Long-term results in childhood rhabdomyosarcoma: A report from the Italian cooperative study RMS 79

Abstract

Background

The results obtained by protocols for children with rhabdomyosarcoma (RMS) have improved in recent decades. Survival curves usually reach a plateau 3 years after the diagnosis, suggesting that long-term survival can be expected, but late events are known to occur. We analyzed the long-term results of the RMS 79 protocol to investigate the type and impact of such events.

Procedure

From 1979 to 1987, 163 children with RMS diagnosed at 21 Italian institutions were registered. Each institution was contacted every year to record patients' status after the end of treatment. When patients were lost to follow-up, their status was checked by inquiring at the Registry Offices of the towns of residence and the cause of death or occurrence of second cancers was investigated by contacting the patients or their family by phone.

Results

Overall, 16 patients had late events, that is, 7 tumor recurrences, 6 second tumors, and 3 deaths due to treatment-related complications. The overall survival rates dropped from 62.6 at 3 years to 52.8 at 20 years. By multivariate analysis, the characteristics influencing long-term survival were histology, tumor site and size, and IRS group. Factors predictive of any kind of late event were tumor site and IRS group.

Conclusions

Major late events can significantly affect the long-term survival of children with RMS. Modern protocols should provide for a much longer follow-up than is usually considered to confirm the results achieved and enable possible correlations between primary treatment and late events to be investigated. Pediatr Blood Cancer 2012; 58: 872–876. © 2011 Wiley Periodicals, Inc.

INTRODUCTION

The outcome of patients with rhabdomyosarcoma (RMS) has greatly improved in the last few decades 1, thanks especially to the use of multidisciplinary treatments involving large numbers of patients. Such trials were initially conducted in the USA, by the Intergroup Rhabdomyosarcoma Study Group (IRSG) 2, and in Europe 3, 4. The Italian Cooperative Group for Pediatric Soft Tissue Sarcomas (now STSC: Soft Tissue Sarcoma Committee) started its first trial in 1979, called the RMS 79. The treatment plan was based on a risk-related approach, depending on the tumor's characteristics at diagnosis and the results of initial surgery. The primary aims of the RMS 79 protocol were: (1) to treat children with RMS and other sarcomas on a national basis; (2) to improve the survival rates (and patients' quality of life, trying to preserve organ function) in the frame of a multimodality treatment.

The event-free survival (EFS) curve for patients entering the RMS 79 reached a plateau 3 years after their diagnosis, suggesting that long-term survival might be expected beyond this time point. Major late events, such as recurrences, second malignant tumors (SMT), or death from treatment-related complications, may occur several years after stopping the treatment, however, so a lengthy follow-up is desirable to confirm the survival rates.

This is an analysis of the long-term outcome of patients included in the RMS 79 protocol, designed to evaluate the type and prevalence of late events in children with RMS and their impact on survival.

METHODS

One hundred and seventy-two patients under 16 years old with histologically confirmed RMS were diagnosed at 21 Italian institutions and registered in the RMS 79 study from March 1979 to April 1987. Three children treated according to a different protocol, five pretreated with chemotherapy or radiotherapy, and one lost to follow-up shortly after diagnosis were ruled out, leaving 163 children considered in this analysis. The boy to girl ratio was 1.7:1 and the age at diagnosis ranged from 7 days to 15 years (median 4.6 years).

All patients were classified after initial surgery according to the IRS staging system, that is: Group I—completely excised tumors, Group II—grossly resected tumors with microscopic residual disease and/or completely excised positive regional lymph nodes; Group III—gross residual disease after incomplete resection or biopsy; Group IV—distant metastases at onset. Patients were also prospectively classified according to the TNM criteria. Tumor sites were retrospectively reassigned soon after a shared definition was agreed internationally 5.

Treatment has been described elsewhere 6. Patients were treated according to their IRS grouping at the time of registration. Chemotherapy had to start within 8 weeks of the initial diagnostic surgery. Patients classified as Groups I and II received 12 courses of alternating CAV (cyclophosphamide 150 mg/m² on days 1–7; adriamycin 35 mg/m² on day 8, vincristine 1.5 mg/m² on day 8) and VAC (with actinomycin D 1.5 mg/m² on day 8, instead of adriamycin) every 4 weeks. Patients in Group III or IV, and those with an alveolar histology received a longer treatment involving 18 alternating courses of CAV/VAC.

All patients were eligible to receive radiotherapy (RT) except for those in Group I and those in complete remission after primary chemotherapy or delayed surgery. RT doses were 40–45 and 45–55 Gy for Group II and Groups III–IV patients, respectively. The maximum dose for children under 6 years old was 40 Gy. The target volume was represented by the tumor bed plus 5 cm and radiation was delivered at a rate of 180–200 cGy a day, 5 days a week. Clinically involved nodes and metastases were treated in the same way as the primary tumor. In cases of lung lesions, bilateral pulmonary irradiation up to 18 Gy was indicated.

After the treatment, all of the institutions that had enrolled patients were contacted at least every year by the Coordinating Center in Padova and their patients' status at their latest follow-up was recorded. For patients no longer in contact with the centers, their status was assessed by making specific inquiries at the Registry Offices of their home towns and the cause of death or occurrence of SMT was investigated by contacting the patients or their family by phone, subject to their consent. Consent to participation in the RMS 79 study had been requested at the moment of each patient's registration in the protocol.

Statistical Methods

Overall and EFS curves were calculated from the start of treatment using the Kaplan–Meier method, considering overall survival (OS) from diagnosis to latest follow-up or death from any cause, EFS from diagnosis to first progression, relapse, occurrence of second tumor or death from any cause, or latest contact for children who never experienced an event. A late event was defined as a disease recurrence, the development of an SMT or death from any cause occurring more than 3 years after the patient's diagnosis. Cumulative incidence curves were calculated to estimate the risk of all late events and SMT in particular. Differences between the curves were analyzed using the log-rank test. Statistical significance was defined as P < 0.05. A multivariate analysis of different patient characteristics and risk factors was performed using Cox's proportional hazards model. All statistical analyses were performed using the SAS statistical package.

RESULTS

All patients eligible for the RMS 79 protocol were included in this analysis. Their clinical characteristics are shown in Table I. The median follow-up for patients who were alive was 22.9 years (range 17.1–27.9) and only four patients had a follow-up shorter than 20 years. Overall, 83 patients were alive with no evidence of RMS at the time of this analysis.

Table I. Characteristics of 163 Patients Included in the Protocol RMS 79

	N	%
Age (years)
<1	12	7.4
1–9	122	74.8
>10 years	29	17.8
Gender
Male	103	63.2
Female	60	36.8
Histology
Embryonal	97	59.5
Botryoid	14	8.6
Leiomyomatous	10	6.1
Alveolar	42	25.8
Initial primary site
Orbit	24	14.7
Head and neck, non-parameningeal	14	8.6
Head and neck, parameningeal	30	18.4
Bladder/prostate	14	8.6
Genito-urinary, non-bladder/prostate	22	13.5
Extremity	20	12.3
Other	39	23.9
Size
≤5 cm	77	47.2
>5 cm	86	52.8
Invasiveness
T1	80	49.1
T2	83	50.9
Clinical nodal involvement
N0	134	82.2
N1	27	16.6
Nx	2	1.2
IRS group
I	32	19.6
II	20	12.3
III	91	55.8
IV	20	12.3

Events are shown in Table II according to the time of their occurrence. Thirty-two patients had tumor progression during their treatment and 48 experienced a tumor relapse. In eight patients the relapse occurred more than 3 years after the initial diagnosis, but after 5 years in only two cases, that is, 8.6 and 8.7 years afterwards. Overall, seven children had a local relapse and one had a combined local and metastatic relapse. A second CR was obtained in seven patients but only one of them, relapsing 4.2 years after the initial diagnosis, was alive 22 years after the event. The clinical characteristics of these patients are shown in Table III.

Table II. Type of Event by Time After Initial Diagnosis

Event	<3 years	3–5 years	>5 years
Tumor progression	33	—	—
Tumor recurrence	40	6	2
Treatment-related death	5	—	3
Second tumor	—	1	6

Table III. Late Events Occurring Among Patients Included in the RMS 79 Protocol

No.	Sex and age at diagnosis (years)	Histotype and tumor site	IRS group	T	RT dose (Gy)	Type of late event	Interval between diagnosis and event (years)	Outcome (years after initial diagnosis)
RT, radiotherapy; LR, local relapse; CR, complete remission; MTS, metastasis; SMT, second malignant tumor; NA, not administered during first-line treatment; DOD, died of disease; NED, no evidence of disease.
1	M (2.7)	Embryonal, abdomen	III	T2b	40	LR	3.1	DOD (3.2)
2	M (15.9)	Embryonal, head and neck	III	T2b	50	LR and SMT	3.2	DOD (12)
3	M (1)	Alveolar, orbit	III	T1a	51	LR	3.3	DOD (7.4)
4	M (8.9)	Alveolar, arm	III	T2b	NA	LR	3.6	DOD (13.9)
5	F (3.7)	Alveolar, leg	I	T1a	NA	LR	4.2	Alive NED (26.4)
6	M (15.8)	Embryonal, thoracic wall	II	T2b	46	LR	4.3	DOD (4.6)
7	M (6.2)	Embryonal, orbit	III	T1a	55	LR	8.6	DOD (14.4)
8	F (8.7)	Alveolar, extremity	I	T1b	NA	LR and MTS	8.7	DOD (9.3)
9	M (5.4)	Embryonal, paraspinal	IV	T2b	44	Progressive respiratory insufficiency	4	Death (11.7)
10	F (1.7)	Embryonal, mediastinum	II	T2b	40	Progressive respiratory insufficiency and cardiac failure	2	Death (22.7)
11	M (1.5)	Botryod, bladder	III	T1a	36	Neurological deterioration	0.2	Death (23.3)
12	F (1.3)	Embryonal, vagina	IV	T1a	40	Rectum carcinoma	4	Alive NED (21.8)
13	N (2)	Embryonal, paraspinal	II	T2b	44	Thyroid carcinoma	25.5	Alive NED (26.5)
14	M (3.3)	Embryonal, bladder	III	T1a	44	Angiosarcoma of the heart	22.5	Dead for SMT (22.7)
15	F (15)	Embryonal, bladder	I	T1a	NA	Breast carcinoma	6.4	Alive NED (23.1)
16	F (10.4)	Embryonal, parameningeal	IV	T2a	30	Thyroid carcinoma	12.4	Alive NED (21.6)
17	F (3.8)	Embryonal, diaphragm	I	T2b	44	Clavicle osteosarcoma	7.8	Dead for SMT (11)

Three patients died of treatment-related complications. Two of them one with a paravertebral tumor with bilateral lung metastases, the other with a mediastinal RMS, underwent aggressive surgery and radiotherapy involving both lungs. They developed progressive restrictive pneumopathy 4 and 2 years after completing their treatment, associated with scoliosis, growth limitation and, in one case, heart failure. Both patients died, 11.7 and 22.7 years after their diagnosis, respectively. One patient developed a tetraparesis after a cardiac arrest occurring during the treatment; he died 23.3 years after his diagnosis of cardiac and respiratory complications correlated with his neurological condition.

Seven patients developed an SMT (see Table III). No conditions predisposing to cancer were reported. All patients received chemotherapy according to the protocol and five of them also had radiotherapy. The SMT occurred in the irradiated field in four cases. Two patients died of their SMT. One patient developed a testicular carcinoma 4 years after his initial diagnosis, while receiving treatment for an RMS relapse, and this eventually led to his death.

The OS rates at 3, 10, and 20 years were 62.6%, 55.8%, and 52.8% respectively, while the EFS rate was 52.8 at 3 years, and dropped to 47.2 and 46 at the 10- and 20-year time points (Fig. 1).

Figure 1. RMS 79 protocol: Event-free (dotted line) and overall (solid line) survival curves.

By multivariate analysis, the characteristics prognostic of long-term survival were histology, tumor site and size, and IRS group. The risk of developing any kind of late event was 12.8 (95%CI: 7.3–21.9) 20 years after the diagnosis, but it rose to 21.2 (95%CI: 12.8–33.8) by 25 years after the diagnosis. The risk of SMT was 4.3 (1.6–11.1) and 7.7 (3.4–16.8), 20 and 25 years after the initial diagnosis, respectively (Fig. 2). The relative rate of recurrence was 1.1 (0.16–7.55) and 6.0 (1.84–18.95), while it was 8.15 (4.16–15.66) at both time points for death due to toxicity. The factors predictive of any kind of late event were tumor site and IRS group. The limited numbers of patients involved precluded any further analysis on the different types of late events.

Figure 2. Cumulative incidence of all late events together (solid line) and second tumors alone (dotted line).

DISCUSSION

Clinical series and case reports from population-based registries indicate that the cure rate for children with soft tissue sarcomas has improved over the last 30 years 7, 8. This has been mainly due to the adoption of well-designed protocols based on a multidisciplinary approach. The first Italian protocol was designed on the strength of the initial IRSG experience 2. This study was the first in which different Italian pediatric centers joined forces and it laid the foundations for the subsequent protocols, RMS 88, RMS 96, and EpSSG RMS 2005.

The minimum follow-up was 17 years and only four patients had a follow-up of less than 20 years. Every effort was made to collect data on patients lost to follow-up by the institutions where they were treated, both by inquiring at the Registry Offices of the patients' home towns and by contacting patients or their families, as necessary. The completeness of follow-up represents a major strength of this study since this allow us to describe the full story of the patients included in the trial. A limitation of our study is the small sample size by comparison with prior publications, particularly the Childhood Cancer Survivor Study 9 and the IRSG cohorts 10, and the consequent relatively large confidence intervals in our analysis of the risk of late events.

The 5-year survival rate observed in the present study is comparable with the results obtained by other cooperative groups in their initial protocols. The first IRSG study reported a 55% 5-year OS, which increased to 63% in the IRS-II study 2. Similarly, the first SIOP study described a 52% OS (EFS 47%) 3.

A recent trial has reported results as 3-year survival rates 8, based on the fact that previous studies demonstrated a plateau in the survival curves after 3 years of observation. As shown by the long-term results of the RMS 79 protocol presented here, however, survival may decline substantially beyond this 3-year cut-off, due to late recurrences or treatment-related events, such as severe toxicity or second tumors. Although an early time point for comparison (3 years) is much more practical and acceptable for randomized clinical trials, a longer follow-up is desirable to fully elucidate the survival rates achieved in RMS patients, taking into account both the effects of the disease and the burden of treatment.

The prognostic factors that we found for long-term survival did not differ significantly from those correlating with survival at 3 or 5 years. The factors predictive of late events were tumor site and IRS group, which correlated closely with the burden of treatment. Similarly, a report from the IRSG associated the risk of a late tumor recurrence, after 5 years, with IRS group and tumor size 10. In contrast with this report, patients treated in the RMS 79 were at greater risk of a second cancer than of a late tumor relapse. This is consistent with the results obtained in a population-based study 11 and is explained by the much longer median follow-up in our series, that is, 22.9 years compared to the 8.9 years for the IRSG series. In addition, no effort was made in the IRSG study to collect data on patients lost to follow-up by the institution where they were treated.

The timing of recurrences seems to be important for predicting the likelihood of a child having a second chance of becoming a long-term survivor. In a previous study, we found that patients relapsing more than 18 months after their initial diagnosis stood a better chance of being rescued 12. More recently, Mattke et al. suggested that survival after relapse is directly related to the length of time between the end of the therapy and the onset of the relapse. Patients relapsing after 12 months had a 3-year OS of 48%, which was significantly better than for patients relapsing after 6 months (13%) or between 6 and 12 months (26%) 13. Overall, studies considering the time of recurrence concluded that the prognosis was better the later the event. Our experience, involving a limited number of patients, suggests that this may not be true in cases relapsing very late: Seven of eight patients who relapsed more than 3 years after their initial diagnosis died.

It is also worth emphasizing that the RMS 79 protocol was based on knowledge gathered in the early 1970s, when more limited diagnostic tools were available and different treatment strategies were used by comparison with more recent protocols. Patient survival has increased in the latest international and national protocols, but the question is whether we can expect fewer treatment-related events in the future. This question is difficult to answer because the treatment for RMS is still rather complex. Technological advances may now afford a more accurate diagnosis and a more focused delivery of therapy. In particular, better RT techniques can avoid some of the late sequelae described here. On the other hand, treatment has become more intensive so that the number of patients with organ dysfunctions or secondary tumors correlated with the administration of alkylating drugs, doxorubicin, and etoposide may increase.

A population-based study demonstrated that the relative risk of developing a second malignancy was higher for children treated for soft tissue sarcomas in more recent times: The observed/expected ratio rose from 4.4 to 12.5 for children treated before versus after 1985 14. In contrast, an analysis of SEER data, including patients surviving more than 5 years after their diagnosis, showed a reduction in the 10-year cumulative incidence of mortality from the 1974–1980 to the 1995–2000 period. This decline was evident for both late relapse and non-relapse mortality 15.

Protocols are designed primarily to assess the results of treatment after 3–5 years of observation. As a consequence, the estimation of late mortality rates generally relies on epidemiological studies. Such studies have the merit of following up large cohorts of patients treated for different tumors with different therapies in different periods of time. They usually rely on information through national health registries 11, 16. Causes of death are consequently grouped according to general classifications, that is, diseases of the cardiovascular or respiratory system, making it impossible to go into more detail. The patients who died of treatment-related complications among those treated according to the RMS 79 experienced a series of adverse events that cannot be properly described by epidemiological studies. Our study shows as the survival figures may decline substantially after the closure of a clinical trial and supports the need to organize these trials so as to systematically follow patients for a much longer time than is usually considered necessary in order to confirm the results achieved and enable investigations into the potential correlations between primary treatment and late events.

Acknowledgements

This work was partially supported by a grant from the Fondazione Città della Speranza. The authors would like to thank Angela Scagnellato for data processing and Ilaria Zanetti and Angela De Paoli for the statistical analysis.