Paediatric Therapeutic Development Workshop on rhabdomyosarcoma Authors: Joseph S. Baxter, Claudia Montiel Equihua, Jan J. Molenaar, Jamie Aye, Gianni Bisogno, Patricia Blanc, Willemijn Breunis, Julia Chisholm, Jacquelyn Crane, Sam Daems, Laura Danielson, Bram de Wilde, Adam D. Durbin, Felipe Galvez-Cancino, Patrizia Gasparini, Birgit Geoerger, Simone Graham, Bass Hassan, Mark E. Hatley, Delphine Heenen, Christine M. Heske, Simone Hettmer, Elizabeth Hookham, Peter Houghton, Pamela Kearns, Charles Keller, Javed Khan, Fiona Kinnis, David Langenau, Geffen Lass, Corinne Mary Linardic, Leo Mascarenhas, Michael T. Meister, Jonathan Metts, Véronique Minard-Colin, Sapna Oberoi, Daniela Palacios, Abbe Pannucci, Seema Patel, Sheena Patel Published: 2026 ABSTRACT The third in a series of Paediatric Therapeutic Development Workshops focused on rhabdomyosarcoma. Rhabdomyosarcoma is the most common soft tissue sarcoma in children, with 90% survival for those with the lowest risk disease, but just 20–30% in children with metastatic disease. An urgent unmet need exists to develop targeted therapeutics for high-risk disease and to reduce the toxicity of treatment. The results of trials of CAR T-cells and ADCs against FGFR4 are awaited with great interest, and developing a FGFR4 degrader is a priority. Directly targeting PAX3::FOXO1 and PAX7::FOXO1 fusion proteins is a high priority. An in vivo study of a MYOD1 degrader approach is required prior to clinical development. Degraders of P300/CBP should be evaluated preclinically with a view to clinical investigation. ROR2 is an interesting target for the L122R mutant MYOD1 rhabdomyosarcoma. Development of a TEAD degrader is a high priority, and this should be evaluated in combination with a Notch inhibitor. Considering targets with existing clinical agents, antibody-drug conjugates targeting cell-surface antigen B7-H3/CD276 are showing preclinical promise in other paediatric cancers and are also deemed a high priority for evaluation in rhabdomyosarcoma. Based on currently available evidence, MEK inhibitors should be evaluated, potentially with BRAF or PI3K inhibitors, in combination with chemotherapy in the maintenance setting. Understanding the mechanism of action underpinning drug combinations, gaining access to therapeutics and optimising clinical trial design will be essential to enable combinatorial testing in patients. Introduction Rhabdomyosarcoma is a high-grade sarcoma resembling developing or embryonic skeletal muscle cells and is the most common form of soft tissue sarcoma in childhood, with 4.5 cases per 1 million people aged 0–20 years [1]. Current treatment for rhabdomyosarcoma is chemotherapy and local therapy (surgery and/or radiotherapy) with the exact regimens being guided by clinical risk groups [1]. Children in the lowest risk group have a good prognosis of >90% overall survival, but the prognosis for the highest risk group with metastatic disease is only 20–30% [2–9]. Despite advances from large, co-operative clinical trials, the outlook for children with metastatic or recurrent disease remains very poor. Moreover, current treatments for rhabdomyosarcoma are associated with severe short and long-term side effects, often due to extensive surgery and intensive radiotherapy and chemotherapy regimens, which seriously impact survivors’ quality of life [1]. The third Paediatric Therapeutic Development Workshop focussed on rhabdomyosarcoma, aiming to understand unmet treatment needs in rhabdomyosarcoma and identify priorities for therapeutic development. The workshop was organised by LifeArc (a self-funded not-for-profit charity), Innovative Therapies for Children and Adolescents with Cancer (ITCC), Cancer Research UK and the Cancer Grand Challenge PROTECT team. The outputs of the workshop will be used to inform future strategies of academia, industry, regulators and charities working in rhabdomyosarcoma. This article summarises the prioritisation decisions and overarching recommendations of the workshop. Background Rhabdomyosarcoma has been divided by the WHO into four subtypes based on its histological features: embryonal rhabdomyosarcoma (ERMS), alveolar rhabdomyosarcoma (ARMS), spindle cell/sclerosing rhabdomyosarcoma (SCRMS) and pleomorphic rhabdomyosarcoma (PRMS) [10]. ERMS has morphological and immunophenotypic features of embryonic skeletal muscle and is the most common rhabdomyosarcoma subtype, accounting for 60% of cases in children [11, 12]. More than a third of ERMS cases occur in children aged <5 years. ARMS is a poor-prognosis subtype and the second most common type of rhabdomyosarcoma, accounting for 20% of all cases in children [11, 13]. Most cases are characterised by a fusion of the Paired box (PAX) 3 or 7 gene with Forkhead box O1 (FOXO1) [12]. SCRMS is a rare form of rhabdomyosarcoma and accounts for 3–10% of all rhabdomyosarcoma cases and affects infants, children and adults [14], and PRMS is exclusively found in adults [15]. Although paediatric rhabdomyosarcoma has been conventionally separated into three histological subtypes (ERMS, ARMS and SCRMS), rhabdomyosarcoma tumours are currently categorised according to their genomic profiles [1]. This consensus has led to molecular categorisation of ARMS as fusion-positive or fusion-negative, based on the fusion status of the PAX3/7 genes with FOXO1, which has replaced histological diagnosis of the alveolar subtype in risk stratification. However, other rarer fusions also exist, such as alternative PAX fusions in ARMS, TFCP fusions, and VGLL2::NCOA2 found in SCRMS [16, 17]. Rhabdomyosarcoma therapy and clinical trial allocation are determined by clinical risk group, and guidance varies between the US and Europe [1, 18–21] (Table 1). The Children’s Oncology Group (COG) Soft Tissue Sarcoma Committee in North America stratifies patients into three risk groups based on factors such as tumour node metastases stage (1–4; site favourability, size, extent of spread and metastases), clinical group (I–IV; extent of spread and resectability), age and fusion status [22]. In the European paediatric Soft Tissue Sarcoma Study Group (EpSSG), there are four risk groups and eight subgroups stratified by fusion status, Intergroup Rhabdomyosarcoma Study (IRS) group (I–IV, extent of spread and resectability), site (favourable or unfavourable), node stage (N0 or N1), size of tumour, and age (favourable or unfavourable). Essentially, EpSSG and COG are using almost identical clinical risk factors and PAX::FOXO1 fusion status for risk stratification, but exactly how they are combined into risk groups varies.Table 1COG and EpSSG risk stratification groups [2, 18–20].Risk groupSubgroupSiteStage (COG; 1–4) or Node stageClinical Groupa or IRS groupb (I–IV)Age (COGS) or Size/Age (EpSSG)Fusion status or groupCOGLowN/AN/A1I, II, III (orbit only)AnyFOXO1−2I, IIIntermediate1III (non-orbit)AnyFOXO1−1, 2, 3I, II, IIIFOXO1+2, 3IIIFOXO1−3I, IIFOXO1−4IV<10 yearsFOXO1−High4IV>10 yearsFOXO1−AnyFOXO1+EpSSGLowAAnyN0IBoth favourableNegativeStandardBAnyN0IOne or both unfavourableNegativeCFavourableN0II, IIIAnyNegativeHighDUnfavourableN0II, IIIAnyNegativeEAnyN1II, IIINegativeFAnyM0I, II, IIIPositiveVery highGAnyN1II, IIAnyPositiveHAnyAnyIVAnyCOG Children’s Oncology Group, EpSSG European paediatric Soft Tissue Sarcoma Study Group.aExtent of spread and resectability, COG.bExtent of spread and resectability, EpSSG. Table 1COG and EpSSG risk stratification groups [2, 18–20].Risk groupSubgroupSiteStage (COG; 1–4) or Node stageClinical Groupa or IRS groupb (I–IV)Age (COGS) or Size/Age (EpSSG)Fusion status or groupCOGLowN/AN/A1I, II, III (orbit only)AnyFOXO1−2I, IIIntermediate1III (non-orbit)AnyFOXO1−1, 2, 3I, II, IIIFOXO1+2, 3IIIFOXO1−3I, IIFOXO1−4IV<10 yearsFOXO1−High4IV>10 yearsFOXO1−AnyFOXO1+EpSSGLowAAnyN0IBoth favourableNegativeStandardBAnyN0IOne or both unfavourableNegativeCFavourableN0II, IIIAnyNegativeHighDUnfavourableN0II, IIIAnyNegativeEAnyN1II, IIINegativeFAnyM0I, II, IIIPositiveVery highGAnyN1II, IIAnyPositiveHAnyAnyIVAnyCOG Children’s Oncology Group, EpSSG European paediatric Soft Tissue Sarcoma Study Group.aExtent of spread and resectability, COG.bExtent of spread and resectability, EpSSG. COG and EpSSG risk stratification groups [2, 18–20]. COG Children’s Oncology Group, EpSSG European paediatric Soft Tissue Sarcoma Study Group. aExtent of spread and resectability, COG. bExtent of spread and resectability, EpSSG. Patients in the lowest-risk groups tend to have a prognosis of 90% overall survival (OS), whereas the OS for the highest-risk groups is 20–30% [18, 9]. Prognosis and outcomes depend not only on patient age, size and location of tumour, and presence or absence of metastases, but also on the histopathological classification, presence of fusion genes and other genomic variants such as MYOD1L122R mutation, TP53 mutation [23–39] (Table 2).Table 2Classification and clinical and molecular features of rhabdomyosarcoma subtypes [23–38].Histopathological subtypeAlveolar (ARMS)Embryonal (ERMS)Spindle cell/sclerosing (SCRMS)Molecular subgroupFusion-positive (70-85%)Fusion-negativeFusion-negativeFusion-positiveFusion-negativeHistopathologyRound cells with diffuse MYOD1/myogeninTwo subclasses: classic or solid variantRound and spindled with heterogeneous MYOD1/myogeninThree subclasses: classic, dense circular or with a spindle cell componentSpindle cell morphology with fascicular patterns.Variable MYOD1/myogenin.Spindle cell morphology with fascicular patterns.Diffuse MYOD1.Developmental originUnknown. Stem cell niches and cellular hierarchy are known to be distinct between fusion-positive and fusion-negative RMS.LocationDeep soft tissue of extremities: head and neck, paraspinal and perineal region50% head and neck, 50% genitourinary. Rarely involves extremity soft tissue.Most commonly, the paratesticular region, followed by the head and neck and other regions.Clinical presentationSymptoms relate to tumour location and size, e.g. affecting cranial nerve if head and neck location.AetiologyNo known lifestyle or environmental causes. Alteration of RAS signalling and other cancer predisposition syndromes in ERMS.Prognosis• Fusion status is the most important prognostic indicator.• ERMS has a superior prognosis to ARMS.• Orbital and genitourinary tumours are the most favourable locations.• Node involvement and presence of bone or bone marrow metastasis are strong negative prognostic factors.Five-year overall survival (OS)29%74%91% (although patients with MYOD1 mutations may have much lower 5-year OS)Relapse rate63%28%UnknownTreatmentSurgery, radiation and chemotherapy (VAC or VAI)EpSSG - Maintenance therapy with cyclophosphamide/vinorelbine in high-risk patients with a complete response to VAC or VAI.MYOD1 myogenic differentiation 1, VAC vincristine, dactinomycin, cyclophosphamide, VAI vincristine, dactinomycin, ifosfamide. Table 2Classification and clinical and molecular features of rhabdomyosarcoma subtypes [23–38].Histopathological subtypeAlveolar (ARMS)Embryonal (ERMS)Spindle cell/sclerosing (SCRMS)Molecular subgroupFusion-positive (70-85%)Fusion-negativeFusion-negativeFusion-positiveFusion-negativeHistopathologyRound cells with diffuse MYOD1/myogeninTwo subclasses: classic or solid variantRound and spindled with heterogeneous MYOD1/myogeninThree subclasses: classic, dense circular or with a spindle cell componentSpindle cell morphology with fascicular patterns.Variable MYOD1/myogenin.Spindle cell morphology with fascicular patterns.Diffuse MYOD1.Developmental originUnknown. Stem cell niches and cellular hierarchy are known to be distinct between fusion-positive and fusion-negative RMS.LocationDeep soft tissue of extremities: head and neck, paraspinal and perineal region50% head and neck, 50% genitourinary. Rarely involves extremity soft tissue.Most commonly, the paratesticular region, followed by the head and neck and other regions.Clinical presentationSymptoms relate to tumour location and size, e.g. affecting cranial nerve if head and neck location.AetiologyNo known lifestyle or environmental causes. Alteration of RAS signalling and other cancer predisposition syndromes in ERMS.Prognosis• Fusion status is the most important prognostic indicator.• ERMS has a superior prognosis to ARMS.• Orbital and genitourinary tumours are the most favourable locations.• Node involvement and presence of bone or bone marrow metastasis are strong negative prognostic factors.Five-year overall survival (OS)29%74%91% (although patients with MYOD1 mutations may have much lower 5-year OS)Relapse rate63%28%UnknownTreatmentSurgery, radiation and chemotherapy (VAC or VAI)EpSSG - Maintenance therapy with cyclophosphamide/vinorelbine in high-risk patients with a complete response to VAC or VAI.MYOD1 myogenic differentiation 1, VAC vincristine, dactinomycin, cyclophosphamide, VAI vincristine, dactinomycin, ifosfamide. Classification and clinical and molecular features of rhabdomyosarcoma subtypes [23–38]. Round cells with diffuse MYOD1/myogenin Two subclasses: classic or solid variant Round and spindled with heterogeneous MYOD1/myogenin Three subclasses: classic, dense circular or with a spindle cell component Spindle cell morphology with fascicular patterns. Variable MYOD1/myogenin. Spindle cell morphology with fascicular patterns. Diffuse MYOD1. • Fusion status is the most important prognostic indicator. • ERMS has a superior prognosis to ARMS. • Orbital and genitourinary tumours are the most favourable locations. • Node involvement and presence of bone or bone marrow metastasis are strong negative prognostic factors. Surgery, radiation and chemotherapy (VAC or VAI) EpSSG - Maintenance therapy with cyclophosphamide/vinorelbine in high-risk patients with a complete response to VAC or VAI. MYOD1 myogenic differentiation 1, VAC vincristine, dactinomycin, cyclophosphamide, VAI vincristine, dactinomycin, ifosfamide. Vincristine and actinomycin-D (VA) is the chemotherapy regimen for low-risk patients. For intermediate- and highest-risk disease, systemic therapy is based on a foundation of intensive, multiagent chemotherapy administered at intervals over 6-9 months [2, 6]. In North America, the combination of vincristine, actinomycin-D and cyclophosphamide (VAC) is the backbone for most rhabdomyosarcoma treatments [2, 6], whilst in Europe, the combination of ifosfamide, vincristine and actinomycin-D (IVA) is the reference regimen. Recently, the EpSSG has added maintenance chemotherapy with low-dose cyclophosphamide and weekly vinorelbine to the treatment of high-risk patients [9]. Patients with metastatic disease continue to have a poor outcome in spite of intensive therapy [3–5]. For patients and parents, quality of life is a crucial aspect of rhabdomyosarcoma therapy. Many long-term survivors of rhabdomyosarcoma have chronic health issues, which may include lifelong severe late effects such as infertility, restricted bone and soft tissue growth, visual and auditory defects, learning difficulties and increased future cancer risk [40–42]. Local control measures, especially radiation, are a major source of long-term sequelae. A parent at the workshop stated that ‘survivors are alive but not living a life’. Moreover, frontline therapy has remained approximately the same over the past 40 years. New and innovative therapeutics for rhabdomyosarcoma are therefore desperately required to both save and improve the quality of patients’ lives. Biology of rhabdomyosarcoma The aetiology of rhabdomyosarcoma is only partially understood, but rhabdomyosarcoma tumours are believed to arise when mesenchymal cells committed to the skeletal muscle lineage fail to fully differentiate into mature myocytes and are arrested at a proliferative rhabdomyoblastic stage [43]. Studies in mouse models suggest that rhabdomyosarcoma tumours can develop from multiple cell lineages, including muscle and non-muscle, but in humans, the precise developmental origins and tumorigenic processes remain unclear [43–48]. Rhabdomyosarcoma is generally driven by a combination of genomic aberrations which function together to block rhabdomyoblast differentiation and promote tumour progression. The fusion genes are major disease-driving aberrations through their encoded transcriptional activity, as discussed below. The phosphoinositide 3-kinase (PI3K) pathway is activated in ~80% of rhabdomyosarcoma tumours overall, including both PAX3/7::FOXO1 positive and negative cases [43]. Mutations in the RAS pathway frequently occur in PAX3/7::FOXO1-negative tumours, whereas PAX3/7::FOXO1-positive tumours have a lower mutational burden [39, 48]. Three key molecular features of SCRMS are used to subclassify this group into subgroups: MYOD1 mutations, TFCP fusions and VGLL2::NCOA2 fusions. Table 3 summarises the different disease drivers for these subtypes [49–70].Table 3Biological features and disease drivers for RMS subtypes [49–70].Histopathological subtypeAlveolar (ARMS)Embryonal (ERMS)Spindle cell/sclerosing (SCRMS)Fusion statusFusion-positiveFusion negativeNoneFusion-positiveFusion-negativeCytogeneticsAlmost identical to ERMS.11p15.5 loss of heterozygosity and high fractional allelic lossOverexpression of IGF2 is nearly universalChromosomal aneuploidies are very common, especially gains of chromosome 8, but also chromosomes 2, 11, 12, 13 and 20Congenital/infantile spindle cell type:VGLL2 (40%). SRF, TEAD1, NCOA2, and CITED2 gene fusions [202]. Occurs in children [99]MYOD1 mutations (~40%) occur almost exclusively in exon 1, with the hot spot mutation p.L122R. Occurs in adults and children [102]Driver eventsTranslocations t(2;13)(q36;q14) and t(1;13)(p36;q14) fusion of the N-terminal DNA binding domain of PAX3 (55%), or more rarely PAX7 (22%) to the C-terminal transactivating domain of FOXO1 [12, 202, 203]PAX3::FOXO1 overexpressed but not increased gene copy number, whereas the PAX7::FOXO1 overexpressed with increases in fusion gene copy number [133]. PAX3/7::FOXO1-positive tumours exhibit global changes in methylation profiles, suggesting they are driven by epigenetic events [39, 133, 204]Unknown but believed to be similar to ERMS. Methylation profiles shared between ERMS and specific normal or progenitor cells from which they might originate. Altered methylation patterns play a key role. Distinct methylation patterns do exist within the ERMS subtypeDisease drivers are incompletely understood.Mutations in RAS/MAPK and/or PI3K occur in >50% patients.RAS pathway mutations occur in ~40% of cases. Clonal evolution studies suggest RAS mutations occur early in the disease process [56] and in vitro studies imply RAS activation is essential for tumour survival and maintenance.But there are other disease drivers yet to be elucidated [205–208].Loss of PTEN, TP53 and/or CDKN2A in >50% patients.FGFR2 hyperactivation in ~20% of cases.Hedgehog pathway activation was observed in ~50% of cases.VGLL2::NCOA2 fusion can promote tumours in zebrafish and was found to upregulate a RAS family GTPase called ARF6 [17], but little is known about other fusions involved in SCRMS.The group is highly heterogeneous in terms of gene rearrangements and fusions.Little is known about the aetiology of fusion-positive SCRMS [10]. TheMYOD1L122R mutations preclude normal differentiation of myogenic progenitor cells by gaining the ability to bind to and activate regulatory motifs typically bound by MYC [209]. Mutant MYOD1 loses the ability to regulate physiological target genes, resulting in a pathology driven by both gain-of-function and loss-of-function mutations, associated with a very poor prognosis [39]. Full mechanism is unknown, but RAS, PI3K and other gene mutations frequently co-occur and are believed to function synergistically with mutant MYOD1 [39, 100, 103]Frequent copy number alterationsMYCN is amplified in ~20% of PAX3/7::FOXO1-positive tumours and shows anti-tumour activity on knockdown [39, 55, 210].MYCN amplification is not prognostic.13q31 amplification is reported in ~25% of patients, resulting in increased copies of the miR-17-92 cluster.Frequently amplified regions:12q13.12, 12q13.3 and 12q13.3-q14.1Frequently deleted regions: 1p21.1, 2q14.1, 5q13.2, 9p12 and 9q12.Loss of chromosomes 10 and 15Polysomy of chromosome 8Loss of heterozygosity at chromosome 11Gains of 11q and 7q regionsFrequent increased MYCN copy numberALK amplification in TFCP2-rearranged tumoursCommonly mutated genesVery low mutational burden.Very low mutational burden.Recurrent mutations in cancer genes, including HRAS, NRAS, KRAS, ALK, FGFR4, PIK3CA, FBXW7, NF1, TP53, CTNNB1, and BCOR, are common but not universal>50% of patients have mutations in RAS/RAF/MAPK and/or PI3K/mTOR pathways~10% of patients have MYOD1 mutations.A small number of patients have no identifiable mutations.TP53, CDKN2A and FGFR4 have been described at relapse.PI3K mutations occur in ~50% of cases with MYOD1 mutations.Germline mutationsVery rare (3%)Very rare (10%)RMS can present in syndromes such as Li Fraumeni, Costello, Noonan, NF1, Dicer, and BW.Can present such as Li Fraumeni, Costellos, Noonan, NF1, Dicer, BWImportant epigenetic playersEZH2BETCHD4Histone modificationsLittle data. Studies usually focus on fusion-positive ARMS or ERMS.EZH2DNMTB3HDACsHMT/HDMsNo dataRMS rhabdomyosarcoma, ALK anaplastic lymphoma kinase, BCOR BCL6 co-repressor, BET bromodomain and extraterminal, BW Beckwith-Wiedemann syndrome, CDKN2A cyclin-dependent kinase inhibitor 2A, CHD4 chromodomain-helicase-DNA-binding protein 4, CITED2 Cbp/p300-interacting transactivator with Glu/Asp-rich carboxy-terminal domain 2, CTNNB1 catenin beta-1, DNMTB3 DNA (cytosine-5)-methyltransferase 3 beta, EZH2 enhancer of zeste homologue 2, FGFR4 fibroblast growth factor receptor 4, FBXW7 F-box/WD repeat domain-containing 7, IGF2 insulin growth factor 2, MAPK mitogen-activated protein kinase, MYOD1 myogenic differentiation 1, NCOA2 nuclear receptor coactivator 2, NF1 neurofibromatosis Type 1, PIK3CA phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit alpha, RAF rapidly accelerated fibrosarcoma, SRF serum response factor, TEAD1 TEA domain family member 1, TFCP2 alpha-globin transcription factor CP2, TP53 tumour protein 53, VGLL2 vestigial like family member 2. Table 3Biological features and disease drivers for RMS subtypes [49–70].Histopathological subtypeAlveolar (ARMS)Embryonal (ERMS)Spindle cell/sclerosing (SCRMS)Fusion statusFusion-positiveFusion negativeNoneFusion-positiveFusion-negativeCytogeneticsAlmost identical to ERMS.11p15.5 loss of heterozygosity and high fractional allelic lossOverexpression of IGF2 is nearly universalChromosomal aneuploidies are very common, especially gains of chromosome 8, but also chromosomes 2, 11, 12, 13 and 20Congenital/infantile spindle cell type:VGLL2 (40%). SRF, TEAD1, NCOA2, and CITED2 gene fusions [202]. Occurs in children [99]MYOD1 mutations (~40%) occur almost exclusively in exon 1, with the hot spot mutation p.L122R. Occurs in adults and children [102]Driver eventsTranslocations t(2;13)(q36;q14) and t(1;13)(p36;q14) fusion of the N-terminal DNA binding domain of PAX3 (55%), or more rarely PAX7 (22%) to the C-terminal transactivating domain of FOXO1 [12, 202, 203]PAX3::FOXO1 overexpressed but not increased gene copy number, whereas the PAX7::FOXO1 overexpressed with increases in fusion gene copy number [133]. PAX3/7::FOXO1-positive tumours exhibit global changes in methylation profiles, suggesting they are driven by epigenetic events [39, 133, 204]Unknown but believed to be similar to ERMS. Methylation profiles shared between ERMS and specific normal or progenitor cells from which they might originate. Altered methylation patterns play a key role. Distinct methylation patterns do exist within the ERMS subtypeDisease drivers are incompletely understood.Mutations in RAS/MAPK and/or PI3K occur in >50% patients.RAS pathway mutations occur in ~40% of cases. Clonal evolution studies suggest RAS mutations occur early in the disease process [56] and in vitro studies imply RAS activation is essential for tumour survival and maintenance.But there are other disease drivers yet to be elucidated [205–208].Loss of PTEN, TP53 and/or CDKN2A in >50% patients.FGFR2 hyperactivation in ~20% of cases.Hedgehog pathway activation was observed in ~50% of cases.VGLL2::NCOA2 fusion can promote tumours in zebrafish and was found to upregulate a RAS family GTPase called ARF6 [17], but little is known about other fusions involved in SCRMS.The group is highly heterogeneous in terms of gene rearrangements and fusions.Little is known about the aetiology of fusion-positive SCRMS [10]. TheMYOD1L122R mutations preclude normal differentiation of myogenic progenitor cells by gaining the ability to bind to and activate regulatory motifs typically bound by MYC [209]. Mutant MYOD1 loses the ability to regulate physiological target genes, resulting in a pathology driven by both gain-of-function and loss-of-function mutations, associated with a very poor prognosis [39]. Full mechanism is unknown, but RAS, PI3K and other gene mutations frequently co-occur and are believed to function synergistically with mutant MYOD1 [39, 100, 103]Frequent copy number alterationsMYCN is amplified in ~20% of PAX3/7::FOXO1-positive tumours and shows anti-tumour activity on knockdown [39, 55, 210].MYCN amplification is not prognostic.13q31 amplification is reported in ~25% of patients, resulting in increased copies of the miR-17-92 cluster.Frequently amplified regions:12q13.12, 12q13.3 and 12q13.3-q14.1Frequently deleted regions: 1p21.1, 2q14.1, 5q13.2, 9p12 and 9q12.Loss of chromosomes 10 and 15Polysomy of chromosome 8Loss of heterozygosity at chromosome 11Gains of 11q and 7q regionsFrequent increased MYCN copy numberALK amplification in TFCP2-rearranged tumoursCommonly mutated genesVery low mutational burden.Very low mutational burden.Recurrent mutations in cancer genes, including HRAS, NRAS, KRAS, ALK, FGFR4, PIK3CA, FBXW7, NF1, TP53, CTNNB1, and BCOR, are common but not universal>50% of patients have mutations in RAS/RAF/MAPK and/or PI3K/mTOR pathways~10% of patients have MYOD1 mutations.A small number of patients have no identifiable mutations.TP53, CDKN2A and FGFR4 have been described at relapse.PI3K mutations occur in ~50% of cases with MYOD1 mutations.Germline mutationsVery rare (3%)Very rare (10%)RMS can present in syndromes such as Li Fraumeni, Costello, Noonan, NF1, Dicer, and BW.Can present such as Li Fraumeni, Costellos, Noonan, NF1, Dicer, BWImportant epigenetic playersEZH2BETCHD4Histone modificationsLittle data. Studies usually focus on fusion-positive ARMS or ERMS.EZH2DNMTB3HDACsHMT/HDMsNo dataRMS rhabdomyosarcoma, ALK anaplastic lymphoma kinase, BCOR BCL6 co-repressor, BET bromodomain and extraterminal, BW Beckwith-Wiedemann syndrome, CDKN2A cyclin-dependent kinase inhibitor 2A, CHD4 chromodomain-helicase-DNA-binding protein 4, CITED2 Cbp/p300-interacting transactivator with Glu/Asp-rich carboxy-terminal domain 2, CTNNB1 catenin beta-1, DNMTB3 DNA (cytosine-5)-methyltransferase 3 beta, EZH2 enhancer of zeste homologue 2, FGFR4 fibroblast growth factor receptor 4, FBXW7 F-box/WD repeat domain-containing 7, IGF2 insulin growth factor 2, MAPK mitogen-activated protein kinase, MYOD1 myogenic differentiation 1, NCOA2 nuclear receptor coactivator 2, NF1 neurofibromatosis Type 1, PIK3CA phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit alpha, RAF rapidly accelerated fibrosarcoma, SRF serum response factor, TEAD1 TEA domain family member 1, TFCP2 alpha-globin transcription factor CP2, TP53 tumour protein 53, VGLL2 vestigial like family member 2. Biological features and disease drivers for RMS subtypes [49–70]. 11p15.5 loss of heterozygosity and high fractional allelic loss Overexpression of IGF2 is nearly universal Chromosomal aneuploidies are very common, especially gains of chromosome 8, but also chromosomes 2, 11, 12, 13 and 20 Congenital/infantile spindle cell type: VGLL2 (40%). SRF, TEAD1, NCOA2, and CITED2 gene fusions [202]. Occurs in children [99] Translocations t(2;13)(q36;q14) and t(1;13)(p36;q14) fusion of the N-terminal DNA binding domain of PAX3 (55%), or more rarely PAX7 (22%) to the C-terminal transactivating domain of FOXO1 [12, 202, 203] PAX3::FOXO1 overexpressed but not increased gene copy number, whereas the PAX7::FOXO1 overexpressed with increases in fusion gene copy number [133]. PAX3/7::FOXO1-positive tumours exhibit global changes in methylation profiles, suggesting they are driven by epigenetic events [39, 133, 204] Disease drivers are incompletely understood. Mutations in RAS/MAPK and/or PI3K occur in >50% patients. RAS pathway mutations occur in ~40% of cases. Clonal evolution studies suggest RAS mutations occur early in the disease process [56] and in vitro studies imply RAS activation is essential for tumour survival and maintenance. But there are other disease drivers yet to be elucidated [205–208]. Loss of PTEN, TP53 and/or CDKN2A in >50% patients. FGFR2 hyperactivation in ~20% of cases. Hedgehog pathway activation was observed in ~50% of cases. VGLL2::NCOA2 fusion can promote tumours in zebrafish and was found to upregulate a RAS family GTPase called ARF6 [17], but little is known about other fusions involved in SCRMS. The group is highly heterogeneous in terms of gene rearrangements and fusions. Little is known about the aetiology of fusion-positive SCRMS [10]. The MYCN is amplified in ~20% of PAX3/7::FOXO1-positive tumours and shows anti-tumour activity on knockdown [39, 55, 210]. MYCN amplification is not prognostic. 13q31 amplification is reported in ~25% of patients, resulting in increased copies of the miR-17-92 cluster. Frequently amplified regions: 12q13.12, 12q13.3 and 12q13.3-q14.1 Frequently deleted regions: 1p21.1, 2q14.1, 5q13.2, 9p12 and 9q12. Loss of chromosomes 10 and 15 Polysomy of chromosome 8 Loss of heterozygosity at chromosome 11 Gains of 11q and 7q regions Frequent increased MYCN copy number Recurrent mutations in cancer genes, including HRAS, NRAS, KRAS, ALK, FGFR4, PIK3CA, FBXW7, NF1, TP53, CTNNB1, and BCOR, are common but not universal >50% of patients have mutations in RAS/RAF/MAPK and/or PI3K/mTOR pathways ~10% of patients have MYOD1 mutations. A small number of patients have no identifiable mutations. EZH2 BET CHD4 Histone modifications EZH2 DNMTB3 HDACs HMT/HDMs RMS rhabdomyosarcoma, ALK anaplastic lymphoma kinase, BCOR BCL6 co-repressor, BET bromodomain and extraterminal, BW Beckwith-Wiedemann syndrome, CDKN2A cyclin-dependent kinase inhibitor 2A, CHD4 chromodomain-helicase-DNA-binding protein 4, CITED2 Cbp/p300-interacting transactivator with Glu/Asp-rich carboxy-terminal domain 2, CTNNB1 catenin beta-1, DNMTB3 DNA (cytosine-5)-methyltransferase 3 beta, EZH2 enhancer of zeste homologue 2, FGFR4 fibroblast growth factor receptor 4, FBXW7 F-box/WD repeat domain-containing 7, IGF2 insulin growth factor 2, MAPK mitogen-activated protein kinase, MYOD1 myogenic differentiation 1, NCOA2 nuclear receptor coactivator 2, NF1 neurofibromatosis Type 1, PIK3CA phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit alpha, RAF rapidly accelerated fibrosarcoma, SRF serum response factor, TEAD1 TEA domain family member 1, TFCP2 alpha-globin transcription factor CP2, TP53 tumour protein 53, VGLL2 vestigial like family member 2. Preclinical models of rhabdomyosarcoma A wide range of preclinical models for rhabdomyosarcoma has been developed, ranging from subcutaneous and orthotopic patient-derived xenografts (PDX and O-PDX) to mesenchymal tumour organoids and genetically engineered murine models [71]. This includes widely available ERMS and ARMS human cell lines, which vary in origin and treatment history (primary tumour vs metastasis; newly diagnosed vs previously treated), location, patient age, treatment history and PAX3/7::FOXO1 status [72]. Several publicly available model repositories and data resources on rhabdomyosarcoma exist, including. St Jude Children’s Research Hospital’s Childhood Solid Tumour Network with access to >45 rhabdomyosarcoma O-PDX models [73] and an Integrated Rhabdomyosarcoma Database comprising genomic, epigenetic, proteomic and drug sensitivity data [74]. The Innovative Therapies for Children with Cancer Paediatric Preclinical Proof-of-concept Platform (ITCC-P4) [75], National Cancer Institute Childhood Cancer Data Initiative (CCDI) Hub [76] and MAPPYACTS Biobank [77] provide repertoires of patient-derived models of paediatric tumours, supported by comprehensive datasets of corresponding molecular and pharmacological characterisation. These models reflect patient-specific rhabdomyosarcoma drug sensitivities and can be genetically modified using CRISPR/Cas9, although they also lack functional human immune systems. Sarcoma_CellMinerCDB is a publicly available interactive database and tool which merges publicly available sarcoma cell line data and newly generated omics data to create a comprehensive database of genomic, transcriptomic, methylomic, proteomic, metabolic, and pharmacologic data [78]. The most commonly used mouse models for rhabdomyosarcoma are cell-line-derived xenografts (CDX), followed by patient-derived xenografts and syngeneic mouse models. However, CDX models do not always fully mimic human rhabdomyosarcoma phenotypes. Neither CDX nor PDX models have functional human immune systems [71]. There is, however, one humanised mouse model made with rhabdomyosarcoma cell lines [79]. An alternative resource is a collection of >30 mesenchymal tumour organoid/tumoroid models covering all major rhabdomyosarcoma subtypes developed by researchers in the Netherlands, which reflect rhabdomyosarcoma drug sensitivities in patients and can be modified by CRISPR/Cas9 [80]. The Rhabdomyosarcoma Paediatric Therapeutic Development Workshop To identify unmet needs and research priorities in rhabdomyosarcoma, a detailed diligence review of rhabdomyosarcoma epidemiology, biology, preclinical tools, potential targets and existing commercial developments was prepared by LifeArc and circulated to experts ahead of the workshop. This included a list of therapeutic targets produced from the GlobalData Drugs Database, supplemented with literature searches, to identify preclinical and clinical evidence for each target’s association with rhabdomyosarcoma (Supplementary Table 1) and the list of drugs in development (Supplementary Table 2). This review was followed by a workshop comprising experts in the biology and therapy of rhabdomyosarcoma and patient advocates. The workshop shared and discussed the scientific evidence and drug development status of the different therapeutic targets, prioritised targets based on their level of validation and tractability and identified potential drug combinations to progress into clinical trials. Workshop attendees were asked to prioritise targets that (i) had some level of preclinical validation in the indication; (ii) were tractable in principle; (iii) for which the optimal therapeutic did not yet exist for rhabdomyosarcoma; (iv) would address a high unmet need within rhabdomyosarcoma; and (v) had the highest potential for therapeutics against them being effective in this indication. Based on the expert opinion of the authors, the six targets in Table 4 were selected, in no particular order, for more detailed discussion. After discussion, each workshop participant numerically scored each target, as well as providing qualitative comments, providing an overall consensus prioritisation score.Table 4Summary of prioritisation of targets.PrioritisationTargetRationale/Validation/tractability/commercial assessmentRecommendations1FGFR4• Downstream target of PAX::FOXO1 fusion protein.• FGFR4 is a transcriptional target of PAX3-FOXO1 in fusion-positive RMS.• In fusion-negative, FGFR4 is highly expressed in nearly all tumours and mutationally activated in approximately 10%.• Overexpression in fusion-negative RMS activates aberrant survival and differentiation signalling cascades• Hotspot mutations (N535K and V550L/M/E) in the tyrosine kinase domain present in 7.5% of primary RMS• Efforts to target N535K led to early resistance• No inhibitors exist for V550L/M/E mutation• CAR T and antibody approaches have shown preclinical promise.• CAR T cells dual-targeting FGFR4 with B7-H3 is an option to be explored.• There is a strong rationale to develop degraders of FGFR4• A combination of degraders and inhibitors is likely to be needed• Evaluation of ADCs is required.2MYOD1• In RMS, MYOD1 is unable to drive transcription of myogenic differentiation genes.• Wild-type MYOD1 is a target in ERMS/ARMS• Mutant MYOD1 is a target in SCRMS.• The L122R mutant MYOD1 has MYC-like properties.• Cancer DepMap data suggests universal dependency of RMS subtypes on wild-type MYOD1.• Indirect evidence suggests restoring MYOD1 activity blocks tumour growth in vitro and in vivo, but no studies have validated that genomic ablation of MYOD1 causes cytotoxicity in RMS models.• No reported academic or commercial programmes pursuing this target.• Recent work shows MYOD1 promotes RMS survival by maintaining cells in a partially differentiated state [105]• Strong consensus that a MYOD1 degrader should be developed.• A new degrader mouse model will demonstrate the effect of MYOD1 knockout during development.3P300/CBP• Known effector of fusion proteins.• Control MYOD1 activation• Epigenetic pathways impact p300 distribution in ERMS, driving oncogenesis.• RMS cells are sensitive to inhibition of the bromodomain and histone acetyltransferase domain of P300/CBP, with preferential sensitivity to bromodomain inhibitors over HAT inhibitors• P300/CBP degraders are in development and show preclinical activity in RMS• P300/CBP inhibitor inobrodib is in Phase 2 trials but has not yet been evaluated in RMS.• Inobrodib should be evaluated in preclinical models of RMS• Therapeutic approaches must consider off-target toxicities.• No additional therapeutic development is recommended given the number of existing therapeutics.4PAX3/7::FOXO1 fusion proteins• Fusions of PAX3 or PAX7 with FOXO1 are found recurrently in a subset of ARMS patients and are associated with poor prognosis• PAX3::FOXO1 is more common and has been extensively validated in cell lines and mouse xenografts.• As fusions of two transcription factors, there are no clear druggable sites.• Multiple attempts to directly target neoantigens immunologically, with mixed results.• Degraders against PAX::FOXO1 are in development and delay growth in preclinical studies.• A subunit vaccine and small-molecule inhibitor are being developed against PAX3::FOXO1.• Small molecule piperacetazine can directly bind to PAX3-FOXO1 and inhibit downstream transcription, possible scaffold for an improved inhibitor.• Lack of complete anti-tumour response with degraders would necessitate a combination approach with chemotherapy.• Intracellular single-domain antibodies are proposed as an alternative approach to directing targeting of PAX3::FOXO1.• Antisense oligonucleotides and CAR T cells were also highlighted, although a lack of class 1 MHC antigens in RMS would require co-treatment with agents that increase MHC expression.5ROR2• Member of the receptor tyrosine kinase orphan receptor family.• Developmental role, generally absent in adult tissue.• Upregulated in a wide range of tumours, including in MYOD1-mutant SCRMS.• ADC and CAR T cell therapies targeting ROR2 have been clinically evaluated for head and neck and renal cancer.• A ROR2-targeted ADC showed in vitro activity in poor-prognosis RMS with L122R mutant MYOD1.• No in vivo investigations in RMS have been undertaken.• Further evaluation of ROR2 is warranted, especially in poor-prognosis RMS and potentially with chemotherapy, which appears to increase activity.6YAP/TAZ• These druggable transcriptional regulators are downstream regulators of the Hippo pathway, which has been implicated in RMS.• Evidence is conflicting, with studies showing evidence in only some RMS subtypes while others suggest relevance to all forms of RMS.• YAP/TAZ gain or loss of function mutations are rare, and amplifications predominate.• Direct modulation of both YAP and TAZ provides some validation for their potential in RMS.• YAP inhibition inhibited cell growth in ERMS cell lines and xenografts, while inducing YAP caused ERMS development in mice.• TAZ suppression prevented proliferation of ARMS cell lines and blocked tumour growth in mouse xenograft models.• YAP/TAZ mediate transcription via TEAD proteins. TEAD inhibitors are in clinical development for mesothelioma, and TEAD degraders are available.• A combination approach of a Notch inhibitor, which upregulates YAP, and a TEAD degrader may be more effective than TEAD degraders alone.• In vitro and in vivo evaluation of TEAD degraders is required to pursue this approach.RMS rhabdomyosarcoma, SCRMS spindle cell/sclerosing rhabdomyosarcoma, ADC antibody-drug conjugates. Table 4Summary of prioritisation of targets.PrioritisationTargetRationale/Validation/tractability/commercial assessmentRecommendations1FGFR4• Downstream target of PAX::FOXO1 fusion protein.• FGFR4 is a transcriptional target of PAX3-FOXO1 in fusion-positive RMS.• In fusion-negative, FGFR4 is highly expressed in nearly all tumours and mutationally activated in approximately 10%.• Overexpression in fusion-negative RMS activates aberrant survival and differentiation signalling cascades• Hotspot mutations (N535K and V550L/M/E) in the tyrosine kinase domain present in 7.5% of primary RMS• Efforts to target N535K led to early resistance• No inhibitors exist for V550L/M/E mutation• CAR T and antibody approaches have shown preclinical promise.• CAR T cells dual-targeting FGFR4 with B7-H3 is an option to be explored.• There is a strong rationale to develop degraders of FGFR4• A combination of degraders and inhibitors is likely to be needed• Evaluation of ADCs is required.2MYOD1• In RMS, MYOD1 is unable to drive transcription of myogenic differentiation genes.• Wild-type MYOD1 is a target in ERMS/ARMS• Mutant MYOD1 is a target in SCRMS.• The L122R mutant MYOD1 has MYC-like properties.• Cancer DepMap data suggests universal dependency of RMS subtypes on wild-type MYOD1.• Indirect evidence suggests restoring MYOD1 activity blocks tumour growth in vitro and in vivo, but no studies have validated that genomic ablation of MYOD1 causes cytotoxicity in RMS models.• No reported academic or commercial programmes pursuing this target.• Recent work shows MYOD1 promotes RMS survival by maintaining cells in a partially differentiated state [105]• Strong consensus that a MYOD1 degrader should be developed.• A new degrader mouse model will demonstrate the effect of MYOD1 knockout during development.3P300/CBP• Known effector of fusion proteins.• Control MYOD1 activation• Epigenetic pathways impact p300 distribution in ERMS, driving oncogenesis.• RMS cells are sensitive to inhibition of the bromodomain and histone acetyltransferase domain of P300/CBP, with preferential sensitivity to bromodomain inhibitors over HAT inhibitors• P300/CBP degraders are in development and show preclinical activity in RMS• P300/CBP inhibitor inobrodib is in Phase 2 trials but has not yet been evaluated in RMS.• Inobrodib should be evaluated in preclinical models of RMS• Therapeutic approaches must consider off-target toxicities.• No additional therapeutic development is recommended given the number of existing therapeutics.4PAX3/7::FOXO1 fusion proteins• Fusions of PAX3 or PAX7 with FOXO1 are found recurrently in a subset of ARMS patients and are associated with poor prognosis• PAX3::FOXO1 is more common and has been extensively validated in cell lines and mouse xenografts.• As fusions of two transcription factors, there are no clear druggable sites.• Multiple attempts to directly target neoantigens immunologically, with mixed results.• Degraders against PAX::FOXO1 are in development and delay growth in preclinical studies.• A subunit vaccine and small-molecule inhibitor are being developed against PAX3::FOXO1.• Small molecule piperacetazine can directly bind to PAX3-FOXO1 and inhibit downstream transcription, possible scaffold for an improved inhibitor.• Lack of complete anti-tumour response with degraders would necessitate a combination approach with chemotherapy.• Intracellular single-domain antibodies are proposed as an alternative approach to directing targeting of PAX3::FOXO1.• Antisense oligonucleotides and CAR T cells were also highlighted, although a lack of class 1 MHC antigens in RMS would require co-treatment with agents that increase MHC expression.5ROR2• Member of the receptor tyrosine kinase orphan receptor family.• Developmental role, generally absent in adult tissue.• Upregulated in a wide range of tumours, including in MYOD1-mutant SCRMS.• ADC and CAR T cell therapies targeting ROR2 have been clinically evaluated for head and neck and renal cancer.• A ROR2-targeted ADC showed in vitro activity in poor-prognosis RMS with L122R mutant MYOD1.• No in vivo investigations in RMS have been undertaken.• Further evaluation of ROR2 is warranted, especially in poor-prognosis RMS and potentially with chemotherapy, which appears to increase activity.6YAP/TAZ• These druggable transcriptional regulators are downstream regulators of the Hippo pathway, which has been implicated in RMS.• Evidence is conflicting, with studies showing evidence in only some RMS subtypes while others suggest relevance to all forms of RMS.• YAP/TAZ gain or loss of function mutations are rare, and amplifications predominate.• Direct modulation of both YAP and TAZ provides some validation for their potential in RMS.• YAP inhibition inhibited cell growth in ERMS cell lines and xenografts, while inducing YAP caused ERMS development in mice.• TAZ suppression prevented proliferation of ARMS cell lines and blocked tumour growth in mouse xenograft models.• YAP/TAZ mediate transcription via TEAD proteins. TEAD inhibitors are in clinical development for mesothelioma, and TEAD degraders are available.• A combination approach of a Notch inhibitor, which upregulates YAP, and a TEAD degrader may be more effective than TEAD degraders alone.• In vitro and in vivo evaluation of TEAD degraders is required to pursue this approach.RMS rhabdomyosarcoma, SCRMS spindle cell/sclerosing rhabdomyosarcoma, ADC antibody-drug conjugates. Summary of prioritisation of targets. • Downstream target of PAX::FOXO1 fusion protein. • FGFR4 is a transcriptional target of PAX3-FOXO1 in fusion-positive RMS. • In fusion-negative, FGFR4 is highly expressed in nearly all tumours and mutationally activated in approximately 10%. • Overexpression in fusion-negative RMS activates aberrant survival and differentiation signalling cascades • Hotspot mutations (N535K and V550L/M/E) in the tyrosine kinase domain present in 7.5% of primary RMS • Efforts to target N535K led to early resistance • No inhibitors exist for V550L/M/E mutation • CAR T and antibody approaches have shown preclinical promise. • CAR T cells dual-targeting FGFR4 with B7-H3 is an option to be explored. • There is a strong rationale to develop degraders of FGFR4 • A combination of degraders and inhibitors is likely to be needed • Evaluation of ADCs is required. • In RMS, MYOD1 is unable to drive transcription of myogenic differentiation genes. • Wild-type MYOD1 is a target in ERMS/ARMS • Mutant MYOD1 is a target in SCRMS. • The L122R mutant MYOD1 has MYC-like properties. • Cancer DepMap data suggests universal dependency of RMS subtypes on wild-type MYOD1. • Indirect evidence suggests restoring MYOD1 activity blocks tumour growth in vitro and in vivo, but no studies have validated that genomic ablation of MYOD1 causes cytotoxicity in RMS models. • No reported academic or commercial programmes pursuing this target. • Recent work shows MYOD1 promotes RMS survival by maintaining cells in a partially differentiated state [105] • Strong consensus that a MYOD1 degrader should be developed. • A new degrader mouse model will demonstrate the effect of MYOD1 knockout during development. • Known effector of fusion proteins. • Control MYOD1 activation • Epigenetic pathways impact p300 distribution in ERMS, driving oncogenesis. • RMS cells are sensitive to inhibition of the bromodomain and histone acetyltransferase domain of P300/CBP, with preferential sensitivity to bromodomain inhibitors over HAT inhibitors • P300/CBP degraders are in development and show preclinical activity in RMS • P300/CBP inhibitor inobrodib is in Phase 2 trials but has not yet been evaluated in RMS. • Inobrodib should be evaluated in preclinical models of RMS • Therapeutic approaches must consider off-target toxicities. • No additional therapeutic development is recommended given the number of existing therapeutics. • Fusions of PAX3 or PAX7 with FOXO1 are found recurrently in a subset of ARMS patients and are associated with poor prognosis • PAX3::FOXO1 is more common and has been extensively validated in cell lines and mouse xenografts. • As fusions of two transcription factors, there are no clear druggable sites. • Multiple attempts to directly target neoantigens immunologically, with mixed results. • Degraders against PAX::FOXO1 are in development and delay growth in preclinical studies. • A subunit vaccine and small-molecule inhibitor are being developed against PAX3::FOXO1. • Small molecule piperacetazine can directly bind to PAX3-FOXO1 and inhibit downstream transcription, possible scaffold for an improved inhibitor. • Lack of complete anti-tumour response with degraders would necessitate a combination approach with chemotherapy. • Intracellular single-domain antibodies are proposed as an alternative approach to directing targeting of PAX3::FOXO1. • Antisense oligonucleotides and CAR T cells were also highlighted, although a lack of class 1 MHC antigens in RMS would require co-treatment with agents that increase MHC expression. • Member of the receptor tyrosine kinase orphan receptor family. • Developmental role, generally absent in adult tissue. • Upregulated in a wide range of tumours, including in MYOD1-mutant SCRMS. • ADC and CAR T cell therapies targeting ROR2 have been clinically evaluated for head and neck and renal cancer. • A ROR2-targeted ADC showed in vitro activity in poor-prognosis RMS with L122R mutant MYOD1. • No in vivo investigations in RMS have been undertaken. • These druggable transcriptional regulators are downstream regulators of the Hippo pathway, which has been implicated in RMS. • Evidence is conflicting, with studies showing evidence in only some RMS subtypes while others suggest relevance to all forms of RMS. • YAP/TAZ gain or loss of function mutations are rare, and amplifications predominate. • Direct modulation of both YAP and TAZ provides some validation for their potential in RMS. • YAP inhibition inhibited cell growth in ERMS cell lines and xenografts, while inducing YAP caused ERMS development in mice. • TAZ suppression prevented proliferation of ARMS cell lines and blocked tumour growth in mouse xenograft models. • YAP/TAZ mediate transcription via TEAD proteins. TEAD inhibitors are in clinical development for mesothelioma, and TEAD degraders are available. • A combination approach of a Notch inhibitor, which upregulates YAP, and a TEAD degrader may be more effective than TEAD degraders alone. • In vitro and in vivo evaluation of TEAD degraders is required to pursue this approach. RMS rhabdomyosarcoma, SCRMS spindle cell/sclerosing rhabdomyosarcoma, ADC antibody-drug conjugates. Prioritised targets discussed at the workshop Fibroblast growth factor receptor 4 (FGFR4) FGFR4 is a downstream transcriptional target of PAX::FOXO1 fusion proteins and is also implicated in PAX3/7::FOXO1-negative rhabdomyosarcoma, where overexpression activates signalling cascades associated with aberrant cell differentiation and survival [81–83]. One study confirmed high FGFR4 expression in primary human rhabdomyosarcoma tumours compared with an independent panel of paediatric tumours and normal tissues and found overexpression was associated with advanced-stage cancer and poor survival [84]. Another study found overexpression in both ARMS and ERMS, but higher FGFR4 levels in ARMS [82]. Rhabdomyosarcoma primary tumours also bear somatic mutations in FGFR4 [84, 85]. Two hotspot mutations (N535K and V550L/M/E) in the tyrosine kinase domain have been identified, which lead to autoactivation of FGFR4 [85]. These mutations are present in up to 7.5% of primary rhabdomyosarcoma tumours (of both ARMS and ERMS subtypes) and are associated with greater metastatic potential and poorer overall survival compared with tumours where FGFR4 is overexpressed in mice [84]. Efforts to target the N535K mutation have led to early drug resistance, and there are currently no specific inhibitors available for the V550L/M/E mutation [83]. Chimeric antigen receptor (CAR) T-cell therapies targeting FGFR4 are being evaluated in clinical trials (NCT06865664), and the results are awaited with great interest, although there may be on-target, off-tumour effects [81, 86–88]. Antibody-drug conjugates (ADCs) targeting FGFR4 have also demonstrated substantial preclinical activity but are earlier in development [89]. While CAR T-cell and ADC therapies targeting FGFR4 are being developed, other options should be explored. These other approaches include dual targeting approaches with B7-H3 (CD276) [88, 90], a cell surface antigen widely expressed on most rhabdomyosarcoma tumours and discussed in detail later [91, 92]. A rationale exists to develop degraders to FGFR4, as these degraders could target both enzymatic and scaffolding functions and might avoid the development of drug resistance. However, degraders were thought unlikely to succeed as monotherapy, and a combination approach of degraders and inhibitors is likely to be needed. Insulin-like growth factor 1 receptor (IGF1R)-directed agents represent a synergistic combination with FGFR4-directed strategies in rhabdomyosarcoma. In relapsed/refractory rhabdomyosarcoma, single-agent IGF1R antibodies such as R1507 and cixutumumab produced occasional objective responses but overall limited and transient activity, establishing both clinical tractability and the insufficiency of monotherapy in this disease [93, 94]. In fusion-positive rhabdomyosarcoma, PAX3-FOXO1 directly regulates both FGFR4 and IGF1R, placing these receptors within a shared oncogenic transcriptional programme [66]. Consistent with this linkage, FGFR4 signalling rescues subsets of ARMS cells from apoptosis induced by compounds targeting the IGF1R-PI3K-mTOR pathway, providing a mechanistic basis for additive or synergistic benefit from dual pathway suppression in selected tumours [95]. This combination rationale is further strengthened by evidence that AKT/mTOR is a key survival node downstream of IGF1R in RMS, that PI3K or mTOR inhibition sensitises IGF1R-antibody-resistant rhabdomyosarcoma cells, and that FGFR4-driven rhabdomyosarcoma models are highly sensitive to PI3K/mTOR inhibition with suppression of phospho-AKT and tumour growth [85, 96, 97]. More broadly, IGF1R blockade has shown synergy with MEK inhibition in RAS-mutant rhabdomyosarcoma models, further supporting combination-based deployment of this target class rather than monotherapy [98]. MYOD1 MYOD1 can be targeted as (i) the wildtype in ERMS/ARMS, and (ii) the mutant MYOD1L122R in SCRMS. The typical physiological function of MYOD1 is as a myogenic transcription factor, promoting the differentiation and cell cycle exit of myogenic precursor cells, but in rhabdomyosarcoma, its function is perturbed, and it is unable to drive transcription of key myogenic differentiation genes, leading to failure to differentiate [66, 99–103]. In ERMS and ARMS, this perturbation occurs without mutation, whereas in SCRMS and some PAX3/7::FOXO1-negative rhabdomyosarcoma [39], MYOD1 mutations occur (exclusively on exon 1, in a hotspot location mutation L122R) and are associated with very poor prognosis [99]. The L122R mutant MYOD1 has MYC-like properties [99] and is present in 40% of SCRMS [17]. RAS and PI3K mutations frequently, as well as other aberrations, co-occur with MYOD1 mutations, and these are believed to function with mutant MYOD1 to drive tumour progression synergistically [14, 99]. Importantly, data from the Cancer Dependency Map (DepMap) shows strong selective dependency on wildtype MYOD1 across different subtypes of rhabdomyosarcoma cell lines when targeted using CRISPR, suggesting universal dependency on this target [104]. Little experimental validation exists for directly targeting mutant or wild-type MYOD1 in rhabdomyosarcoma. For wildtype MYOD1, de-differentiation paradoxically makes cells more vulnerable to stress, and there is a benefit from enhancing with chemotherapy [101, 102, 105–109]. MYF5 and MYOD1 are required for human ERMS xenograft growth [103], and there is a strong association of loss of MYOD1 resulting in cell death in rhabdomyosarcoma [104]. Overall, considering a direct MYOD1-targeted approach, there was a strong consensus among workshop participants that a MYOD1 degrader should be developed, as there are no reported academic or commercial programmes pursuing this target. The concerns regarding potential toxicity were partially allayed by observations that MyoD1 knockout mice do not display any somatic effects, and there is redundancy in the MyoD1 pathway, specifically with Myf5. This will be further informed by a degrader mouse model that is in development and will demonstrate the effect of MyoD1 KO at different developmental stages. ROR2 Normal skeletal muscle development is partially driven by canonical WNT/β-catenin signalling. Whilst activating mutations in CTNNB1 in patient samples of rhabdomyosarcoma [100] support involvement of WNT signalling, other evidence has indicated only a minor role for β-catenin and canonical WNT signalling in rhabdomyosarcoma growth or differentiation [110, 111]. ROR2 is a receptor tyrosine kinase-like type 1 transmembrane receptor involved in non-canonical WNT signalling via WNT5A and regulation of cell polarity, proliferation and migration [112]. Despite being often referred to as a pseudokinase, ROR2 can homodimerize and autophosphorylate when bound by WNT5A [113, 114]. Its other binding partners include Frizzled and ROR1, which induce other WNT signalling pathways [115]. There are a few publications linking ROR2 with rhabdomyosarcoma, but ROR2 is implicated in several other [116, 117] studies, and studies have shown that upregulated WNT5A/ROR2 signalling promotes cancer invasiveness and stemness [118–120]. Moreover, evidence suggests ROR2 signalling drives chemotherapy resistance in ovarian cancer [121]. ADC and CAR T-cell treatments targeting ROR2 have been evaluated in the clinic for head and neck and renal cancer [122–124]. A ROR2-targeted ADC has shown in vitro activity against very poor prognosis rhabdomyosarcoma with L122R-mutant MYOD1 [125], but no in vivo investigations have been undertaken to date. Workshop participants agreed that further evaluation of ROR2, including ADCs, is warranted, especially in poor-prognosis rhabdomyosarcoma, and potentially in combination with chemotherapy, which appears to increase activity. P300/CBP P300 and CBP are ubiquitously expressed multi-domain paralogous proteins which interact with numerous transcription factors and act to increase the expression of their target genes [126]. In PAX3::FOXO1 fusion-positive tumours, P300/CBP is known to be recruited by the fusion protein to shape downstream transcriptional oncogenic pathways and include important roles such as controlling activation of MYOD1 [127]. Similarly, alternative fusion proteins in SCRMS have been proposed to drive tumorigenesis through recruitment of P300 [128], while in ERMS, epigenetic pathways impact on P300 distribution, driving oncogenesis through inhibition of WNT signalling pathways [129]. Rhabdomyosarcoma cells have been shown to be sensitive to inhibition of both the bromodomain and histone acetyltransferase domain of P300/CBP [130], and degraders of P300 and CBP have been evaluated in rhabdomyosarcoma and show signs of preclinical activity [127]. The most clinically advanced P300/CBP inhibitor is inobrodib, which is in Phase II trials in adults with haematological cancers [131]. Inobrodib has not yet been evaluated in rhabdomyosarcoma and should be explored in preclinical models alongside other developed inhibitors and degraders with a longer half-life. The numerous P300/CBP degraders in development for oncology, although not specifically for rhabdomyosarcoma, could be of great interest for rhabdomyosarcoma. As CBP and P300 are both ubiquitously expressed at RNA and protein levels, therapeutic approaches should consider possible off-target toxicities in healthy tissue, although CRISPR knockout of either P300 or CBP appears to be well tolerated across cancer cell lines [132]. Given the number and breadth of existing therapeutics targeting these proteins, engagement with existing pipelines was encouraged, and no additional academic de novo drug development was recommended at this stage. PAX3::FOXO1 and PAX7::FOXO1 fusion genes The PAX3::FOXO1 and PAX7::FOXO1 fusions generate potent transcription factors with similar but distinct downstream transcription profiles, which include pro-oncogenic pathways [1, 67, 133]. PAX3::FOXO1 is more common [51] and has been the subject of extensive target validation studies using antisense oligonucleotides [134], siRNA, shRNA [135, 136], small molecules [137–139], CRISPR [126, 140] and degradation [141] in cell lines and mouse xenografts. Based on extensive validation, the PAX3::FOXO1 fusion protein is a very high-priority yet challenging target. As fusions of two transcription factors, PAX3/7::FOXO1 are nuclear proteins with no clear druggable sites for inhibitors. However, PAX3/7::FOXO1 is amenable to targeting by other modalities. PAX3::FOXO1 degraders are in active preclinical development, for example, by the KOODAC Cancer Grand Challenge and the NIH FusOnC2 and TFCC Consortia. In preclinical studies, degraders result in growth delay and differentiation in cell lines with one or many PAX3/7::FOXO1 loci tagged with FK506-Binding Proteins (FKBPs) [141], but as they are unlikely to provide a complete response as monotherapy, they would need to be used in combination in the clinic. A subunit vaccine [142] and small molecule inhibitor are also being developed against the PAX3::FOXO1 fusion protein [138], but no clinical-grade therapeutic has yet been developed. An alternative approach to direct targeting of PAX3::FOXO1 discussed at the workshop is the use of intracellular single domain antibodies (iDAbs), which would allow the fusion protein to be targeted with bispecific antibodies. iDAbs can be conjugated to any cytotoxic warhead and delivered by a nanoparticle or virus, although such approaches have not yet been used in the clinic. Other approaches discussed included molecular glues, antisense oligonucleotides (ASOs) and T cell receptor (TCR) T-cells. One issue with TCR T-cell therapy would be the lack of class 1 major histocompatibility complex (MHC) antigen presentation in fusion-positive rhabdomyosarcoma [143, 144], but several agents could potentially be used to increase MHC antigen expression, such as HDAC inhibitors [145, 146]. YAP/TAZ YAP and TAZ are two closely related, druggable transcriptional regulators encoded by the YAP1 and WWTR1 genes, respectively, that sense how cells perceive their shape, polarity and tissue environment and communicate accordingly with the surrounding extracellular matrix [147]. YAP and TAZ are downstream regulators in the Hippo pathway, which has been implicated in rhabdomyosarcoma, although studies have been somewhat contradictory, with one study showing increased expression in ERMS but not ARMS [148], and others showing expression in both [149, 150] and relevance to all forms of rhabdomyosarcoma [151]. Across different cancers, gain or loss of function mutations in YAP/TAZ are rare, and amplifications predominate [152, 153]. Moreover, in SCRMS, fusion genes involving the downstream transcription factor TEAD1 (specifically TEAD1::NCOA2) have been observed [154]. Direct modulation of both YAP and TAZ has provided some validation for their potential as targets in rhabdomyosarcoma. YAP suppression with shRNA inhibited cell growth and induced apoptosis in human ERMS cell lines and ERMS xenografts [155]. Inducing YAP expression caused ERMS development in mice while knockdown in ERMS cells decreased tumour proliferation and encouraged differentiation [148]. In a PAX::FOXO1-upregulated mouse model, RNAi-mediated suppression of TAZ prevented ARMS proliferation and induced apoptosis in vitro, while in murine xenograft models, TAZ suppression blocked tumour growth and prolonged survival [149]. Following nuclear translocation, YAP/TAZ regulate gene transcription through association with the DNA-binding TEAD 1–4 proteins [147]. A small molecule that targets the hydrophobic pocket of TEADs is being investigated in a Phase I/II study in adults with mesothelioma and metastatic solid tumours with NF2 mutations [156]. TEAD1 and 4 degraders are available and may be more effective. Moreover, a combination approach of a Notch inhibitor and TEAD degrader could, in theory, be effective for all rhabdomyosarcoma subtypes, on the basis that a Notch inhibitor could prevent activation of Notch transcription by the Hippo signalling pathway [157, 158]. Specifically, Notch signalling transcriptionally upregulates YAP1 gene expression and YAP activity [159], while YAP, in turn, transcriptionally upregulates the Notch ligand genes JAG1 and DLL1 and the core Notch transcription factor RBPJ [159]. This approach could be valuable for intermediate risk rhabdomyosarcoma and could reduce the toxicity of current treatments. In vitro and in vivo evaluation of TEAD degraders is required to pursue this approach. MYCN In addition to the top six prioritised targets, targeting MYCN should be a high priority in PAX3/7::FOXO1 rhabdomyosarcoma. The challenges of directly therapeutically targeting MYCN have been discussed extensively in an earlier Paediatric Therapeutic Development Workshop focused on medulloblastoma, and investigation of the effectiveness of agents targeting MYCN dependencies in rhabdomyosarcoma models is also required. Cell surface targets Like many paediatric cancers, rhabdomyosarcoma has a low mutational burden, resulting in few targetable neoantigens [39] and is immunologically cold. Checkpoint blocking antibodies (including PD-L1 blockade) in rhabdomyosarcoma have shown little clinical benefit [160–162], and descriptions of expression of PD-L1 in rhabdomyosarcoma are contradictory, reporting either minimal expression [163] or higher levels, which are associated with poor overall survival [164]. There are several potential cell surface targets: IGF1R, the cell surface gangliosides GD2 and GD3, glypican GPC2, GPC3, human epidermal growth factor receptor 2 and 3 (HER2/3), epidermal growth factor receptor (EGFR), EpHA2 [165], B7-H3 (CD276) and FGFR4 (discussed previously). Tumour target expression heterogeneity, normal tissue expression profile, preclinical evidence and clinical experience guide the development of therapeutics for these cell surface targets in rhabdomyosarcoma. IGF signalling is implicated in the growth, proliferation and metastasis of rhabdomyosarcoma [166]. IGF1R-targeted antibodies have been evaluated in several early clinical studies, with an aggregate overall response rate in rhabdomyosarcoma ranging from 0-11% [93, 94, 167–171], despite a wealth of supporting preclinical data [172–176]. Multiple effective preclinical combinations with IGF1R targeting agents have [177, 178] been reported in rhabdomyosarcoma models, including with small molecule inhibitors, as well as dual-targeting CAR T-cells [150]. CAR T-cells targeting IGF1R and tyrosine kinase-like orphan receptor 1 (ROR1), both of which are strongly expressed in ARMS and ERMS, have been shown to significantly reduce tumour growth in xenograft models [178]. Targeting IGF1R by CAR T-cells may have several advantages over antibodies as they: (i). directly and specifically kill IGF1R-positive cells even at a low level of surface expression; (ii) produce TNF-α and IFN-γ cytokines; and (iii) they persist for months and years. Antibody-based therapies targeting GD2 and GD3, which are expressed by many tumours, including rhabdomyosarcoma, are showing preclinical promise, with studies ongoing [179]. GPC2 is a recently described surface antigen expressed on childhood solid tumours, including a subgroup of rhabdomyosarcoma. An ADC targeting GPC2 is in development, although there is no preclinical data in rhabdomyosarcoma [180]. HER2 is another potential target for rhabdomyosarcoma despite heterogeneous expression levels and initial safety concerns [181]. HER2-directed CAR T-cell therapy is now being applied safely and has shown some evidence of activity in patients with sarcomas, including rhabdomyosarcoma [182, 183]. Also, ADCs against HER2 have been effective in treating advanced breast cancer and may be effective for rhabdomyosarcoma. A related approach is patritumab deruxtecan (HER3-DXd), an investigational HER3-targeted ADC [184]. B7-H3 (CD276) is a cell surface antigen widely expressed on a majority of rhabdomyosarcoma tumours [92, 185] and positively regulated by the PAX3::FOXO1 fusion protein [186, 187]. CAR T-cells targeting B7-H3 have shown significant anti-tumour activity in in vivo orthotopic models of paediatric solid tumours, with some reaching regression or complete response [90]. A threshold level of expression was required for a complete response. In a Phase I study in paediatric, teenage and young adult patients with solid tumours (two with rhabdomyosarcoma), B7-H3-targeted CAR T-cell therapy was well tolerated [188]. Other Phase I trials with different B7-H3 CAR T-cell designs are ongoing or in preparation [189–191], and studies with ADCs targeting B7-H3 are in clinical trials in adults [187, 192], but have not yet been evaluated in rhabdomyosarcoma, although there is supportive preclinical data [193]. Bicistronic CAR T-cells targeting FGFR4 and B7-H3 have shown potent activity against rhabdomyosarcoma [88]. Based on workshop discussion of these data, B7-H3 fulfils the key requirements as priority targets for further exploration in rhabdomyosarcoma. Drug combination strategies Before the workshop, participants identified the following inhibitors as priorities to be evaluated in the relapsed rhabdomyosarcoma setting, in combination with radio/chemotherapy (ranked): mitogen-activated protein kinase (MEK) and mitogen-activated protein kinase (MAPK), polo-like kinase 1 (PLK1), DNA damage response pathways (e.g. poly(ADP-ribose) polymerase [PARP]), HDACs, enhancer of zeste homologue 2 (EZH2), apoptotic pathways, PI3K and cyclin-dependent kinases 4 and 6 (CDK4/6). HDAC inhibitors have been shown to have preclinical activity in in vitro rhabdomyosarcoma cell line models and in in vivo rhabdomyosarcoma xenograft models. The HDAC inhibitor entinostat targeting HDAC1, 2, 3 and 11 demonstrated synergistic activity in PAX3::FOXO1 ARMS mouse models with vinorelbine [194]. Early results of an early phase trial of entinostat and vinorelbine in relapsed or refractory rhabdomyosarcoma demonstrated a PR in 4 of 8 patients [195]. The rationale for MEK inhibition is to target the cell population that escapes standard chemotherapy and to deliver a MEK inhibitor during maintenance therapy. There is strong preclinical support for this approach: MEK inhibition with trametinib promotes late myogenic differentiation via MYOG and, in combination with an IGF1R inhibitor, trametinib potently decreases rhabdomyosarcoma cell viability and slows tumour growth in xenograft models [196]. The EpSSG are currently exploring the option to add a MEK inhibitor to standard maintenance therapy. Administering a MEK inhibitor with either a BRAF inhibitor or a PI3K inhibitor may be more effective, based on potential pathway crosstalk and the fact that CRAF activation limits the activity of single-agent MEK inhibition [197, 198], potentially considering stratification by TP53 mutation status. The issue of toxicity (typically haematological or mucositis) is substantial when a targeted agent is either combined with another targeted agent (novel-novel) or with chemotherapy; an intermittent schedule may overcome this toxicity. Combining a targeted agent with chemotherapy may be the easiest route to the clinic. It was agreed that knowledge of the biology of refractory disease is crucial as resistance mechanisms need to be overcome. Trial designs Targeting IGF1R by CAR T-cells may have several advantages over antibodies as they: (i). directly and specifically kill IGF1R-positive cells even at a low level of surface expression; (ii) produce TNF-α and IFN-γ cytokines; and (iii) they persist for months and years. The consensus of participants was that PAX3/7::FOXO1-positive rhabdomyosarcoma should be considered separately from PAX3/7::FOXO1-negative disease, and any predictive biomarkers should be considered. There is a challenge in identifying the best trial design in view of the small number of patients and the heterogeneous nature of rhabdomyosarcoma with multiple molecular subgroups. Novel agents have traditionally been introduced in the relapse setting, but where a high-risk population can be defined, where survival with front-line therapy is very poor – e.g. PAX3/7::FOXO1-positive metastatic disease in bone and bone marrow, very similar to diffuse midline glioma – novel agents could be evaluated upfront. The EpSSG Frontline and Relapsed Rhabdomyosarcoma (FaR-RMS) trial [19, 199] incorporates PAX3/7::FOXO1 status instead of histology into risk stratification and has several benefits that enable evaluation of frontline and relapse therapy and its multi-arm, multi-stage design provides scope and capacity to evaluate novel agents in the protocol [19]. Once a new therapeutic is developed against one of the high-priority targets, an early phase study in children, determining the recommended Phase II dose, pharmacokinetics and safety profile, would be undertaken [200]. If there is a signal of activity, the therapy could then progress into FaR-RMS for relapse patients and, if there is a further positive signal, into the frontline. However, many existing drugs that are high priority for investigation by academic cooperative groups prove to be difficult to access for clinical trials from pharmaceutical companies. This is often due to them being relatively late in their (adult) development lifecycle, and there is no regulatory incentive for companies to undertake these additional clinical trials. Overall, there is a paucity of drugs that can be evaluated in the relapse setting owing to insufficient preclinical evidence, the lack of preclinical data translating efficiently to new treatment approaches for patients, and challenges accessing therapeutics from pharmaceutical companies for preclinical and clinical evaluation. However, new approaches must be developed to accelerate the introduction of innovative treatments. Conclusion Although survival rates for children with localised, low-risk rhabdomyosarcoma have improved, survival remains at 20–30% for patients with metastatic disease, and all children with rhabdomyosarcoma still face intensive treatment with poor short- and long-term quality of life. There is an urgent need for new therapeutic options for metastatic and recurrent disease, and for treatments with fewer side effects. To make progress, both new therapeutics need to be developed and existing drugs used optimally, and in view of the different biology, evaluating novel agents in PAX3/7::FOXO1-positive and PAX3/7::FOXO1-negative cohorts should be considered. FGFR4 was considered the highest-priority target. Currently, an FGFR4-targeted CAR T-cell (NCT06865664) is underway, and initiatives to develop ADCs against FGFR4 are ongoing; clinical evaluation of these is awaited with great interest. Developing an FGFR4 degrader is a high priority because it could be superior to a specific inhibitor targeting the V550L/M/E mutation. However, with either a degrader or an inhibitor, a combination strategy, such as with an IGF1R inhibitor, may be required. Directly targeting PAX3::FOXO1 and PAX7::FOXO1 fusion proteins is a high priority but demanding, and this is ongoing by the Cancer Grand Challenge and NIH-based groups. Depending on the results of in vivo studies, a MYOD1 degrader may also be a priority, and should be evaluated in preclinical rhabdomyosarcoma models. Degraders of P300/CBP are available (inobrodib), and newer compounds currently in development have a longer half-life [201], and these should be evaluated preclinically with a view to clinical investigation. ROR2 is a very interesting target for L122R mutant MYOD1 rhabdomyosarcoma, but the development of therapeutics is in early stages. To target YAP/TAZ, development of a TEAD degrader is a high priority, and this should be evaluated in combination with a Notch inhibitor. Existing ADCs targeting B7-H3 have been extensively evaluated in adults, and their evaluation in rhabdomyosarcoma should be expedited. ADCs targeting FGFR4 that show promising preclinical activity should be moved to clinical trials. New therapeutic agents should undergo rapid evaluation in an early phase trial, with a dose finding or confirmation cohort and then an expansion cohort. Defining early go or no-go decisions is crucial. A coordinated global development, with international trans-Atlantic trials, is mandatory to avoid duplication and accelerate regulatory approval. MEK inhibitors, combined with BRAF or PI3K inhibition, should be evaluated in combination with chemotherapy in the maintenance setting in PAX3/7::FOXO1 negative rhabdomyosarcoma. Understanding the biology to underpin the rational combination of these agents will be essential. Pursuing these avenues should present opportunities to improve survival in high-risk patients with rhabdomyosarcoma. Supplementary information Supplementary Tables Supplementary Tables