Current Open Trials and Molecular Update for Pediatric Embryonal Tumors

Background: Embryonal tumors are highly malignant cancers of the central nervous system, with a relatively high incidence in infants and young children. Even with intensive multimodal treatment, the prognosis of many types is guarded, and treatment-related toxicity is significant. Recent advances in molecular diagnostics allowed the discovery of novel entities and inter-tumor subgroups, with opportunities for improved risk-stratification and treatment approaches. Summary: Medulloblastomas separate into four distinct subgroups with distinct clinicopathologic characteristics, and data from recent clinical trials for newly diagnosed medulloblastoma support subgroup-specific treatment approaches. Atypical teratoid rhabdoid tumor (ATRT), embryonal tumor with multilayered rosettes (ETMR), and pineoblastoma, as well as other rare embryonal tumors, can be distinguished from histologically similar tumors by virtue of characteristic molecular findings, with DNA methylation analysis providing a strong adjunct in indeterminate cases. Methylation analysis can also allow further subgrouping of ATRT and pineoblastoma. Despite the dire need to improve outcomes for patients with these tumors, their rarity and lack of actionable targets lead to a paucity of clinical trials and novel therapeutics. Key Messages: (1) Embryonal tumors can be accurately diagnosed with pediatric-specific sequencing techniques. (2) Medulloblastoma risk stratification and treatment decisions should take into account molecular subgroups. (3) There is a dire need for a novel collaborative clinical trial design to improve outcomes is rare pediatric embryonal tumors.


Introduction
Tumors of the central nervous system (CNS) are the second most common form of cancer and the leading cause of cancer-related death in children and adolescents [1].Among these, embryonal tumors consist of about 16% [2], including entities such as medulloblastoma, atypical teratoid rhabdoid tumor (ATRT), embryonal tumor with multilayered rosettes (ETMR), and others, even rarer entities.
As a group, embryonal tumors are characterized by a relatively high incidence in infants and young children, histopathologic appearance of highly cellular tumors consisting of small, blue, round cells, and a high tendency for dissemination.The use of intensive, multimodal treatment approaches has led to improved survival for some children with embryonal tumors, but this group of tumors nonetheless contains some of the most aggressive and difficult to treat pediatric cancers [3][4][5].
Significant advances in our understanding of the genetic and molecular drivers of CNS tumors have led to improved diagnostics and the recognition of specific tumor subgroups that carry clinical implications.For medulloblastoma, such advances have revealed 4 distinct molecular subgroups.For other tumors that were previously difficult to diagnose or termed primitive neuroectodermal tumors (PNET), distinct entities can now be recognized based on specific molecular alterations and/or methylation patterns [6].For example, ATRTs harbor a near-universal loss of SMARCB1, and ETMRs show an amplification of the C19MC miRNA cluster.
The incorporation of molecular advances into clinical trial design can be challenging for several reasons.First, embryonal tumors are rare, and additional molecular subgrouping requires enrolling a larger number of patients to reach statistically significant results.Second, clinical trials can take many years to complete and analyze, with the rapid pace of molecular advances making some designs potentially obsolete while trial is still ongoing.Lastly, despite the robust diagnostic and prognostic implications of many molecular level changes, most are not actionable for therapeutic purposes as of yet.
We highlight specific embryonal tumors entities in the following review, with an emphasis on current molecular findings and their implication for prognosis and treatment.Clinical trials were collated from www.clinicaltrials.gov.Search terms used included all the specific disease entities detailed, as well as "embryonal tumors" and "PNET."The trial list was reviewed and filtered for actively recruiting, phase 2 or higher, interventional clinical trials.Phase 1 trials and "all comer" trials were excluded, as shown in Figure 1.The final list of diseasespecific trials is summarized in Table 1.

Sequencing and Molecular Classification
Since pediatric CNS tumors differ in their genetic makeup from adult tumors or other pediatric solid tumors, it is vital that specific molecular profiling assays be used for this population.As the differential diagnosis of embryonal tumors includes other high-grade CNS tumors, diagnostic sequencing panels need to capture a broad set of driver events that define most pediatric CNS tumors, including somatic mutations, copy number changes, rearrangements with intronic breakpoints, and RNA-level amplifications [7].
In recent years, methylation profiling has emerged as a powerful tool to aid in the diagnosis of pediatric CNS tumors by comparing the epigenetic landscape of the tumor methylome to a predefined bank of known tumor profiles [6].Such analysis can aid in distinguishing between histologically similar tumor types and decrease Fig. 1.Summary of the process identifying clinically relevant trials.The Clinicaltrials.Gov dataset was searched by the disease type, with search results subsequently filtered according to (1) recruitment status and participant age; (2) study phase; and (3) individual review of study eligibility to identify trials that are disease specific, excluding all-comer trials.the risk of misdiagnosis [8], allow for the detection of novel entities [9], and for some tumor types provide additional subgrouping information of potential clinical utility [10].
The growing role of tumor genetics and molecular classification is reflected in the incorporation of specific alterations into the diagnostic criteria of many such tumors within the recently published WHO 2021 classification of CNS tumors [11].In addition, large-scale efforts that offer molecular characterization of childhood tumors, such as the Children's Oncology Group "Project: EveryChild" (NCT02402244), are paving the way for both improved diagnostics and potential novel therapies for children newly diagnosed with cancer.

Medulloblastoma
Medulloblastoma is the most common malignant tumor of childhood, historically risk-stratified according to patients' age, residual disease, metastatic status, and tumor histology [12].Using gene sequencing, medulloblastomas have been shown to subgroup into four distinct molecular variants -Wingless-activated (WNT), Sonic Hedgehogactivated (SHH), and the non-WNT, non-SHH groups 3 and 4 [13].These molecular groups have distinct demographics, immunohistologic and clinical correlations, leading to their universal acceptance and incorporation within the WHO classification of CNS tumors [13].In-depth analysis reveals heterogeneity within each group, suggesting that tumors can be even further subclassified [14].Such data, however, are yet to be integrated within prospective clinical trials.
Despite these advances in molecular diagnostics, the mainstay of treatment for medulloblastoma remains unchanged over the past several decades and is comprised of maximal safe surgery, followed by craniospinal irradiation and maintenance chemotherapy.Such approaches can lead to long-term survival of over 80% for average risk disease and about 60% for high risk disease [4,15,16].Patients too young to receive craniospinal irradiation due to its deleterious effect on the developing brain are often treated with protocols that incorporate more intensive chemotherapy to delay or defer its use [17].As trial data from the past decade mature, the prognostic and potential therapeutic implications of the molecular subgroups are being extensively analyzed, and evidence supporting subgroup-specific treatment approaches is starting to accumulate.The recently published COG trial for average-risk medulloblastoma, ACNS0331 (NCT00085735), randomly assigned young patients ages 3-7 years to either standard dose or reduced dose craniospinal irradiation, with inferior results for the reduced dose cohort [15].However, post hoc analysis shows excellent results for the WNT-activated subgroup, providing a rational for therapy de-escalation for that group, which is explored in the currently open COG ACNS1422 trial (NCT02724579).Another COG trial, ACNS0332 for newly diagnosed high-risk medulloblastoma (NCT00392327), tested the addition of carboplatin concomitantly with radiation as well as the role of isotretinoin as a proapoptotic agent.While isotretinoin did not affect patient outcomes, the study did include a subanalysis of molecular subgroups for both interventions, showing a survival advantage with the addition of carboplatin to radiation in patients with group 3 medulloblastoma [18].The recently published results from the SJMB03 trial for newly diagnosed medulloblastoma (NCT00085202) also include an extensive subanalysis of the different molecular subgroups, further validating their clinical significance [4].In the actively enrolling SJMB12 clinical trial for newly diagnosed medulloblastoma (NCT01878617), patients are treated according to their molecular subgroup, but further risk stratified using additional clinical and molecular features specific to each subgroup.
Infants with medulloblastoma have historically been considered high risk for recurrence due to the necessary avoidance of radiation therapy [19].We now know that this population is enriched for SHH-activated tumors, and these tumors were recently shown to further separate into distinct biological groups by methylation profiling [20,21].The clinical implications of this nuanced risk stratification require further prospective validation.The currently enrolling HeadStart4 trial (NCT 02875314) incorporates histological subtypes, molecular subgroups, and clinical response into its risk-adapted therapy design.Dhall et al. [22] recently presented excellent outcomes with this approach for infants with localized, nonmetastatic SHH medulloblastoma.
Despite the advances made in the diagnosis and treatment of medulloblastoma in the molecular era, recurrent disease remains a challenging diagnosis with poor prognosis, with no consensus on standard of care [23,24].A recently completed COG phase 2 trial showed improved median survival with a combination of bevacizumab, irinotecan, and temozlomide, but long-term OS is still unacceptably low [25].Despite that, there are only a few actively enrolling clinical trials that are specific for recurrent medulloblastoma, which for the most part do not take subgroup data into account, making this an area in a dire need of further research.

Atypical Teratoid Rhabdoid Tumor
The aggressive embryonal tumor known shorthand as ATRT is the most common malignant CNS tumor among children younger than 1 year of age.ATRT was first defined as a separate disease entity in 1996, with the recognition of recurrent chromosome 22q abnormalities in these tumors [26].These abnormalities were later recognized to lead to a biallelic loss of the SMARCB1 gene, manifesting with loss of INI-1 protein expression [27].SMARCB1 is a core component of the SWI/SNF chromatin remodeling complex and acts as a tumor suppressor.SMARCB1 loss confers a high risk of developing histologically similar tumors like ATRT, renal malignant rhabdoid tumors, and renal medullary carcinoma.Rarely, ATRT can be driven by a loss of the SMARCA4 gene, leading to a similar phenotype [28].
Early reports of ATRT cases had dismal prognosis, with such tumors treated with either medulloblastoma or PNETdirected therapies [29].However, identification of this unique entity and focus on the rare survivors has allowed for the development of disease specific, multimodal approaches with improved survival outcomes.While there is no accepted standard of care, different treatment approaches all utilize maximal safe surgery followed by intensive chemotherapy.Various treatment protocols include anthracycline-based chemotherapy and radiotherapy [5]; high-dose chemotherapy with autologous stem cell support followed by radiotherapy [30]; and the utilization of anthracycline, optional radiotherapy, and intrathecal chemotherapy [31].The COG study ACNS0333 (NCT00653068) achieved an overall survival of 43% at 4 years, using an intensive multimodal approach, results that are comparable with other published studies of such approaches.While these results are vastly improved compared to historical cohorts, many patients still suffer disease progression and death.Moreover, the long-term toxicities among survivors of such intensive therapy are significant [32].
Although the genetic makeup of ATRT is simple, with SMARCB1 or SMARCA4 being the only recurrent alterations, it is associated with wide changes on the epigenetic level.Potential tumor vulnerabilities include upregulation of Aurora Kinase A expression and tumor cell dependence on enhancer of zeste homolog-2 (EZH2).Earlyphase clinical trials utilizing the Aurora Kinase inhibitor Alisertib and the EZH2 inhibitor tazemetostat were recently concluded, showing only a limited potential as single agents in the recurrent setting [33,34].However, both agents were safe and show clinical promise, and so are being explored in further combination clinical trials (NCT 02114229, 05407441).
The use of methylation profiling has revealed ATRT to be a heterogeneous group of tumors, with three consensus molecular subtypes that differ in their expression profiles and clinical characteristics.One subgroup shows an upregulation of melanosomal proteins such as tyrosinase and is dubbed ATRT-TYR, with evidence to suggest an improved prognosis.A second group shows overexpression of SHH and notch signaling, termed ATRT-SHH.A third group shows expression of the MYC oncogene and is more prevalent among older patients [10,35,36].While the rarity of ATRT makes incorporating further subgrouping into trial design challenging, it is nonetheless expected that the next generation of clinical trials will include prospective evaluation of molecular subgroups in the hope of generating robust data used for risk stratification.

Embryonal Tumor with Multilayered Rosettes
ETMR is a novel disease entity introduced in the 2016 revision of the WHO classification of CNS tumors, and another rare and aggressive embryonal tumor affecting predominantly young children.Epidemiological data are limited since such tumors were previously diagnosed as embryonal tumor with abundant neuropil and true rosettes, ependymoblastoma, medulloepithelioma, or PNET.However, gene sequencing analysis and methylation profile have shown these tumors to have a unifying molecular signature of Chr.19 miRNA cluster (C19MC) amplification [37].Together with a high expression of the LIN28 gene, which is often found but is not specific to ETMR, an accurate diagnosis can more readily be made, with methylation profiling helpful for indeterminate cases.Of note, about 5% of cases harbor a germline DICER1 mutation, playing the role of oncogenic driver in the absence of C19MC amplification [38].
Historical data for patients with ETMR show a very poor prognosis when treated with various PNET-directed therapies [37].However, more recent reports using molecularly guided diagnostics and structured multimodal therapy approaches show improved survival outcomes.Hanson et al. [39] reported prolonged survival in 4/5 patients treated with a modified IRS-III chemotherapy regimen following a gross total resection.Within the prospective P-HIT trial (NCT 00303810) patients with ETMR received postoperative induction with chemotherapy, additional intrathecal methotrexate, and a HDCT consolidation with autologous stem cell support, followed by optional radiation therapy.For children treated with carboplatin/ etoposide induction followed by HDCT, 5-year PFS and OS were 35% and 47%, significantly superior to other treatment regimens [3].In their retrospective review of 159 patients with molecularly confirmed ETMR, Khan et al. report a 2-year over survival rate of 31% for patients treated with curative intent.Achieving a gross total resection, use of HDCT, and radiotherapy were found to be statistically significant factors for improved survival [40].
Only two current clinical trials list ETMR as a target disease: the ongoing registry trial ETMR One (NCT 04794686) is attempting to collect more clinical data while also suggesting an IRS-III therapy approach; SJDAWN (NCT03434262) is a phase 1 molecular-driven therapy trial, which includes recurrent ETMR in one of the study arms.

Other Embryonal Tumors
The following disease entities are all characterized by embryonal histology and an aggressive disease course.Taken together, about 50 such tumors are diagnosed annually in the USA [2].No current disease-specific clinical trials are available for patients diagnosed with these rare diseases, highlighting an area of dire clinical need.

Pineoblastoma
Pineoblastoma is a malignant embryonal tumor of the pineal gland, histologically difficult to distinguish from other tumors of the pineal region as well as other embryonal tumors.Its rarity, along with nonspecific histology, has led to its treatment within "umbrella" PNET protocols.Available data suggest that older age, localized disease, a gross total resection, and the inclusion of radiotherapy are favorable prognostic markers [41,42].
Pineoblastoma can be associated with a number of genetic alterations.About 5% of cases have been reported in the setting of germline RB1 and the presentation of "trilateral retinoblastoma" [43], and additional cases have been described in the setting of mutually exclusive DICER1, DROSHA, or DGCR8 mutations [44].Recent integrated molecular analysis shows that pineoblastoma can be segregated into five distinct molecular subgroups, based on the defective miRNA synthesis pathway (groups 1, 2), altered protein degradation (group 3), and RB or MYC alterations (groups 4 and 5, respectively).These groups also show distinct clinicopathologic differences, but their clinical implications require further validation [45].

CNS Neuroblastoma, FOXR2-Activated
This tumor type emerged as a novel, distinct entity from a large study utilizing whole-genome methylation profiling on a large set of previously diagnosed CNS PNET and was incorporated to the 2021 WHO classification [9,11].The molecular subgroup CSN neuroblastoma, FOXR2activated constituted 14% of tumors within the group and was so named due to a histologic appearance in line with a diagnosis of CNS neuroblastoma or ganglioneuroblastoma.Additional genetic alterations include Chr.1q gain, CHr.16q loss, and Chr.8 gain, and most tumors were shown to overexpress the FOXR2 gene.The oncogenic mechanism behind FOXR2 overexpression is yet to be fully elucidated, but recent studies suggest that it plays a role in stabilizing MYCN as well as interacting with the ETS family of transcription factors [46,47].
Due to the rarity of this novel tumor entity, there is a paucity of outcome data and no suggested standard of care.At this time, no disease specific clinical trial is enrolling newly diagnosed patients.

CNS Tumor with BCOR Internal Tandem Duplication
This is another novel entity defined using whole-genome sequencing and methylation profiling of previously diagnosed CNS PNET, though with a distinct histology not reminiscent of other embryonal tumors [9].The recurrent molecular alteration in these tumors is an in-frame internal tandem duplication of BCOR, which is a similar oncogenic driver to that found in clear cell sarcoma of the kidney.Additional genetic alterations have been described, but none are currently amenable for targeted therapy [48].Prognosis for patients with this tumor entity is poor, based on limited numbers, and no standard of care or disease-specific clinical trials are available.

Conclusions
CNS embryonal tumors are an aggressive group of tumors affecting mostly infants and young children.Modern molecular diagnosis techniques such as whole-gene sequencing and methylation analysis have advanced our understanding of the landscape of these tumors: previously termed PNET and embryonal tumors not otherwise specified can now more accurately be diagnosed [6,9], while histologically similar but biologically distinct tumors can be excluded [8].Novel tumor entities such as ETMR have been described based on their unique molecular signature, paving the way for disease-specific clinical research efforts.The heterogeneity within specific tumors continues to be uncovered, such as with medulloblastoma molecular subgrouping, allowing for a higher resolution of risk stratification.
Our remarkable biological advances are slow to transform treatment approaches, owing to the overall rarity of embryonal tumors, the slow pace of clinical trial design and execution, and the lack of actionable targets identified in most tumors.Our lack of clinical progress is evident in the paucity of actively recruiting clinical trials for pediatric patients with embryonal tumors, which is in stark contrast to the guarded prognosis of many diseases and the significant toxicity associated with conventional therapies.
Moving forward, researchers and clinicians will need to harness their creativity and joint efforts to translate molecular findings into novel clinical approaches, with an eye toward disease-specific, biologydriven treatment approaches.The rarity and increasing number of tumor subgroups call for large consortium and multi-institutional collaborations to ensure that patients are treated according to best available practice and are offered participation in clinical trials designed to improve both survival outcomes and survivors' quality of life.