Molecular Fingerprints Reshape the Fight Against Childhood Brain Tumors Above the Tentorium

Among the many cancers that strike children, brain tumors carry the heaviest burden of suffering and death, and within that group ependymoma occupies a particularly stubborn position. Accounting for roughly 6 to 12 percent of pediatric brain tumors, it ranks third in frequency behind pilocytic astrocytoma and medulloblastoma, yet its behavior has long defied the neat categories that pathologists once drew under the microscope. A sweeping new review published in the Journal of Neuro-Oncology now pulls together what modern genomics has revealed about the supratentorial form of the disease, the tumors that arise above the tentorium in the cerebral hemispheres, and the findings are reshaping how surgeons, radiation oncologists, and neuro-oncologists plan treatment for some of the most vulnerable patients in medicine.

Supratentorial ependymoma, or ST-EP, makes up approximately 30 to 40 percent of all ependymoma cases and tends to affect slightly older children than its infratentorial counterpart. The 2021 World Health Organization classification fundamentally reorganized the field by defining three molecular entities: ZFTA-fusion tumors, YAP1-fusion tumors, and a residual category labeled not elsewhere classified, or NEC. This reorganization matters because tumors that look nearly identical under the microscope can behave in radically different ways depending on which gene fusion drives them. The old histologic grading system, which divided ependymomas into grades 1 through 3, suffered from high interobserver variance even among experienced neuropathologists, and prognosis did not always track with grade. Molecular testing, through DNA methylation profiling and RNA sequencing, has now become the decisive diagnostic tool.

The dominant player is the ZFTA-fusion subtype, formerly known as RELA-fused ependymoma, which accounts for roughly 70 percent of supratentorial cases. The ZFTA gene, residing on chromosome 11q13 and formerly designated C11orf95, fuses most commonly with RELA, producing a chimeric oncoprotein that shuttles constitutively into the nucleus and drives transcription through NF-κB signaling, a pathway normally involved in inflammation and proliferation. These tumors typically affect children around eight years of age on average and carry a relatively worse prognosis than their YAP1 counterparts. Intriguingly, recent work has shown that NF-κB upregulation alone does not explain the malignancy: CRISPR knockout and pharmacologic inhibition of NF-κB fail to halt tumor growth in experimental models. Instead, ZFTA fusion proteins appear to hijack transcription factor programs involving PLAGL1 and PLAGL2, factors implicated in normal neuronal development, pushing cells along MAPK, SHH, and SWI/SNF pathways toward a dedifferentiated, anaplastic state.

The biology of ZFTA tumors extends even deeper into cellular metabolism. A recent study in Nature found that ZFTA-RELA ependymomas sustain their oncogenic fusion expression through epigenetically modulated metabolic pathways involving itaconate, an immune-signaling metabolite, suggesting that the tumor’s fuel supply is intertwined with its genetic engine. High levels of HDAC4 and SIRT1 messenger RNA have been associated with significantly worse outcomes, and computational drug screens have flagged HDAC4 and ABCG2 inhibitors as candidate therapeutics. On the cell surface, the adhesion molecule L1CAM is highly expressed and has been used as a surrogate marker for ZFTA fusion status, though prospective validation remains incomplete. These converging lines of evidence are beginning to sketch the outline of targeted therapies for the subtype that needs them most.

At the opposite end of the prognostic spectrum sit YAP1-fused tumors, which comprise fewer than 10 percent of supratentorial ependymomas but carry a dramatically more favorable outlook. These tumors arise in children with a median age under two years, and the largest study to date, published in 2019, reported near-complete five-year progression-free survival, with some patients requiring no post-surgical treatment at all. The molecular driver is most commonly a YAP1-MAMLD1 fusion: YAP1 is a principal downstream effector of the Hippo tumor-suppressor pathway, and when fused with MAMLD1 the resulting protein becomes a potent oncogenic transcriptional regulator that depends on TEAD and nuclear factor I-mediated programs. Because these tumors are rare, the evidence base remains thin, and the authors of the review caution that the spectacular survival figures come from a single, relatively small series pending confirmation in larger cohorts.

Between these two defined entities lies the NEC category, representing roughly 20 percent of supratentorial cases. These tumors resemble classic ependymomas histologically but carry molecular profiles matching no established subtype, encompassing both lesions with no identifiable defining alteration and incompletely characterized emerging entities. Their heterogeneity poses a genuine clinical problem, since they may not respond to treatment regimens calibrated for ZFTA or YAP1 biology. Better molecular characterization of this group is one of the field’s most pressing unmet needs, and the review’s authors call for standardized genomic interrogation across institutions to dissect it.

Even the radiology suite can now offer early molecular hints. ZFTA-fused tumors preferentially arise in the cerebral cortex and parenchyma, most often in the frontal lobe, extending from the pial surface to the ventricular margin as predominantly cystic masses with eccentric cysts, calcifications, hemorrhagic foci, and marked perilesional edema, showing heterogeneous contrast enhancement and restricted diffusion on MRI. At recurrence they tend to re-emerge along the dural flap. YAP1 tumors, by contrast, are characteristically intra- or periventricular, with multinodular solid components and a distinctive garland-shaped pattern of peripheral enhancement, and less prominent edema. These signatures are not substitutes for molecular analysis, but they help multidisciplinary teams anticipate the expected clinical course before genetic results return.

Across every molecular subtype, one principle towers above all others: the extent of surgical resection is the most important modifiable prognostic factor, and gross total resection offers the best chance of cure. Surgeons now deploy neuronavigation, ultrasound, and intraoperative MRI in nearly every case, and second-look operations have proven their worth, with a pan-Italian study showing that children who underwent a second procedure before radiotherapy achieved overall survival comparable to those completely resected at first surgery, without significantly higher morbidity. In some cases, chemotherapy given after the first operation has reduced tumor vascularity and facilitated complete resection at second look. Once resection is complete, patients undergo staging with cerebrospinal fluid cytology and gadolinium-enhanced spine imaging at least 10 to 14 days postoperatively before adjuvant decisions are made.

Radiation therapy follows maximal resection, with focal treatment of the tumor bed at cumulative doses classically ranging from 54 to 59.4 Gy, and the clinical target volume margin has shrunk from 10 millimeters in older trials to a standard 5 millimeters based on the excellent outcomes of the ACNS0831 study. Proton beam therapy is recommended where available because radiation dose to uninvolved brain structures is linked to long-term cognitive decline, and the conformal nature of protons can deliver full prescription dose to the tumor bed while leaving the rest of the brain at near-zero exposure. Perhaps the most consequential finding concerns observation alone: in the ACNS0831 trial, 37 children with grossly totally resected grade 2 supratentorial tumors were monitored without immediate radiotherapy, achieving 66.9 percent five-year event-free survival with no deaths from recurrence, since nearly all relapses were salvageable with repeat surgery and radiation. Notably, no patient in that observation arm harbored a CDKN2A deletion, a homozygous loss that the German HIT cohort identified as an ominous marker, conferring five-year progression-free survival of just 19 percent versus 81 percent in wild-type tumors within the E-HIT-2000 analysis.

Chemotherapy, by contrast, has repeatedly failed to earn its place. The ACNS0831 randomized comparison of radiotherapy alone versus maintenance chemotherapy after radiotherapy showed no significant difference in five-year event-free or overall survival, and a multi-institutional analysis of 108 ZFTA-fused tumors likewise found no benefit. For infants too young for radiation, management remains heterogeneous, with many clinicians bridging with chemotherapy until an appropriate age despite sparse supporting data. Roughly half of children with intracranial ependymoma eventually relapse, most often locally at the original tumor bed, and salvage hinges on repeat gross total resection and re-irradiation, while salvage chemotherapy has been associated with worse outcomes. With the SIOP Ependymoma II trial reporting soon and the prospective SIOP Ependymoma study continuing to enroll, the field is converging on a future in which each child’s tumor is profiled molecularly at diagnosis and treatment is tailored accordingly, a vision that seemed remote when these tumors were classified by appearance alone.

Subject of Research: Molecular classification and management of pediatric supratentorial ependymoma

Article Title: Supratentorial ependymoma: a comprehensive review of molecular classification, management strategies, and clinical outcomes (Part II of pediatric ependymomas across compartments)

Article References: Vallejo, F. A., Monsour, M. A., Fredricks, N. S., Tsang, D. S., Koutsouras, G., Ritzmann, T. A., Aquilina, K., Ramaswamy, V., & Dewan, M. C. (2026). Supratentorial ependymoma: a comprehensive review of molecular classification, management strategies, and clinical outcomes (Part II of pediatric ependymomas across compartments). Journal of Neuro-Oncology, 180(1), Article 8. https://doi.org/10.1007/s11060-026-05818-7

Image Credits: AI Generated

DOI: 10.1007/s11060-026-05818-7

Keywords: supratentorial ependymoma, ZFTA fusion, YAP1 fusion, pediatric brain tumor, WHO 2021 classification, gross total resection, proton beam therapy, CDKN2A deletion, NF-kB signaling, radiotherapy, NEC ependymoma, SIOP Ependymoma II