Extremity rhabdomyosarcoma in children, adolescents and young adults: A report from Children's Oncology Group trials

Abstract

Background

Long‐term survival and prognostic factors for patients with extremity rhabdomyosarcoma (RMS) treated on contemporary Children’s Oncology Group (COG) trials are unknown.

Methods

Data of extremity RMS patients enrolled on COG trials from 1998 to 2014 were analyzed to estimate event‐free survival (EFS) and overall survival (OS), and factors associated with survival.

Results

The authors identified 264 extremity RMS patients, 159 (60%) localized and 105 (40%) metastatic. The 5‐year EFS and OS of patients with localized disease was 62.6% (54%, 71.2%), and 78.7% (71.4%, 85.9%) and of those with metastatic disease was 7.7% (2.2%, 13.2%) and 22.7% (13.9%, 31.4%). Age at diagnosis was associated with EFS, whereas both sex and age at diagnosis were associated with OS. In metastatic patients, a lower Oberlin score was associated with improved EFS and OS, and females had improved OS. The 5‐year EFS of patients with localized disease treated on the D9803 trial was not statistically different from those treated on ARST0531, but 5‐year OS was statistically superior (adjusted hazard ratio [adjHR] 0.43 [0.21, 0.86]) after adjusting for potential confounders. The 5‐year EFS of group 3 patients undergoing delayed primary excision (DPE) with R0 margins was better than those with DPE with positive margins (adjHR 0.31 [0.11, 0.87]) and comparable to group 2 RMS patients.

Conclusions

Outcomes for metastatic extremity RMS remain poor. Among patients with localized disease, younger age, female sex, and treatment on D9803 were associated with improved OS, whereas DPE with R0 margins led to better EFS compared to DPE with positive margins in group 3 patients.

Short abstract

Among patients with localized rhabdomyosarcoma (RMS), young patients and females had improved overall survival (OS). The OS of localized extremity RMS treated on D9803 was superior to ARST0531. Survival of those with group 3 disease and undergoing delayed primary excision (DPE) with negative surgical margins was comparable to those with group 2 disease.

INTRODUCTION

Extremity rhabdomyosarcoma (RMS) accounts for 15%–20% of RMS cases. ¹ , ² Extremity RMS is more prevalent in older children, often presenting with alveolar histology and regional nodal involvement, and is associated with poorer survival compared to RMS at favorable sites. ¹ , ² , ³ , ⁴ , ⁵ , ⁶ , ⁷ , ⁸ Sentinel lymph node biopsy is recommended due to the high incidence of regional lymph node involvement, which, if overlooked, may result in undertreatment and poor survival. ⁹ , ¹⁰ , ¹¹ An international cohort study of children treated between 1983 and 2004 reported a 5‐year overall survival (OS) of 67% for localized extremity RMS. Inferior OS was independently associated with age >3 years, T2 and N1 stage, incomplete initial surgery, pre‐1995 treatment, and treatment on European cooperative group trials. ¹ Notably, patients with grossly resected tumors had better survival rates in North American trials compared to European trials in this analysis, possibly due to differences in chemotherapy backbone and tumor location. ¹ The study, although extensive, was limited by its heterogeneous treatment protocols, with most patients treated before the year 2000, lack of lymph node sampling data, absence of FOXO1 fusion status information, and the exclusion of metastatic patients. With the recent publication on delayed primary excision (DPE) and cumulative cyclophosphamide dose or intensity potentially impacting survival outcomes of patients with RMS, a thorough analysis of more recent patient cohorts is required to provide updated survival rates and to identify the treatment strategies that may impact the outcomes of extremity RMS patients. ¹² , ¹³

Therefore, we sought to analyze data of patients with extremity RMS treated on the Children’s Oncology Group (COG) trials from 1998 to 2014. ⁶ , ⁸ , ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ Our objectives were to determine the 5‐year event‐free survival (EFS) and OS of the patients with localized and metastatic extremity RMS, identify the factors associated with EFS and OS, compare the outcome of localized extremity RMS treated on two low‐risk RMS trials (D9602 and ARST0331) and two intermediate‐risk (IR) RMS trials (D9803 and ARST0531), and ascertain the effect of DPE on EFS and OS of Intergroup Rhabdomyosarcoma Study (IRS) group 3 extremity RMS patients. ⁶ , ⁸ , ¹⁴ , ¹⁵

MATERIALS AND METHODS

Study design and patient selection

The study employed a retrospective cohort design and identified newly diagnosed patients with RMS of the upper or lower limbs, including shoulder or buttocks, regardless of age at diagnosis and risk‐stratification treated on D9602 (NCT00002995), D9802 (NCT00003955), D9803 (NCT00003958), ARST0331(NCT00075582), ARST0431 (NCT00354744), ARST0531 (NCT00354835), and ARST08P1(NCT01055314) RMS trials. ⁶ , ⁸ , ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ Table S1 outlines the chemotherapy regimens and key local control modalities employed in the respective trials. All trials received approval from the institutional review board of each participating institution or the Pediatric Central Institutional Review Board of the National Cancer Institute. For each eligible patient, we extracted various disease‐, patient‐, and treatment‐related variables, including age at diagnosis, sex, race, tumor site, tumor invasiveness, size, stage, IRS group, histology, FOXO1 fusion status, presence of anaplasia on pathology, and clinical/radiological/pathological regional lymph node involvement, site and number of distant metastatic sites, as well as the treatment protocol. Two investigators independently reviewed each patient's case report form in duplicate to gather information on the type of primary surgical procedures and surgical margins, regional lymph node evaluation, DPE, surgical margins of DPE, and radiotherapy details. Resection was classified as R0 (complete tumor removal with negative margins), R1 (microscopically positive margins), or R2 (macroscopic residual disease). Biopsy, performed for diagnostic purposes, involved obtaining a small tissue sample without intent for total tumor removal. Histological diagnoses were established through a central pathologic review, and all radiation therapy data underwent a centralized quality assurance review.

Statistical analysis

Demographics and disease characteristics were summarized descriptively. Local failure was defined as the progression or relapse at the primary site as the initial event, with or without concurrent regional and/or distant failure. Regional failure was characterized by recurrence in tissue adjacent to the primary site or in regional lymph nodes, whereas the emergence of distant metastases defined distant failure. The Oberlin score was based on four factors: 1) age <1 year or ≥10 years, 2) primary tumor in unfavorable site, such as extremity or other site, 3) metastases in three or more organ systems, and 4) presence of bone or bone marrow metastases. ¹⁹ EFS was defined as the interval from study enrollment to disease recurrence, developing a second malignant neoplasm (SMN), or death from any cause, whichever occurred first. OS was defined as the interval from study enrollment to death from any cause. Data were censored at the date of last contact for patients who were event‐free or still alive. Patient follow‐up information was current as of December 2018. The Kaplan–Meier method was employed to estimate EFS and OS distributions, with 95% confidence intervals (CIs) calculated using the Peto and Peto methods. Differences between survival curves based on patient, disease, or treatment characteristics were analyzed using the log‐rank test. Factors associated with EFS or OS in the univariable analysis at p < .1 were entered into a Cox proportional hazards regression model to evaluate the factors statistically significantly independently associated with EFS and OS of localized and metastatic extremity RMS patients. Additionally, the Cox proportional hazards regression model was employed to adjust for the covariates that differed statistically significantly between the comparisons of EFS and OS between patients with IRS group 3 by DPE status and IR‐RMS trials (D9803 vs. ARST0531). All the tests were two‐tailed with statistical significance defined as p < .05, and analysis was performed using SAS 9.4 (SAS Institute, Cary, North Carolina).

RESULTS

Clinical characteristics

We identified 264 patients with extremity RMS eligible for this study, of which 159 (60%) had localized disease (Table S2). The median age at diagnosis was 7.9 years (range, 0.1–32.4 years). The three most common sites of the primary tumor were thigh (N = 50, 19%), forearm (N = 48, 18%), and leg (N = 48, 18%). The most common RMS histology in this cohort was alveolar RMS (ARMS) (N = 200, 76%). The FOXO1 fusion status was available for 165 patients: ARMS = 150, embryonal RMS (ERMS) = 9, RMS not otherwise specified (NOS)/mixed = 5, and spindle cell RMS = 1. Among 150 patients with ARMS, 133 were FOXO1 fusion‐positive, and among five patients with RMS NOS/mixed histology, three were FOXO1 fusion‐positive. Among 136 patients with FOXO1 fusion, 85 (62%) had a PAX3 partner, 34 (25%) had PAX7, whereas the fusion partner was unknown in 17 (12%) (Table S2).

Among the whole cohort, 43 patients (27%) with localized disease and 71 patients (68%) with metastatic disease exhibited clinical or radiological evidence of regional lymph node involvement. Regional lymph node sampling was performed in 102 patients (64%) with localized disease and 37 patients (35%) with metastatic disease. The most common method was random lymph node sampling via incisional or excisional biopsy (75% in localized and 70% in metastatic patients), followed by sentinel node biopsy (17%) in localized cases and needle biopsy 5% in metastatic patients. Pathological regional lymph node evaluation altered lymph node involvement status in 12 (12%) of the 102 patients with localized disease, with seven patients initially classified as positive by clinical or radiological assessment found to be negative on pathological evaluation, and five patients initially classified as negative clinically or on imaging found to have positive regional lymph nodes on pathology. As a result, regional lymph node involvement was confirmed in 39 (24%) patients with localized disease. In the localized cohort, regional lymph node involvement was more prevalent in patients with ARMS than localized ERMS (35% vs. 6%). Similarly, in patients with metastatic disease, pathological regional lymph node evaluation altered lymph node involvement status in four (11%) of 37 patients undergoing lymph node sampling, with two reclassified from positive on clinical or radiological assessment to negative on pathological evaluation and two cases reclassified from negative to positive. In total, pathological lymph node evaluation changed the lymph node involvement status in 12% (16 of 139) of patients with extremity RMS. Among patients with metastatic disease, bone marrow (54%) was the most common site of metastases, followed by bone (46%) and lungs (30%).

Treatment

Among patients with localized disease, 17 had LR‐RMS and were treated on two LR trials, D9602 and ARST0331, and 142 had IR‐RMS and were treated on D9803 (N = 85) and ARST 0531 (N = 57) trials. Among patients with localized disease, 80 (50%) underwent primary tumor resection at diagnosis, with 60 (75%) patients achieving gross total resection: 25 (16%) group 1 and 35 (22%) group 2. Primary tumor resection included amputation in three patients at diagnosis. Among 99 patients with group 3 disease, 51 (52%) underwent DPE, and among those undergoing DPE, 40 (78%) achieved R0 (microscopically negative surgical margins) resection, five (10%) had R1 (microscopically positive surgical margins) surgical resection (5%), three (6%) had R2 (surgically macroscopic residual disease), and for two patients (4%), it was unclear whether the resection was R1 or R2. The DPE procedure involved amputation of the affected limb in four patients. The median time to DPE in group 3 patients was 13.6 weeks from enrollment.

Among patients with metastatic disease, 12 (11%) underwent upfront primary tumor resection, with seven (58%) achieving gross total resection. Additionally, nine (8%) patients underwent DPE, with all but one achieving R0/R1 margins. Overall, 121 (76%) patients with localized disease received RT and 81 (77%) of those with metastatic disease had RT. Among 43 patients with localized disease with regional lymph node involvement, 77% (N = 23) received both primary tumor and regional lymph node radiation, and among those with metastatic disease, 47 (58%) received both primary and metastatic site radiation.

Survival outcomes

Localized disease

Of 159 patients with localized disease, 52 (33%) experienced relapse/progression of the disease, with 22 (13.8%) having local or loco‐regional, 26 (16.3%) having metastatic, and three (1.9%) having both local and metastatic relapse/progression. Five patients developed SMN. The 5‐year EFS of patients with localized disease was 62.6% (95% CI, 54%–71.2%), and the 5‐year OS was 78.7% (95% CI, 71.4%–85.9%) (Figures S1A,B). Table 1 presents the univariable analysis exploring the association of demographics and disease factors associated with EFS and OS. In the multivariable analysis including factors such as age, sex, IRS group, fusion status, and regional lymph node status associated with EFS on univariable analysis (p < .1), only age demonstrated independent statistically significant association with EFS with patients ≥1 and <10 years old having lower hazards of event than those ≥10 years old (HR, 0.54; 95% CI, 0.29–0.99; p = .046).

TABLE 1.

Univariable analysis of factors associated with 5‐year EFS and OS of patients with localized extremity RMS (N = 159).

Variable	No. of patients	5‐year EFS	95% CI	p	5‐year OS	95% CI	p
Age (years)
<1	9	55.6%	(13.6%–97.5%)	.04	66.7%	(29%–100.0%)	.001
≥1 to <10	99	70.1%	(60.1%–80.1%)		88.6%	(81.6%–95.5%)
≥10	51	47.3%	(31.0%–63.7%)		59.4%	(43.2%–75.6%)
Sex
Male	87	55.9%	(43.6%–68.2%)	.04	73.9%	(63.2%–84.6%)	.04
Female	72	70.2%	(58.4%–81.9%)		84.2%	(74.8%–93.5%)
Race
Non‐White	23	66.7%	(42.4%–90.1%)	.78	84.5%	(67.1%–100.0%)	.88
White	122	61.3%	(51.7%–71%)		77.9%	(69.7%–86.1%)
Unknown	14	68.1%	(37.3%–98.9%)		75.0%	(45%–100.0%)
Primary site
Lower limb	78	66.3%	(54.2%–78.4%)	.60	85.7%	(76.9%–94.5%)	.06
Upper limb	81	59.3%	(47.1%–71.4%)		72.5%	(61.4%–83.6%)
T status
T1	118	65.5%	(55.7%–75.2%)	.30	82.2%	(74.4%–90%)	.06
T2	41	54.31%	(36.3%–72.3%)		68.5%	(52.1%–85%)
Tumor size
≤5 cm	75	66.7%	(54.3%–79.1%)	.13	85.9%	(76.8%–95.1%)	.01
>5 cm	84	58.8%	(46.9%–70.6%)		72.0%	(61.3%, 82.7%)
IRS group
1	25	78.2%	(60.3%–96.1%)	.08	90.9%	(78.6%–100.0%)	.02
2	35	66.6%	(48.3%–84.9%)		84.6%	(70.4%–98.8%)
3	99	57.6%	(46.3%–68.8%)		73.7%	(63.9%–83.5%)
IRS stage
II	59	66.0%	(52.2%–79.8%)	.23	89.1%	(80.0%–98.2%)	.005
III	100	60.4%	(49.4%–71.4%)		72.5%	(62.6%–82.4%)
Regional lymph node involvement a
Yes	39	50.5%	(31.8%–69.1%)	.09	65%	(47.2%–87.7%)	.70
No	119	66.3%	(56.7%–76%)		83%	(75.4%–90.6%)
Histology b
ARMS	106	58.9%	(47.8%–70.1%)	.34	74.4%	(64.5%–84.2%)	.23
ERMS	35	64.7%	(47.4%–82%)		88.2%	(76.8%–99.6%)
Spindle cell	12	83.3%	(61.1%–100.0%)		91.7%	(75.3%–100.0%)
Anaplasia
No	122	61.0%	(50.9%–71.1%)	.33	76.1%	(67.3%–84.9%)	.31
Yes	31	74.2%	(58.1%–90.3%)		90.3%	(79.5%–100.0%)
Fusion status c
Fusion status unknown	17	73.3%	(48.6%–98.1%)	.093	92.9%	(78.8%–100.0%)	.028
Negative	61	69.8%	(57.2%–82.3%)		84.8%	(75.1%–94.5%)
PAX3‐ FOXO1	43	50.5%	(32.5%–68.5%)		62.6%	(44.9%–80.3%)
PAX7‐FOXOI	29	68.2%	(49.4%–87.1%)		89.5%	(77.1%–100.0%)
Fusion positive but fusion partner unknown	9	33.3%	(2.5%–64.1%)		50.8%	(0%–100.0%)
Low‐risk trials
ARST0331	7	71.4%	(38%–100.0%)	.90	85.7%	(59.8%–100.0%)	.54
D9602	10	70.0%	(41.6%–98.4%)		90.0%	(71.4%–100.0%)
Intermediate‐risk trials
ARST0531	57	51.0%	(32.3%–69.7%)	.023	64.4%	(47.1%–81.7%)	.007
D9803	85	68.9%	(58.2%–79.5%)		86.2%	(78.3%–94.1%)
Radiation
Yes	121	62.0%	(52%–72.0%)	.87	78.9%	(70.6%–87.3%)	.55
No	38	64.2%	(47.4%–81.1%)		77.8%	(63.4%–92.2%)

Abbreviations: ARMS, alveolar RMS; CI, confidence interval; EFS, event‐free survival; ERMS, embryonal RMS; IRS, Intergroup Rhabdomyosarcoma Study; NOS, not otherwise specified; OS, overall survival; RMS, rhabdomyosarcoma.

^a Lymph node involvement defined as “yes” if lymph nodes were either clinically and pathologically involved, clinically involved but pathological status was unknown, or clinically not involved but pathologically involved on biopsy. Lymph node status was unknown for one patient.

^b Histology was RMS NOS in two, and mixed RMS mixed in three; these patients not included in this analysis due to small numbers.

^c Fusion status imputed as negative for those with ERMS and spindle cell RMS.

Among the baseline characteristics, age at diagnosis, sex, tumor size, IRS group, fusion status, stage, and regional lymph node status were statistically significantly associated with OS in univariable analysis. However, only sex and age at diagnosis had an independent statistically significant association with OS in multivariable analysis, with females having improved OS than males (HR, 0.45; 95% CI, 0.22–0.95; p = .04) and patients ≥1 and <10 years old having lower hazards of mortality than those ≥10 years old (HR, 0.40; 95% CI, 0.19–0.84; p = .02) (Table 2).

TABLE 2.

Multivariable analysis of factors associated with 5‐year EFS and OS of patients with localized extremity RMS (N = 158). ^a

Covariate	5‐year EFS			5‐year OS
	HR	95% CI	p	HR	95% CI	p
Age (years)
<1	0.88	0.28–2.82	.83	1.0	0.24–4.19	>.99
≥1 to <10	0.54	0.29–0.99	.046	0.4	0.19–0.84	.02
≥10 (Ref)	—	—	—	—	—	—
Sex
Female	0.63	0.36–1.10	.11	0.45	0.22–0.95	.04
Male (Ref)		—	—	—	—	—
IRS group
1	0.48	0.19–1.23	.12	0.54	0.15–1.97	.35
2	0.59	0.3–1.16	.13	0.49	0.19–1.27	.14
3 (Ref)	—	—	—	—	—	—
Fusion status b
PAX3	1.22	0.6–2.51	.57	1.35	0.56–3.28	.51
PAX7	0.95	0.42–2.17	.96	0.58	0.20–1.75	.34
Fusion‐positive but fusion partner unknown	2.10	0.74–5.79	.15	1.97	0.55–7.06	.30
Unknown	0.61	0.2–1.84	.38	0.62	0.16–2.37	.49
Negative (Ref)	—	—	—	—	—	—
Regional lymph node involvement c
No	0.93	0.49–1.74	.82	0.84	0.35–2.03	.70
Yes (Ref)	—	—	—	—	—	—
Primary tumor size
≤5 cm	NA	NA	NA	0.58	0.19–1.84	.36
>5 cm (Ref)	NA	NA	NA	—	—	—
Stage
II	NA	NA	NA	0.73	0.18–2.96	.66
III (Ref)	NA	NA	NA	—	—	—
Tumor site
Lower limb	NA	NA	NA	0.71	0.33–1.54	.39
Upper limb (Ref)	NA	NA	NA	—	—	—
Tumor invasiveness
T2	NA	NA	NA	0.86	0.38–1.94	.71
T1 (Ref)	NA	NA	NA	—	—	—

Abbreviations: CI, confidence interval; EFS, event‐free survival; HR, hazard ratio; IRS, Intergroup Rhabdomyosarcoma Study; NA, not applicable; OS, overall survival; RMS, rhabdomyosarcoma.

^a One patient with not available regional lymph node status was not included.

^b Fusion status imputed as negative for those with ERMS and spind560le cell RMS.

^c Lymph node involvement defined as “yes” if lymph nodes were either clinically and pathologically involved, clinically involved but pathological status was unknown, or clinically not involved but pathologically involved on biopsy. Lymph node status was unknown for one patient.

Metastatic disease

Of 105 patients with metastatic disease, 88 (84%) experienced relapse/progression of the disease, with 59 (56.2%) having a metastatic site, eight (7.6%) having local or loco‐regional, and 19 (18.1%) having both local and metastatic relapse/progression. The 5‐year EFS of patients with metastatic extremity RMS was 7.7% (95% CI, 2.2%–13.2%), and the 5‐year OS was 22.7% (95% CI, 3.9%–31.4%) (Figures S2A,B). In the univariable analysis, the Oberlin score was statistically significantly associated with EFS and OS (p value <.05) (Table 3; Figure 1). Because the Oberlin score is a composite measure that incorporates age at diagnosis, and these variables exhibit strong collinearity, multivariable analysis for EFS was not conducted for metastatic patients. Multivariable analysis of factors associated with OS among patients with metastatic extremity RMS demonstrated that the Oberlin score is independently associated with OS and that females have a lower hazard of mortality compared to males after adjusting for the Oberlin score (HR, 0.62; 95% CI, 0.40–0.97; p value = .04) (Table S3).

TABLE 3.

Univariable analysis of factors associated with 5‐year EFS and OS of extremity RMS with metastatic disease (N = 105).

Variable	No. of patients	5‐year EFS	95% CI	p	5‐year OS	95% CI	p
Age, years
<10	37	16.5%	(3.3%–29.7%)	.09	31.1%	(15.1%–47.0%)	.19
≥10	68	3.3%	(0%–7.8%)		18.2%	(8.0%–28.4%)
Sex
Male	47	7.4%	(0%–15.5%)	.11	14.6%	(2.8%–26.5%)	.04
Female	58	8.0%	(0.5%–15.4%)		29.2%	(16.8%–41.6%)
Race
White	74	6.6%	(0.3%–12.9%)	.97	23.6%	(12.8%–34.5%)	.62
Non‐White	23	9.2%	(0%–21.4%)		18.2%	(2.1%–34.4%)
Unknown	8	12.5%	(0%–35.4%)		25.0%	(0%–55.0%)
Site of the primary
Lower limb	65	10.9%	(2.7%–19.2%)	.72	25.3%	(14.2%–36.4%)	.48
Upper limb	40	2.8%	(0%–8.4%)		18.1%	(3.7%–32.5%)
Tumor site
Hand and feet	35	2.9%	(0.0%–8.4%)	.82	15.9%	(1.6%–30.2%)	.62
Other sites	70	10.7%	(2.6%–18.8%)		26.2%	(15.2%–37.3%)
Tumor size
≤5	36	5.6%	(0%–13.0%)	.84	23.2%	(8.1%–38.3%)	.93
>5	69	9.0%	(1.5%–16.6%)		22.4%	(11.7%–33.2%)
Oberlin score
1	13	34.6%	(7.2%–62.0%)	.005	51.3%	(22.6%–79.9%)	.007
2	25	8.9%	(0%–20.6%)		40.4%	(18.8%–62.0%)
3	33	4.2%	(0%–12.1%)		13.6%	(0%–27.9%)
4	34	0%	(.–.)		9.4%	(0%–19.5%)
Study
ARST0431	19	5.6%	(0%–16.4%)	.55	27.8%	(7.1%–48.5%)	.41
ARST08P1	47	2.8%	(0%–8.2%)		23.7%	(9.4%–38.1%)
D9802	37	11.3%	(0.9%–21.8%)		16.8%	(4.5%–29.1%)
D9803	2	50.0%	(0%–100.0%)		50.0%	(0%–100.0%)
Radiation
Yes	81	6.7%	(1.0%–12.3%)	.18	23%	(13.1%–32.9%)	.32
No	24	12.5%	(0.00%–28.7%)		22.9%	(3.2%–42.6%)

Abbreviations: CI, confidence interval; EFS, event‐free survival; OS, overall survival; RMS, rhabdomyosarcoma.

FIGURE 1.

(A) Five‐year EFS of the extremity RMS patients with metastatic disease by Oberlin score. (B) Five‐year OS of the extremity RMS patients with metastatic disease by Oberlin score. EFS indicates event‐free survival; OS, overall survival; RMS, rhabdomyosarcoma.

Survival outcomes by the clinical trial type

No statistical differences were observed in the survival outcomes of the LR extremity RMS patients treated on the D9602 and ARST0331 trials (5‐year EFS: 70.0% [95% CI, 41.6%–98.4%] vs. 71.4% [38%–100.0%]; and 5‐year OS: 90% [95% CI, 71.4%–100.0%] vs. 85.7% [95% CI, 59.8%–100.0%] log‐rank test: p = .90 and .54, respectively) (Figures S3A,B). Of 142 localized extremity IR‐RMS patients treated on IR‐RMS studies D9803 and ARST0531, the relapse/progression rate was 24.7% (n = 21) in D9803 (loco‐regional: 47.6%; distant: 52.4%) and 45.6% (n = 26) in ARST0531 (loco‐regional: 32%; distant: 68%), respectively. The survival outcomes of the IR‐localized extremity RMS patients treated on IR‐RMS D9803 and ARST0531 trials differed significantly with patients treated on the D9803 trial exhibiting superior 5‐year EFS and OS compared to those treated on ARST0531 (5‐year EFS: 68.9% [95% CI, 58.2%–79.5%] vs. 51%; 5‐year OS: 86.2% [95% CI, 78.3%–94.1%] vs. 64.4% [95% CI, 47.1%–81.7%], log‐rank test: p = .023 and .007, respectively) (Figure 2A,B). On adjustment of patient‐ and disease‐factors that differed significantly between the D9803 and ARST0531 studies, including histology, fusion status, stage, and regional lymph node involvement (Table S4), the EFS was not statistically significantly different between the two trials (adjusted hazard ratio [adjHR], 0.58 [95% CI, 0.32–1.03]; p = .06) but OS between the two trials was statistically different (Figure 2C,D) (adjHR, 0.43 [95% 0.21–0.86]; p = .018). In addition to the factors mentioned above, when EFS and OS were adjusted for differences in the receipt of DPE and radiation between D9803 and ARST0531, the EFS of IR‐RMS patients was not statistically different between the two trials (p = .16), but the OS remained statistically superior in the D9803 trial compared to ARST0531 (adjHR, 0.37 [95% CI, 0.17–0.81]; p = .013) (Figure 2C,D). The EFS and OS of the metastatic extremity RMS patients did not differ significantly by the type of trial (Figures S4A,B).

FIGURE 2.

(A and B) Unadjusted 5‐year EFS and OS of the intermediate‐risk extremity RMS patients treated on D9803 and ARST0531. (C and D) Adjusted 5‐year EFS and OS of the intermediate‐risk extremity RMS patients treated on D9803 and ARST0531 (adjusted for fusion status, histology, stage, and regional lymph node involvement). (E and F) Adjusted 5‐year EFS and OS of the intermediate‐risk RMS patients treated on D9803 and ARST0531 (adjusted for fusion status, histology, stage, regional lymph node involvement, DPE and radiation). EFS indicates event‐free survival; OS, overall survival; RMS, rhabdomyosarcoma.

Survival of group 3 patients by DPE status

The 5‐year EFS of group 3 patients undergoing DPE with R0 margins versus those without DPE versus those with DPE with positive margins was statistically different (71.2% vs. 49.3% vs. 40%, p = .048) (Figure 3A). After adjusting for tumor invasiveness, tumor site, regional lymph node involvement, radiation, and trial type that differed significantly among the patients undergoing DPE versus no DPE (Table S5), there was no statistically significant difference in EFS among the group 3 extremity RMS patients undergoing DPE with R0 resection compared to those not undergoing DPE (adjHR, 0.46 [95% CI, 0.20–1.03]; p = .06). However, patients who underwent DPE with R0 margins demonstrated significantly better EFS compared to those with DPE with positive margins (adjHR, 0.31 [95% CI, 0.11–0.87]; p = .027) (Figure 3C). Both the adjusted and unadjusted 5‐year OS were not statistically significantly different between these three subgroups, with 5‐year unadjusted OS being DPE with R0 margins: 86.9% (95% CI, 75.0%–98.8%) versus no DPE: 66.1% (95% CI, 50.0%–82.2%) versus DPE with positive margins: 60% (95% CI, 29.6%–90.4%), p = .20 (Figure 3B,D). Additionally, the survival outcomes of the group 3 patients undergoing DPE with R0 margins were comparable to group 2 extremity RMS patients with 5‐year EFS of 71.2% (95% CI, 54.9%–87.5%) versus 66.6% (95% CI, 48.3%–84.9%)] and 5‐year OS of 86.9% (95% CI, 75.0%–98.8%) versus 84.6% (95% CI, 70.4%–98.8%) (Figure 3E,F). The survival outcomes of group 3 with DPE R0 were also not statistically different from group 1 RMS patients with 5‐year EFS of 71.2% (95% CI, 54.9%–87.5%) versus 78.2% (95% CI, 60.3%–96.1%) and 5‐year OS of 86.9% (95% CI, 75.0%–98.8%) versus 90.9% (95% CI, 78.6%–100%) (Figure 3G,H).

FIGURE 3.

(A and B) Five‐year EFS and OS of group 3 extremity RMS patients by DPE status. (C and D) Adjusted 5‐year EFS and OS of the group 3 extremity RMS patients by DPE status (adjusted for tumor invasiveness, tumor site, regional lymph node involvement, and trial type and radiation). (E and F) Five‐year EFS and OS of group 3 extremity RMS patients undergoing R0 DPE and group 2 extremity RMS patients. (G and H) Five‐year EFS and OS of group 3 extremity RMS patients undergoing R0 DPE and group 2 extremity RMS patients. DPE indicates delayed primary excision; EFS, event‐free survival; OS, overall survival; RMS, rhabdomyosarcoma.

DISCUSSION

This study of a large cohort of extremity RMS patients treated over 16 years on COG trials demonstrated several key findings. Among patients with localized disease, age at diagnosis was associated with EFS and OS and sex was associated with OS. Patients with IR‐RMS treated on the D9803 trial had superior OS. Additionally, patients with group 3 disease who underwent DPE achieving R0 margins exhibited survival outcomes comparable to those with group 2 disease. Outcomes for patients with metastatic disease remain poor, and an Oberlin score of 1 and female sex were associated with better OS in this subgroup.

The 5‐year EFS rate of 63% and OS rate of 79% among patients with localized disease are comparable to the 5‐year EFS and OS rates of 59% and 72%, respectively, reported for 162 patients with localized extremity RMS treated on the EpSSG‐RMS2005 study. ⁵ These outcomes, however, compare favorably with the findings of Oberlin et al., ¹ who reported a 5‐year OS of 67% in an international cohort of 643 patients with localized extremity RMS treated between 1983 and 2004. Within the limitations of historical comparisons, the 5‐year OS for localized group 3 patients in our study appears superior to that of the entire Oberlin’s cohort (74% vs. 59%) and COG subcohort of Oberlin’s cohort (74% vs. 60%), as is the 5‐year EFS compared to the 3‐year failure‐free survival for group 3 patients treated in the IRS‐IV trial (58% vs. 50%). ¹⁰ The observed improvement in survival of localized patients may be attributed to advancements in imaging techniques, improvements in regional lymph node evaluation and increased utilization of local treatment modalities such as surgery and radiation. ²⁰ , ²¹

Consistent with previous studies, patients between 1 and 10 years had favorable EFS and OS than older patients in our study. ⁵ , ²² The observation that females demonstrated better OS among localized and metastatic patients, even after adjusting for confounding variables, is noteworthy and may be due to unmeasured confounders, including the absence of fusion status data for all patients. This finding warrants further investigation in future trials and pooled analyses of patients with extremity RMS using the international RMS data set through the International Soft Tissue Sarcoma Consortium (INSTRuCT). ²³ Notably, Oberlin et al. ¹ identified T status, regional lymph node involvement, surgery, cooperative group, and treatment era as independent prognostic factors associated with OS in patients with localized extremity RMS. The variation in identified prognostic factors between our analysis and that of Oberlin et al. ¹ may be attributed to our focus on determining the association of baseline variables with OS rather than treatment‐related factors, a relatively smaller sample size of our cohort, or the possible mitigation of adverse baseline prognostic factors with the current risk‐stratified treatment.

The prognosis for patients with metastatic extremity RMS remains poor, with a 5‐year EFS of 8% and a 5‐year OS of 23%. These outcomes are comparable to those recently reported for 67 metastatic extremity RMS patients treated according to the Cooperative Weichteilsarkom Studiengruppe protocol, with a 5‐year EFS of 14% and OS of 23%. ²⁴ As demonstrated in prior studies, the Oberlin score emerged as an independent predictor of EFS and OS in patients with metastatic RMS, with an Oberlin score of 1 associated with improved survival. ¹⁹ , ²⁴ Notably, female patients exhibited better OS than males, even among those with metastatic disease, which warrants further investigation in future studies.

When evaluating the outcomes of IR localized extremity RMS patients by trial type, the 5‐year OS was significantly better for those treated in the D9803 trial than the ARST0531 trial, even after adjusting for key covariates that differed significantly between the trials. This superiority in OS persisted after further adjustments for DPE and radiation therapy, suggesting that the higher dose of cyclophosphamide used in the D9803 trial may have contributed to the improved overall survival. ¹² Although the retrospective nature of the comparison limits this conclusion, it underscores the importance of future trials to determine the optimal cyclophosphamide dose needed to achieve the best patient outcomes.

An important observation from our analysis is that achieving either a group 1 or 2 upfront resection or an R0 or R1 DPE allowed for a reduced radiation dose in 105 of 159 (66%) patients with localized disease. Notably, among group 3 RMS patients, those who underwent DPE and achieved negative surgical margins had superior unadjusted EFS compared to those who did not undergo DPE, despite receiving a lower radiation dose (50.4 Gy vs. 36 Gy). Moreover, the EFS and OS of group 3 patients undergoing DPE with R0 margins were comparable to those of group 2 patients. Although the survival outcomes of the group 3 patients undergoing DPE with R0 margins are not statistically superior to those of group 1 patients—this analysis is limited by a small sample size. A recent analysis from EpSSG RMS2005 suggests that omitting radiotherapy for local control does not compromise OS in patients with localized group 2/3 ERMS at favorable sites, except for those with orbital tumors. Furthermore, among patients undergoing DPE, the omission of RT did not impact EFS and OS, underscoring the potential importance of DPE in improving survival. ²⁵ Similarly, findings from the SIOP MMT89 and MMT95 trials suggest that primary tumor surgery may contribute to better OS. In MMT95, surgery was performed more frequently (90%) compared to MMT89 (60%), whereas RT use remained similar in both studies (30%). Notably, MMT95 demonstrated improved OS, indicating that surgical intervention may be important in improving survival. ⁵ , ²⁶ , ²⁷ Additionally, studies suggest that the likelihood of radiation‐related complications, including secondary cancers, increases with the radiation dose. ²⁸ , ²⁹ , ³⁰ As a result, reducing RT doses—such as decreasing from 50.4 Gy to 36 Gy could mitigate some of the long‐term toxicities associated with RT in this population. ³¹ However, surgical interventions may still result in notable short or long‐term morbidity, ³² especially when critical organs are involved or removed, as exemplified by the four patients in our cohort who underwent amputation as part of the DPE. It is crucial for future trials to prospectively collect data on the feasibility of upfront and DPE based on multidisciplinary discussion and to document the reasons for pursuing or not pursuing upfront or DPE in patients. Such information can provide valuable insights into the patient selection criteria for upfront or DPE, physician decision‐making process, potential benefits or risks that influence this choice, and short‐ and long‐term patient outcomes, as well as reduce the biases in understating the impact of surgery on survival outcomes. Our data also emphasize that in cases where DPE is expected to result in positive margins, it should be avoided as it does not improve outcomes.

Another notable finding from our study is that regional lymph node sampling altered nodal status in 12% of patients with localized or metastatic extremity RMS. This is consistent with prior reports, which have shown regional lymph node sampling either upstaging or downstaging the nodal status in 16%–17% of patients with extremity RMS. ⁵ , ³³ Regional lymph node sampling was required in the included COG trials for patients with extremity RMS, and our data suggest that omission of this step could have resulted in under‐ or overtreatment, potentially compromising patient outcomes. These findings reinforce the importance of pathological regional lymph node evaluation in all patients with extremity RMS, regardless of clinical or radiographic evidence of nodal involvement. ³⁴

Several limitations of this analysis warrant consideration. First, patients were treated over 16 years across sequential trials, employing different approaches to chemotherapy and DPE. Notably, there were significant differences in the cumulative dose and intensity of cyclophosphamide and the implementation of DPE across the IR RMS trials. Moreover, the absence of prospective data on factors influencing the decision to pursue or avoid DPE introduces potential bias in the analysis. It is conceivable that patients with smaller tumors were more likely to undergo DPE than those with larger tumors, although adjusted analyses may mitigate some of these differences. It remains impossible to fully account for unknown or unmeasured variables that may have affected patient survival outcomes despite adjusting for known confounders.

In conclusion, this study of a large cohort of patients with extremity RMS highlights the suboptimal outcomes of patients with metastatic disease, underscoring the urgent need for treatments to improve their outcomes. Given the favorable outcomes of IR‐localized extremity RMS patients treated on D9803 and those undergoing DPE, future trials should prospectively evaluate the impact of cyclophosphamide dose and intensity and the role of DPE in improving the survival of these patients.

AUTHOR CONTRIBUTIONS

Sapna Oberoi: Conceptualization, formal analysis, writing—original draft, methodology, and investigation. Wei Xue: Data curation, methodology, formal analysis, and writing—review and editing. Amira Qumseya: Data curation, formal analysis, and writing—review and editing. Thomas Scharschmidt: Conceptualization, methodology, investigation, and writing—review and editing. Odion Binitie: Conceptualization, investigation, methodology, and writing—review and editing. Joel I. Sorger: Conceptualization, methodology, investigation, and writing—review and editing. Kiran A. Kumar: Conceptualization, methodology, investigation, and writing—review and editing. Kenneth Wong: Conceptualization, methodology, writing—review and editing, and investigation. Sarah S. Donaldson: Conceptualization, methodology, investigation, and writing—review and editing. Lisa Teot: Conceptualization, methodology, investigation, and writing—review and editing. Erin R. Rudzinski: Conceptualization, methodology, and investigation, writing—review and editing. Rajkumar Venkatramani: Conceptualization, formal analysis, supervision, writing—review and editing, and methodology.

CONFLICT OF INTEREST STATEMENT

Odion Binitie reports fees for professional activities from the American Academy of Orthopaedic Surgeons; consulting fees from Onkos Surgical, Inc; and participation as a fiduciary officer with Childrens Oncology Group, the Florida Orthopedic Society, the Musculoskeletal Tumor Society, and the National Comprehensive Cancer Network. Amira Qumseya reports fees for participation on a data and safety monitoring board for the University of Florida. The other authors declare no conflicts of interest.

Supporting information

Supplementary Material

Oberoi S, Xue W, Qumseya A, et al. Extremity rhabdomyosarcoma in children, adolescents and young adults: a report from Children's Oncology Group trials. Cancer. 2025;e35929. doi: 10.1002/cncr.35929

DATA AVAILABILITY STATEMENT

The Children’s Oncology Group Data Sharing policy describes the release and use of COG individual subject data for use in research projects in accordance with National Clinical Trials Network (NCTN) Program and NCI Community Oncology Research Program (NCORP) Guidelines. Only data expressly released from the oversight of the relevant COG Data and Safety Monitoring Committee are available to be shared. Data sharing will ordinarily be considered only after the primary study manuscript is accepted for publication. For phase 3 studies, individual‐level de‐identified data sets that would be sufficient to reproduce results provided in a publication containing the primary study analysis can be requested from the NCTN/NCORP Data Archive at https://nctn‐data‐archive.nci.nih.gov/. Data are available to researchers who wish to analyze the data in secondary studies to enhance the public health benefit of the original work and agree to the terms and conditions of use. For nonphase 3 studies, data are available following the primary publication. An individual‐level de‐identified data set containing the variables analyzed in the primary results article can be expected to be available on request. Requests for access to COG protocol research data should be sent to: datarequest@childrensoncologygroup.org. Data are available to researchers whose proposed analysis is found by COG to be feasible and of scientific merit and who agree to the terms and conditions of use. For all requests, no other study documents, including the protocol, will be made available and no end date exists for requests. In addition to above, release of data collected in a clinical trial conducted under a binding collaborative agreement between COG or the NCI Cancer Therapy Evaluation Program and a pharmaceutical/biotechnology company must comply with the data sharing terms of the binding collaborative/contractual agreement and must receive the proper approvals.