection of the urinary bladder and diversion of the colon. In these patients, assessment of the pelvis can only be performed by imaging modalities less dependent upon anatomic integrity, such as CT and gallium scans.21
Rhabdomyosarcomas may arise from the chest wall, hemidiaphragm, heart, mediastinum, or pulmonary parenchyma. Except for the difficulty of evaluating diaphragmatic masses, which are better assessed by ultrasound,22 CT is the modality of choice for the initial evaluation and follow-up of tumors originating in the chest.
Skeletal scintigraphy probably has little applicability in evaluating bone tumors of the chest wall that have been demonstrated by other imaging modalities. Still, a bone scan can be helpful in identifying rib involvement by hone metastases.
Figure 2A. Pyelogram of a 2-year-old boy with rhabdomyosarcoma of the bladder.
Rhabdomyosarcomas may originate in the muscle of the abdominal wall or from retroperitoneal structures. Computed tomography, ultrasound, or both may be used to determine the location and extent of the tumor. But in those tumors arising in the retroperitoneum, CT and ultrasound may not be helpful in differentiating a rhabdomyosarcoma from neuroblastoma or lymphoma.
Skeletal scintigraphy may demonstrate osseous involvement as well as abnormal extraosseous uptake similar to that seen in neuroblastomas. Retroperitoneal rhabdomyosarcomas may demonstrate uptake of gallium, which is an unusual finding in neuroblastomas. Hepatobiliary scintigraphy may demonstrate an intrahepatic mass and the presence of bile duct obstruction.
Rhabdomyosarcomas of the extremities are clearly demonstrated by imaging procedures. Magnetic resonance imaging is the best modality for imaging in this location and is helpful to the surgeon in deciding on extent of surgical excision. CT scans have better resolution than MRI in demonstrating bone involvement, which can also be evaluated by skeletal scintigraphy. 2 } Gallium scans are valuable in extremity lesions, primarily to detect regional nodal metastases, a frequent finding in tumors of this location.
Figure 2B. CT examination following intravenous enhancement shows the tumor fills most of the entire bladder.
Figure 2C. Examination one year after shows total resolution of the mass.
Metastases to the lungs are frequent, and they tend to be subpleural and frequently located at the base of the lungs. Computed tomography of the chest is the best method for detection of these metastases. When they involve the chest wall, and particularly when they are associated with pleural effusions, ultrasound becomes an excellent method of evaluation.
Regional nodal metastases are best evaluated by gallium scintigraphy, and osseous metastases are better studied by conventional bone scans. Hepatic metastases can be detected using a combination of ultrasound, CT scan, and liver scintigraphy. Metastases to the brain should he evaluated by CT scan or MRI. Cerebral scintigraphy is reserved for use in those patients who have equivocal or negative cranial CT scans and symptoms suggestive of intracranial metastases.
Biological Markers in Rhabdomyosarcomas
Biological markers have not as yet been identified for all pediatric malignancies. Although for some tumors biological markers are considered to be of extreme importance in diagnosis and prognosis, in others they are of questionable value or of no use at all. Even though no biological markers have yet been identified tor rhabdomyosarcomas, the growing interest in the production and release of cellular molecules by cancer cells may lead to discovery of specific rhabdomyosarcoma markers. Nevertheless, several biological markers including enzymes, antigens, proteins, and cytogenetic abnormalities have been found, which may be useful in diagnosing and identifying rhabdomyosarcomas (Table 2).
Creatinine kinase is a widely distributed enzyme generally associated with the energy metabolism of contractile systems. The enzyme is a dirtier. Thus, the twosubunits, designated M (muscle type) and B (brain type) form three different dimeric isoenzymes (ie, MM, MB, and BB). Skeletal muscle has the greatest concentration of creatinine kinase activity, almost exclusively the MM isoenzyme. Not surprisingly, creatinine kinase subunit M has been found to be a useful marker tor distinguishing poorly differentiated rhabdomyosarcomas from other types of small round cell tumors in children.24 Most recently, high plasma activity of creatinine kinase MB isoenzyme has been observed in association with rhabdomyosarcoma.2"' Although it is of limited use as a biological marker, creatinine kinase may prove to be of significant value in the differentiation of soft tissue sarcomas of childhood.
The detection of tumor associated antigen (TAA) in the urine of patients having any of a variety of sarcomas has attracted attention to these antigens. By postoperative measurements of TAA in sarcoma patients, Huth and colleagues were able to identity those patients at high risk for the development of recurrence.20 Unfortunately, tumor associated antigens lack specificity since they can be demonstrated in a variety of human malignant neoplasms.
It has recently been shown that desmin, an intermediate protein type, is a very useful marker tor rhabdomyosarcoma. 2/ This intermediate protein is readily characterized by the immunoperoxidase method using antibodies against desmin. It yields few, it any, false positives or false negatives and can thus be used to distinguish rhabdomyosarcomas from other round cell tumors of children.
Although chromosome abnormalities, primarily chromosomal aneuploides, have been observed in rhabdomyosarcomas, a single consistent chromosomal abnormality associated with this tumor has not been found. But frequent alterations of the short arm of chromosome 3 between regions 14 and 21 were recently noted.2* Also, the finding of translocation t(2:13) in an alveolar rhabdomyosarcoma may help in discriminating between the embryonal and alveolar subtypes of this tumor.29 Finally, a mutant allele locus was shown to be present on chromosome 11 in rhabdomyosarcomas. ,0 Cytogenetic abnormalities may become important prognostic biological markers in rhabdomyosarcomas, as they are with other malignancies.
A malignant tumor of skeletal muscle histogenesis, the rhabdomyosarcoma is the most common soft tissue sarcoma of childhood and occurs relatively rarely in adults.
On gross examination, rhabdomyosarcomas are rubbery and gray-white to tan in color. They are usually associated with extensive necrosis and hemorrhage and have ill-defined margins. The characteristic histologic feature for the diagnosis of rhabdomyosarcomas is the presence of malignant rhabdomyoblasts, malignant cells with cross-striations, or both. The pathological classification oí Horn and Enterline, i which is widely accepted, subclassifies tumors with recognizable myogenesis into four histologic groups: embryonal, botryoid, alveolar, and pleomorphic.
The embryonal type may contain round or spindle cells, with the round cells sometimes being difficult to differentiate from lymphocytes, Ewing's sarcoma cells, and neuroblasts. The median age of patients with the embryonal type is 8 years. The embryonal variant accounts for 56% oí the patients51 and is most commonly found in the head and neck, and genitourinary tract. Sarcoma botryoides is a form of embryonal rhabdomyosarcoma and usually has a grape-like appearance. The botryoid histology occurs in any submucosal site, but is most frequently seen in the genitourinary tract or in the head and neck region. These polypoid tumors consist of small, round cells, with an edematous and mixoid matrix in a submucosal location. The botryoid variant is commonly seen in younger children and accounts for about 5% of all histologic types of rhabdomyosarcomas.
The alveolar rhabdomyosarcoma variant is characterized by a prominent alveolar arrangement of stroma and cells. In this variant, rhabdomyoblasts have a tendency to form syncytial multinucleated giant cells, and cross-striations can be found in about 40% of the tumors. The alveolar variant is most frequently seen in tumors originating in the extremities, trunk, and perianal region; they account for 20% of the rhabdomyosarcomas seen in children. The pleomorphic rhabdomyosarcoma is mainly characterized by its extensive pleomorphism and the large size ot the tumor cell.
Other variants of soft tissue sarcomas have been identified, including the extraosseous Ewing's sarcoma52 and most recently the group called undifferentiated rhabdomyosarcomas. The extraosseous Ewing's sarcoma is composed of uniform, small, round or oval cells with scanty cytoplasm. These tumors arise next to bone, but do not have their origin within the bone. However, they have the same morphologic and cytologic features as Ewing's sarcoma. These tumors should not be interpreted as a variant of rhabdomyosarcoma as they are considered analogous to the large cell Ewing's sarcoma. 5 5
Another tumor variant with a cytology different from any oi the rhabdomyosarcoma subtypes has been identified. *3 This variant is referred to as small, round cell sarcoma, type indeterminate. It is thought to represent a mesenchymal tumor arising from a primitive cell type where its cell line of differentiation could not be defined.
The histologic variant of rhabdomyosarcoma present in a given patient markedly influences survival. 34 The proportion of patients surviving as long as three years in IRS I and II was as follows: botryoid 88%, extraosseous Ewing's sarcoma 75%, embryonal 68%, and alveolar and small round cell type indeterminate 53%. Furthermore, the alveolar variant of rhabdomyosarcoma has been shown to have the highest incidence oí distant metastases and regional node involvement, and a low rate of local recurrence.35
Recently, using cytologic and histologic characteristics, Palmer introduced a set of criteria that identifies features associated with an unfavorable prognosis. î6·37 Using these criteria, he was able to identify two different subtypes, the anaplastic and monomorphous forms, both oí which have an unfavorable outcome. However, these findings have not yet been confirmed by other pathologists.
Several other pathological classifications of rhabdomyosarcoma have been proposed, and they have added to the difficulty in comparing patients studied in various clinical trials. At present, a large international study in pathological classification sponsored by the National Institutes of Health is being conducted. This study may result in unification oí the pathological classification of rhabdomyosarcomas.
CLINICAL GROUPING AND STAGING
Despite major improvement in classification of clinical groups, rhabdomyosarcomas continue to pose difficult problems in staging and assessing their behavior. The clinical grouping system introduced by the IRS group in 1972 is the one most commonly used (Table 3). 5S This system is based primarily on the extent of disease and the type of surgery performed. However, variations oí this system have been used by several institutions. Others have adopted individual staging criteria, using other important biological characteristics oí the tumor, such as those related to primary site, evidence of lymph node involvement, tumor size, etc.
The major difficulty with the IRS clinical group system and other variants is that they depend on data obtained after surgery has been performed or attempted. The International Union Against Cancer (UICC) and the American Joint Committee on Cancer have proposed staging criteria for all pediatric tumors, }9 with a subdivision for soft tissue sarcomas (Tables 4 and 5). These criteria are based on tumor invasiveness (T), nodal involvement (N) and distant metastases (M). This system is called TNM. The TNM system also divides tumor histopathology in two major groups: H-I favorable histology (such as embryonal, botryoid, mixed, and undifferentiated) and H-2 unfavorable histology (namely, alveolar, monomorphus, and anaplastic).
Obviously, there is great need for a unique and reproducible pretreatment staging system that will allow accurate evaluation and meaningful comparison of various treatment programs. Donaldson and Belli emphasized a need for a system that can predict the patient's prognosis based on biology and natural history of the tumor.40 They have proposed a clinical staging system also based on TNM status, and they give consideration to the primary site of the tumor as a factor influencing staging. This system would prevent the bias of the IRS system, which relies on the therapeutic strategy initially used, rather than the natural history of the disease. Recently, various pretreatment factors in patients entered in the IRS II study were reviewed in an effort to document the basis for tumor invasiveness, tumor size, and metastatic aspects of the UICC system.41 However, using the same data they failed to document the significance of nodal involvement as proposed by Donaldson and Belli.
Surgical-Pathological Grouping System for Rhabdomyosarcomas
Categories in the Pretreatment Clinical Staging System Proposed by the International Union Against Cancer (UICC)
Rhabdomyosarcoma is an example of another pediatric malignancy in which the introduction of a multidisciplinary comprehensive approach has resulted in significant improvements in therapy outcome. Rhabdomyosarcomas are relatively rare tumors. It is evident that the intergroup study mechanism has been an effective way of studying and treating this malignancy. The overall cure rate for childhood rhabdomyosarcoma has improved from 55% in children treated in the IRS I,42 to about 65% in those treated in the IRS II.43
Meanwhile, treatment has been redefined for ditferent primary sites and clinical groups as these factors have been shown to be most important in determining survival (Figures 3 and 4)- These refinements in the therapy of rhabdomyosarcoma have been achieved in all three therapy modalities: chemotherapy, surgery, and radiotherapy. The following recommendations for treatment have evolved.
The combination of vincristine, actinomycin-D and cyclophosphamide (VAC) continues to be the most effective chemotherapy regimen for rhabdomyosarcoma.44 The role of adriamycin in the treatment of childhood rhabdomyosarcoma is still not clear. It has not yet been proven that giving adriamycin as an adjuvant to the combination VAC is better than giving VAC alone. A recent IRS trial alternated pulse VAC and pulse VadrC (vincristine, adriamycin, and cyclophosphamide) tor the treatment of group III and IV patients. This study failed to improve the results obtained with pulse VAC alone.45 Cisplarin (Platinol®) and the etoposide (VP-16) have shown antitumor activity against rhabdomyosarcoma and undifferentiated sarcomas.46-4' As a result, these drugs are now included in various chemotherapy regimens in the IRS III study.
Ifostamide is an alkylating agent with activity similar to cyclophosphamide, but has decreased myelosuppressive effects. The International Society of Pediatric Oncology (SlOP) reported a complete response rate of 79% with use of the combination of ifosfamide, vincristine, and actinomycin-D. 4,s Als·.), the combination ot ifosfamide and etoposide induced remissions in nine ot ? children with recurrent rhabdomyosarcomas. 49 Ifosfamide may soon become part of the frontline chemotherapy armamentarium tor rhabdomyosarcomas. However, its superiority over cyclophosphamide will have to be defined.
TNM* Staging System Proposed by the International Union Against Cancer (UICC)
Figure 3. Rhabdomyosarcoma of childhood survival by primary site
Figure 4. Rhabdomyosarcoma of childhood survival by clinical group
The introduction ot a xenograft animal model has provided a means to predict the therapeutic effectiveness of several chemotherapeutic agents on childhood rhabdomyosarcomas. 50 Among the active agents identified by these xenograft models is melphalan, another alkylating agent. The activity of melphalan has recently been confirmed in a study of newly diagnosed children with rhabdomyosarcoma.51
The dose of radiation therapy required to achieve local and regional control of rhabdomyosarcomas was shown to be 4000 to 4500 cGy.32 Doses below 4000 cGy (rad) have been associated with higher rates of relapse in children who are over 6 years of age ox who have large tumors measuring over 5 cm in diameter.
However, two questions are still unanswered: When is the most effective time to introduce this therapy? Should the total dose of radiation therapy and the field of administration be modified to take into account response to previous chemotherapy? The status of the regional lymph nodes may influence the efficacy of radiation therapy in achieving local control. But unfortunately, the true influence of regional node involvement in preventing local elimination of tumor has not been clearly established.
The efficacy of surgery in the treatment of childhood rhabdomyosarcoma was documented many years ago. Many children showed prolonged survival when complete excision of the tumor was the only therapy used. Unfortunately, because of the nature of this tumor, complete surgical excision is not often possible. Also, clear documentation showing absence of residual microscopic disease is not always found in early published reports. Nevertheless, surgical resection as the initial phase of treatment still constitutes standard practice in the treatment of childhood rhabdomyosarcomas. With the introduction of effective adjuvant chemotherapy, a new approach is being tried. Chemotherapy is given prior to surgery in an attempt to decrease tumor size. With shrinkage of the tumor, surgery can be less radical, with less damage to adjacent organs and tissues, as in the case of pelvic rhabdomyosarcomas. 53
TREATMENT: SPECIFIC SITES
Head and Neck
In children, rhabdomyosarcomas frequently involve the head and neck. Thus, a great deal of experience has been gained in the management of tumors in these sites, and some guidelines have already been established. By grouping rhabdomyosarcomas of the head and neck region according to anatomic location, they can be divided in two major groups: the parameningeal and the non-parameningeal.
Parameningeal sites include the nasopharynx, nasal cavity, paranasal sinuses, pterygopalatine and infratemporal fossae, and the middle ear and mastoid. Lesions originating in these sites extend directly into the meninges and central nervous system in 35% of the patients, and 90% of these patients eventually die of this complication.54 The high mortality rate, and the apparent frequency of overt meningeal extension, have led to the use of CNS prophylaxis. This consists of giving cranial radiation plus intrathecal chemotherapy with methotrexate, hydrocortisone, and cytosine arabinoside. Use of this additional therapy has increased the relapse-free survival rate from 33% at three years to 63%. In a later study (IRS II), the meningeal relapse rate declined from 23% to 6%.55 Patients who showed no evidence of intracranial extension of tumor, regardless of whether there was erosion of the base of the skull or cranial nerve palsy, are now being evaluated to determine if the size of the irradiated field can be diminished without jeopardizing disease control and duration of survival.
Non-parameningeal sites include the eyes, orbits, scalp, parotids, oral cavity, larynx, oropharynx, and cheek. This group of patients has an excellent prognosis. 56 Approximately 90% of these patients, including those classified as group III, are relapse-free at two years after postoperative radiation and chemotherapy. However, significant late effects of treatment on the eye, orbit, and other bony structures suggest a need to minimize the amount of radiation therapy given.57 This need is being evaluated in IRS III. This protocol includes postoperative radiation to the tumor bed, plus a one-year course of intermittent chemotherapy with cyclic-sequential vincristine and dactinomycin.
Rhabdomyosarcoma is the most common malignant neoplasm involving the pelvis of children. The common recommendation that rhabdomyosarcomas be treated by surgical excision, followed by radiation and chemotherapy, is based on the historical time sequence in which these modalities of treatment were developed. For primary tumor of the bladder, vagina, and uterus, which have a favorable prognosis, organ salvage is an important goal.
A different order of treatment, which is equally successful in terms of survival and preserves pelvic organs by avoiding pelvic exenteration, has been used by different investigators.58'60 This program was tried in a small series of patients and was later tested by the IRS II in a larger number of patients. Patients with primary tumors of the vulva, vagina and uterus, and paratesticular region show better local control following treatment than those with primaries in the bladder-prostate region. This difference in outcome is probably because complete surgical resection can be accomplished less often in children with tumors in the bladder and prostate. The increased frequency of regional lymph node involvement in patients with primary rhabdomyosarcoma of the genitourinary system has led to the recommendation that lymph node exploration and biopsy be done at the time of the initial surgery.
The best time to begin radiation therapy in the treatment of patients with bladder-prostate primaries has not been determined. It has been observed that when radiation therapy is given later than 16 weeks into the course of the disease, local eradication of the tumor often is not achieved.61 Patients with primary tumors arising from the vagina and uterus often do not require radiation therapy since complete surgical excision is more often performed with these tumors than with tumors in any other pelvic site. Those tumors arising in the paratesticular region can be effectively treated by radical orchiectomy with high ligation of the spermatic cord. Routine exploration of the regional lymph nodes in these patients is not recommended by all investigators. However, we still favor biopsy of regional nodes, while we await further documentation of the efficacy and sensitivity of diagnostic imaging and scanning techniques in detecting tumors in the retroperitoneum.
Trunk and Retroperitoneum
Soft-tissue sarcomas of the trunk are an uncommon type of childhood cancer. They tend to occur in the chest wall, abdominal wall, or in the paraspinal region. Histologically, the tumors are often of the alveolar subtype. The prognosis is between that of tumors of the more favorable genitourinary site and the less favorable rhabdomyosarcomas of the extremities. These patients should be treated with complete excision of the tumor whenever possible. Surgery may also include "late" excision of persistent sarcoma after shrinking the tumor by chemotherapy, radiation, or both.62·63
Soft tissue sarcomas arising from the retroperitoneum account for only 6% to 11% of all patients with rhabdomyosarcoma.64 Children with retroperitoneal sarcomas usually appear with large tumors that are difficult to resect completely and difficult to reach with optimal radiation therapy.64 The threeyear relapse-free survival rate of only 42% and the overall survival rate of only 28% in this group of patients reflect the inadequacy of available treatment for patients with tumors in the retroperitoneal area.
Primary rhabdomyosarcomas of the extremities constitute a group with specific characteristics that require special consideration. This group of tumors represents 18% of all childhood rhabdomyosarcomas and has a poor prognosis.65 Extremity lesions are often of the alveolar histologic type, which has a higher incidence of distant metastases and, thus, requires more intensive systemic therapy. Patients with tumors having histology other than alveolar very often present with local or regional spread of the tumor in spite of ongoing therapy. Amputation does not seem to improve the survival rate for these patients.66 Aggressive local therapy is imperative in this group if the survival rate is to be improved.
Involvement of the regional lymph nodes of patients with extremity rhabdomyosarcomas is not routinely determined. But the early experience of the Intergroup Rhabdomyosarcoma Study showed a 17% incidence of lymph node involvement at the time of diagnosis.67 Patients found to have nodal involvement should benefit from irradiation of the regional lymph nodes. In general, extremity lesions should be initially treated with wide surgical excision, with extensive margins free of tumor. Primary re-excision procedures may be performed as long as limb function can be preserved. Biopsy of regional lymph nodes is mandatory as it will identify patients requiring irradiation of regional nodal areas.
CHALLENGES FOR THE FUTURE
Childhood rhabdomyosarcoma continues to be a significant challenge. This challenge extends not only to those involved in the management of these patients, but also involves those in basic research who are interested in further defining the biology of this malignancy.
The introduction of new effective chemotherapeutic agents such as ifosfamide, and new chemotherapy protocols containing combinations of proven agents, may further improve the survival rate of these children. However, common failings of clinical chemotherapy studies in the past have been the addition of lower doses of effective drugs, or the inclusion of nonactive drugs to an already proven chemotherapy regimen. These practices often prevent delivery of the most effective drugs at the optimal dosage and frequency. Factors that may influence the success of chemotherapy, such as dose intensity and timing, may also affect treatment outcome independent of tumorrelated factors. Thus, too widely spaced administration of active drugs (poor timing) may result in significant tumor regression followed by equally significant regrowth between courses. Also, giving a combination of two or more drugs to which the tumor is not crossresistant should be evaluated in future clinical trials of childhood rhabdomyosarcoma.
Radiation therapists need to evaluate different techniques of administering radiation therapy. Optimal use of these techniques, including hyperfractionated doses, split courses, intraoperative irradiation, and use of radio-enhancers and radioblockers constitute a challenge for the future.
Many surgical aspects need to be addressed. The influence of regional lymph node involvement on disease-free survival and the need for local treatment of these nodes must be determined. Innovations and modification of some standard surgical procedures need to be introduced in an effort to further reduce the extent of excision of adjacent normal organs and tissues, especially for patients with primary tumors in pelvic sites. Finally, we need to determine the validity of delaying surgery until the maximum effect of chemotherapy has been achieved.
Rapid progress is being made in the areas of tumor biology and molecular biology. Exciting new information is now available on biological markers, cytogenetic abnormalities, presence of oncogenes, influence of oncogenes in tumor behavior, and mechanisms of drug resistance. These advances will undoubtedly improve our knowledge of the nature of this tumor and may make it feasible to design more effective comprehensive therapeutic plans.
1. Pack OT. Eberhart WF: Rhabdomyosarcoma of the skeletal muscle: Report of 100 cases. Surgery 1985; 32:1023.
2. Stobbe GC, Dargeon HW: Embryonal rhabdomyosarcoma of the head and neck in children and adolescents. Cancer 1950; 3:826.
3. Horn RC, Enterline HT: Rhabdomyosarcoma; clinicopathological study and classification of 39 cases. Cancer 1958; 11:181.
4. Maurer HM, Moon T, Donaldson M, et al: The lntergroup Rhabdomyosarcoma Study: A preliminary report. Cancer 1977; 40:2015.
5. D'Angio GJ, Färber S, Maddock CL: Fbtentiation of x-ray effects by actinomycin D. Radiology 1959; 73:175.
6. Pinkel D, Pickren J: Rhabdomyosarcoma in children. JAMA 1961; 175:293.
7. Heyn RM1 Holland R, Newton WA, et al: The role of combined chemotherapy in the treatment of rhabdomyosarcoma in children. Cancer 1974; 34:2128.
8. Wilbur JR: Combination chemotherapy of embryonal rhabdomyosarcoma. Cancer CnemotKerapy Repon 1974; 58:281.
9. YoungJL, Miller RW: Incidence of malignant tumors in U.S. children. J Pedían 1975; 86:254.
10. Li FP, Fraumeni JF Jr: Soft tissue sarcomas, breast cancer and other neoplasms: A familial syndrome? Ann J ntem Med 1969; 71:747.
11. McKeen EA, Bodurtha J, Meadows AT, et al: Rhabdomyosarcoma complicating multiple neurofibromatosis. J Pediatr 1978; 93:992.
12. Margaret AT, D'Angio GJ, Boice JD, et al: Bone sarcoma linked to radiotherapy and chemotherapy in children. N Engl. J Med 1987; 317:588.
13. Ruymann FB, Gaiger AM, Newton W: Congenital anomalies in rhabdomyosarcoma. Proceeding of Conference on Congenital Birth Defects, Memphis, Tenn, June 9-11, 1977.
14. Maurer HM, Donaldson M, Gehan EA, et al: The lntergroup Rhabdomyosarcoma Study: Update, November 1978. Natl Cancer Inst Monogr 1981; 56:61.
15. Gasparini M, Lombardi F, Gianni C, et al: Childhood rhabdomyosarcoma With meningeal extension: Results of combined therapy including central nervous system prophylaxis. AmJ CIm Oncol 1983; 6:393.
16. Lawrence W, Hays D, Moon T: Lymphatic metastasis with childhood rhabdomyosarcoma. Cancer 1977; 39:556.
17. Miller JH, Ortega JA: Rhabdomyosarcoma and soft tissue tumors, in Miller JH (ed): Imaging in Pediatric Oncology. Waverly Press. Baltimore, 1985, ? 355.
18. Scotti G, Harwood-Nash DC: Computed tomography of rhabdomyosarcoma of the skull base in children. ) Comfnu Assist Tomogr 1982; 6:33.
19. Weinblatt ME, Miller JH: Radionuclide scanning in children with rhabdomyosarcoma. Med Pediatr Oncol 1981; 9:293.
20. Bree RL, Green B-. The Gray Scale sonographic appearance of intra-abdominal mesenchymal sarcomas. Radiology 1978; 128:193.
21. Carter BL, Kahn PC, Wolpert SM, et al: Unusual pelvic masses: A comparison of computed tomographic scanning and ultrasonography. Radiology 1976; 121:383.
22. Miller JH, Reid BS, Kemberling CR: Automated water path ultrasound of chest disease in childhood. Radiology 1984; 152:401.
23. Levine E, Lee KR, Neff JR, et al: Comparison of computed tomography and other imaging modalities in the evaluation of musculoskeletal tumors. Radiology 1979; 131:431.
24. Dejong ASH, van Kessel M, Albus-Lutter CE, et al: Creatine kinase subunits M and B as markers in the diagnosis of poorly differentiated rhabdomyosarcomas in children. Hum ftunol 1985; 16:924-928.
25. Hickman PE, Messina S, Trotter JM, et al: High creatinine kinase MB isoenzyme activity associated with a rhabdomyosarcoma. Clin Chem 1983; 29:1549-1550.
26. Huth JF, Gupta RK, Eilber FR, et al: A prospective post-operative evaluation of urinary tumor-associated antigens in sarcoma patients: Correlation with disease recurrence. Cancer 1984; 53:1306-1310.
27. Actmansberger M, Weber K, Droste R, et al: Desmin is a specific marker for rhabdomyosarcomas of human and rat origin. Am J Pathol 1985: 118:85-95.
28. Trent J, Casper J, Mekzer P, et al: Non-random chromosome alterations in rhabdomyosarcomas. Cancer Genet Cytogenet 1985; 16:189-197.
29. Ture -Carel C, Lizard-Wacol S, Justrabu E, et al: Consistent chromosomal translocation in alveolar rhabdomyosarcomas. Cancer Genet Cytogenet 1986; 19:361-362.
30. Koufos A, Hansen M, Copeland NG, et al: Loss of heterozygosity in 3 embryonal tumors suggests a common pathogenic mechanism. Nature 1985; 316:330-334.
31. Maurer HM, Beltangady M, Gehan EA, et al: The lntergroup Rhabdomyosarcoma Study I: A final report. Cancer 1988; to be published.
32. Angervall L, Enzinger FM: Extraskeletal neoplasm resembling Ewing's sarcoma. Cancer 1975; 36:240-251.
33. Nascimento AG, Unni KK, Pritchard DJ, et al: A ciinicopathologic study of 20 cases of large cell (artificial) Ewing's sarcoma of bone. Am I Surg Pathol 1980; 4:29-36.
34. Newton WA, Soule EH. Hamoudi AB, et al: Histopathology of childhood sarcomas, lntergroup Rhabdomyosarcoma Studies I and 11. Ciinicopathologic correlation, to be published.
35. Shimada H. Newton WA Jt, Soule EH, et al: For the lntergroup Rhabdomyosarcoma Committee: Pathology of fatal rhabdomyosarcoma: Reports from lntergroup Rhabdomyosarcoma Study (IRS I-ll). Cancer 1987; 59:459-465.
36. Palmer N. Sacks N, Foulkes M: Histopathology and prognosis in rhabdomyosarcoma IRS-I. Abstracted. Proc ASCO 1982; 1:170.
37. Palmer N, Foulkes M: Histopathology and prognosis in the second intergroup rhabdomyosarcoma study (IRS-H). Abstracted. Proc ASCO 1983; 2:229.
38. Maurer HM: The Intergroup Rhabdomyosarcoma Study: Objectives and clinical staging classification. J Pediatr Surg 1975; 10:977-978.
39. Harmer MH: TNM classification of paediarric tumours. Geneva, UlCC International Union Against Cancer, 1982.
40. Donaldson SS, Belli JA: A rational clinical staging system for childhood rhabdomyosarcoma. ; CIm Oncol 1987; 2:135-139.
41 · Lawrence W, Gehan EA, Hays DM, et al: Prognostic significance of staging factors of the UICC staging system in childhood rhabdomyosarcomas: A report from the lntergroup Rhabdomyosarcoma Study (IRS III). J CIm Oncol 1987; 5:46-54.
42. Maurer HM. Beltrangady M, Gehan EA, et al: The Intergroup Rhabdomyosarcoma Study-1: A final report. Cancer 1988; to be published.
43. Maurer HM. Foulkes M, Gehan E: For the IRS Committee of CCSG and POG: Intergroup Rhabdomyosarcoma Study (IRS-II): A preliminary report. Abstracted. Proc ASCO 1983, 2:70.
44. Ortega JA, Rivard GE, Isaacs H, et al: The influence of chemotherapy on the prognosis of rhabdomyosarcoma. Med Pediatr Oncol 1975; 1:227-234.
45. Raney RB, Gehan EA, Maurer HM, et al: Evaluation of intensified chemotherapy in children with advanced rhabdomyosarcoma (Clinical Groups III and IV). Clinical Cancer Trials 1979; 3:1928.
46. Baum ES, Gaynon P. Greenberg L, et al: Phase II trial of cis-dichlorodiamineplatinum 11 in refractory childhood cancer: Children's Cancer Study Group report. Cancer Treat Rep 1981; 65:815.
47. Chard RL, Krivit W. Bleyer WA, Hammond D: Phase Il study of VP-16-215 in childhood malignant disease; a CCSG report. Cancer Treat Rep 1979; 63:1755.
48. Often J, Flamant F, Rodary C, et al: Effectiveness of combination of ifosfamide, vincristine and actinomycin-D in inducing remission in rhabdomyosarcoma in children (Abstract). Proc ASCO 1985; 4:236.
49. Miser JS, Kinsella TJ1 Triche TJ, et al: Ifosfamide with mensa uroprotection and etoposide: An effective regimen in the treatment of recurrent sarcomas and other tumors of children and young adults. ] Clin Oncol 1987; 5:1191-1198.
50. Houghton PJ, Houghton JA, Brodeur GM, et al: Characterization, chemosensitivity and drug resistance in xenografts of human childhood rhabdomyosarcomas (RMS). Abstracted. Proc AACR/ASCO 1981; 22:245.
51. Horowitz M. Etucubanas E. Christenson N. et al: Melphalan (LPAM) a clinically effective agent for rhabdomyosarcoma (RMS) as predicted by the xenograft model. Abstracted. Proc ASCO 1987; 5:806.
52. Tefft M, Wharam M, Ruymann F, et al: Radiotherapy (RT) for rhabdomyosarcoma (RMS) in children: A report from the Intergroup Rhabdomyosarcoma Study #2 (IRSII). Abstracted. Proc ASCO 1985. 4:234.
53. Rivard G, Ortega JA, Hittle R, et al: Intensive chemotherapy as primary treatment for rhabdomyosarcoma of the pelvis. Cancer 1978; 36:1593-1597.
54. Tefft M, Fernandez C, Donaldson M, et al: Evidence of meningeal involvement by rhabdomyosarcoma of the head and neck in children. Cancer 1978; 42:253-258.
55. Raney RB, Tefft M, Maurer HM, et al: Results of intensive treatment of children with cranial parameningeal sarcoma: A report from the lntergroup Rhabdomyosarcoma Study (IRSK). Proc Am Assoc Cancer Research 1982; 23:120.
56. Wharam M, Foulkes M, Lawrence W, et al: Soft tissue sarcoma of the head and neck in childhood: Nonorbital and nonparameningeal sites. Abstracted. Proc ASCO 1983; 160.
57. Heyn R, Ragab A, Raney B. et al: Late effects of therapy in orbital rhabdomyosarcoma: A report from the Intergroup Rhabdomyosarcoma Study. Abstracted. Proc ASCO 1983; 66.
58. Ortega JA: A therapeutic approach to childhood pelvic rhabdomyosarcoma without pelvic exenteration. J Pediatr 1979; 94:205-209.
59. Voute PA, Jos A: Combination chemotherapy as primary treatment of children with rhabdomyosarcoma to avoid mutilating surgery or radiotherapy. Abstracted. Proc AACR 1977; 18:327.
60. Kumar APM, Wren EL, Fleming ID. et al: Combined therapy to prevent pelvic exenteration for rhabdomyosarcoma of the vagina or uterus. Cancer 1979; 37:118-122.
61. Raney G, Hays D. Maurer H. et al: Primary chemotherapy, plus radiation therapy (RT) and/or surgery for children with sarcoma of the prostate, bladder, or vagina: Preliminary results of the Intergroup Rhabdomyosarcoma Study (IRS II. 1978-1982). Abstracted. Proc ASCO 1983; 243.
62. Malagoni MA, Ofstcin LC, Grosfeld JL. et al: Survival and pulmonary function following chest wall resection and reconstruction in children. J Pediatr Surg 1980; 15:906-912.
63. Raney RB. Ragab AH. Ruymann FB, et al: Soft-tissue sarcomas of the trunk in childhood: Results of the lntergroup Rhabdomyosarcoma Study. Cancer 1982; 49:2612-2616.
64- Crist WM, Raney RB, et al: Soft-tissue sarcoma arising in the retroperitoneal space in children: A report from the Intergroup Rhabdomyosarcoma Study Committee. Cancer 1985; 56:2125-2132.
65. Hays DM. Soule EH, Lawrence W, et al: For the IRS Committee: Extremity lesions in the Intergroup Rhabdomyosarcoma Study (IRSI); a preliminary report. Cancer 1982; 48:1-8.
66. Maurer HM, Beltangady M, Gehan EA, et al: The Intergroup Rhabdomyosarcoma Study I: A final report. Cancer 1988; to be published.
67. Heyn R, Hays D. Lawrence W: For the IRS Committee: Extremity alveolar rhabdomyosarcoma and lymph node spread: A preliminary report from the lntergroup Rhabdomyosarcoma Study (IRS II). Abstracted. Proc ASCO 1984; 3:80.
Diagnostic Imaging Methods: Effective Uses and Limitations
Biological Markers in Rhabdomyosarcomas
Surgical-Pathological Grouping System for Rhabdomyosarcomas
Categories in the Pretreatment Clinical Staging System Proposed by the International Union Against Cancer (UICC)
TNM* Staging System Proposed by the International Union Against Cancer (UICC)