YAP1::MAML2 fusion, a newly identified genetic anomaly in a posterior fossa infant tumor, associated with the methylation class “embryonal tumor with multilayered rosettes, non-C19MC-altered” in the Heidelberg central nervous system tumor classifier and “embryonal tumor with multilayered rosettes-DICER” in the NIH classifier Authors: Catherine Godfraind, Nabil Djebbour, Romain Appay, Anne Pagnier, Raia Doumit, Julien Masliah-Planchon, Fabien Forest, Fanny Burel-Vandenbos, Jean Boutonnat, Berengere Dadone-Montaudié Published: 2025 ABSTRACT AbstractEmbryonal tumors with multilayered rosettes (ETMR) are rare embryonal tumors that usually affect children under two years old. They are characterized histologically by the presence of multilayered rosettes and by the immunohistochemical expression of LIN28A. Their genetic hallmarks include C19MC amplification, which is most common, followed by DICER1 mutation. Each of these alterations correlates with a specific methylation class in the Heidelberg central nervous system tumor classifier and the National Institutes of Health brain tumor classifier. Meanwhile, 2.5 percent of LIN28A-positive embryonal tumors lack the previously mentioned genetic alterations associated with ETMR. Here, we present and discuss a case of this type. A posterior fossa tumor was found in a five-month-old infant. Histologically, the lesion appeared as an embryonal tumor, lacking multilayered rosettes but showing focal positivity for LIN28A. It did not show C19MC amplification or DICER1 mutation, yet it clustered within the ETMR non-C19MC-altered methylation class of the Heidelberg classifier (V12.5) and the ETMR-DICER class of the NIH classifier. Additionally, a YAP1::MAML2 fusion was identified, a finding not yet associated with these methylation classes.Keywords: ETMR, YAP1::MALM2 fusion, DICER1 mutation, ETMR non-C19MC methylation class, ETMR DICER1-altered Introduction Embryonal tumors with multilayered rosettes (ETMR) are a rare type of tumor that usually affects the cerebral hemisphere of infants under two years of age. The entity was introduced in the 2016 WHO update of Central Nervous System tumors (CNS [1],). ETMR results from a combination of previously known tumors: embryonal tumor with abundant neuropil and true rosettes (ETANTR), ependymoblastoma (EBL), and medulloepithelioma (MEPL), as they share common markers: LIN28 immunohistochemical expression [2] and, for most, a genomic alteration first identified in ETANTR, which is the amplification of the C19MC miRNA cluster located at 19q13.42 ([3, 4]). The 2021 World Health Organization Central Nervous System classification identifies two distinct subtypes of ETMR [5]: the ETMR C19MC-altered, characterized by amplification of the C19MC locus, and the ETMR DICER1-altered, which usually has a DICER1 mutation, linked or not to a hereditary cancer syndrome ([6, 7]) or, in rarer cases, an amplification of the miR-17–92 cluster (MIR17HG) or a remodeling of the C19MC genomic region [8]. Each subtype has its methylation class (MC) within the Heidelberg Central Nervous System tumor (CNS) classifier [9]: the MC Embryonal tumor with multilayered rosettes, C19MC-altered, and the MC Embryonal tumor with multilayered rosettes, non-C19MC-altered, as well as in the National Institutes of Health (NIH) classifier (Methylscape analysis, [10]): the MC Embryonal tumor with multilayered rosettes, C19MC-altered, and the MC Embryonal tumor with multilayered rosettes, DICER. However, some embryonal tumors expressing LIN28A lack the reported genomic alterations, except for the methylation profile associated with the MC embryonal tumor with multilayered rosettes, non-C19MC-altered, according to the Heidelberg CNS tumor classifier (Table 1). Here, we describe such a tumor in a five-month-old girl. The lesion was located in the posterior fossa and exhibited an embryonal appearance similar to EBL, with LIN28A protein expression. No multilayered rosettes, C19MC amplification, DICER1 mutation, or MIR17HG amplification were detected. However, a YAP1::MAML2 fusion and an LZTR1 mutation were present; the significance of the latter remains uncertain. This is the first report of a YAP1 fusion in the MC embryonal tumor with multilayered rosettes, non-C19MC-altered, as classified by the Heidelberg CNS tumor classifier (V12.5) [9], and in the MC embryonal tumor with multilayered rosettes, DICER, from the NIH. It highlights the importance of testing for YAP1 fusion in ETMR tumors lacking C19MC alterations and DICER1 mutations, as well as MIR17HG amplifications, thereby broadening the spectrum of genetic alterations associated with ETMR and of tumors related to YAP1 fusion. Table 1EMTR non-C19MC-amplified: literature reviewAuthorAgeSexLocationLIN28A IHCOlig2 IHCGFAP IHCEMA IHCMutationAdditional mutationMethylation analysisTreatmentFellow-upReferenceUro-Coste et al., 201911 monthsFPosterior fossaPositiveNANANADICER1 c.5437G > A, p.E1813KNAInstitutional tSNE: close to ETMR clusterTwo cycles of etoposide/carboplatin(post operation)Not reportedPMID: 30446821PMID: 30,446,8218 monthsFPosterior fossaPositiveNANANADICER1 c.5113G > A, p.E1705KNAInstitutional protocol for infant with high-risk CNS PNET(post operation)Not reportedLambo et al., 2019NANA13 posterior fossa, 2 ST midline, 1 parietalNANANANA8 DICER1 mutation, 2 miR-17–92 amplification, 3 breakpoints of C19MC, 3 NANACluster on tSNE with ETMR-C19MC amplifiedNANAPMID: 31802000PMID: 31,802,000de Kock et al., 20202 monthsFPosterior fossaPostiveNANANADICER1 c.823G > T, p.E275*NANANANot reportedPMID: 30953130PMID: 30,953,130Liu et al., 20201.42 yearsMPineal region-posterior fossaNANANANADICER1 c.2771 T > G, p.L924* and c.5425G > A, p.G1809RDNA methylation profilingtSNE analysis: clustered with Uro-Coste et al. casesNADOD, 7 monthsPMID: 31802236PMID: 31,802,2364.73 yearsMPineal region-posterior fossaNANANANADICER1 c.4300del, p.E1434fs and c.5113G > A, p.E1705KDNA methylation profilingNADOD, 1.3 yearsWang et al., 20222 yearsFPosterior fossaPositiveFocalFocalNegDICER1 c.5113G > A, p.E1705K and c.3747delA, p.K1249Nfs*9HIST1H3B (H3.1) K27MNATwo cycles of irinotecan and temozolomide(post operation)Alive (4 months follow up)PMID: 34729860PMID: 34,729,860Kim et al., 202330 yearsMCerebellopontine anglePositiveNAPositiveNADICER1 splicing c.5365-2A > G & D1709E c.5127 T > ANot reportedETMR,non-C19MC-alteredInitial radiosurgery followed by gross total surgical ressection.After additional radiotherapy (30 Gy) and & cycle of vincristine, cyclophosphamide and prednisoloneDOD, 8 months with leptomeningeal dissiminationPMID: 37927133PMID: 37,927,133Campos Mármol et al., 202418 yearsFSellar and supra-sellarNANANegNegDICER1 c.5127 T > A,p.Asp1709GluNot reportedETMR, non-C19MC-altered (V12.5, calibrated score:0.97)6 cycles of etoposide/cisplatin and sequential radiation therapy with radical intentNot reportedPMID: 38979591PMID: 38,979,591Sheriff et al., 202436 yearsMInternal auditory canal extending to the cerebellar-pontine anglePositivePositiveNANADICER1 c.5438A > T, p.E1813VNot reportedETMR,non-C19MC-altered, (V12.8; calibrated score: 0.97251)Three-dimensional conformal radiation delivered to whole brain et spinal cord (40 Gy) plus 14 Gy to the posterior fossa. Followed by 3 cycles of ifosfamide, cisplatin, and etoposide (ICE), 25 Gy on frontal lobe, additional ICEDOD, 22 months with supra and infra-tentoriel tumoral disseminationPMID: 37920133PMID: 37,920,133Yamazaki et al., 20253 monthsFPosterior fossaPositivePositiveNegNADICER1 c.5125G > A, p.D1709NNot reportedNATwo partial resection followed by a complete resection. After first surgery: 2 cycles of vincristine (VCR), cyclo-phosphamide (Cy), cisplatin (CDDP), etoposide (VP-16),and double intrathecal methotrexate therapy (DIT). After second surgery: 4 cycles of VCR, Cy, CDDP, VP-16, and DIT and proton therapy (local, 54 Gy). After third surgery: 4 cycles of ifosfamide, carboplatin, VP-16,and DITDOD, 26 monthsPMID: 39809299 PMID: 39,809,299Current case5 monthsFPosterior fossaPositivePositivePositiveNegYAP1::MAML2 fusionLZTR1 c.1922 T > A/p.(Leu641*)ETMR non-C19MC-altered (V12.5; calibrated score: 0.92) & ETMR DICER (NIH; score 0.995)Complete surgical resection followed by 2 cycles of carboplatin-etoposide, 2 cycles of thiotepa associated to autograft and 6 cycles of topotecan -temozolomideStill alive (5 years)NANA: not done or not reported or patient to patient data difficult to assess EMTR non-C19MC-amplified: literature review Two cycles of etoposide/carboplatin (post operation) Institutional protocol for infant with high-risk CNS PNET (post operation) Two cycles of irinotecan and temozolomide (post operation) DICER1 c.5127 T > A, p.Asp1709Glu NA: not done or not reported or patient to patient data difficult to assess Case presentation Clinical presentation A five-month-old baby girl experienced vomiting and drowsiness for six weeks before hospitalization at the University Hospital of Grenoble because of further development of right torticollis and strabismus with loss of eye tracking. A posterior fossa mass measuring 55 × 51 × 48 mm was identified. It caused severe hydrocephalus with amygdala engagement. It appeared well-demarcated, with minimal or no perilesional edema. On T2-weighted imaging, the lesion appeared heterogeneous with mixed hyper- and hypo-intense regions (Fig. 1A, B). Echo gradient sequences revealed hemorrhagic areas and calcifications. Slight contrast enhancement was seen after gadolinium injection. ADC ratio indicated diffusion restriction. The child had the tumor completely removed through surgery, followed by chemotherapy and an autograft. Specifically, she was treated according to the French PNET-HR protocol, adapted for her age, which included two cycles of carboplatin (160 mg/m2) and etoposide (100 mg/m2). Since the MRI showed no tumor growth and the CSF was free of tumor cells, the treatment continued with two cycles of thiotepa (200 mg/m2), each followed by an autograft, and six cycles of TOTEM (topotecan at 0.75 mg/m2 and temozolomide at 150 mg/m2). She did not receive radiotherapy or intraventricular methotrexate. Five years after surgery, she remained relapse-free, and the MRI showed the surgical cavity was still clear. A–B: T2-weighted MRI showing a heterogeneous lesion with hyper- and hypo-intense regions. C: Methylscape UMAP for the proband. D: H&E stain revealing a poorly differentiated tumor lacking an architectural pattern. E: H&E stain showing a perivascular anucleated region. F: H&E stain with hyperchromatic nuclei demonstrating molding. G, J: IHC for LIN28A, GFAP showing partial positivity. H and I: IHC for IN1, BRG1 showing nuclear positivity. K: IHC for Ki67 indicating a heterogeneous proliferation index, with regions reaching 60%. L: IHC for YAP-C showing loss of nuclear expression in tumor cells Pathological findings At histology, the lesion resembled an embryonal tumor. Regions lacking specific architectural patterns (Fig. 1C) alternated with areas showing broad pseudo-perivascular rosettes, reminiscent of EBL (Fig. 1D). No rosettes, canals, or multilayered rosettes were observed despite extensive sampling of the lesion. The nuclei appeared hyperchromatic and often demonstrated molding (Fig. 1E). Mitoses were numerous and mostly atypical. Apoptotic bodies were seen. Necrosis, sometimes calcified, was noted. At immunohistochemistry (IHC, Fig. 1F to L), INI1 (Cell Marque, MRQ-27) and BRG1 (Abcam, ab110641) nuclear expressions were preserved. Vimentin (Dako, clone V9) was the only marker positive in all tumor cells. They also focally expressed Olig 2 (Sigma, HPA003254), GFAP (Cell Marque, EP672Y), synaptophysin (Ventana, SP11), LIN28 (Cell Signalling Technology, A177), and SALL4 (Cell Marque, 6E3). No labelling for EMA (Roche, E29) was observed. H3K27me3 was maintained (Diagenode, C15410195). Ki67 (Ventana, 30–9) showed heterogeneity, reaching up to 60% in the most proliferative areas. Based on RNA-seq results, additional IHC was performed using an antibody targeting the C-terminal of YAP1 (Abcam, ab52771). It revealed a loss of nuclear expression in tumor cells, while the nuclei of endothelial cells remained positive (Fig. 1L). Molecular findings A SNP array (GeneChipR Oncoscan Array, Affymetrix) showed complete gains of chromosomes 2, 3, 5, 6, 7, 8, 10, 11, 12, and 18, as well as loss of heterozygosity without copy number variation for chromosomes 1 and 21. A custom NGS panel of 571 genes involved in oncogenesis, including DICER1, NF1, and NF2 (DRAGON: Detection of Relevant Alterations in Genes involved in Oncogenetics by NGS; SureSelect CD Curie CGP by Agilent), detected only an LZTR1 c.1922 T > A/p.(Leu641*) probably pathogenic mutation with an allelic frequency of 49.4%. A custom RNA-seq panel dedicated to cerebral tumors (ArcherDx, Integrated DNA Technologies) revealed an in-frame YAP1::MAML2 fusion between exon 6 of YAP1 (11q22.1, breakpoint: chr11:102,080,295) and exon 2 of MAML2 (11q21, breakpoint: chr11:95,826,681). At that time, the DNA methylation profile (V1, EPIC Human Methylation microarray, Illumina, San Diego, USA) did not match any of the MCs of the Heidelberg CNS tumor classifier (V11.4) [9]. However, the UMAP analysis performed on the Epidip website clustered the lesion into ETMR [11]. An integrated diagnosis of ETMR was suggested. Later, the V12.5 of the Heidelberg CNS tumor classifier proposed an embryonal tumor with multilayered rosettes, non-C19MC-altered as MC (score: 0.92), and the NIH brain tumor classifier suggested, as family, the neuroblastic embryonal tumors (score 0.992) with ETMR DICER as MC (score 0.995). The NIH UMAP clustered the lesion in the ETMR-DICER cluster (Fig. 1C). Discussion and conclusion Here, we report an unusual posterior fossa tumor in a five-month-old girl. Histologically, the tumor exhibited an embryonal appearance reminiscent of EBL and showed focal positivity for LIN28A. Multilayered rosettes, the hallmark feature of ETMR, were absent despite extensive sampling of the tumor. No C19MC amplification or DICER1 mutation was detected. The final diagnosis was based on the methylation profile, which classified the lesion in the MC of EMTR non-C19MC-altered with a confidence score of 0.92 using the Heidelberg classifier (V12.5), and within the neuroblastic embryonal tumor family (score 0.992) and the ETMR DICER class (score 0.995) of the NIH classifier. NIH UMAP confirmed the methylation class. Most publications reporting an embryonal tumor with LIN28A positivity and no C19MC amplification also used methylation class and/or t-SNE clustering to support the diagnosis (Table 1). This highlights the importance of methylation profiling in the integrated diagnosis of tumors with unusual clinical, histological, or genomic features. Focal positivity for LIN28A is a key IHC hallmark of ETMR [12]. LIN28A is an RNA-binding protein that acts as an oncogene by inhibiting the tumor suppressor miRNA let-7 family ([13–15]). Interestingly, this microRNA family is also affected by DICER1 mutations, which are the most common genomic alterations in LIN28A-positive embryonal tumors lacking C19MC amplification [16]. In this group, other genetic changes such as C19MC structural variations and MIR17HG amplifications have been reported [3]. None of these changes were observed in our case, a situation reported in 2.5% of ETMRs expressing LIN28A [3]; instead, a YAP1::MAML2 fusion was identified. Meanwhile, their genomic instability has been well-documented; the only fusion previously reported in ETMR is a TTYH1::C19MC fusion ([17, 8]). In the current tumor, we identified a YAP1::MAML2 fusion. Oncogenic fusions involving YAP1 and various partners, such as MAML2, MAMLD1, FAM118B, or ZFTA, have been described in different CNS pediatric brain tumors [18]. The YAP1::MAML2 fusion is most often reported in meningiomas associated with neurofibromatosis type 2 (NF2) [19]. YAP1, the main effector of the Hippo signaling pathway, plays a critical role in cell proliferation, survival, and stem cell maintenance, acting as a tumor suppressor [20]. By interacting with transcriptional enhanced associate domain (TEAD) transcription factors, the C-terminal part of YAP1, which is the final nuclear effector of the Hippo pathway, turns this pathway on “OFF” [21]. When lost, as in YAP1 fusion demonstrated here by the IHC negativity of the tumoral cell nuclei for the YAP1-C terminal, the Hippo pathway remains in an “ON” state, thus promoting oncogenesis [11]. Other pathways also interact with TEAD, such as the WNT pathway, which is known to dysregulate the miRNA machinery in ETMR, along with the NOTCH and SHH pathways [8]. The current lesion also had a pathogenic LZTR1 mutation, a gene known to be mutated in schwannomatosis and Noonan syndrome (OMIM: 600,574). This mutation is thus the second identified in the current lesion associated with the neurofibromatosis spectrum [22]. LZTR1 mutations are increasingly found in various cancers, including sporadic schwannomas [23] and glioblastomas [12]. With an allelic frequency of 49.4%, the LZTR1 mutation is likely germline. Meanwhile, our proband showed no symptoms of Noonan syndrome or schwannomatosis, which may be due to the likely incomplete or very low penetrance of this mutation. Therefore, its significance remains uncertain. ETMRs lacking C19M amplification are mainly reported in the posterior fossa [3], a location similar to that of our patient (Table 1). The observed symptoms were related to intracranial hypertension, a common feature of ETMRs, such as visual impairment and torticollis ([24, 3]). Additionally, ETMRs without C19MC amplification tend to have worse overall survival compared to tumors with amplification, often around one year [3]. In contrast, our patient was still alive and free of relapse five years after diagnosis. Shah et al. reported that the 5-year survival rate ranges from 0 to 30% [25]. Although cases with long-term survival are documented in the literature, the most critical factor appears to be the treatment the patient received. Many published studies do not include the genetic status of the tumors for all cases. The largest series with molecularly confirmed ETMR is the prospective P-HIT trial [26]. In this study, 47% of patients achieved a 5-year overall survival after treatment with carboplatin/etoposide induction and high-dose chemotherapy, compared to only 8% with other treatments. Our patient received similar therapy. The importance of intensive chemotherapy is often emphasized by various authors, as it seems to help avoid radiotherapy in generally young patients [27]. Meanwhile, the beneficial role of radiotherapy has been demonstrated in multiple trials ([28, 3, 29]). Currently, there are no international guidelines for treating ETMR; the literature indicates that the standard treatment goal is complete gross surgical resection combined with chemotherapy and either intensification or radiotherapy when feasible. In the prospective P-HIT trial, infratentorial ETMRs were also linked to poorer survival. The authors do not provide a clear explanation for this difference; like others, they note that there is currently no clue for predicting long-term survival. A hypothesis for differences in survival may relate to proteomic heterogeneity both within and between ETMRs, with common proteomic findings seen among tumors of similar histology ([30, 8]). In conclusion, we report an unusual case of a posterior fossa tumor in an infant with a radiological appearance consistent with ETMR [31]. Histologically, the tumor exhibited an embryonal appearance reminiscent of EBL, with focal expression of LIN28A, but lacked the characteristic multilayered rosettes of ETMR. The tumor also lacked C19MC amplification, DICER1 mutation, or other less common genetic alterations associated with ETMR; however, a YAP1::MAML2 fusion was identified. The methylation profile matched the ETMR non-C19MC-altered MC in the Heidelberg classifier and the ETMR DICER MC in the NIH. This, along with existing literature, suggests the need to revise the WHO terminology for the “ETMR DICER1-mutated” subtype to account for the other genetic alterations observed in this ETMR subgroup. Additionally, it expands the spectrum of tumors linked to YAP1 fusions and genomic alterations associated with ETMR non-C19MC-altered MC, by reporting, for the first time, a YAP1::MAML2 fusion in this context. Acknowledgements CG, JB, FF, and FV are members of the French neuropathological network, the RENOCLIP-LOC. Abbreviations Central nervous system Ependymoblastoma Embryonal tumor with abundant neuropil and true rosettes Embryonal tumor with multilayered rosettes Hemalum, eosin, safran staining Immunohistochemistry Methylation class Medulloepithelioma Next generation sequencing Neurofibromatosis type 2 National Institutes of Health RNA sequencing Single nucleotide polymorphism array Author contributions All authors participated in editing the manuscript and approved the final version. Additionally, CG, FF, FV, and JB contributed to diagnosis discussions. CG and ND authored the manuscript. AP assisted with the patient's clinical management and collected relevant clinical data. CG, JM, RA, and BD were involved in genetic analyses. RD help for the radiological interpretation. ND provided editorial feedback on the manuscript. Funding Methylation analysis was made possible thanks to support from Canceropole Lyon Auvergne-Rhône-Alpes (CLARA). Data availability IDAT files were introduced in the DKFZ and NIH classifiers, with agreement to be used for research purposes. Idats files can be transferred upon request to CG. Declarations Ethics approval and consent to participate Not applicable. Consent for publication The patient’s parents agree to introduce their child into the French RENOCLIP network, which is linked to a research agreement. Competing interests The authors declare no competing interests. Footnotes Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement IDAT files were introduced in the DKFZ and NIH classifiers, with agreement to be used for research purposes. Idats files can be transferred upon request to CG.