High-yield points
- DMG, H3 K27-altered = midline location and loss of H3K27me3. It is always WHO grade 4, whatever the histology .
- Routes to the alteration: H3.3 or H3.1 K27M, EZHIP overexpression, or EGFR. In children, H3.3 does worse than H3.1 (median 10.1 vs 14.2 months) .
- Children's DMGs are mostly pontine, with median OS about 12 months. Adults' are thalamic or spinal and do better .
- Thalamic DMG: major debulking was associated with longer survival (18.5 vs 11.4 months) in HERBY. Leptomeningeal spread predicts worse outcome .
- Always obtain tissue (biopsy or resection), and image the whole neuraxis.
- ONC201 (dordaviprone): 20% response in recurrent H3 K27M DMG (pontine and spinal tumors excluded; mostly adults). It is the first approved systemic agent, though the evidence is debated .
Landmark papers
- Mackay et al., Cancer Cell 2017: an integrated molecular meta-analysis of more than 1,000 pediatric high-grade gliomas and DIPGs. It defined the histone-mutant subgroups and their partner mutations .
- ONC201 integrated analysis (Arrillaga-Romany et al., J Clin Oncol 2024): 50 patients (46 adults) with recurrent H3 K27M DMG, excluding pontine and spinal tumors, with an objective response rate of 20%. It led to the first approved systemic therapy for DMG .
Diffuse midline glioma (DMG), H3 K27-altered is a WHO 2021 tumor type defined by a diffuse glial tumor in a midline location with the molecular H3 K27 alteration, rather than by histological grade .
- Where it arises: midline structures, chiefly the pons, thalamus and spinal cord.
- Grade: always CNS WHO grade 4, even when the histology looks low grade.
- Which topic covers what:
- Pontine DMG (DIPG): its clinical diagnosis, biopsy and treatment are covered in DIPG.
- This topic: the entity as a whole, and thalamic, spinal and other non-pontine DMG, where surgery plays a larger role.
- Wider context: see High-Grade Glioma for pediatric high-grade glioma generally.
Definition and molecular subtypes
- Defining feature: loss of H3K27 trimethylation (H3K27me3) in a diffuse midline glioma. It can arise in several ways :
- H3 K27M mutation:
- in H3.3 (H3F3A), the most common;
- or in H3.1 (HIST1H3B/C) or, rarely, H3.2 (HIST2H3 genes).
- EZHIP overexpression in H3-wild-type tumors;
- EGFR-mutant midline gliomas.
- Subgroup biology :
- each histone-mutant subgroup carries co-segregating mutations:
- H3.1 K27M with BCOR mutations (and ACVR1 in pontine tumors; see caveats);
- H3.3 K27M with TOP3A rearrangements.
- Infants are biologically distinct.
- H3.1 vs H3.3: age changes the meaning (individual patient data from 26 studies: 102 H3.1 and 529 H3.3 tumors) :
- H3.1 is more common in younger patients;
- in children, H3.3 carried shorter survival (median 10.1 vs 14.2 months);
- in adults the reverse held (H3.3 14.4 vs H3.1 1.7 months).
Sites and outcomes
- Children vs adults :
- children's DMGs are mostly pontine (64.9%), with median overall survival 12.0 months;
- adults' are mostly thalamic (41.5%) or spinal (27.0%), with better survival (30.3 months).
- Thalamic DMG (HERBY trial, 42 children with thalamic-based midline high-grade glioma) :
- patterns:
- partial thalamic;
- whole thalamus;
- unithalamic with diffuse contiguous extension;
- bithalamic.
- outcomes were poor whatever the H3 K27 subtype;
- major debulking or near-total or total resection was associated with longer overall survival (18.5 vs 11.4 months);
- leptomeningeal dissemination before treatment predicted worse outcome (OS 11.4 vs 18.5 months).
- Spinal DMG: presents with myelopathy or pain. Biopsy or limited resection, then radiotherapy; resection is limited by the risk to cord function (see caveats).
Role of the neurosurgeon
- Tissue is essential. H3 K27 status, the H3.1 vs H3.3 subtype and co-mutations define the diagnosis, prognosis and trial eligibility .
- Stereotactic biopsy of thalamic or brainstem lesions is generally safe in experienced hands (see DIPG and caveats).
- Resection where it can be done safely. For a thalamic DMG that can be safely debulked, more extensive resection was associated with longer survival in HERBY .
- The data are observational, and bithalamic or diffusely infiltrating tumors aren't candidates for resection.
- Hydrocephalus: ETV with biopsy in the same sitting, or a shunt (see caveats).
- Screen the neuraxis: MRI of the whole spine before treatment, because leptomeningeal spread changes prognosis .
Treatment
- Radiotherapy (about 54–59.4 Gy) is the standard (see caveats). It gives transient benefit but is not curative.
- ONC201 (dordaviprone), an imipridone :
- integrated analysis of recurrent H3 K27M DMG (50 patients, mostly adults; pontine and spinal tumors excluded):
- objective response rate 20% (RANO-HGG);
- median duration of response 11.2 months;
- grade 3 treatment-related adverse events in 20%.
- FDA accelerated approval for progressive H3 K27M DMG is the first regulatory approval of a systemic therapy for DMG .
- Whether the evidence is enough for widespread use is debated, because it comes from single-arm early-phase studies .
- Clinical trials should be offered whenever possible.
Related topics
- DIPG: pontine DMG: diagnosis, biopsy and treatment.
- Pediatric High-Grade Glioma: the wider group of pediatric-type high-grade gliomas.
- Spinal Cord Gliomas: spinal DMG and other intramedullary gliomas.
- Endoscopic Third Ventriculostomy: CSF diversion with biopsy for hydrocephalus.
- Low-Grade Gliomas of the Brainstem in Children: focal brainstem low-grade gliomas.
References
- Louis DN, Perry A, Wesseling P, et al. The 2021 WHO Classification of Tumors of the Central Nervous System: a summary. Neuro Oncol 2021;23(8):1231–1251. doi:10.1093/neuonc/noab106. PMID 34185076. verified
- Rodriguez D, Calmon R, Aliaga ES, et al. MRI and molecular characterization of pediatric high-grade midline thalamic gliomas: the HERBY phase II trial. Radiology 2022;304(1):174–182. doi:10.1148/radiol.211464. PMID 35412366. verified
- Mackay A, Burford A, Carvalho D, et al. Integrated molecular meta-analysis of 1,000 pediatric high-grade and diffuse intrinsic pontine glioma. Cancer Cell 2017;32(4):520–537.e5. doi:10.1016/j.ccell.2017.08.017. PMID 28966033. verified
- Vuong HG, Ngo TNM, Le HT, et al. The prognostic significance of HIST1H3B/C and H3F3A K27M mutations in diffuse midline gliomas is influenced by patient age. J Neurooncol 2022;158(3):405–412. doi:10.1007/s11060-022-04027-2. PMID 35606633. verified
- Sim Y, McClelland AC, Choi K, et al. A comprehensive multicenter analysis of clinical, molecular, and imaging characteristics and outcomes of H3 K27-altered diffuse midline glioma in adults. J Neurosurg 2025;142(5):1307–1318. doi:10.3171/2024.8.JNS241180. PMID 39793011. verified
- Arrillaga-Romany I, Gardner SL, Odia Y, et al. ONC201 (dordaviprone) in recurrent H3 K27M-mutant diffuse midline glioma. J Clin Oncol 2024;42(13):1542–1552. doi:10.1200/JCO.23.01134. PMID 38335473. verified
- Khasraw M, Dun MD, Arrillaga-Romany I, et al. Point/counterpoint: dordaviprone for diffuse midline glioma—a landmark approval or a premature step? Neuro Oncol 2026;28(8):1829–1839. doi:10.1093/neuonc/noag117. PMID 42172613. verified
Verification (2026-11-18):
- Each reference's authors, title, journal, year, volume, pages, DOI and PMID were checked on PubMed. Numbers were checked against the PubMed abstracts.
- Reference 1 was previously verified in the High-Grade Glioma topic.
Caveats:
- The following are standard teaching, not drawn from the cited abstracts:
- the WHO grade-4 assignment and the full list of alteration routes, which come from reference 1's full text (the HERBY abstract lists the H3 K27M, EZHIP and EGFR subtypes);
- ACVR1 co-mutation in H3.1 pontine tumors;
- spinal DMG presentation and management;
- the safety of stereotactic biopsy (see the DIPG topic's references);
- that bithalamic or diffusely infiltrating tumors aren't resection candidates;
- ETV with biopsy;
- radiotherapy dose.
- Reference 5 is an adult series compared with a pediatric meta-analysis dataset.
- Reference 6 enrolled mostly adults.
- The HERBY resection association is observational.
- The figure is an original schematic.