• rhabdomyosarcoma;
  • nonalveolar histology;
  • recurrence;
  • prognostic factors


  1. Top of page
  2. Abstract
  6. Acknowledgements


Although > 90% of children with nonmetastatic rhabdomyosarcoma (RMS) achieve complete remission with current treatment, up to one-third of them experience a recurrence. Survival rates are not always poor in patients who develop recurrences; thus, prognostic factors are needed to tailor salvage treatment.


The current analysis included 125 children who were affected by localized RMS and were enrolled in 3 consecutive Italian protocols (RMS79, RMS88, and RMS96) who developed recurrences after complete remission. Patient, tumor, and treatment characteristics were studied in univariate and multivariate analyses to determine the independent significance of different factors.


The median time from diagnosis to recurrence was 17.8 months. Most patients had local recurrences (72%). The 5-year overall survival (OS) rate was 28.3% ± 8.7%. Multivariate analysis identified 4 factors that were associated with poor survival: 1) alveolar subtype (relative risk [RR], 2.0), 2) parameningeal or “other” sites (RR, 2.6), 3) systemic recurrence (RR, 3.1), and 4) recurrence on therapy (RR, 2.3). The absence of any of these risk factors identified a “favorable risk” group (12% of patients) with a 5-year OS rate of 71.8% ± 23.5%. Patients with a single risk factor (32%) had an OS rate of 37.5% ± 17.2%. Combining patients with 0 or 1 risk factor, the OS rate was 66.5% in the subgroup who had not received radiotherapy compared with an OS rate of 30.3% in the subgroup who had received radiotherapy; this difference was significant (P = 0.03).


The results of the current analysis demonstrated that groups with a different prognosis can be identified among patients with recurrent RMS. Patients with a nonalveolar histology, a primary site other than the parameningeal or “other” sites, local recurrence, and recurrence off therapy had a better prognosis. First-line treatment may have an impact on prognostic variables. In fact, patients who had no or only one risk factor and patients who had tumors with a nonalveolar histology benefited more from salvage therapy if they had not received radiotherapy for their initial treatment. These data may be useful in planning risk-adapted salvage protocols. Cancer 2005;. © 2005 American Cancer Society.

The long-term survival rates for patients with rhabdomyosarcoma (RMS) improved from 50% in the 1970s1 to 70% in the 1990s.2 Greater than 90% of children with nonmetastatic RMS now achieve complete tumor remission with current multidisciplinary treatment, which includes chemotherapy, radiotherapy (XRT), and surgery. Up to one-third of patients experience recurrence,3–5 however. Survival after recurrence is poor,6 and new salvage therapy strategies are needed. The first step in planning second-line treatment is to identify prognostic factors in these patients, so that risk-adapted protocols can be developed. With this objective, we retrospectively analyzed data on children with RMS who were enrolled on consecutive protocols of the Associazione Italiana di Ematologia e Oncologia Pediatrica (AIEOP) Soft Tissue Sarcoma Committee (STSC) (formerly the Italian Cooperative Group [ICG]) who developed recurrences after complete remission.


  1. Top of page
  2. Abstract
  6. Acknowledgements


Between October 1979 and December 1998, a total of 445 patients with localized RMS were enrolled for the RMS79, RMS88, and RMS96 protocols. Among the 401 patients (90.7%) who achieved a first complete remission, 125 patients (31.1%) developed recurrences. The clinical characteristics of this population, the treatment modalities, and the types of recurrence were analyzed and correlated with the patients' survival. Of the 445 patients with nonmetastatic RMS who were enrolled in AIEOP protocols, 41 patients (9.2%) experienced disease progression; these children were excluded from our analysis, because the definition of recurrence implies the prior achievement of complete remission. However, the outcome of patients who did not achieve complete remission is presented.

Patient variables included age at diagnosis and recurrence and gender. Primary tumor site, histologic subtype (alveolar or nonalveolar), regional lymph node involvement, and tumor size at the time of initial diagnosis also were considered.

The first-line treatment was assigned on the basis of the initial surgery. Consequently, we included the extent of resection in this analysis, classified according to the Intergroup Rhabdomyosarcoma Study (IRS) grouping system, i.e., completely excised tumor (Group I); macroscopically resected tumor with microscopic residual disease (Group II), and macroscopic residual disease after incomplete resection or biopsy (Group III).

The administration of XRT during first-line therapy and the initial protocol also were considered in the analysis. The different chemotherapeutic regimens used for the initial treatment were not analyzed because of the limited number of patients in each subset. Among the 125 patients who had recurrences, only 4 patients were treated initially with 2 drugs (vincristine and actinomycin D), whereas the others had received at least 3 drugs (vincristine, actinomycin D, and cyclophosphamide or ifosfamide with or without doxorubicin). Various chemotherapeutic regimens were used after recurrence, thus preventing any further analysis.

For the purpose of this study, recurrences were classified as local (local recurrence alone and/or regional recurrence) or systemic (distant metastases alone or combined with local/regional recurrence) and were defined as early (< 18 months after first diagnosis) or late (≥ 18 months after first diagnosis). Time to recurrence was classified further according to whether the recurrence occurred before (on therapy) or after (off therapy) treatment was completed.

Statistical Analysis

Overall survival (OS) after recurrence was estimated from the time of first disease recurrence to death or last follow-up evaluation. Follow-up was updated at September, 2003. Death due to any cause was considered an event. Prognosis was defined according to the duration of OS. Survival curves were calculated using the Kaplan–Meier method.7 The log-rank test was used to compare the survival curves of patient subgroups by univariate analysis to ascertain the potential value of prognostic factors.8

Variables with P values < 0.25 in the log-rank test were entered in a multivariate analysis to determine the independent significance of different factors. Multivariate assessment of OS was performed using a Cox proportional hazards model.9


  1. Top of page
  2. Abstract
  6. Acknowledgements

Demographic and Clinical Features of the Study Group

The initial characteristics of the 125 patients with RMS who experienced a recurrence are listed in Table 1. Most of them were classified in IRS Group III (72.8%). The primary tumor site was orbit in 15.2% of patients, nonparameningeal head and neck (non-PM HN) in 12.8% of patients, parameningeal (PM) in 14.4% of patients, extremity in 15.2% of patients, genitourinary bladder/prostate (GU B/P) in 6.4% of patients, genitourinary nonbladder/prostate (GU non-B/P) in 4.8% of patients, and “other” sites in 31.2% of patients (trunk, paraspinal region, intrathoracic and intraabdominal organs, pelvis, and perineal region).

Table 1. Initial Characteristics of 125 Patients with Nonmetastatic Rhabdomyosarcoma who Experienced Recurrence after Complete Remission
CharacteristicNo. of patients%
  1. RMS: rhabdomyosarcoma; IRS: Intergroup Rhabdomyosarcoma Study.

 < 1 yr1512.0
 1–5 yrs5140.8
 > 5–9 yrs3124.8
 ≥ 10 yrs2822.4
Primary tumor site  
 Head and neck nonparameningeal1612.8
 Genitourinary bladder/prostate86.4
 Genitourinary nonbladder/prostate64.8
 Other sites3931.2
 Embryonal RMS6753.6
 Alveolar RMS4737.6
 Botryoid RMS64.8
 Leiomyomatous RMS21.6
 Unclassified RMS32.4
Involvement of regional lymph nodes  
Tumor size  
 ≤ 5 cm4737.6
 > 5 cm7560.0
IRS group  
Initial protocol  

Tumor size at diagnosis was > 5 cm in 60% of patients, and regional lymph node involvement was evident in 19.2%. of patients. The embryonal subtype was the most common, as expected.

Recurrence Pattern and Survival

The median time from diagnosis to recurrence was 17.8 months (range, from 1.7 months to 12.0 years) and was 14.0 months for patients with systemic recurrences and 19.3 months for patients with local recurrences. Recurrences occurred while on therapy in 22% of patients, within 18 months of diagnosis in 50.4% of patients, and within 5 years of diagnosis in 95% of patients.

The recurrence was local in 90 patients (72%) and systemic in 35 patients (28%). Patients who had alveolar RMS had a higher rate of systemic recurrence compared with patients who had nonalveolar RMS, but the difference was not significant (33% vs. 25%; P = 0.29).

The 5-year OS rate after recurrence was 28.3% (Fig. 1), with a median survival of 15.4 months. No toxic deaths were recorded. The median follow-up for survivors was 83.7 months after initial diagnosis (range, 5.4–240.3 months) and 41.0 months after recurrence (range, 0.0–226.7 months). Among the 41 patients who had disease progression, 39 patients died of disease, and 2 patients achieved a long-term remission and were alive 4 years and 10 years after progression.

thumbnail image

Figure 1. This is the overall survival curve for all patients with nonmetastatic rhabdomyosarcoma who developed recurrent disease after achieving a complete remission.

Download figure to PowerPoint

Prognostic Factors for Survival after Recurrence

Univariate analysis

The 5-year OS rates per prognostic variable are given in Table 2. Primary tumor site (P < 0.0001), tumor size (P = 0.0008), histology (P = 0.05), and prior XRT (P = 0.0051) were associated significantly with survival after recurrence.

Table 2. Univariate Analysis: Prognostic Factors for Survival after Recurrence
CharacteristicNo. of patientsFive-yr OS (%)95% CIRRP valuea
  • OS: overall survival; 95%CI: 95% confidence interval; RR: relative risk; GU: genitourinary; B/P: bladder/prostate; PM: parameningeal; HN: head and neck; RMS: rhabdomyosarcoma.

  • a

    Log-rank test.

Variables at initial diagnosis      
  < 1 yr15828.40.0–57.8  
  1–9 yrs825333.022.0–44.0 0.2933
  > 10 yrs282214.40.0–31.3  
 Primary tumor site      
  Orbit, GU-non-B/P251252.131.6–72.71.0 
  Non-PM HN, GU-B/P, extremity442534.117.8–50.41.6< 0.0001
  PM, other sites564611.62.0–21.2  
 Regional lymph node involvement      
  N01006431.721.6–41.8 0.1163
 Tumor size      
  ≤ 5 cm472539.823.9–
  > 5 cm755521.811.6–32.02.2 
 IRS Group      
  II171310.50.0–28.4 0.5233
  Not given553042.928.8–
 Initial protocol      
  RMS 79433425.612.5–38.6  
  RMS 88644624.713.5–36.0 0.0744
  RMS 9618377.554.2–100.0  
Variables at recurrence      
  < 1 yr210.0  
  1–9 yrs855529.418.5–40.2 0.4557
  > 10 yrs382727.512.4–42.6  
 Type of recurrence      
  Local905535.024.2–45.71.0< 0.0001
 Time from diagnosis to recurrence      
  Early (< 18 mos)635017.47.3–27.41< 0.0002
  Late (≥ 18 mos)623340.326.4–54.20.3 
 Recurrence and treatment      
  Recurrence on therapy28247.80.0–8.32.9< 0.0001
  Recurrence off therapy975934.023.4–44.61.0 

Univariate analysis was conducted assembling primary tumor sites with a similar outcome into three groups: favorable (GU non-B/P, orbit), intermediate (non-PM HN, extremity, GU B/P), and unfavorable (PM and “other” sites). In fact, the patients who had most favorable prognosis were those with GU non-B/P RMS (OS, 60.0%; 95% confidence interval [95%CI], 17.1–100.0%) or orbit RMS (OS, 55.9%; 95%CI, 33.0–78.8%). Patients who had the worst prognosis had tumors arising at “other” sites (OS, 18.8%; 95%CI, 5.7–31.9%) or PM sites (none of these children survived > 3 years). The OS rates for patients with RMS arising in non-PM HN, extremity, and GU B/P sites were 48.8%, 42.1%, and 32.4%, respectively.

Among the variables at initial diagnosis, age, gender, regional lymph node involvement, and IRS group did not affect 5-year survival. Among the variables at recurrence, the type and time of recurrence and its relation with therapy were associated strongly with survival (see Table 2). To be more precise, a patient with a PM RMS or with RMS at “other” sites who had received XRT during first-line treatment had very little chance of surviving a systemic, early recurrence.

Histology also proved significant, with nonalveolar RMS associated with a longer survival (P = 0.05). Further analysis of the nonalveolar group showed that OS was significantly better after a local recurrence than after a systemic recurrence (46.4% vs. 12.3%; P = 0.003) and when XRT had not been given during first-line treatment (OS, 53.4% vs. 24.7%; P = 0.01).

The survival rate in patients with alveolar RMS also was significantly better after local recurrences compared with systemic recurrences, but it was very poor in both groups (OS, 14.4% vs. 7.6%; P = 0.0001). Patients with alveolar RMS survived longer if they had not received XRT: The survival rate was 20.8% (95%CI, 0.0–43.1%) versus 6.2% (95%CI, 0.0–17.2%) in the subgroup of patients who received XRT, but the difference was not statistically significant (P = 0.40). No significant differences emerged when patients were analyzed by protocol (see Table 2).

Multivariate analysis

Four prognostic factors that were associated with a longer survival emerged from the multivariate analysis (Table 3): histology, primary tumor site, type of recurrence, and its relation with therapy. Patients who had tumors with nonalveolar histology, a primary tumor site different from PM or “other” sites, local recurrence, and recurrence off therapy had a better prognosis.

Table 3. Multivariate Analysis: Significant Prognostic Factors
VariableHRHR 95%CLP value
  1. HR: hazard ratio; 95%CL: 95% confidence limits; GU: genitourinary; PM: parameningeal; HN: head and neck; B/P: bladder/prostate.

Primary site   
 Orbit and GU non-B/P1.0  
 Non-PM HN, GU B/P and extremity1.00.5–2.10.0005
 PM and other sites2.61.3–5.1 
Type of recurrence   
 Systemic3.11.9–5.1< 0.0001
Recurrence and treatment   
 Off therapy1.0  
 On therapy2.31.4–3.80.001
Risk-stratification model

The results of multivariate analysis were used to develop a risk-stratification model, and survival was calculated on the basis of the number of prognostic factors for each child (Fig. 2). The 5-year survival rate was 71.8% for children who had no risk factors (12% of patients with recurrences) and 37.5% and 16.1% for children who had 1 or 2 risk factors, respectively. None of the children with 3 or 4 risk factors were alive after 5 years (P < 0.0001).

thumbnail image

Figure 2. This chart illustrates overall survival (OS) according to the risk stratification model in patients with recurrent rhabdomyosarcoma. Among 15 patients with 0 risk factors (solid line), there were 5 treatment failures (5-year OS, 71.8%; 95% confidence interval [95%CI], 48.3–95.3%). Among 40 patients with 1 risk factor (long dashed line), there were 23 treatment failures (5-year OS, 37.5%; 95%CI, 20.3–54.7%). Among 46 patients with 2 risk factors (short dashed line), there were 34 treatment failures (5-year OS, 16.1%; 95%CI, 3.4–28.8%). Among 24 patients with 3 or 4 risk factors (dashed and dotted line), there were 21 treatment failures (5-year OS not applicable).

Download figure to PowerPoint

Among the children who had 0 risk factors or 1 risk factor (44% of patients), the 5-year OS rate was 66.5% (95%CI, 47.5–85.5%) in the 26 patients who were not treated initially with XRT versus 30.3% (95%CI, 11.0–49.6%) in the 29 patients who received initial XRT; this difference was significant (P = 0.03). In the subset of children who had > 1 risk factor, the survival rate was slightly better in the 29 patients who were not treated initially with XRT (19.1%; 95%CI, 2.2–36.0%) compared with the 41 patients who received initial XRT (6.6%; 95%CI, 0.0–17.1%), but the difference was not statistically significant (P = 0.15).

This mathematical model does not take single risk factors into account. The relation of each factor (i.e., alveolar histology) to prognosis in the absence of other factors and with respect to XRT could not be assessed because of the small number of patients in each subgroup.


  1. Top of page
  2. Abstract
  6. Acknowledgements

In the current series, approximately one-third of patients with nonmetastatic RMS who achieved a complete remission subsequently developed a recurrence, which is consistent with the data reported in the literature.3–5 Local recurrence was the most common cause of treatment failure; the rate of local (and/or regional) recurrence (72%) observed in the current study was similar to that reported in the International Society of Pediatric Oncology (SIOP) malignant mesenchymal (MMT84) study (75%),10 which considered patients with Group I–III disease, whereas Pappo et al.6 reported a lower rate (51% of patients who developed a recurrence), probably because of the initial local therapy (most patients received XRT) and, above all, because they included patients with Group IV disease and/or progressive disease, characteristics that give rise to a relatively higher frequency of systemic recurrence.

Local therapy is difficult to manage in patients with RMS because of the site of the primary tumor and the possible late effects of XRT, because the majority of children with RMS are very young. In an attempt to reduce the potential sequelae of this treatment, the administration of XRT in first-line therapy has been tailored in Italian and European protocols. In the ICG RMS 79 and RMS 88 protocols, XRT was avoided in patients with histologic complete remission at secondary surgery after primary chemotherapy. XRT was not recommended in children age < 3 years, and primary reexcision was the treatment of choice, wherever feasible. The purpose of surgical exploration after chemotherapy alone is to avoid XRT or use it in lower doses. Unfortunately, the value of histologic assessment is uncertain: In the SIOP MMT84 study, the local recurrence rate remained high even when biopsies apparently confirmed clinical remission,11 and this has been confirmed by the ICG experience (our unpublished data). Whereas the best timing and modulation of local tumor control still is debated, the development of methods for ascertaining the presence of minimal residual disease is desirable.

In the current retrospective study, we identified several factors that are important in determining the likelihood of survival for patients with recurrent RMS. Multivariate analysis found that two characteristics of the initial tumor (histology and primary site) and two characteristics of the recurrence (type and temporal relation with therapy) were associated significantly with prognosis. Patients with a nonalveolar histology, primary tumor site different from parameningeal and “other” sites, local recurrence, and recurrence off therapy had a better prognosis.

It is well known that patients with alveolar RMS have a higher risk of recurrence. The unfavorable role of the alveolar histology also has been confirmed after recurrence, and survival is very poor in these patients (only 11% by comparison with 38% in patients with nonalveolar RMS). This is consistent with the experience of Pappo et al.6 who, considering children with Group I–IV disease who were treated on the IRS-III, IRS-IV pilot, and IRS-IV protocols, reported a 5-year survival rate after recurrence of 5% for patients with an alveolar histology compared with 26% for patients with the embryonal subtype. It should be noted that children with botryoid tumors had an encouraging 64% 5-year survival rate. A specific analysis for botryoid tumors was impossible in our study because of the small number of patients with this histologic subtype (n = 6 patients). Klingebiel et al.12 confirmed that patients with recurrent embryonal tumors fared significantly better.

Initial tumor site also is an important variable after recurrence. Patients with GU non-B/P RMS and orbit RMS had the best outcomes (60% and 56%, respectively), whereas patients with “other” and PM sites had the worst outcomes (18.8% and 0%, respectively).

Children with RMS who have systemic recurrences are more difficult to rescue, and only 10% survive at 5 years. Similar results have been described in the SIOP MMT84 study: The 5-year survival rate in that study was 0% after distant recurrences and 41% after local recurrences of nonmetastatic RMS.10

In the current study, the timing of the recurrence also was important, as also noted in the study by Raney et al.,13 who reported that children who developed recurrent disease after completing chemotherapy had a significantly higher survival rate compared with patients who had developed recurrences while they were receiving chemotherapy (19% vs. 2.7%; P < 0.05). Recurrence during treatment indicates a biologically more aggressive tumor or the selection of chemoresistant clones that make retrieval therapy very difficult.

Little information has been published to date on the role of treatment after recurrence, and this aspect could not be analyzed in our series because of the variety of chemotherapeutic regimens administered. In the report by Raney et al., complete excision of the recurrent tumor was associated with a better outcome.13 Klingebiel et al. found that patients who received salvage regimens that included XRT had a significantly better survival (3-year event-free survival rate, 46% vs. 10%; P = 0.002).12 This is confirmed by our own experience: Patients who had no or only one risk factor and patients who had RMS with nonalveolar histology and had not received XRT during their initial treatment had better outcomes after recurrence: This presumably was due to the XRT given during second-line treatment.

This finding also is in agreement with the SIOP experience,10 in which patients with locally recurrent RMS who had been treated with chemotherapy alone had a better survival than patients who also had received XRT (46% vs. 29%, respectively). In the large retrospective analysis by Pappo et al., the influence of initial XRT on survival after recurrence was not considered, because most patients in the IRS studies received XRT as part of their first-line treatment.6 This is because the IRS Group's strategy differs from that of the European Groups on this matter: In IRS studies, only Group I patients with nonalveolar RMS were not given XRT (except for the IRS-I study), whereas the SIOP strategy tries to treat patients successfully without significant local therapy.

We used the prognostic factors identified by multivariate analysis to develop a risk-stratification model for children with recurrent RMS. Patients with no risk factors had a satisfactory outcome, with 71% of them surviving, whereas children who had ≥ 2 negative factors had a very unfavorable outcome. Unlike the patients described in the IRS-G report by Pappo et al.,6 the patients who experienced disease progression were excluded from our analysis, although, in our experience, their outcome was as poor as in that patients with three or four risk factors who developed recurrent disease.

Pappo et al. estimated a 5-year survival rate after recurrence of approximately 50% for patients who were diagnosed initially with 1) botryoid tumors, 2) Stage I or Group I tumors (nonalveolar or alveolar) with local or regional recurrence, and 3) Group I alveolar or undifferentiated RMS.6 Our results differed to some degree, because the IRS clinical group does not emerge as a risk factor even in univariate analysis, probably because our sample was less numerous and included fewer botryoid RMS, and the analysis was limited to localized RMS. Univariate analysis showed a higher 5-year OS rate for Group III patients compared with Group II patients, but the difference was not statistically significant. A possible explanation may be the higher percentage of alveolar tumors in Group II than in Group III (47% vs. 34%), whereas no difference was documented in the use of XRT as first-line therapy in these patient subsets (63% and 60% in Groups II and III, respectively).

The data emerging from this study may be useful in planning risk-adapted salvage protocols, such as the Children's Oncology Group ARST 0121 randomized Phase II window study concerning new drug combinations.

In salvage protocols, patients with no risk factors may benefit from intensive standard chemotherapy and XRT, whereas experimental treatments should be reserved for patients with more than one unfavorable risk factor. Unfortunately, only 12% of patients who develop recurrent disease belong to the group with a highly favorable prognosis. The use of experimental therapies may be questionable for children with a single risk factor (32%), whose prognosis is not so poor. In this subgroup, intensive standard chemotherapy may be appropriate combined with XRT in patients who have not received it previously.

In conclusion, the results of the current study demonstrated that groups of patients with different prognoses can be identified at the time of RMS recurrence. Patients who have a nonalveolar histology, primary sites that differ from parameningeal and “other” sites, local recurrences, recurrences off therapy, and no previous XRT have the best chance of achieving a second, long-term complete remission.

When disease recurrence occurs, further treatment should be planned in the light of the first-line therapy and the risk factors identified. The current study demonstrated that patients with no risk factors may be cured by intensive chemotherapy and appropriate local treatment, whereas experimental therapy should only be proposed for patients who have more than one risk factor.


  1. Top of page
  2. Abstract
  6. Acknowledgements

The authors thank Ilaria Zanetti and Angela Scagnellato for data processing and Gloria Tridello and Gianluca De Salvo for the statistical analysis.


  1. Top of page
  2. Abstract
  6. Acknowledgements
  • 1
    Carli M, De Bernardi B, Castello M, et al. Long-term results in childhood rhabdomyosarcoma: a retrospective study in Italy. Pediatr Hematol Oncol. 1986; 3: 371378.
  • 2
    Crist W, Gehan EA, Ragab AH, et al. The Intergroup Rhabdomyosarcoma Study-III. J Clin Oncol. 1995; 13: 610630.
  • 3
    Stevens MCG, Rey A, Bouvet N, et al. Treatment of nonmetastatic rhabdomyosarcoma in childhood and adolescence. Results of the third study of the International Society of Paediatric Oncology—SIOP Malignant Mesenchymal Tumor 89. J Clin Oncol. 2005; 23: 26182628.
  • 4
    Koscielniak E, Harms D, Henze G, et al. 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. 1999; 17: 37063719.
  • 5
    Crist WM, Anderson JR, Meza JL, et al. Intergroup Rhabdomyosarcoma Study-IV: results for patients with nonmetastatic disease. J Clin Oncol. 2001; 19: 30913102.
  • 6
    Pappo AS, Anderson JR, Crist WM, et al. Survival after relapse in children and adolescents with rhabdomyosarcoma: a report from the Intergroup Rhabdomyosarcoma Study Group. J Clin Oncol. 1999; 17: 34873493.
  • 7
    Kaplan E, Meier P. Non-parametric estimation from incomplete observation. J Am Stat Assoc. 1958; 53: 457481.
  • 8
    Peto R, Pike M, Armitage P, et al. Design and analysis of randomized clinical trials requiring prolonged observation of each patient: I-Analysis and examples. Br J Cancer. 1977; 35: 139.
  • 9
    Cox DR. Regression models and life tables. J R Stat Soc B. 1972; 34: 187220.
  • 10
    Flamant F, Rodary C, Rey A, et al. 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. 1998; 34: 10501062.
  • 11
    Godzinsky J, Flamant F, Rey A, et al. Value of postchemotherapy bioptical verification of complete clinical remission in previously incompletely resected malignant mesenchymal tumors in children: SIOP 84 malignant mesenchymal tumours study. Med Pediatr Oncol. 1994; 22: 2226.
  • 12
    Klingebiel T, Pertl U, Hess C, et al. Treatment of children with relapsed soft tissue sarcoma: report of the German CESS/CWS REZ 91 trial. Med Pediatr Oncol. 1998; 30: 269275.
  • 13
    Raney RB Jr., Crist WM, Maurer HM, Foulkes MA, for the Intergroup Rhabdomyosarcoma Committee. Prognosis of children with soft tissue sarcoma who relapse after achieving a complete response. A report from the Intergroup Rhabdomyosarcoma Study I. Cancer. 1983; 52: 4450.