Gianni Bisogno
Christophe Bergeron
Introduction
Soft tissue sarcomas comprises a heterogeneous group of tumours derived from mesenchymal cells. As these cells normally mature into muscle, fibrous structures, fat, etc., the different histiotypes of soft tissue sarcomas are designated according to the line of differentiation that can be recognized in the tumour. Soft tissue sarcomas comprise ~8 per cent of all paediatric malignancies. Rhabdomyosarcoma (RMS) is the most common subtype in the first two decades of life, accounting for ~60–70 per cent of these diagnoses. RMS is rare in adults and arises before the age of 6 years in two-thirds of cases. Non-rhabdomyosarcoma soft tissue sarcomas (NRSTSs) comprise 20–30 per cent of all sarcomas in children and have a histology similar to that of adult sarcomas. The most common NRSTSs seen in children include synovial sarcoma, fibrosarcoma, and malignant peripheral nerve sheath tumour.
Rhabdomyosarcoma
Rhabdomyosarcoma can develop in any anatomic location of the body where mesenchymal tissue other than bone is present, and it can spread to the lungs, bone marrow, bones, lymph nodes, and other sites. The most common primary sites are the head and neck (40 per cent), genitourinary sites (20 per cent), and extremities (20 per cent). The strong chemosensitivity of RMS has been proved by the studies promoted by cooperative groups from North America and Europe, and the current approach to treatment includes a variable combination of chemotherapy, surgery, and radiotherapy. A number of prognostic factors have been identified, and defining the therapeutic strategy according to these factors creates complexity. The comparison between the results reported by the main international collaborative groups (IRS, SIOP, CWS, ICG)1 is complicated by the use of different staging systems. However, recent collaboration between these groups has begun to resolve some of these difficulties, and agreement has been found on a standard approach to the criteria used for staging1 and the pathologic classification.2
Aetiology
Rhabdomyosarcoma cannot be simply considered as a cancer arising from skeletal muscle; it is more appropriate to define it as a tumour derived from primitive mesenchyme and exhibiting a profound tendency towards myogenesis. Disease aetiology is unclear, but an association with familial cancer risk in both the Li–Fraumeni syndrome and neurofibromatosis type 1 has created interest in possible genetic causal factors. Additional aetiologic theories may emerge from data suggesting links between RMS and various environmental factors although no consistent relationships have been proved.
Pathology and biology
Rhabdomyosarcoma is characterized by a greater or lesser degree of myogenic differentiation, a feature that results from biologic forces related to aberrant transcription signals and the resultant production of myogenic proteins.3 Classically, RMS is histologically distinguished into two main forms: embryonal (which accounts for 80 per cent of all RMS) and alveolar subtypes (15–20 per cent of RMS). A third form, pleomorphic RMS, is described in adults but almost never encountered in children and no longer forms part of the paediatric classification.
Desmin and muscle-specific actin are relatively sensitive immunohistochemical markers, but they are also expressed by a variety of cells other than rhabdomyoblasts and detection of muscle transcription factors such as MyoD and myogenin is the most sensitive diagnostic indication. Other technologies, particularly the molecular genetic detection of the expression of myogenic transcription factors (e.g. MYFgenes from the MyoD protein family) and the presence of cytogenetic abnormalities representing abnormal fusion genes, are likely to become increasingly important in clarifying the diagnosis of RMS, in distinguishing it from other soft tissue or small round cell tumours, and in confirming histologic subtype. Most alveolar RMSs display a t(2,13)(q35;q14) translocation, with genetic breakpoints which result in the fusion of genes PAX3 and FKHR, or, less frequently, a variant t(1,13)(p36;q14) translocation with fusion of genes PAX7 and FKHR, whilst many embryonal RMSs exhibit a loss of heterozygosity at chromosome 11p.3
Embryonal rhabdomyosarcoma
Embryonal rhabdomyosarcoma, described by Bérard in 1894 as ‘tumeur embryonnaire du muscle strié’ is characterized by a spindle or spindle and round cell tumour by a loose myxoid or dense collagenous stroma. Rhabdomyoblastic differentiation is expressed by the presence of strap-like cells, but cellularity, pleomorphism, and the number of mitoses vary considerably. Cross-striations are seen in more differentiated forms, and ultrastructural examination with electron microscopy demonstrates the presence of features such as sarcomeric Z bands and thin and thick filaments. The botryoid subtype is typically found at vaginal or nasopharyngeal sites where tumour grows into organ cavities. It is histologically similar to embryonal RMS with the additional feature of a condensed layer of tumour cells under the overlying mucosa, the so-called cambium layer. The spindle cell variant presents as either a collagen-poor leiomyomatous form or a collagen-rich form with a storiform pattern arising in paratesticular locations. Distinction from other forms of soft tissue sarcoma relies on the presence of well-differentiated rhabdomyoblasts in the spindle cell population.
Alveolar rhabdomyosarcoma
Alveolar rhabdomyosarcoma was first described by Riopelle and Theriault in 1956. Classically, this form displays an alveolar architecture, i.e. well-defined alveolar-like spaces separated by thick collagenous bands and lined with round tumour cells showing variable myogenic features. It is now generally accepted that the percentage of cells showing an alveolar pattern is unimportant and that even a focal presence is sufficient to justify the diagnosis. However, attention has also been paid to the cytologic features of alveolar RMS which are distinct from embryonal RMS, and a diagnosis of alveolar RMS can be made in the absence of an overt alveolar pattern when characteristic cytologic features are present. This is the basis for the diagnosis of the so-called solid alveolar variant.
International Classification of Rhabdomyosarcoma
Recently, pathologists from the major international soft tissue sarcoma groups published a consensus for a new International Classification of Rhabdomyosarcoma (ICR).2 This has been tested in multivariate analysis and shown to be strongly predictive of survival in addition to established clinical risk factors. Three risk groups have been identified:
· superior prognosis: botryoid and spindle cell RMS
· intermediate prognosis: embryonal RMS
· poor prognosis: alveolar RMS (including the solid variant).
This classification system should now be used by all pathologists and cooperative groups in order to provide comparability between and within multi-institutional studies (Table 18.1). However, there remain areas of uncertainty, particularly with tumours that do not demonstrate clear cytologic evidence for myogenic differentiation (undifferentiated sarcoma) or cannot be adequately classified (sarcoma, not otherwise specified). There are also tumours, such as those with rhabdoid features, where it is not clear whether they form a distinct and separate group or represent morphologic variants of the major subtypes.
Clinical presentation and diagnosis
RMS is encountered at almost all anatomic sites, although the head and neck and genitourinary locations are the most common (Fig. 18.1). Presentation is strongly influenced by site. For example, tumours within the orbit tend to present early with obvious displacement of the globe and are rarely associated with regional lymph node or distant metastatic spread (Fig 18.2), whilst tumours in the nasopharynx may result in a relatively long history of nasal discharge and obstruction and frequently involve local extension to the base of the skull or the posterior aspect of the orbits with the potential for associated cranial nerve palsies or visual loss (Fig. 18.3). The definition of certain head and neck sites as ‘parameningeal’ relates to the risk of direct tumour extension into the meninges and beyond. Such tumours carry a risk of intracranial extension and, in some cases, involvement of the cerebrospinal fluid (CSF). Tumours within the genitourinary tract may present with urinary obstruction in bladder and prostate sites (Fig. 18.4), as a scrotal mass (paratesticular), or as a vaginal polyp or discharge (vaginal and uterine tumours). Elsewhere, presentation is usually associated with the development of a mass, and the child is often not unwell unless there is metastatic disease. In rare cases, widespread metastatic disease is encountered without clear evidence of a primary tumour and the diagnosis is confirmed by bone marrow examination.
|
Table 18.1. Classifications of rhabdomyosarcoma: comparison of different systems |
|||||||||||||||||||||
|
|||||||||||||||||||||
|
|
|||||||||||||||||||||
|
Fig. 18.1 Distribution of sites of disease. |
|||||||||||||||||||||
Diagnostic and staging investigations must include adequate imaging of the primary site (CT or MRI) and accurate assessment of sites of potential metastatic spread (lung, bone, and bone marrow). The CSF should be sampled in the case of parameningeal tumours. The involvement of regional lymph nodes depends on the primary site, and the frequency with which positive lymph node spread is reported also depends on the manner in which this is investigated. This remains a source of some controversy and, for example, the strategies promoted by the North American Intergroup Rhabdomyosarcoma Study (IRS) Group have encouraged greater systematic use of surgicopathologic lymph node assessment than in European (particularly SIOP) studies, in which clinical and radiologic evaluation has always been the standard approach.
Diagnosis must be confirmed histologically. Although needle biopsy may be the simplest approach favoured by some clinicians, it has the disadvantage of limiting tissue available for conventional histologic examination, including immunohistochemistry, and may restrict access to fresh and frozen tissue for cytogenetic and molecular genetic investigation, both of which may be of considerable importance in guiding diagnosis in difficult cases. Therefore open biopsy is frequently preferred and, when possible, should be undertaken at an oncology center where optimal use of diagnostic material can be achieved and the initial surgical approach determined by the multidisciplinary team responsible for the child's subsequent treatment. As site is such an important determinant of prognosis and treatment strategy, classification by site has been standardized by international agreement into seven major groups (Table 18.2).4
|
|
|
Fig. 18.2 Orbital tumour: (a) clinical presentation with proptosis and deviation of globe. (b) CT scan shows a large anteromedial soft tissue mass without bone erosion. |
Treatment
Chemotherapy, which became an option in the 1970s, has proved efficient for the treatment of RMS, contributing to the remarkable increase in survival of such patients from 25 to 70 per cent. Advances in surgery and radiotherapy have paralleled those of chemotherapy.
The multidisciplinary clinical management of RMS promotes systemic treatment for both local and metastatic lesions, with further local treatment using surgery and/or radiotherapy. As RMS often occurs in young children, the treatment strategy must balance the best chance of cure with the lightest possible burden of therapy in order to decrease sequelae and risk to future health. Treatment strategies are based on the prognostic factors developed over the past 20 years.
|
|
|
Fig. 18.3 Nasopharyngeal tumour: (a) CT scan showing a large soft tissue mass filling the nasopharynx and nasal cavities with destruction of the pterygoid plate; (b) CT scan reconstruction showing tumour extension through the base of skull. |
|
|
|
Fig. 18.4 Bladder tumour: (a) CT scanogram showing a hugely distended bladder filled with lobulated tumour; (b) cross-sectional CT image showing a lobulated tumour filling the bladder and outlined by contrast excretion into surrounding urine. |
Prognostic factors
Histology It is now clear that alveolar RMS conveys a poorer prognosis than the embryonal subtype. This was apparent in the early IRS and SIOP trials and confirmed by results of the IRS IV study which showed 88 per cent relapse-free survival in embryonal RMS compared with 66 per cent in alveolar disease.5
Site Site of disease has long been shown to be of prognostic value, both in North American and European studies. International consensus associates good prognosis with orbit, non-bladder prostate genitourinary, and non-parameningeal head and neck disease sites; all other sites are considered to have poor prognosis (Table 18.2).
|
Table 18.2. Definition of sites of involvement in childhood rhabdomyosarcoma |
|||||||
|
Stage Ever since 1972, North American teams have favoured a classification of rhabdomyosarcomas into four groups, taking into account the extent of the initial surgical excision and the spread of the tumour tolocal tissues and/or the lymphnodes [IRS Grouping System (Table 18.3)]. The same approach has been used in Europe by the German and Italian cooperative groups, whereas centers associated in the SIOP group chose to use a tumour, nodes, and metastases (TNM) classification relying on the clinical description of the disease before (Table 18.4) and after (Table 18.5) initial surgery (Figs 18.5 and 18.6). The distinction between these classification systems has made the comparison of results difficult. However, collaboration between the groups has clarified the importance of preand postsurgical stage in prognostic significance.
|
Table 18.3. IRS Clinical Grouping System |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Table 18.4. SIOP pretreatment clinical (TNM) staging |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Age The IRS IV study demonstrated that three age groups could be associated with prognosis in localized RMS: <1 year, 1–9 years, and >10 years of age.5 This also applies to subgroups of patients: the IRS III and IV trials demonstrated that age >10 years was a factor of poor prognosis in paratesticular RMS and identical results were obtained from the SIOP MMT 84 and MMT 89 studies.6 Age is also predictive of prognosis in metastatic RMS with a poorer survival associated with age >10 years.
It is important to note that factors predictive of prognosis (histology, postsurgery stage, lymph node involvement, tumour size, and patient age) are often interdependent (Fig. 18.7): limb tumours are generally of the alveolar type, whereas vaginal or bladder RMSs are known to be botryoid, and paratesticular RMSs more frequently present the spindle cell variants. It is also well known that alveolar subtypes tend to spread more frequently to both regional nodes and distant metastatic sites. Finally, the status after surgery (complete microscopic surgical excision, incomplete microscopic excision, gross residual tumour) is strongly determined by the site of the disease.
Therapeutic strategy
Both North American and European groups determine treatment strategy according to disease histology (alveolar versus non-alveolar), postsurgical status, IRS Group, disease site (favourable versus unfavourable), tumour size, and age.
Distinct risk groups can be defined:
· low-risk RMS with an average of 90 per cent event-free survival (EFS) at 5 years
· intermediate-risk RMS with 60–80 per cent survival
· high-risk RMS with <60 per cent survival
· very-high-risk RMS with <40 per cent survival.
|
Table 18.5. SIOP postsurgical histopathologic (pTNM) classification |
||||||||||||||||
|
||||||||||||||||
|
|
||||||||||||||||
|
Fig. 18.5 Pie chart showing distribution of clinical (TNM) stages at diagnosis. |
||||||||||||||||
|
|
|
Fig. 18.6 Pie chart showing the distribution of postsurgical stages (pTNM) before chemotherapy. |
|
|
|
Fig. 18.7 Histogram showing the relationship between site and clinical stage. |
Local tumour control is central to the possibility of cure for patients with non-metastatic disease. Controversies here relate to the method and timing of local treatment, and, more specifically, to the place of radiotherapy in guaranteeing local control for patients who appear to achieve complete remission with chemotherapy with or without surgery. This represents an important philosophical difference between the SIOP MMTstudies and those of the IRS Group and, to some extent, of the German and Italian Cooperative groups. Local relapse rates are higher in the SIOP studies than those experienced elsewhere, although the SIOP experience has also made it clear that a significant number of patients who relapse may be cured with alternative treatment.7 In the context of such differences in approach to local treatment, overall survival rather than disease-free or progression-free survival may become the most important criterion for measuring outcome. However, the ‘cost’ of survival must take into account the predicted late effects of treatment and the total burden of therapy experienced by an individual patient. This must include an assessment of all the treatments necessary to cure the child, including those used after relapse.
Surgery Primary tumour resection should be undertaken only if there is no evidence of metastatic disease and the tumour can be excised with good margins without danger, functional impairment, or mutilation. If this is not possible, a diagnostic biopsy is required. An attempt at surgical resection which leaves microscopic residual disease makes treatment decisions more complicated because the patient is unassessable for the efficacy of chemotherapy and may still require further local treatment. Primary re-excision (i.e. a second surgical resection before chemotherapy) may be worthwhile in a minority of cases if there is confidence that clear margins of excision can be achieved without functional or cosmetic disadvantage.8 This applies particularly to trunk, limb, and paratesticular tumours. Histological assessment of tumour margins is not always consistent with the surgical findings because of the diffuse infiltration of some tumours into adjacent tissues.
The importance of surgical evaluation of lymph nodes at diagnosis is controversial. Clinically and radiologically suspicious nodes should be sampled (fine-needle aspiration may be a useful technique in such circumstances), but radical lymph node dissections are rarely justified and routine surgical staging of regional (para-aortic) nodes in patients with localized paratesticular tumours is not considered necessary by the European groups.6,9
The value of secondary operations to achieve complete remission should be distinguished from procedures undertaken merely to confirm clinical or radiologic complete response. Secondary operations, and even multiple biopsies for verification of local control, are not indicated if there is no visible tumour clinically, endoscopically, or on CT or MRI scanning. A significant minority of patients thought to be in partial remission after initial chemotherapy in the IRS III study were shown actually to be in complete remission at second-look surgery. Secondary surgery to achieve local control after initial chemotherapy remains an important aspect of treatment but depends on the site of disease. For example, surgery has little or no role in the primary management of orbital RMS and has only a selective place in the local control of head and neck tumours in general. Surgery at this point in treatment should generally be conservative, whatever the site of disease, anticipating that the morbidity of radiotherapy may be more acceptable than radical operations which result in important functional (e.g. total cystectomy) and/or cosmetic (e.g. amputation) consequences. However, in some circumstances the morbidity of radical surgery to achieve local control may be preferred, for example, to avoid pelvic irradiation in very young children.
Chemotherapy Chemotherapy is an essential component of treatment for all children with RMS. Experience since the 1970s has defined the efficacy of a variety of chemotherapeutic agents, the value of multi-agent chemotherapy combinations, and the importance of adjuvant therapy in patients without macroscopic residual disease after initial surgery. It has long been recognized that treatment of large unresectable tumours with chemotherapy could reduce the extent of subsequent surgery or radiation therapy. However, the role of intensified chemotherapy in reducing or avoiding the need for local therapy remains controversial. This has been most consistently explored by the SIOP MMT studies.7,10 The strategies of the IRS and of the other European (German and Italian) cooperative groups still tend to retain a systematic approach to local therapy regardless of chemotherapy response, except in patients who achieve complete primary tumour resection with initial surgery. Overall, it seems likely that some patients who achieve complete tumour control with chemotherapy can be spared local treatment, but it is important to recognize that local recurrence is the predominant pattern of relapse in non-metastatic disease and clinicians must not disregard the relevance of local therapy for many patients.
Vincristine (V), actinomycin D (A), cyclophosphamide (C) and doxorubicin [Adriamycin1 (Adr)] have been the most frequently utilized agents in the treatment of RMS and have been used in various combinations (VA; VAC; VACAdr) in the sequential IRS studies. Doxorubicin is an active agent when used alone, but its role remains controversial and concern about potential cardiotoxicity justifies caution in its use as part of primary treatment. The introduction of newer drugs has not always been accompanied by clear evidence of their benefit as single agents within phase II studies. Such data are available for cisplatin, etoposide, and DTIC, all of which were introduced into IRS III. However, it was not possible to show that cisplatin, with or without etoposide, offered any survival advantage, although the combination of cisplatin with doxorubicin in MMT 84 produced significant response rates in patients who failed to show an adequate response to IVA.7
The substitution of cyclophosphamide by ifosfamide in combination with vincristine and actinomycin D (with or without doxorubicin) has been the hallmark of all recent European studies. Ifosfamide appears to convey some advantages over its analogue cyclophosphamide, showing a lack of cross-resistance and a lower myelotoxicity profile, thus permitting the possibility of administering larger doses. The rate of response to ifosfamide-containing regimens appeared favourable in the historical comparison of the SIOP and CWS studies.11,12 A prospective randomized trial comparing an ifosfamide-based combination (IVA) with a cyclophosphamide-based combination (VAC) was conducted in the IRSG IV study. The doses of cyclophosphamide and ifosfamide in each cycle were respectively 2.2g/m2 for 1 day and 1.8g/m2/day for 5 days (these doses were previously found to produce comparable myelosuppression). No differences in either 3-year survival or failure-free survival rates were seen between different regimens (84 per cent and 75 per cent for VAC compared with 84 per cent and 77 per cent for IVA).5 Both drugs require concurrent administration of mesna to avoid haemorrhagic cystitis, but ifosfamide carries a risk of renal toxicity not experienced with cyclophosphamide and VAC remains the combination of choice for future North American studies. Nevertheless, the European groups have decided to keep IVA as standard combination because data suggest that there is only a small risk of significant renal toxicity at cumulative ifosfamide doses < 60 g/m2 and a higher risk of gonadal toxicity with cyclophosphamide.
A collaborative European protocol for patients with metastatic disease introduced carboplatin and epirubicin (Epiadriamycin1) into first-line therapy as part of an intensive six-drug schedule denoted CEVAIE (with ifosfamide, vincristine, actinomycin D, and etoposide) designed to overcome drug resistance. The choice of carboplatin and epirubicin was based on preferential toxicity profiles compared with cisplatin and Adriamycin1. This chemotherapy strategy was also incorporated into MMT 89 for the treatment of high-risk patients with lymph node disease and produced a significant improvement in outcome compared with historical data from similar patients treated in the previous study (MMT 84). Current European protocols are exploring this six-drug combination in a direct randomized comparison with conventional IVA (SIOP Group) or VAIA (German and Italian Groups) for patients with nonmetastatic disease. Ifosfamide and doxorubicin given as a phase II ‘window’ in children with newly diagnosed metastatic RMS showed a response rate (complete remission and partial response) of 63 per cent and should be considered for inclusion in front-line therapy for children with intermediateor high-risk RMS.13
Topotecan was studied in a classic phase II study in 24 relapsed patients with no response and 48 chemotherapy-naive patients in a window study with 46 per cent response rate.14 Results of treatment with irinotecan are awaited.
High-dose therapies The place of high-dose chemotherapy strategies necessitating autologous bone marrow or peripheral blood stem cell rescue remains unclear. Some experience has been gained in single institutions utilizing a variety of chemotherapy schedules and, predominantly, in patients with relapsed disease. More recently, the European collaborative groups agreed a shared strategy for the treatment of newly diagnosed patients with metastatic disease. This study was initially intended to explore the value of high-dose chemotherapy only amongst patients with incomplete response to initial chemotherapy, but a modification to the study design in 1991 encouraged the use of high-dose melphalan as consolidation therapy for all patients who achieved complete remission after six courses of CEVAIE. Preliminary analysis suggests that there is no survival advantage for those who received consolidation chemotherapy with melphalan compared with those, treated in the earlier phase of the study, who did not.15
Radiotherapy Early experience with radiation therapy demonstrating local control in up to 90 per cent of patients included in the IRS studies has confirmed that doses >50 Gy are not usually required when given by conventional (once daily) fractionation. However, there is also evidence that doses <40 Gy may be insufficient, particularly in patients with macroscopic residual disease. The dose used in the SIOP studies is 45 Gy regardless of site or age (although particular efforts are made to avoid irradiation in young children), with a possible boost to 50 or 55 Gy to a reduced field when there is bulky residual macroscopic disease at initiation of therapy. Randomization studies within IRS I–III studies have established that radiotherapy is unnecessary for patients with embryonal histology and tumour completely resected at diagnosis (IRS Clinical Group I). However, analyses from the same studies indicate that radiotherapy does offer an improved failure-free survival in patients with completely resected alveolar RMS.16 Current guidelines for therapy within the IRS Group vary the prescribed dose from 40 to 55 Gy depending on the site, size, and histology of the tumour, as well as on the age of the child.
Studies from the European groups have attempted to relate the use of radiotherapy to the response to initial chemotherapy. The most radical approach has been used in the SIOP protocols, where patients in IRS Clinical Group III (SIOP pT3b) disease avoided radiotherapy if complete remission had been achieved with initial chemotherapy, with or without second surgery, except for parameningeal RMS. This approach has proved feasible,17 but the psychologic impact of relapse and the burden of second-line therapy are important and the definition of such favourable patients should be refined.
Treatment must always be given using megavoltage equipment. Electron treatment may be useful for superficial tumours, either as a direct electron field or as a boost to a linear accelerator planned field. Adequate margins must be used (usually 2–3 cm), and treatment for patients with parameningeal disease is normally planned to the initial tumour volume. In tumours at other sites that show a good response to initial chemotherapy, treatment can be planned to the residual volume (plus margins).
Conventional treatment is usually given as a single daily fraction of 1.8 Gy. Interest in hyperfractionated schedules has been explored in both the IRS IV and (to a more limited extent) MMT 89 studies. Overall, the data suggest that no benefit in disease control can be expected from the use of hyperfractionation,18 and standard conventional fractionation is still used by most cooperative groups.
Early experience in the treatment of parameningeal tumours was discouraging. Local failure rates were high and there was a high incidence of local extension into the adjacent meninges, often with spinal subarachnoid spread and high mortality. Investigation suggested that these patients were receiving inadequate dose and volume of radiation treatment, and the IRS studies were modified to include earlier introduction of radiotherapy, wider fields (extending to whole brain and spine in some cases), increased dose to the site of bulk disease, and the concurrent administration of intrathecal chemotherapy. This resulted in a much improved survival rate. In fact, radiotherapy to the target volume with systemic chemotherapy are successful treatments for the majority of patients with localized parameningeal sarcomas, and guidelines have been liberalized, particularly in relation to the volume of treatment, so that whole-brain treatment is avoided whenever possible. However, all groups agree that patients with parameningeal disease require mandatory radiotherapy regardless of response to chemotherapy. This is especially important as assessment of complete response can be difficult at these sites and surgery rarely offers a valid alternative approach to local control. Intrathecal chemotherapy has never been used routinely in the SIOP studies and there seems to be little justification to do so in the majority of patients who do not demonstrate evidence of CSF or spinal dissemination.
Interstitial radiotherapy (brachytherapy) using intracavitary moulds or implanted wires may be of particular relevance for small tumours at selected sites, notably in the vagina and perineum. Occasionally this technique is utilized at other genitourinary sites, including tumours of the bladder base and prostate, and there is limited experience of its application to head and neck sites.10
Outcome
The most recently published IRS Group study (IRS IV) reported the outcome for 883 patients recruited between 1991 and 1997.5 Overall, 3-year failure-free survival and overall survival were 77 per cent and 86 per cent, respectively, and did not differ from those of similar patients treated in IRS III with a 5-year progression-free survival of 65 per cent. This compares with the 5-year EFS and overall survival of 57 per cent and 71 per cent, respectively, reported in SIOP MMT 89 and of 59 per cent and 69 per cent, respectively, reported in CWS 86.19 As discussed previously, these studies used significantly different approaches to local treatment, and the larger difference between overall survival and EFS in the SIOP studies reflects a higher relapse rate with successful salvage therapy for some patients.
Table 18.6 gives details of the outcome of treatment according to site in four cooperative group studies. These confirm the prognostic effect of site with an obvious difference between the favourable outcome associated with orbital and genitourinary sites, and the poor results achieved with tumours presenting in the limbs and at ‘other’ sites.
Late effects of treatment
‘Cure at what cost?’ is the difficult, yet essential, issue to be addressed when reviewing the outcome of survivors of all forms of cancer in childhood, particularly when survival relates to different philosophies and treatment modalities. The importance of accurate prognostic assessment at diagnosis is as much to ensure that patients with good prognosis are not overtreated as to identify those with a poorer prognosis who require a more aggressive approach. Much concern has been focused on the late sequelae of local treatment for RMS, particularly after radiotherapy and the types of aggressive surgery that result in significant functional or cosmetic problems (e.g. orbital exenteration, retroperitoneal lymph node dissection, and total cystectomy). Chemotherapy is also associated with significant sequelae in some patients, and the concept that more intensive chemotherapy may reduce the use of local treatment must be balanced against the additional toxicity that it may cause. The more recent use of ifosfamide has raised concern about long-term renal damage, whereas the continuing use of high doses of alkylating agents and, more recently, etoposide may result in second malignancies. Long-term follow-up and prospective evaluation of all survivors is required in order to document the frequency and functional significance of all possible late effects of therapy.
|
Table 18.6. Five-year survival by primary sites according to collaborative group trials |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Non-rhabdomyosarcoma soft tissue sarcoma
The term nonrhabdomyosarcoma soft tissue sarcoma (NRSTS) includes a group of rare tumours with marked clinical, histologic, and biologic heterogeneity. Most histiotypes are more common in adults and occur sporadically in children. However, in some cases, clinical behaviour maybe peculiar to a paediatric presentation, justifying a different treatment approach.
Overall, NRSTS show two peaks of incidence, the first in children aged <5 years and the second in early adolescence. The relative frequencies of different histologic subtypes vary with age; for example, fibrosarcomaiscommoninchildrenaged <2yearsandsynovialsarcomaandmalignant peripheral nerve sheath tumour (MPNST) most commonly affect adolescents.20 Other NRSTS include extraosseous primitive neuroectodermal tumour (PNET)/extraosseous Ewing sarcoma (EOE), undifferentiated sarcoma (US), vascular tumors (haemangiosarcoma, haemangiopericytoma), epithelioid sarcoma and alveolar soft part sarcoma. More recently, new entities such as intraabdominal desmoplastic small round cell tumour have been added to the NRSTS group. Very rare entities include liposarcoma, malignant fibrous histiocytoma, and leiomyosarcoma. The incidence of NRSTS as registered by the Italian cooperative group is shown in Figure 18.8.
NRSTSs share several clinical characteristics with the more frequent RMSs. They can occur anywhere in the body, usually as a painless growing mass which shows symptoms only when an adjacent organ or structure is invaded. They mostly arise in the extremities, the wall of the trunk, and the retroperitoneum, while head and neck or genitourinary locations are extremely rare. Evidence of regional and distant metastases at diagnosis is less frequent than in RMS, with involvement of regional lymph nodes present in 10 per cent of cases and distant metastases in about 10–12 per cent.
Open biopsy is recommended and sufficient material should be analysed by an expert pathologist to confirm the diagnosis of NRSTS and to grade the tumour appropriately. Molecular biology is of increasing importance as a number of specific translocations have been identified. This can be an important tool to assist the diagnosis of these rare tumours and for research purposes. Detection of fusion gene product is used for the primary diagnosis of Ewing sarcoma and PNET and for the detection of metastatic or residual disease. A list of the more common genetic abnormalities in NRSTS is presented in Table 18.7.
In general, the diagnostic workup and staging systems are the same as those adopted for RMS. The IRS grouping system and the TNM staging system are commonly used. A histologic grading system has been developed for paediatric NRSTSs which identifies three prognostic groups based on histology, amount of necrosis, number of mitoses, and cellular pleomorphism.21 The prognostic importance of this has not yet been confirmed in large clinical trials.
Generally, management of NRSTS has been derived from treatment used for RMS and from adult experience. Increasing data from single institutions and cooperative groups should allow the design of more specific clinical trials for at least some histiotypes in the near future.
|
|
|
Fig. 18.8 Incidence of different NRSTSs in childhood (data from the Italian Co-operative Studies). |
Prognostic factors
The search for prognostic variables in such rare tumours presents difficulties. However, some factors are consistently present when analysing data from larger series containing different histiotypes or small series dedicated to a single histiotype.22,23,24,25,26,27,28,29 These can be summarized as follows.
· Disease extension at diagnosis: survival is very poor in children with disseminated disease.
· Initial surgery: complete tumour resection gives 70–80 per cent chance of cure in most histiotypes.
· Tumour site: influences tumour resection.
· Tumour size: lesions >5 cm have an increased risk of local and distant relapse.
· Tumour invasiveness (T stage): this is often associated with tumour size.
· Histology: different histiotypes have different biologic behaviour and response to treatment.
· Grading.
· Age: tumours in older children behave more like adult sarcomas and carry a worse prognosis.
It has been suggested that, as in adults, risk factors for local recurrence may be different from those for metastatic relapse, with the latter being more correlated with tumour size, invasiveness, and high histology grading.22 This may determine the selection of patients for adjuvant chemotherapy after complete surgical resection.
|
Table 18.7. Specific chromosomal abnormalities identified in soft tissue sarcoma |
|||||||||||||||||||||||||||||||||||||
|
Treatment
A complete tumour resection with wide margins is advocated in adults. Radiotherapy is added if resection margins are close to the tumour and sometimes may be administered preoperatively. Adjuvant chemotherapy is a contentious issue because response rate is quite low and few studies have shown benefit in terms of disease-free and overall survival.
Surgery is also considered the mainstay of NRSTS treatment in children, although a more conservative approach is usually adopted. Limb-sparing procedures and non-compartmental resections are preferred, and therefore the concept of radical surgery in children allows closer surgical margins than are generally acceptable in adults. A complete resection may be the only treatment needed for some histiotypes. If resection is considered unlikely at diagnosis, an incisional biopsy is required for diagnostic purposes (fine-needle biopsy does not usually provide sufficient accuracy) and other treatment modalities need to be considered before referring the child back to the surgeon for a definitive procedure. However, this conservative approach must be balanced against the higher risk of local relapse of NRSTS compared with RMS, and therefore a more aggressive approach to local treatment is generally required.
Radiotherapy is administered when residual disease is present or suspected after resection. Several reports suggest that radiotherapy is effective for maintaining local control only when the residual disease is minimal. In some cases radiotherapy may be considered preoperatively to target the volume better and to make the subsequent tumour resection easier. Doses of at least 40–50 Gy are used but concern about late effects can constrain the use of irradiation, especially in young children. Occasionally, alternative techniques, such as brachytherapy, may be considered.
The benefit of chemotherapy is far from established for NRSTS and represents one of the main therapeutic problems, particularly for the less common entities. Sporadic reports of chemotherapy response exist for several histiotypes, but in general response rates are much lower than those reported in RMS. Some recent data suggest that NRSTS can be divided in three different groups.
· Tumours with proven chemosensitivity: PNET/EOE and US. These tumours respond to treatments adopted for RMS.
· Tumours with possible chemosensitivity: synovial sarcoma, malignant fibrous histiocytoma, liposarcoma, and congenital–infantile fibrosarcoma. Although encouraging chemotherapy responses have been reported, the benefit in terms of improved survival is not well established.
· Tumours with unproven chemosensitivity: juvenile or ‘adult type’ fibrosarcoma, MPNST, alveolar soft part sarcoma. Responses to chemotherapy have been reported for these entities, and therefore a trial of chemotherapy may be justified when conservative surgical surgery is not considered feasible at diagnosis.
Peripheral neuroectodermal tumours and extraosseous Ewing sarcoma
PNET and Ewing sarcoma are histologically similar and share the same chromosomal translocation, typically t(11;22). Therefore they are grouped in the Ewing family of tumours. Traditionally, paediatric PNET and Ewing sarcoma arising outside bone have been included in soft tissue sarcoma protocols in many countries. However, there is ongoing debate to establish whether or not they should be treated differently from osseous Ewing tumours. Extraosseous PNET/Ewing tumours are considered to be aggressive, but the advent of effective multimodality therapy has improved prognosis and 5-year survival is reported to be 60 per cent.23 Unfortunately, survival in patients with metastatic disease at diagnosis remains very poor despite aggressive treatment. Regimens including alkylating agents (ifosfamide or cyclophosphamide) plus actinomycin D and vincristine have been used. Anthracylines are considered important drugs in the treatment of osseous Ewing tumours. However their role has been questioned in soft tissue PNET/Ewing tumours.23 Aggressive surgery and radiotherapy is recommended for local control.
Fibrosarcoma
Fibrosarcoma represents 10 per cent of paediatric NRSTSs and is the most common soft tissue sarcoma in infants aged <1 year. Two forms of fibrosarcoma are recognized: congenital-infantile fibrosarcoma that occurs in children aged <2 years, and juvenile or ‘adult-type’ fibrosarcoma that is more typical of children aged 10–15 years. They are not distinguishable on the basis of histology, as both forms are composed of spindle-shaped fibroblasts which may show different growth patterns (solid, haemangiopericytoma-like, and herringbone), exhibiting variable collagen production and showing no evidence of other differentiation. The clinical characteristics of the tumour, patient age, and cytogenetic differences are considered the main factors in distinguishing the infantile and adult forms. A recurring t(12;15); (p13;q25) translocation has been documented in congenital–infantile fibrosarcoma. Interestingly, the same chromosomal anomaly has been found in congenital mesoblastic nephroma, a usually benign renal tumour occurring in infants. Other translocations [t(2,5) and t(7,22)] have been described in adult-type fibrosarcoma.
Congenital–infantile fibrosarcoma
Typically, this diagnosis is applied only to children aged <2 years although some authors extend this limit up to 4 or 5 years. The tumour is usually located in the distal region of the extremities and presents as a painless rapidly enlarging mass. Despite the rapid growth, evolution may be indolent with a tendency to recur locally without metastatic spread. Spontaneous regression has also been described. Complete tumour resection is the only treatment necessary for most patients, but a trial of chemotherapy is warranted for inoperable tumours. Responses have been reported using regimens including vincristine, dactinomycin, and ifosfamide or cyclophosphamide. Despite a risk of relapse, ~80 per cent of patients with congenital fibrosarcoma are long-term survivors24 and recently it has been suggested that the presence of translocation (2;5) conveys a better prognosis.30
Adult-type fibrosarcoma
In older children fibrosarcoma has clinical features similar to those found in adults. The tumour usually grows slowly and the typical locations are the proximal region of extremities, deep trunk, and cavity sites. Aggressive surgery is advocated to obtain a complete tumour resection. Radiotherapy is used when postsurgical residues are evident or suspected. Response to chemotherapy has been reported, but its role remains uncertain. Unfortunately, local relapse is frequent and this may precede the development of distant metastases. Long-term survival rates of ~50–60 per cent are usually achieved.24
Synovial sarcoma
Synovial sarcoma is a malignant undifferentiated mesenchymal tumour resembling normal synovial tissue. It represents 5–10 per cent of all soft tissue tumours in adults and is one of the most common non-rhabdomyosarcoma cancers in the paediatric age range (15–25 per cent),20 occurring primarily in adolescents. Histologically two major subtypes are described: the monophasic form, characterized by spindle cells, and the more frequent biphasic form, characterized by spindle and epithelial cells organized in glandular structures. Cytogenetic studies have reported the presence of a characteristic translocation t(X;18)(p11;q11) in ~90 per cent of cases. Two fusion transcripts, SYT–SSX1 and SYT–SSX2, are derived from fusion of the SYT and SSX genes localized on chromosomes 18q11 and Xp11, respectively.
Synovial sarcoma usually arises in extremities, and the thigh and knee are the most frequently affected sites. The lungs are by far the most common metastatic site. Prognosis is related to the presence of metastatic lesions, tumour size (>5 cm) and invasiveness, tumour site (distal locations fare better than proximal), and complete resection. Histologic subtype influences outcome, with the biphasic form being more favourable. The type of fusion transcript has also been correlated with survival (patients with SYT–SSX2 tumours do better than those with SYT–SSX1 tumours).
Wide excision of the primary tumour is the mainstay of treatment. The use of radiotherapy and chemotherapy is controversial. Irradiation is usually recommended to improve local control when resection is incomplete. Chemotherapy has been shown to be active in paediatric trials, with response rates of up to 60–70 per cent,25 and is used in inoperable and disseminated cases. However, the role of chemotherapy after a complete resection is questionable, although survival rates in paediatric series, where chemotherapy has been used systematically, is usually better than in adult series, where chemotherapy is used less frequently and the metastatic relapse rate is higher. A recent multicenter multivariate analysis has shown overall 80 per cent 5-year survival, 88 per cent for patients with tumours grossly resected at diagnosis (IRS Groups I and II) and 75 per cent for patients with localized unresectable tumor (IRS Group III). Adjuvant chemotherapy did not seem to have an impact on survival in patients in IRS Groups I or II, and tumour size appeared to be the most relevant prognostic factor.25
Also based on experience in adults, ifosfamide and doxorubicin seem to be the most active drugs and there is a need to test this combination in a randomized international trial.
Malignant peripheral nerve sheath tumour (malignant Schwannoma)
MPNST is strongly associated with neurofibromatosis type I (NF-1): ~20 per cent of patients with MPNST present NF-1 features and up to 15 per cent of children with NF-1 are at risk of developing a MPNST. The NF-1 gene, localized at 17q11.2, is thought to have a role as a tumour suppressor gene, and its inactivation or loss may be involved in the neoplastic transformation of cells. The extremities, retroperitoneum, and trunk are the primary sites most commonly involved, and MPNST tends to be locally aggressive with only few patients showing metastases (mainly in the lungs) at diagnosis.
The most important prognostic factor is the completeness of tumour resection. Radiotherapy, in doses up to 50–60 Gy, is used in cases of microscopic postsurgical residual. The role of chemotherapy is uncertain, although a response (complete or partial) was documented in 30 per cent of a series of 64 patients enrolled by the Italian and German Cooperative groups. In some cases the tumour shrinkage achieved with chemotherapy facilitates its removal, and a trial of chemotherapy (based on ifosfamide and doxorubicin) is warranted in patients with initially unresectable tumour. The 10-year overall survival in this series was 45 per cent, but survival was much worse (5 per cent) in patients with NF-1 in whom the tumour tended to be larger, unresectable, and highly chemoresistant.
Vascular tumours
Vascular neoplasms represent the most common mesenchymal lesion of subcutaneous and deep soft tissue in the paediatric age range. However, ~90 per cent of them are benign and a subset is considered to have limited local aggressiveness. Malignant lesions include angiosarcoma, malignant haemangioendothelioma, and Kaposi's sarcoma. Haemangiopericytoma is also included in this group, although it originates not from the endothelium but from pericytes.
Angiosarcoma is an aggressive tumour with a high propensity to recur locally and to metastatize to local lymph nodes, lung, and liver. It can occur anywhere in the body, but mostly arises in the skin and superficial soft tissue. The head and neck region is the commonest site of origin. Previous radiotherapy, chronic lymphoedema, or environmental toxins (thorotrast, steroids) are all linked to the occurrence of angiosarcoma. Complete tumour resection is important, but this is not generally sufficient to ensure local control. Therefore radiotherapy is recommended in adults. The use of drugs usually administered for soft tissue sarcomas has not improved outcome, and long-term survival is ~30 per cent.26
The term haemangioendothelioma includes several different entities. Malignant haemangioendothelioma and epithelioid haemangioendothelioma exhibit an aggressive or borderline malignant potential, respectively. Malignant haemangioendothelioma is considered similar to angiosarcoma. Epithelioid haemangioendothelioma may present as a single lesion of the limbs or with multifocal lesions involving bone, lung, or liver. Surgery is of paramount importance, but no response to chemotherapy has been reported. New approaches are needed for vascular tumours, and the use of paclitaxel in angiosarcoma and interferon-α in haemangioendothelioma have both shown promising results.
Haemangiopericytoma derives from mesenchymal cells with pericytic differentiation. Two distinct clinical entities are described in the paediatric age group: the infantile type, occurring in young infants, and the adult type, occurring in children aged >1 year. Infantile haemangiopericytoma typically occurs in the subcutis and oral cavity. Multifocal lesions may occur and spontaneous regression has been described. Adult-type haemangiopericytoma has a more aggressive behaviour, similar to that seen in adults, with metastatic potential to lung and bone. Late relapses have been described. Wide tumour excision is the treatment of choice. However, both forms of haemangiopericytoma have shown good response to chemotherapy, and this should be implemented in cases where primary resection is not feasible. Radiotherapy is also recommended after incomplete tumour removal in adult-type haemangiopericytoma. Prognosis appears good for infantile-type haemangiopericytoma, with >80 per cent of children surviving at 5 years. The outcome for adult-type haemangiopericytoma in children seems to be better than that in adults, with 5 year EFS rates >60 per cent.26
Alveolar soft part sarcoma
Alveolar soft part sarcoma accounts for 1–5 per cent of paediatric NRSTS and occurs more frequently between 15 and 35 years of age. Females outnumber males, especially during the first two decades of life. Histologically, alveolar soft part sarcoma is composed of aggregates of polygonal cells separated by vascular spaces. The degeneration of the central cells produces the alveolar pattern, although the most distinctive feature is the presence of intracytoplasmic periodic acid, diastase-resistant inclusions of unknown nature. There is still considerable uncertainty as to the exact histiogenesis of this tumour.
Alveolar soft part sarcoma usually arises in the extremities, although the head and neck region, including the orbit and tongue, is a more common region in children. Metastatic lesions are present at diagnosis in ~20 per cent of patients; these involve lung, bone, and, less frequently, the central nervous system.
The clinical course is often indolent and patients may survive for several years with evidence of disease. Patients cured of localized lesions may present with metastases after prolonged disease-free intervals, sometimes exceeding 10 years.27 Complete surgical resection is the strongest prognostic factor and may represent the only treatment for localized disease. The role of adjuvant chemotherapy and radiotherapy is not well defined. Because of the rarity of the tumour and the need for long-term follow-up, it is not clear whether chemotherapy reduces the rate of metastatic relapse. However, chemotherapy responses have been reported, and a trial of chemotherapy seems appropriate when the tumour is not resectable. Radiotherapy may improve the local control when there is postsurgical residual disease.
Overall, the prognosis seems favourable when metastatic lesions are not present at diagnosis. In a recent paediatric series the overall survival was 80 per cent, with 93 per cent diseasefree survival for patients with localized disease. Favourable outcome was related to the completeness of tumour resection in most cases.28 These figures are better than those reported for adults, where alveolar soft part sarcoma seems to have a higher propensity to metastasize.
Desmoplastic small round cell tumour
Since its first description in 1989 by Gerald and Rosai, desmoplastic small round cell tumour (DSRCT) has been increasingly identified but its histiogenesis remains uncertain. It is distinguished from the other small round cell tumours by a characteristic histologic appearance marked by nests of cellular growth within abundant desmoplastic stroma, and a specific polyphenotypic differentiation with coexpression of epithelial, mesenchymal, and neural markers. A recurrent specific chromosomal translocation, t(11,22)(p13;q12), involving the EWS and the WT1 genes has been identified.
The tumour predominantly affects young males, usually in their second decade of life. DSRCT typically presents as a large abdominal mass, often widely disseminated at the time of diagnosis, with extensive spread to the regional lymph nodes, peritoneal seeding, and distant metastases to liver, lung, and bone. Other, less frequent, primary sites are the paratesticular region and the thoracic cavity, sometimes with extensive involvement of the pleura.
DSRCT seems to be chemosensitive, and aggressive multimodality treatment including surgery, radiotherapy, and high-dose chemotherapy based on alkylating drugs has been used. However, early relapses after completing the treatment are common, and the survival is <40 per cent at 30 months.29
Future treatment for non-rhabdomyosarcoma soft tissue sarcoma
The experience achieved by the Italian and German Cooperative groups with various histologic types is detailed in Tables 18.8 and 18.9. In most of them surgery is still the best and often the only proven treatment. The role of radiotherapy and chemotherapy is difficult to establish because of the small number of studies published. However, irradiation is widely recommended when there is evidence of residual tumour in the surgical margins. In some cases radiotherapy may be considered in the preoperative phase to improve the possibility of obtaining a complete resection.
|
Table 18.8. Non-rhabdomyosarcoma soft tissue sarcoma: retrospective analysis of single histotypes from Italian and German group studies |
||||||||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||||||||
|
Table 18.9. Non-rhabdomyosarcoma soft tissue sarcoma: retrospective analysis for single histiotypes from Italian and German group studies |
||||||||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||||||||
Chemotherapy responses have been documented in leiomyosarcoma, liposarcoma, epithelioid sarcoma, and sporadically in other histiotypes. However, it is not clear whether this translates into a higher survival rate.
In the past, paediatric non-RMS soft tissue sarcomas have been treated according to strategies developed for RMS, but there is increasing evidence that this should not be the case. The characteristics of many of these tumours seem to be similar to those of the same diagnoses treated by adult oncologists, although survival seems better in children and some entities show differences in biology, clinical behaviour, and sensitivity to chemotherapy (e.g. infantile fibrosarcoma and haemangiopericytoma). There is a need for trials specifically designed for these rare sarcomas. In view of their rarity, wider multinational collaboration will be needed to perform meaningful studies to identify prognostic factors and effective treatments.
References
1. Rodary C, Flamant F, Donaldson SS (1989). An attempt to use a common staging system in rhabdomyosarcoma. A report of an international workshop initiated by the International Society of Paediatric Oncology. Med Pediatr Oncol 17, 210–15.
2. Newton WA, Jr., Gehan EA, Webber BL, et al. (1995). Classification of rhabdomyosarcomas and related sarcomas. Pathologic aspects and proposal for a new classification—an Intergroup Rhabdomyosarcoma Study. Cancer 76, 1073–1085.
3. Barr FG (1997). Molecular genetics and pathogenesis of rhabdomyosarcoma. J Pediatr Hematol Oncol 19, 483–91.
4. Donaldson SS, Draper GJ, Flamant F, et al. (1986). Topography of childhood tumors: pediatric coding system. Pediatr Hematol Oncol 3, 249–58.
5. Crist WM, Anderson JR, Meza JL, et al. (2001). Intergroup rhabdomyosarcoma study—IV: results for patients with nonmetastatic disease. J Clin Oncol 19, 3091–102.
6. Stewart RJ, Martelli H, Oberlin O, et al. (2003) Treatment of children with nonmetastatic paratesticular rhabdomyosarcoma: results of the Malignant Mesenchymal Tumors studies (MMT 84 and MMT 89) of the International Society of Pediatric Oncology. J Clin Oncol 21, 793–8.
7. Flamant F, Rodary C, Rey A, et al. (1998). Treatment of non-metastatic rhabdomyosarcomas in childhood and adolescence. Results of the second study of the International Society of Paediatric Oncology: MMT84. Eur J Cancer 34, 1050–62.
8. Cecchetto G, Guglielmi M, Inserra A, et al. (2001). Primary re-excision: the Italian experience in patients with localized soft-tissue sarcomas. Pediatr Surg Int 17, 532–34.
9. Ferrari A, Bisogno G, Casanova M, et al. (2002). Paratesticular rhabdomyosarcoma: report from the Italian and German cooperative group. J Clin Oncol 20, 449–55.
10. Martelli H, Oberlin O, Rey A (1999). Conservative treatment for girls with nonmetastatic rhabdomyosarcoma of the genital tract: a report from the Study Committee of the International Society of Pediatric Oncology. J Clin Oncol 17, 2117–22.
11. Stevens M, Flamant F. (1990) Ifosfamide for children with solid tumours. Lancet 335, 1398–1400.
12. Treuner J, Koscielniak E, Keim M. (1989). Comparison of the rates of response to ifosfamide and cyclophosphamide in primary unresectable rhabdomyosarcomas. Cancer Chemother Pharmacol 24, S48–50.
13. Sandler E, Lyden E, Ruymann F, et al. (2001). Efficacy of ifosfamide and doxorubicin given as a phase II ‘window’ in children with newly diagnosed metastatic rhabdomyosarcoma: a report from the Intergroup Rhabdomyosarcoma Study Group. Med Pediatr Oncol 37, 442–8.
14. Pappo AS, Lyden E, Breneman J, et al. (2001). Up-front window trial of topotecan in previously untreated children and adolescents with metastatic rhabdomyosarcoma: an Intergroup Rhabdomyosarcoma study. J Clin Oncol 19, 213–19.
15. Carli M, Colombatti R, Oberlin O, et al. (1991). High-dose melphalan with autologous stem-cell rescue in metastatic rhabdomyosarcoma. J Clin Oncol 17, 2796–803.
16. Wolden SL, Anderson JR, Crist WM, et al. (1999). Indications for radiotherapy and chemotherapy after complete resection in rhabdomyosarcoma: a report from the Intergroup Rhabdomyosarcoma Studies I to III. J Clin Oncol 17, 3468–75.
17. Oberlin O, Rey A, Anderson J, et al. (2001). Treatment of orbital rhabdomyosarcoma: survival and late effects of treatment–results of an international workshop. J Clin Oncol 19, 197–204.
18. Donaldson SS, Meza J, Breneman JC, et al. (2001). Results from the IRS-IV randomized trial of hyperfractionated radiotherapy in children with rhabdomyosarcoma—a report from the IRSG. Int J Radiat Oncol Biol Phys 51, 718–28.
19. Koscielniak E, Harms D, Henze G, et al. (1999). Results of treatment for soft tissue sarcoma in childhood and adolescence: a final report of the German Cooperative Soft Tissue Sarcoma Study CWS-86. J Clin Oncol 17, 3706–19.
20. Stiller C. (2002). Epidemiology of cancer in adolescents. Med Pediatr Oncol 39, 149–55.
21. Parham DM, Webber BL, Jenkins JJ 3rd, et al. (1995). Non-rhabdomyosarcomatous soft tissue sarcomas of childhood: formulation of a simplified system for grading. Mod Pathol 8, 705–10.
22. Spunt SL, Poquette CA, Hurt YS, et al. (1999). Prognostic factors for children and adolescents with surgically resected nonrhabdomyosarcoma soft tissue sarcoma: an analysis of 121 patients treated at St Jude Children's Research Hospital. J Clin Oncol 17, 3697–705.
23. Raney RB, Asmar L, Newton WA Jr, et al. (1997). Ewing's sarcoma of soft tissues in childhood: a report from the Intergroup Rhabdomyosarcoma Study, 1972 to 1991. J Clin Oncol 15, 574–82.
24. Cecchetto G, Carli M, Alaggio R, et al. (2001). Fibrosarcoma in pediatric patients: results of the Italian cooperative group studies (1979–1995). J Surg Oncol 78, 225–31.
25. Okcu MF, Munsell M, Treuner J, et al. (2003). Synovial sarcoma of childhood and adolescence: a multicenter, multivariate analysis of outcome. J Clin Oncol 21, 1602–11.
26. Ferrari A, Casanova M, Bisogno G, et al. (2002). Malignant vascular tumors in children and adolescents: a report from the Italian and German soft tissue sarcoma cooperative group. Med Pediatr Oncol 39, 109–14.
27. Pappo AS, Parham DM, Cain A, et al. (1996). Alveolar soft part sarcoma in children and adolescents: clinical features and outcome of 11 patients. Med Pediatr Oncol 26, 81–4.
28. Casanova M, Ferrari A, Bisogno G, et al. (2000). Alveolar soft part sarcoma in children and adolescents: A report from the Soft-Tissue Sarcoma Italian Cooperative Group. Ann Oncol 11, 1445–9.
29. Gerald WL, Ladanyi M, de Alava E, et al. (1998). Clinical, pathologic, and molecular spectrum of tumors associated with t(11;22)(p13;q12): desmoplastic small round-cell tumor and its variants. J Clin Oncol 16, 3028–36.
30. McCahon E, Sorensen PH, Davis JH, et al. (2003). Non-resectable congenital tumors with the ETV6-NTRK3 gene fusion are highly responsive to chemotherapy. Med Pediatr Oncol 40, 288–292.
31. Ferrari A, Bisogno G, Casanova M, et al. (2001). Childhood leiomyosarcoma: a report from the Soft Tissue Sarcoma Italian Cooperative Group. Ann Oncol 12, 1163–8.
32. Ferrari A, Casanova M, Bisogno G, et al. Clear cell sarcoma of tendons and aponeuroses in pediatric patients: a report from the Italian and German Soft Tissue Sarcoma Cooperative Group. Cancer94, 3269–76
33. Ferrari A, Casanova M, Bisogno G, et al. (2001) Hemangiopericytoma in pediatric ages: a report from the Italian and German Soft Tissue Sarcoma Cooperative Group. Cancer 92, 2692–8.
34. Ferrari A, Casanova M, Meazza C, et al. (2001). Vascular tumours in pediatric age. Ital J Pediatr 27, 774–8.
35. Bisogno G, Roganovich J, Sotti G, et al. Desmoplastic small round tumour in children and adolescents. Med Pediatr Oncol 34, 338–42.
