Implications of Implementing Children’s Oncology Group Risk Stratification to Patients With Rhabdomyosarcoma Treated on European Paediatric Soft Tissue Sarcoma Study Group Clinical Trial

Abstract

Background:

Prognostic factors are crucial in tailoring treatments for patients with rhabdomyosarcoma (RMS). The European paediatric Soft tissue sarcoma Study Group (EpSSG) and the Children’s Oncology Group (COG) employ similar prognostic factors, but utilize them differently resulting in diverse stratification systems. This diversity may result in dissimilar treatment approaches for comparable patients and hinder the comparison of clinical trial results.

Procedure:

We reclassified 1993 patients enrolled in the EpSSG RMS 2005 and MTS 2008 studies based on the risk stratification used in current EpSSG and COG trials, and compared the type and cumulative doses of chemotherapy recommended to the different risk groups. Alkylating agents were compared using the cyclophosphamide equivalent dose formula. Metastatic RMS with high-risk features were excluded because no standard recommended treatment exists.

Results:

Patients were variably distributed across EpSSG and COG risk stratifications. Notably, 34.2% of EpSSG standard-risk patients fell into three different COG risk groups (very low, low, and intermediate), and 66.8% of the total population, classified as standard, high, and very high risk by EpSSG, would all be considered intermediate risk by COG. Consequently, only 57.3% of the study population would receive comparable intensive chemotherapy under both EpSSG and COG protocols. Disparities emerged, with 16.5% undergoing more intensive and 17.2% receiving less intensive treatment in COG protocols compared to EpSSG studies.

Conclusions:

Our study shows the complexities of the current RMS risk stratification systems, emphasizing the need for a global consensus. A unified approach would reduce the risk of disparate treatments for similar patients and facilitate more straightforward cross-study comparisons.

1 |. Introduction

Since the initial clinical trials in the early 1970s, the survival rates for patients diagnosed with rhabdomyosarcoma (RMS) have seen notable improvements. This progress can be attributed to the continuous enhancement of multimodal therapeutic approaches that typically combine chemotherapy with surgical intervention, and in most instances include radiotherapy [1].

The likelihood of successfully treating RMS varies from patient to patient and is influenced by an array of prognostic indicators. These encompass patient-specific attributes such as age, as well as tumor-specific factors like the presence or absence of metastatic spread, site, and size. More recently, molecular markers like PAX–FOXO1 fusions and MyoD1 mutations have been integrated into prognostic stratification [2].

However, there exists a complexity in the treatment landscape. Multiple cooperative research and clinical groups, have each historically formulated their own stepwise systems for risk stratification, utilizing a different set of prognostic factors or differently weighting the same factors [2–4]. This divergence results in two important outcomes: first, patients with analogous conditions may undergo divergent treatments depending on the stratification system employed; second, these disparities pose challenges when comparing the effectiveness of treatments across various clinical trials.

Amongst the largest cooperative entities coordinating RMS treatment protocols are the European paediatric Soft tissue sarcoma Study Group (EpSSG) and the Children’s Oncology Group (COG). Both organizations employ similar prognostic variables, but diverge in how these factors are weighted within each of their risk stratification systems.

These differences have been greatly influenced by the treatment strategies utilized in the subsequent studies coordinated by the two groups. The combination of vincristine, actinomycin D, and cyclophosphamide (VAC) was adopted in the initial Intergroup Rhabdomyosarcoma Study (IRS) Group studies and, although the dosage of the drug has varied in the different trials, it remains the backbone for COG studies. VAC was also adopted in the initial European studies, but when ifosfamide demonstrated a higher response rate in comparison to cyclophosphamide-containing regimens [5], the IVA regimen (ifosfamide/vincristine/actinomycin D) became the standard in Europe. This was in line with the European strategy that for many years tried to intensify chemotherapy in order to limit or even avoid aggressive local treatment and radiotherapy in particular [1].

The objective of this analysis is two-fold: (I) to contrast the risk-classification methodologies employed by the EpSSG and COG; and (II) to assess the resulting differences in (a) patient categorization, and (b) chemotherapy regimens applied.

2 |. Methods

For our analysis, we utilized data from the patients enrolled in the EpSSG RMS 2005 [4] and MTS 2008 [6] studies, reclassifying these patients based on the risk stratification used in the current EpSSG trial and comparing to current COG risk classification. The existing EpSSG risk stratification is operational in the ongoing Frontline and Relapsed Rhabdomyosarcoma (FaR-RMS) study (refer to Table S1) and incorporates slight adjustments compared to the RMS 2005 criteria, that is the use of FOXO1 fusion status instead of histology and the inclusion of bladder-prostate and biliary RMS in the group of favorable sites.

COG’s risk classification system, much of which has been in effect for approximately two decades, is also employed with minor variations in the current ARST studies (as detailed in Tables S2 and S3).

The ongoing cooperative group studies include randomized clinical trials; however, to specifically examine potential treatment variations arising from different risk stratifications, we focused on the regimens being used as the standard or comparator chemotherapy arms in the current FaR-RMS and the recently completed ARST studies.

While fusion status is a classification metric in the ongoing protocols, it was not used in the RMS 2005 study. To address this, for cases where fusion status was unavailable, embryonal RMS was designated as fusion-negative and alveolar RMS as fusion-positive, consistent with previous studies [7].

Table 1 outlines the standard chemotherapy regimens for both EpSSG and COG, complete with cumulative dosages for each agent. To enable a comparison of the alkylating agents cyclophosphamide and ifosfamide, we employed the cyclophosphamide equivalent dose (CED) as formulated by Green et al. [8]. This measure takes into account the myelotoxic effects of each of these agents, and is calculated as follows: CED (g/m²) = 1.0 (cumulative cyclophosphamide dose in mg/m²) = 0.244* (cumulative ifosfamide dose in g/m²). According to this metric, 6 g/m² of ifosfamide in the EpSSG IVA regimen equates to 1.464 g/m² of cyclophosphamide.

TABLE 1 |.

EpSSG and COG standard chemotherapy regimens.

Regimen	Patent risk group	No. of cycles	Drug cumulative doses	Duration of CT (weeks)	CED (g/m2)
EpSSG
VA	Low	8	VCR: 24 mg/m2 ACT: 12 mg/m2	24	0
IVA/VA	Standard	9 (IVA ×4 + VA ×5)	IFO: 24 g/m2 VCR:19.5 mg/m2 ACT: 13.5 mg/m2	27	5.8
IVA + CPM/VNL maintenance	High	15 (IVA ×9+ Maintenance ×6)	IFO: 54 g/m2 VCR: 19.5 mg/m2 ACT: 13.5 mg/m2 VNL 450 mg CPM: 4.2 g	51 (27 standard + 24 maintenance chemotherapy)	17.3
IVADo + maintenance	Very high	15 (IVADO ×4 + IVA ×5 + maintenance ×6)	IFO: 54 g/m2 VCR: 19.5 mg/m2 ACT: 13.5 mg/m2 DOXO: 240 mg VNL: 450 mg CPM: 4.2 g	51 (27 standard + 24 maintenance chemotherapy)	17.3
COG
VA	Very low	8	VCR: 27 mg/m2 ACT: 0.36 mg/kg = 10.8 mg/m2	24	0
VAC/VA	Low	8 (VAC ×4 + VA ×4)	VCR: 27 mg/m2 ACT: 0.36 mg/kg = 10.8 mg/m2 CPM: 4.8 mg/m2	24	4.8
VAC	Intermediate	14	VCR: 45 mg/m2 ACT: 0.54 mg/kg = 16.2 mg/m2 CPM: 16.8 mg/m2	43	16.8

Note: Each cycle lasts 4 weeks.

Abbreviations: ACT, actinomycin D; CED, cyclophosphamide equivalent dose; COG, Children’s Oncology Group; CPM, cyclophosphamide; CT, chemotherapy; DOXO, doxorubicin; EpSSG, European paediatric Soft tissue sarcoma Study Group; IFO, ifosfamide; IVA, ifosfamide–vincristine–actinomycin D; IVADo, ifosfamide–vincristine–actinomycin D–doxorubicin; VA, vincristine–actinomycin D; VAC, vincristine–actinomycin D–cyclophosphamide; VCR, vincristine; VNL, vinorelbine.

We gauged the intensity of the various chemotherapy regimens based on their planned cumulative dosages. Greater emphasis was placed on the cumulative dose of alkylating agents compared to vincristine and actinomycin D, given their higher associated risks of acute toxicity and long-term effects. Furthermore, considering the increased risks of acute and long-term cardiotoxicity linked to the inclusion of anthracyclines, we deemed the IVADo (ifosfamide–vincristine–actinomycin D–doxorubicin) regimen more intensive, regardless of the cumulative alkylating agent dose. Patients with metastatic RMS presenting high-risk features, such as alveolar histology or embryonal histology in patients older than 10 years, were not included in this analysis, as there are no standard recommended treatments from either COG or EpSSG for these cases.

The RMS 2005 and MTS 2008 are registered with EUDRACT Number 2005-000217-35 and ClinicalTrials.gov number NCT00339118. All participating centers were required to obtain written approval from their local authorities and ethical committees, and written informed consent from patients and/or their parents or legal guardians.

3 |. Results

The RMS 2005 study, active from 2005 to 2016, enrolled 1733 patients diagnosed with localized RMS. The MTS 2008 study, conducted between 2008 and 2016, enrolled 270 patients with metastatic RMS. Due to missing data, 10 patients—six with localized and four with metastatic RMS—were excluded, leaving a total of 1993 patients for this analysis. Patient classifications based on the latest EpSSG and COG risk stratification systems are summarized in Table 2.

TABLE 2 |.

COG versus FaR-RMS patients classification.

	Low	Standard	High	Very high
	Subgroup A	Subgroup B-C	Subgroup D-E-F	Subgroup G-H
FaR-RMS COG	VA ×24 weeks	IVA/VA ×25 weeks	IVA ×25 weeks + CPM/VNL ×24 weeks	IVADo/IVA ×25 weeks + CPM/VNL ×48 weeks	Total
Very low VA ×24 weeks	72	102	–	–	174
Low VAC/VA ×24 weeks	10	263	35	–	308
Intermediate VAC ×43 weeks	–	318	808	206	1332
High VAC ×43 weeks	–	–	–	179	179
Total	82	683	843	385	1993

Note: In gray: patients who would receive treatment of similar intensity in both protocols. In green patients who would receive less intensive treatment in FaR-RMS. In red patients who would receive more intensive treatment in FaR-RMS. In blue patients without a standard treatment.

Abbreviations: COG, Children’s Oncology Group; CPM/VNL, cyclophosphamide–vinorelbine; EpSSG, European paediatric Soft tissue sarcoma Study Group; IVA, ifosfamide–vincristine–actinomycin D; IVADo, ifosfamide–vincristine–actinomycin D–doxorubicin; VA, vincristine–actinomycin D; VAC, vincristine–actinomycin D–cyclophosphamide.

The data revealed an important variability in patient distribution across the different risk categories. This is due in particular to the fact that 683 (34.2%) EpSSG standard-risk patients are included in three different COG risk groups (very low, low, and intermediate), and 1332 patients (representing 66.8% of the whole population) included in the standard-, high-, and very high-risk groups by EpSSG would all be classified as intermediate risk in COG.

Across the entire treated cohort, 1143 patients (comprising 57.3% of the study population) would receive similarly intensive chemotherapy regimens under either the EpSSG or COG protocols. These patients were divided into three main categories:

• Category 1: 72 patients identified as COG very low risk and EpSSG low risk would receive VA (vincristine–actinomycin D) chemotherapy for 24 weeks. Minor treatment differences were noted: COG patients would receive a moderately higher cumulative dose of vincristine (12% more) and a slightly lower dose of actinomycin D (9% less).

• Category 2: 263 patients in the COG low-risk and EpSSG standard-risk group would undergo VAC/VA or IVA/VA treatment for 24–25 weeks. In this group, treatment duration and cumulative doses of vincristine and actinomycin D were similar to Category 1. COG patients would, however, receive a moderately reduced CED (approximately 17% less).

• Category 3: 808 patients, the largest group, in the COG intermediate-risk and EpSSG high-risk group would receive differently administered alkylating agents, but a comparable CED: 16.3 g/m² in EpSSG versus 16.8 g/m² in COG.

The remaining patients would receive regimens with different intensities under the COG or EpSSG protocols. Specifically, 328 patients (or 16.5% of the total) would undergo more intensive treatment under COG protocols compared to those recommended by the EpSSG FaR-RMS study. This includes a small subgroup (10 patients) assigned to the low-risk group both in COG and EpSSG who would receive alkylating drugs only if treated in the COG protocol, and an additional subgroup of 318 patients included to the intermediate group by COG but assigned to the standard-risk group by EpSSG. According to EpSSG FaR-RMS recommendations, this group would receive a lower cumulative dose of ifosfamide (CED 5.8 g/m² in EpSSG vs. 16.8 g/m² in COG) and lower cumulative doses of vincristine (19.5 mg/m² in EpSSG vs. 45 mg/m² in COG) and actinomycin D (13.5 mg/m² in EpSSG vs. 16.2 mg/m² in COG).

Conversely, 343 patients (or 17.2% of the total) would receive less intense treatment under the COG protocols compared to the EpSSG FaR-RMS guidelines. This group includes three subsets of patients. First, 102 patients who COG would classify as very low risk and treat with VA only, whereas EpSSG classifies these patients in the standard-risk group and would treat them with a limited dose of ifosfamide. Second, a small group of 35 EpSSG high-risk patients would be classified as low risk by COG and would receive only four cycles of cyclophosphamide, instead of nine courses of ifosfamide plus maintenance provided by FaR-RMS. Finally, 206 patients would be classified as very high risk according to EpSSG (119 localized and 87 metastatic), but as intermediate risk by COG, and if treated according to the EpSSG protocol would receive anthracyclines and a similar CED.

A final subgroup consisted of metastatic patients with high-risk features (179 in this study) was not included in this analysis because in consideration of their grim prognosis, no standard treatment exists; these patients are treated in both COG and EpSSG protocols with intensive chemotherapy and/or experimental drugs.

4 |. Discussion

Advancements in the treatment of pediatric RMS have been significant, largely thanks to the collective efforts of national and international research collaborations that have refined treatment approaches. These refinements were predicated on a series of prognostic factors linked to varying levels of treatment failure risk [1, 3]. Metastasis at diagnosis remains the most formidable risk factor, with survival rates plummeting to just one-third at 3 years post diagnosis [9].

Both the EpSSG and COG leverage the same prognostic factors for patient stratification in their respective protocols: postsurgical IRS group, fusion status, tumor site, nodal involvement, tumor size, and patient age. However, nuances in their application of these factors have led to a heterogeneous treatment landscape. For instance, the definition of “low risk” varies between the two groups: the low-risk group in EpSSG includes only small, embryonal RMS tumors completely resected at diagnosis (IRS Group I) in patients of favorable age, whereas COG defines very low- and low-risk patients to include the same patients as EpSSG but also includes IRS Group II and orbital RMS tumors. This difference is strictly related to the treatment provided for this group, which includes a limited dose of alkylating agents in COG studies but systematic irradiation for non-IRS I patients. In EpSSG studies, this group has been selected to include only those who could be treated without alkylating drugs or radiotherapy.

Differences in patient stratification may make it difficult to directly transfer new findings from one population to another. This is the case for instance for the maintenance treatment that was found effective in the RMS 2005 trial, but remains difficult to translate in the COG context due to the different risk stratification and treatment strategies [10].

Our study reveals a notable variation in patient classification when comparing EpSSG and COG protocols, particularly in the COG intermediate-risk group and the EpSSG standard-risk group. This incongruity leads to variations in chemotherapy regimens for patients, depending on whether the diagnosis was made in Europe or the United States. While only a modest 3.7% of the whole cohort of patients would receive almost identical treatment by both cooperative groups, this percentage climbs to nearly 60% when considering the CED of alkylating agents.

Recent evidence from COG has suggested improved local control with higher cumulative doses of cyclophosphamide (30.8 g/m²) [11]. If higher cyclophosphamide doses are again adopted for future COG studies, the differences in CED between the COG and EpSSG regimes will increase [12].

However, the use of CED as a comparative measure presents limitations. Although it offers a comparable index for alkylating agents based on myelotoxicity, it does not account for other long-term effects, like risk of infertility or nephrotoxicity. In addition, it does not take into consideration different methods of administering alkylating agents, that is, high dose versus continuous low dose (maintenance). A correlation between a higher CED score and fertility impairment and Leydig cell function has been reported, however, supporting our choice to use this indicator to compare chemotherapy regimens that use different types of alkylators [13, 14].

Relatedly, an additional limitation is the comparison of differences in the intensity of various regimens, such as the VAC regimen given for 43 weeks versus the IVA regimen plus maintenance given for 51 weeks. In an EpSSG study, maintenance was very tolerable, with only 45% of patients experiencing grade 4 neutropenia [15]. Similarly, the VAC regimen with reduced cyclophosphamide dosing of 1.2 g/m² was also well tolerated by patients, with grade 4 neutropenia observed in 44.9% [16]. The long-term sequelae of the different regimens are difficult to compare, and in particular the administration of low-dose cyclophosphamide after ifosfamide remains to be determined. Apart from alkylator doses, the cumulative dose of vincristine is higher in the COG protocols, but vinorelbine is added in high-risk patients in the EpSSG protocols and this may add neurotoxicity, although this was not observed in the RMS 2005 study [15]. The cumulative dose of actinomycin D recommended by COG is lower in very low- and low-risk patients, but higher in intermediate-risk patients in comparison to FaR-RMS. Ideally, a direct comparison of acute toxicity should be performed (although veno-occlusive disease [VOD], the most worrisome actinomycin-related side effect, is not related to actinomycin cumulative dose), and there is less concern about long-term actinomycin toxicity.

Finally, it is important to understand if patients with similar characteristics who are treated differently by the two cooperative groups experience differences in event-free and overall survival (OS).

Common EpSSG–COG analyses for some subgroups of patients have already been done. The outcome of patients with alveolar RMS and regional nodal involvement (very high-risk according to EpSSG, but intermediate-risk group according to COG) showed no differences in the 5-year event free survival (44.05% and 49.03%), and OS (53.65% and 50.89%) on ARST0531 and RMS 2005, respectively [17]. These results suggest that anthracyclines used in the EpSSG protocol may be avoided in this group of patients, but their unsatisfactory prognosis makes difficult any attempt to reduce the treatment intensity.

Another interesting finding suggesting that EpSSG may reduce the intensity of its treatment regimens is the subset of EpSSG standard-risk patients who are treated by COG without alkylating agents. A substantial proportion of patients included in this group present with orbital RMS and present a very good outcome, so the systematic use of radiotherapy may allow a reduction in the cumulative doses of alkylating agents. EpSSG has already begun to move in this direction, as these patients were historically treated with nine ifosfamide-containing cycles but this was reduced to four in RMS 2005 [3] and in the ongoing FaR-RMS study. On the other hand, the EpSSG experience may suggest that the dose of alkylating agents may be reduced by COG in a small group considered low risk by both protocols but treated without alkylating agents by EpSSG and a larger group included in the intermediate group by COG but treated with only four IVA cycles and five VA in EpSSG standard group.

Making a direct correlation just between the intensity and cumulative doses of chemotherapy treatment and survival is difficult and may be misleading as the effect of local control needs to be considered. For example, in the past, the lower progression-free survival (PFS) (but not OS) presented in some European studies was explained by the attempt to avoid radiotherapy in a portion of patients [18]. The recent RMS 2005 study implemented radiotherapy more systematically and showed results comparable to the COG studies [3]. Another example of potential interrelationship between local therapy and cumulative alkylator dose is the COG experience in the ARST0331 trial, where subset 2 low-risk RMS patients showed inferior local control with reduced cyclophosphamide dose [19]. The current debate on optimal cyclophosphamide dose for COG intermediate-risk patients is the subject of planned future COG randomized trial that will compare the two different doses of cyclophosphamide [12].

5 |. Conclusions

Our study shows the complexities of the stratifications currently in use, and underline the need for a global unified risk stratification system for RMS patients. A harmonized approach would mitigate the risk of similar patients receiving varied treatments, and would enable more straightforward cross-study comparisons. The newly founded INSTRuCT consortium, an outcome of longstanding collaborations between European and North American Cooperative Groups, may serve as a catalyst in this regard [20].

In summary, our study underscores the imperative for international collaboration to harmonize treatment protocols for RMS, thereby enhancing our understanding and management of this complex pediatric malignancy.

Supplementary Material

Additional supporting information can be found online in the Supporting Information section.

Abbreviations:

COG

Children’s Oncology Group

EpSSG

European paediatric Soft tissue sarcoma Study Group

FaR-RMS

Frontline and Relapsed Rhabdomyosarcoma study

IRS

Intergroup Rhabdomyosarcoma Study

IVA

ifosfamide–vincristine–actinomycin D

IVADo

ifosfamide–vincristine–actinomycin D–doxorubicin

OS

overall survival

PFS

progression-free survival

RMS

rhabdomyosarcoma

VA

vincristine–actinomycin D

VAC

vincristine–actinomycin D–cyclophosphamide

VOD

veno-occlusive disease

Data Availability Statement

Data sharing not applicable to this article as no datasets were generated or analyzed during the current study.