Prognosis of children and young adults with newly diagnosed rhabdomyosarcoma metastatic to bone marrow treated on Children Oncology Group Studies

Abstract

Background

Rhabdomyosarcoma (RMS) is the most common soft tissue sarcoma in children. Metastatic disease occurs in 16% of all RMS cases and has a poor prognosis. There are limited studies examining the outcomes specific to patients with RMS metastatic to bone marrow despite an incidence of 6% at diagnosis. Our study aims to document the outcomes, prognostic factors and clinical courses of children presenting with RMS metastatic to bone marrow treated on Children’s Oncology Group (COG) cooperative trials.

Methods

We performed a retrospective analysis of the patients diagnosed with RMS metastatic to bone marrow between 1997 and 2013 enrolled on one of four COG RMS clinical trials of D9802, D9803, ARST0431 and ARST08P1.

Results

We identified 179 cases with RMS metastatic to bone marrow. Patients had a median age of 14.8 years, 58% were male, predominantly alveolar histology (76%), extremity was the most common primary site (32%), and 87% had metastatic disease to additional sites. 83% (n=149) received radiation as a treatment modality. The 3- and 5-year Event-Free survival was 9.4% and 8.2% respectively. The 3- and 5-year Overall Survival was 26.1% and 12.6% respectively. Age ≥ 10 years, alveolar histology, FOXO1 fusion presence, unfavorable primary location, higher Oberlin score, and lack of radiation were identified as poor prognostic/predictive characteristics.

Conclusions

This study represents the largest analysis of RMS metastatic to bone marrow defining the poor prognostic outcome for these patients. These patients may be eligible for therapy deintensification or early pursuit of novel treatments/approaches which are desperately needed.

INTRODUCTION

Rhabdomyosarcoma (RMS) is the most common soft tissue sarcoma of childhood. It is a malignancy of mesenchymal origin that can present with different histologies and driver mutations at different ages, primary locations, and patterns of metastatic spread. This variable presentation leads to complex risk stratification and treatment algorithms with diverse outcomes and responses to therapy. Localized disease outcomes have improved with the selective combination of multi-agent chemotherapy, surgery and radiation. Localized RMS has a 3-year overall survival (OS) rates of approximately 85%.¹ Unfortunately, metastatic disease comprises 16% of all RMS² and has a comparatively poor prognosis with a 3-year OS of only 34–56%.^3,4 Despite attempts at therapy intensification, metastatic RMS has demonstrated minimal outcome improvements.^1,3–7 Patients with bone marrow metastases comprise 6% of all new RMS diagnoses² and have been identified as a particularly poor prognostic group.^3,8,9 Studies of metastatic RMS frequently do not subdivide outcomes by metastatic site involvement and combine bone or bone marrow involvement as a single category.^2,4,10–16 While some studies have found the presence of bone marrow metastases and bone metastases highly correlated^3,17 others have demonstrated limited concomitant involvement at presentation.² Further, multiple studies have demonstrated that not all metastatic presentations have uniformly poor outcomes.^3,4,9,17,18 Specific subpopulations may be amenable to therapy intensification while other presentations with dismal prognosis with intensification may be candidates for therapy de-escalation to reduce toxicity and improve quality of life or early pursuit of experimental approaches. There are limited studies that have examined the outcomes specific to patients with RMS metastatic to bone marrow at presentation.

It is essential to determine associated factors and subpopulations affecting the survival and clinical course of patients with RMS metastatic to bone marrow. Insight into presentation patterns, response to therapy, and locations/timing of recurrence can offer guidance for staging, therapy and surveillance. We therefore examined the outcomes of children presenting with RMS metastatic to bone marrow who were treated on the Children’s Oncology Group (COG) trials of D9802, D9803, ARST0431 and ARST08P1.

METHODS

Patients and Included Trials:

Patients diagnosed with metastatic RMS between 1997 and 2013 enrolled on one of four COG clinical RMS studies (D9802, n = 111¹⁹, D9803, n = 617¹³, ARST0431, n = 109⁴, or ARST08P1, n = 168¹⁴)(Supplemental Table 1). Clinicopathologic variables including age, sex, race, histology subtype, primary site, tumor invasiveness, tumor size, regional lymph node involvement, and sites of metastatic disease were analyzed with the clinical outcome and treatment modality data. FOXO1 and fusion partner status was available for a subset of patients. Patients with non-alveolar histology that did not have documented fusion status assessment were analyzed with the FOXO1 fusion negative group when performing assessments based on fusion status. For this study the receipt of radiation was characterized as either receipt or no receipt of radiation not subdivided to locations, dose, modality, or pattern.

D9802 assessed the outcomes for patients with HR rhabdomyosarcoma treated with window therapies of Irinotecan alone or Irinotecan/Vincristine followed by alternating VAC/VI if they responded compared to standard VAC.¹⁹ D9803 compared the outcomes of patients with Intermediate Risk (IR) rhabdomyosarcoma treated with standard 42 weeks of VAC to VAC/VTC. No difference was noted in outcomes across risk groups.¹³ Patients receiving radiation to the primary and sites of metastatic disease began in week 15 for D9802 and week 12 for D9803. There was exception for parameningeal disease with intracranial extension or emergent radiation which was defined as initiation within the first 2 weeks of treatment. 50.4 Gy was delivered to the primary and metastatic sites in 28 fractions with multiple dose exceptions including orbital disease, negative margin resection, microscopic residual resection, or complete response of lymph node disease. ARST0431 administered 54 weeks of alternating cycles of VI, interval compressed VDC/IE, and VAC.⁴ ARST08P1 assessed the addition of either temozolomide or cixutumumab to the ARST0431 backbone with no improvements in outcomes noted.¹⁴ In both ARST0431 and ARST08P1 patients receiving radiation to the primary and sites of metastatic disease began in week 20. Additional metastatic sites not feasible to include at week 20 treatments were deferred and delivered at week 47. There was exception for parameningeal disease with intracranial extension or emergent radiation which was defined as initiation within the first 2 weeks of treatment. 50.4 Gy was delivered to the primary and metastatic sites in 28 fractions with multiple dose exceptions including orbital disease, negative margin resection, or microscopic residual resection. All four protocols allowed for radiation modalities of photon, proton, electron, brachytherapy and intensity modulated radiation therapy. Histologic diagnosis was determined by central pathology review and radiation therapy underwent centralized quality review as part of these trials. Bilateral bone marrow aspirate/biopsy for histologic assessment of bone marrow involvement was a required pre-study observation for all trials. Tumor site was categorized as favorable or unfavorable and Oberlin score were classified as per Oberlin et al., 2008.³ Unfavorable prognostic factors for Oberlin score include 1) Age < 1 or ≥ 10, 2) Unfavorable site of disease (Limb and other), 3) Bone or bone marrow involvement, and 4) Number of metastatic sites ≥ 3.

Statistical methods:

Event-free survival (EFS) was defined as time from study entry to disease recurrence, progression, or death (from any cause) as a first event. Overall survival (OS) was defined as time from study entry to death from any cause or censored at the time of the last follow-up. Follow-up is current as of December 31, 2018. A p value <0.05 was considered significant. The Kaplan-Meier method was used to estimate survival distributions and the Peto-Peto method was used to estimate the standard error of the Kaplan-Meier estimate. The proportional hazard model was utilized for multiple variable analysis including variables that were significant in univariate analysis and eliminating ones not significant with a p>0.05. Software program SAS 9.4 was used for the analysis.

RESULTS

Clinical Characteristics:

A total of 979 patients with a new diagnosis of RMS were enrolled on the included studies. There were 429 (43.8%) patients with metastatic disease and 179 (18.3%) with RMS metastatic to bone marrow at diagnosis (Table 1). 129 (72%) patients were ≥ 10 years of age, 47 (26%) age 1–9, and only 3 (2%) age less than 1 year for a median age of 14.8 years (0.02 to 26.3 years) (Table 1). Alveolar histology was present in 136 (76%) patients of which 71 (52%) had FOXO1 fusion status determined, 68 (96%) of which were positive. Of the 68 patients with a FOXO1 fusion there were 47 (69%) partnered with PAX3, 5 (7%) partnered with PAX7, and 16 (24%) with partner not determined/reported. The most common primary site was extremity with 57 (32%) patients, followed by retroperitoneal (14%) and trunk/paravertebral (10%). While 156 (87%) patients had metastatic disease to additional locations, only 100 (56%) patients with RMS metastatic to bone marrow had concurrent metastatic disease to bone. The 22 (12%) patients with RMS metastatic only to bone marrow had a median age of 10.66 years (0.8 – 20.6 years) were predominantly alveolar histology (n=16, 73%) and T2 tumor invasiveness (n = 20, 91%).

TABLE 1.

Baseline patient and tumor characteristics

	Bone Marrow Metastatic
Variable	Number	(%)
Total Number Patients	179	100
Age median, range	14.8 (0.02, 26.3)
Sex
Male	103	57.5
Female	76	42.5
Age at diagnosis, years
< 1	3	1.7
1 – 9	47	26.3
≥ 10	129	72.1
Race
Nonwhite	29	16.2
White	132	73.7
Unknown	18	10.1
Histology Subtype
Alveolar	136	76.0
Alveolar Fusion Status
Positive FOXO1 Fusion	68	50.0
PAX3-FOXO1	47	69.1
PAX7-FOXO1	5	7.4
Unknown Partner-FOXO1	16	23.5
Negative FOXO1 Fusion	3	2.2
Unknown Fusion Status	65	47.8
Embryonal	25	14.1
Spindle cell	1	0.6
Other	17	9.5
Primary Site
Extremity	57	31.8
Retroperitoneal	25	14.0
Trunk/paravertebral	18	10.1
Bladder/Prostate	15	8.4
Parameningeal	14	7.8
Perineum	13	7.3
GU non-Bladder/Prostate	12	6.7
Other	9	5.0
Head and Neck, Non-Parameningeal	7	3.9
Unknown	7	3.9
Intrathoracic	1	0.6
Orbit	1	0.6
Tumor Invasiveness
T1	11	6.2
T2	167	93.3
Unknown	1	0.6
Tumor Size, cm
≤ 5	51	28.5
>5	128	71.5
Regional Lymph Node
N0	51	28.5
N1	118	65.9
Nx	10	5.6
Number of Metastatic Sites
1 – Bone marrow Only	22	12.3
2	46	25.7
≥ 3	110	61.5
missing	1	0.6
Additional Metastatic Sites
Bone	100	55.9
Lung	70	39.1
Lymph node	81	45.3
Oberlin Score
1	10	5.6
2	34	19.0
3	65	36.3
4	63	35.2
Missing/Not calculable	7	3.9

Treatment and Outcomes:

The median duration of follow-up for patients who were alive was 7.97 years. One hundred forty-nine (149) patients (83%) received radiation as a treatment modality. 163 (91%) patients experienced an event with 150 (84%) relapse/progressions, eleven (6%) deaths, and two (1%) second malignancies as the first events. Relapse/progression was local/regional in 18 (12%) patients, metastatic in 93 (62%) patients, and combined local/regional and metastatic in 36 (20%) of patients. (Table 2). The 3- and 5-year Event-Free survival was 9.4% (4.9 – 13.9%) and 8.2% (3.9 – 12.4%) respectively. The 3- and 5-year Overall Survival was 26.1% (19.5 – 32.8%) and 12.6% (7.3 – 17.9%) respectively. (Figure 1A and 1B).

TABLE 2.

Outcomes of patients with rhabdomyosarcoma metastatic to bone marrow (n=179)

	Bone Marrow Metastatic
Variable	Number	(%)
First Event Type
No event	16	8.9
Relapse/progression	150	83.8
Second malignancy	2	1.1
Death	11	6.2
Relapse / Progression Site
Local/Regional	18	10.1
Metastatic	93	52.0
Concurrent Local/Regional + Metastatic	36	20.1
Not applicable	29	16.2
Unknown	3	1.7

FIGURE 1.

Event-free and overall survival of rhabdomyosarcoma metastatic to bone marrow

Event-free (A) and overall survival (B) of 179 patients with rhabdomyosarcoma metastatic to bone marrow at diagnosis.

Table 3 reports the univariate factors and their respective associated 3-year EFS and OS. Increased age (≥10 years) (p=0.0008), alveolar histology (p=0.0122), FOXO1 fusion presence (p=0.0367) (fusion positive alveolar histology compared to non-alveolar histology inclusive of fusion negative and fusion unknown), unfavorable primary site (p=0.0013), higher Oberlin score (p=0.0011), and lack of radiation (p<0.0001) were all associated with significantly inferior EFS (Table 3). Increased age (≥10 years) (p <0.0001), alveolar histology (p=0.0377), unfavorable primary site (p=0.0260), higher Oberlin Score (p=0.0012), and lack of radiation (p=0.0005) were also associated with significantly inferior OS (Table 3). Additionally, ≥ 3 metastatic sites also reached significance for association with inferior OS (p=0.0476) but not for EFS (p=0.0662). FOXO1 fusion positive alveolar histology vs non-alveolar histology (inclusive of fusion negative and fusion unknown) did not achieve significance for OS (p=0.0764). The presence of PAX7-FOXO1 did not achieve significance as an inferior prognostic marker vs PAX3-FOXO1 for both EFS (p=0.0638) and OS (p=0.0840). All four patients with FOXO1 fusion positive disease who demonstrated greater than 3 years of EFS received radiation, had Oberlin scores of 2 or 3, and had a PAX3-FOXO1 fusion. These patients did not have a consistent age, primary site, number/pattern of metastases, or treatment protocol. Independent EFS (Figure 2) and OS (Figure 3) curves for age (A), fusion status (B), primary site (C), number of metastatic sites (D), Oberlin score (E) and receipt of radiation (F) are displayed. Outcomes by enrolled protocol and histology are displayed in Supplemental Figure 1 and Supplemental Figure 2 respectively. Detailed secondary malignancy data was available the patients enrolled in ARST0431 and ARST08P1. Of the patients with bone marrow metastatic disease treated on these studies (n=121) there were 5 secondary malignancies (myelodysplastic syndrome = 1, acute myeloid leukemia = 1, thyroid carcinoma = 1, and secondary sarcomas = 2).

TABLE 3.

Univariate analysis of factors associated with EFS and OS

		Event-Free Survival			Overall Survival
Variable	Number at risk	3-year Event-Free Survival	95% Confidence Interval	Log rank test P-value	3-year Overall Survival	95% Confidence Interval	Log rank test P-value
Age as a categorical variable
< 1	3	66.7%	( 13.3%, 100.0% )	0.0008	66.7%	( 13.3%, 100.0%)	<0.0001
1 – 9	47	25.0%	( 11.6%, 38.4% )		43.7%	( 28.9%, 58.4% )
≥ 10	129	2.5%	( 0.0%, 5.3% )		18.8%	( 11.9%, 25.8% )
Race
Nonwhite	29	6.9%	( 0.0%, 16.1% )	0.2322	24.1%	( 8.6%, 39.7% )	0.1233
White	132	10.6%	( 4.9%, 16.2% )		28.6%	( 20.6%, 36.6% )
Unknown	18	5.6%	( 0.0%, 16.1% )		11.1%	( 0.00%, 25.63% )
Histology Subtype
Alveolar	136	4.8%	( 1.1%, 8.5% )	0.0122	23.9%	( 16.6%, 31.3% )	0.0337
Alveolar Fusion Status
Positive FOXO1 Fusion	68	4.5%	( 0.0%, 9.5% )	0.0367#	28.4%	( 17.6%, 39.2% )	0.0764#
PAX3-FOXO1	47	6.6%	( 0.0%, 13.7% )	0.1039	34.8%	( 21.0%, 48.6%)	0.1184
PAX7-FOXO1	5	0.0%	( ., . )		0.0%	( ., . )
Unknown Partner-FOXO1	16	0.0%	( ., . )		18.8%	(0 0%.,37.9%)
Negative FOXO1 Fusion	3	0.0%	( ., . )		0.0%	( ., . )
Unknown Fusion Status	65	5.4%	( 0.0%, 11.3% )		20.2%	( 10.0%, 30.4% )
Non-alveolar (Embryonal, Botryoid, Spindle, Other)	43	24.6%	( 10.6%, 38.5% )		33.4%	( 18.6%, 48.2% )
Primary Site
Favorable (Orbit, Head and Neck-non-PM, PM, GU B/P, GU non-B/P	49	25.1%	( 12.3%, 38.0% )	0.0013	35.6%	( 21.6%, 49.5% )	0.0260
Unfavorable (Extremity, other)	123	3.8%	( 0.2%, 7.4% )		24.0%	( 16.3%, 31.8% )
Unknown	7	0.0%	( ., . )		0.0%	( ., . )
T status
T1	11	9.1%	( 0.0%, 26.1% )	0.8266	18.2%	( 0.0%, 41.0% )	0.8935
T2	167	9.5%	( 4.7%, 14.2% )		26.83%	( 19.9%, 33.8% )
Size of the tumor
≤ 5 cm	51	8.0%	( 0.5%, 15.6% )	0.8420	24.01%	( 12.2%, 35.9% )	0.7630
>5 cm	128	10.0%	( 4.4%, 15.6% )		27.0%	( 19.0%, 35.0% )
Number of Metastatic Sites
≤ 2	68	15.4%	( 6.6%, 24.2% )	0.0662	32.8%	( 21.6%, 44.1% )	0.0476
≥ 3	110	5.5%	( 0.8%, 10.1% )		21.7%	( 13.7%, 29.8% )
Additional Metastatic Sites
Bone Marrow Only	22	9.4%	( 4.9%, 13.9% )		26.1%	( 19.5%, 32.8% )
Bone	100	6.4%	( 1.0%, 11.8% )	0.4526	20.7%	( 12.4%, 29.0% )	0.1204
Lung	70	7.2%	( 0.4%, 14.1% )	0.9109	26.1%	( 15.1%, 37.0% )	0.8038
Lymph node	81	6.8%	( 1.1%, 12.6% )	0.1969	21.3%	( 12.0%, 30.5% )	0.0814
Oberlin Score
1	10	40.0%	( 9.6%, 70.4% )	0.0011	50.0%	( 19.0%, 81.0% )	0.0012
2	34	26.1%	( 10.6%, 41.7% )		40.5%	( 23.5%, 57.5% )
3	65	6.0%	( 0.0%, 12.6% )		27.1%	( 15.8%, 38.5% )
4	63	0.0%	( ., . )		16.5%	( 7.2%, 25.9% )
Radiation
Yes	149	10.9%	( 5.7%, 16.1% )	<0.0001	28.9%	( 21.4%, 36.3% )	0.0005
No	30	0.0%	( ., . )		11.9%	( 0.0%, 24.5% )

^#Fusion Positive compared to non-alveolar histology (Inclusive of fusion negative and fusion unknown)

FIGURE 2.

Event Free Survival Univariate Factors

EFS of patients with bone marrow metastatic rhabdomyosarcoma at diagnosis by age (A), fusion status (B), primary site (C), number of metastatic sites (D), Oberlin score (E) and receipt of radiation (F).

FIGURE 3.

Overall Survival Univariate Factors

OS of patients with bone marrow metastatic rhabdomyosarcoma at diagnosis by age (A), fusion status (B), primary site (C), number of metastatic sites (D), Oberlin score (E) and receipt of radiation (F).

Multiple Variable Analysis:

Variables significant in univariate analysis were fit in the multivariable proportional hazard model. Proportional hazard assumption was assessed. Oberlin score was defined based on the prognostic factors and thus not included in the analysis. In multivariable analysis patients age (≥ 10 years), primary sites (unfavorable vs favorable), and radiation therapy (yes vs no) were independently significantly associated with EFS after backward elimination. (Supplemental Table 2A). Patients age (≥ 10 years) and radiation therapy (yes vs no) were independently significantly associated with OS after backward elimination. (Supplemental Table 2B).

DISCUSSION

In this study, we describe the clinical characteristics and outcomes of 179 patients with RMS metastatic to bone marrow at diagnosis treated on the COG clinical trials between 1997 and 2013. There are limited studies examining the outcomes specific to patients with RMS metastatic to bone marrow. Reported outcomes frequently have short follow-up, are single institution or are combined with RMS metastatic to bone.^{3,8,9,20–22} However, these studies raise important questions including the impact of the FOXO1 fusion partner, histology, and response to treatment for the outcomes on this patient population.^{2,3,11,15,17,18,21–25}

Our analysis demonstrates patients with RMS metastatic to bone marrow at diagnosis had predominantly alveolar histology, were older than age 10, with primary tumors that were > 5cm, T2 invasive, and at an extremity location. Frequently these patients co-presented with metastatic disease in additional locations (87%), including the bones (56%), lymph nodes (45%), or lungs (39%). Patients less that 1 year of age comprised a very small proportion of the RMS metastatic to the bone marrow (1.7%). These findings confirm the analysis by Weiss et al that found no bone marrow involvement in patients with ERMS and T1 invasiveness compared to a 23% involvement in patients with ARMS and T2 disease.² Our findings support their recommendation that a bone marrow (BM) evaluation is not required at diagnosis for a patient with localized ERMS with a T1 invasive tumor while supporting the continued BM evaluation for patients with alveolar histology or T2 invasive tumors.

Once diagnosed, the survival rates for metastatic RMS have remained consistently poor despite intensive therapy and high initial response rates.^19,21,26,27 Bone marrow metastatic disease has been infrequently reported as a separate category, often combined with bone metastatic disease, but has been identified as a particularly poor prognostic factor.^3,15,17,23 We demonstrated concomitant bone marrow and bone metastatic disease at presentation in only 56% of patients. Our population displayed very little change in either EFS or OS outcomes when bone was an additional metastatic site of disease (Table 3). In contrast, patients with RMS metastatic only to bone enrolled on these same trials (n=15) have demonstrated superior 3-year EFS (42.4%; 95% CI 6.0% - 78.9%) and 3-year OS (49.4%; 95% CI 18.6% - 80.15%) (Personal Communication). These disparate outcomes and limited co-presentation support recommendations to divide them into separate risk categories though several groups have previously noted high correlation and included both as a single prognostic category.^2,3,17

Consistent with the outcomes in prior studies^{3,8,9,17,20–22}, our analysis demonstrates the poor prognosis for EFS and OS for patients with RMS bone marrow metastasis at diagnosis with early relapse largely occurring in the first two years off therapy (Figure 1). We verified existing RMS poor prognostic markers within the cohort of patients metastatic to bone marrow for EFS including age ≥ 10 years, alveolar histology, FOXO1 fusion positivity, unfavorable primary site, and elevated Oberlin score. We newly identified absence of radiation as a poor predictive marker within this cohort. Similarly, existing prognostic markers were confirmed for OS including age ≥ 10 years, alveolar histology, FOXO1 fusion positivity, unfavorable primary site and higher Oberlin Score. Number of metastatic sites ≥3 was also associated with inferior OS after only demonstrating marginal significance for impaired EFS. Absence of radiation continued to be a novel negative predictive marker for OS. We did not confirm the finding that PAX7-FOXO1 has superior outcomes to PAX3-FOXO1 fusions as shown in Sorenson et al. which interestingly had no patients with a PAX7-FOXO1 fusion and bone marrow metastases.¹¹ In contrast, our metastatic to bone marrow population demonstrated a non-significant trend in the opposite direction with no survivors in the PAX7-FOXO1 fusion group, albeit small numbers (n=5). This cohort did not have a uniform site or pattern of presentation with exception that four of five presented with concurrent bone metastases, all had T2 invasive primary tumors, and all received radiation prior to all five progressing/relapsing. These findings merit continued analysis in future studies to determine if there is a particular etiology for a negative prognosis associated with this PAX7 fusion partner when patients develop bone marrow metastases or a biologic rationale limiting metastases to the bone marrow.

Multiple variable analysis confirmed age ≥ 10 years, unfavorable primary location, and absence of radiation therapy as negative prognostic/predictive factors for EFS and age ≥ 10 years and absence of radiation therapy as negative prognostic/predictive factors for OS. The positive prognosis correlation with radiation therapy does not prove causality as this may indicate very aggressive disease that progressed prior to radiation being able to be given. However, there are emerging data that implicate aggressive management of metastatic sites with radiation contributing to event-free and overall survivals in metastatic RMS.²⁹ This finding supports continued prospective investigation in the role of radiation therapy utilization for patients with RMS metastatic to bone marrow. Association of radiation therapy and positive prognosis, despite absence of bone marrow irradiation, can be viewed as validation for aggressive local control management of other metastatic sites despite bone marrow involvement.

The dismal prognosis of the larger group and in particular the subgroups as noted above (Age ≥ 10 years, alveolar histology, FOXO1 fusion presence, unfavorable primary location, increased metastatic sites, higher Oberlin score, and inability to deliver radiation) appropriately lead to optimizing quality-of-life discussions. Current VAC based protocols represent significant intensity and duration of treatment. Patients/families in these subgroups may wish to consider early pursuit of novel therapies/approaches or therapy de-escalation rather than current standard of care. Novel treatments and approaches are needed to provide this patient group with alternatives if desired. It remains to be seen the impact of vinorelbine can have on this population with its inclusion in ARST2031 for HR-RMS. Additionally, given the frequent early relapses, targeting this window with low dose maintenance regimens may also be of interest.

Recurrences were largely metastatic or combined metastatic and local/regional recurrences while local/regional only recurrences were rare despite no patients achieving an R0 resection. The limited isolated localized recurrences (6.7%) when RMS is metastatic to bone marrow at diagnosis are in contrast to localized disease.²⁸ Combined with the EFS and OS nearly convergent at the 10-year timeframe (EFS 6.0% vs OS 8.9%), this implies that while salvage therapy can control recurrent disease, there are very few patients that are ultimately long-term survivors of recurrent disease.

Multiple studies have investigated the larger issue of the survival benefit to this surveillance imaging in rhabdomyosarcoma and this remains an area of research.^29,30 Our data adds to this debate of prolonged surveillance with a 3-year EFS of 9% declining to only 6% at 10 years. Further, existing prospectively collected data did not delineate if metastatic recurrences occurred at prior locations of metastatic disease and if those metastatic lesions had received locally directed therapy. Future prospective collection of this data would be helpful in determining utility of surveillance imaging and local management of metastatic lesions even in the setting of bone marrow disease. This is particularly interesting in light of our finding of EFS and OS advantages with radiation therapy combined with the survival benefits noted recently by Ferrari et al. in the larger metastatic RMS population.²⁸. Future detailed assessment of the patterns of radiation (primary vs metastatic), number of radiation sites, and compliance with prescribed radiation is recommended in this population.

This study was limited by absence of data on repeat bone marrow assessments which prevents determination of the response rate of the bone marrow disease to verify the findings of Bailey and Wexler who reported rapid clearance of marrow to initiation of chemotherapy.²¹

Summary

This study represents the largest analysis of the characteristics and outcomes of pediatric RMS metastatic to bone marrow. Patients demonstrated dismal outcomes with early metastatic relapse in the first two years off therapy and few long-term salvages of recurrent disease. We have validated current screening algorithms and supported the division of metastatic disease to the bone and bone marrow as separate entities. We have also identified several groups that with particularly poor prognosis who may wish to consider early pursuit of alternative/novel therapies rather than current high intensity and low quality of life regimens that are currently the standard of care. Development of novel treatments and access to those treatments for these patient populations should be a priority.

Included Abbreviations

ARMS

Alveolar rhabdomyosarcoma

BM

Bone marrow

COG

Children’s Oncology Group

ERMS

Embryonal rhabdomyosarcoma

EFS

Event-Free Survival

HR

High risk

IE

Ifosfamide, Etoposide

IR

Intermediate Risk

PET

Positron emission tomography

OS

Overall Survival

VAC

Vincristine, Actinomycin D, Cyclophosphamide

VDC

Vincristine, Doxorubicin, Cyclophosphamide

VI

Vincristine, Irinotecan

VTC

Vincristine, Topotecan, Cyclophosphamide

DATA AVAILABILITY STATEMENT

Research data are not shared