Fore Biotherapeutics · Regulatory · plixorafenib (PLX120-03 / FORTE)
GBM & eGBM — Histology Overview
Structure follows: "Patient Story Boards — Sub B, A1 CNS, PLX120-03" (M. Winkler, 03 Aug 2026)
Prepared: 11 Aug 2026
Classification: Internal — Confidential
About this format — supports CTD Module 2.5.1
This is a disease-background reference, and its role in a marketing application is to be the evidence base behind Section 2.5.1, Product Development Rationale — the opening section of the Module 2.5 Clinical Overview under ICH M4E(R2). 2.5.1 must identify the disease and population, summarise current therapy and its limitations, state the unmet medical need, and give the scientific rationale linking mechanism to need.
2.5.1 itself is short and argumentative; it cannot carry this level of detail, but every figure it uses has to be traceable to something that can. That is what this document is. The consolidated rationale drawing on it lives at 2.5.1 Product Development Rationale.
2.5.1 Product Development Rationale
← evidence base: this document
citation register2.7.3 Efficacy2.7.4 Safety
⚠ Decision support — not medical or regulatory advice. Every figure below is traced to a named source and its exact location. Medical + regulatory HITL review required before any external use.
A disease-characterization reference for glioblastoma, IDH-wildtype and its epithelioid morphology, written to the same structure the deck uses for DLGNT and PXA: description → subtypes → localization → imaging → epidemiology → genetic profile → grading → WHO diagnostic criteria → diagnostic checklist. Sources are tagged In folder when they come from GBM/GBM references.docx and Added when the folder list did not cover the section and a primary source was retrieved to fill it.
Why eGBM gets its own half. Epithelioid GBM is the reason this histology is in scope for a pan-RAF program: BRAF-V600E is rare in glioblastoma overall but concentrates in the epithelioid morphology. Slide 6 of the deck already points at it — V600E "can also be found in gangliogliomas, EGBM, HGAP." §C addresses the diagnostic boundary problem that creates with the deck's existing PXA section.
A · Glioblastoma, IDH-wildtype (CNS WHO grade 4)
A1 · Description
A diffusely infiltrating, IDH-wildtype astrocytic glioma of adults, CNS WHO grade 4 — the most common malignant primary CNS tumor. Under WHO 2021 the diagnosis is no longer purely histologic: an IDH-wildtype diffuse and astrocytic glioma in an adult qualifies as glioblastoma on either classic histology or a defining molecular alteration, so tumors that would once have been called "diffuse astrocytoma" are now graded 4 on genotype alone In folder Louis 2021.
Histologically: a diffusely infiltrative astrocytic tumor with marked nuclear atypia and brisk mitotic activity, plus microvascular proliferation and/or necrosis — classically palisading necrosis. Margins are poorly defined; infiltrating tumor cells extend well beyond the enhancing mass.
A2 · Subtypes
WHO 2021 does not list morphologic subtypes as separate classification entries — "subtypes (eg, Gliosarcoma and Giant cell glioblastoma) are not listed in the classification, but these classic variants are discussed in their respective chapters" In folder Louis 2021. Epithelioid glioblastoma is not named in the classification summary at all (see §B1 and §C).
The division that carries real clinical weight is histological vs molecular glioblastoma. In a 191-patient real-world cohort applying the WHO 2021 criteria, 146 (76%) were histological GBM (diagnosed on necrosis/microvascular proliferation) and 45 (24%) were molecular GBM (diagnosed on genotype without those histologic features) In folder Guo 2023. The two behave differently — see §A9.
A3 · Localization
- Supratentorial in ~85% of cases Added StatPearls.
- Frontal lobe ~25% — the single most common lobe; temporal, parietal and occipital follow Added StatPearls.
- Brainstem, cerebellum and spinal cord each <5% Added StatPearls. Brainstem GBM is disproportionately pediatric.
- Spread along white-matter tracts is characteristic; extension across the corpus callosum produces the bilateral "butterfly" pattern.
Provenance note. No supplied reference contains localization data. The only locational statement in the folder list is incidental — 3 of 5 BRAF-mutant tumors "found adjacent to the ventricular system" In folder McNulty 2021 — which is an n=3 observation about a mutational subset, not a localization claim for glioblastoma, and should not be cited as one.
A4 · Imaging
MRI is the modality of record. The characteristic appearance Added StatPearls:
- T1 post-contrast: a poorly circumscribed mass with irregular peripheral ("ring") enhancement around a non-enhancing centre — the enhancing rim is viable tumor, the centre is necrosis.
- T1 pre-contrast: central hypointensity corresponding to necrosis.
- T2 / FLAIR: peripheral hyperintensity representing peritumoral oedema and infiltrating tumor — the two cannot be separated on conventional sequences, which is why the FLAIR abnormality always exceeds the enhancing volume.
- Mass effect and midline shift are common at presentation.
- Advanced sequences: perfusion-weighted imaging shows increased cerebral blood flow (elevated rCBV) in the enhancing rim; MR spectroscopy shows a choline peak with reduced NAA.
Diagnostic caution: ring enhancement is not specific — abscess, metastasis, demyelination and lymphoma all produce it.
Provenance note. No supplied reference contains any imaging content. This section is built entirely on an added source. Both the DLGNT and PXA slides carry a full imaging paragraph; if this section is to match them, an imaging reference must be added to the folder list.
A5 · Epidemiology
| Measure | Value | Source & exact location |
|---|
| Incidence (US, age-adjusted) | 3.23 per 100,000
(95% CI 3.20–3.26) | In folder Ostrom 2021 — Table 5 (n = 61,699 cases, 2014–2018) |
| Share of malignant CNS tumors | 49.1% | In folder Ostrom 2021 — Executive Summary |
| Share of all CNS tumors | 14.3% | In folder Ostrom 2021 — Executive Summary |
| Median age at diagnosis | 65 years | In folder Ostrom 2021 — Table 5 (64 y, peak incidence 75–84 y — Added StatPearls) |
| Sex | M:F ≈ 1.6:1
M 4.04 vs F 2.53 /100,000 | In folder Ostrom 2021 — Table 5; ratio calculated. Independently stated as 1.6:1 Added StatPearls |
| Median overall survival | ~15 months with standard therapy
12.6 months real-world | Added StatPearls; real-world figure In folder Guo 2023 (n=191) |
| 5-year survival | 5.5% | Added StatPearls. Ostrom 2021 gives 35.6% for all malignant CNS tumors combined — do not use that figure for GBM. |
Classification caveat — state this if Ostrom and Louis appear on one slide. CBTRUS 2014–2018 data were coded before WHO 2021. Its "glioblastoma" is the histologic entity, which does not map cleanly onto the molecularly defined GBM of Louis 2021 — the molecular definition pulls in tumors CBTRUS would have counted as diffuse astrocytoma. Incidence figures and molecular-era definitions should not be presented as if they describe the same population.
A6 · Genetic profile
Defining alterations (any one confers grade 4 in an IDH-wildtype adult diffuse astrocytic glioma) In folder Louis 2021:
| Alteration | Frequency | Source & location |
|---|
| +7/−10 (whole-chromosome 7 gain + 10 loss) | 73/77 — 94.8% | In folder Guo 2023 — Table 1 |
| TERT promoter mutation | 136/182 — 74.7% | In folder Guo 2023 — Table 1 |
| EGFR amplification | 46/80 — 57.5% | In folder Guo 2023 — Table 1 |
| IDH1/2 | Wildtype by definition | In folder Louis 2021 — an IDH mutation excludes this diagnosis |
| BRAF p.V600E | 3/312 — 0.96% | Added Behling 2016 — Table 1 (969 CNS neoplasms screened) |
Other recurrent alterations guiding stratification: MGMT promoter methylation (the predictive marker for temozolomide benefit), TP53, PTEN loss, and NF1 Added StatPearls. No frequency is asserted for MGMT here — no retrieved source gave one for a conventional GBM cohort. Supply a source before quoting a number.
The BRAF finding that matters for this program. V600E is present in roughly 1% of glioblastomas Added Behling 2016 — but in that 969-tumor screen, all three V600E-mutant glioblastomas showed epithelioid features. The mutation is rare in GBM as a whole and concentrates in the epithelioid morphology. That single observation, not any cohort-level frequency, is the honest basis for the eGBM rationale.
On McNulty 2021, the folder's BRAF reference. It reports 5 BRAF-mutant tumors among 91 specimens / 89 patients (Washington University, referred for molecular testing 2012–2015). Only one is canonical p.V600E; the others are p.K483E, p.H574Q, p.G596R (class III) and a likely loss-of-function frameshift. BRAF-mutant patients were younger (mean 41.8 ± 15.9 y vs 53.2 ± 11.3 y for EGFR-mutant) and lived longer (48.0 ± 24.4 vs 16.1 ± 11.5 months, log-rank p = 0.02) — but that survival comparison rests on three patients, and this is a referral cohort, so its mutation frequencies are enriched and are not population rates. Critically for §B: the paper states "None were characterized as epithelioid GBM." It cannot be cited for anything eGBM-related.
A7 · Grading
CNS WHO grade 4 — the only grade assigned to glioblastoma. Grade is no longer determined by histology alone: under WHO 2021 an IDH-wildtype adult diffuse astrocytic glioma is grade 4 if it shows microvascular proliferation or necrosis or TERT promoter mutation or EGFR amplification or +7/−10 In folder Louis 2021. A tumor lacking all histologic hallmarks is still grade 4 on genotype.
A8 · WHO 2021 diagnostic criteria
Diffuse astrocytic glioma, in an adult, IDH-wildtype and H3-wildtype, with one or more of:
- Microvascular proliferation
- Necrosis
- TERT promoter mutation
- EGFR gene amplification
- +7/−10 chromosome copy-number changes
In folder Louis 2021. Criteria 3–5 are the molecular route; criteria 1–2 are the classic histologic route.
A9 · Diagnostic checklist
| Step | What to establish |
|---|
| 1 · Confirm lineage | Diffuse and astrocytic glioma. GFAP-positive; exclude non-glial mimics. |
| 2 · IDH status | Must be IDH-wildtype. An IDH mutation moves the tumor to astrocytoma, IDH-mutant — a different entity with different prognosis. |
| 3 · H3 status | Exclude H3-altered diffuse midline glioma, particularly in midline/paediatric tumors. |
| 4 · Adult context | The WHO 2021 definition is framed for adults. Paediatric-type diffuse high-grade gliomas are classified separately. |
| 5 · Apply the grade-4 criteria | Histologic (microvascular proliferation / necrosis) or molecular (TERTp / EGFR amp / +7−10). Record which route was used — this is the histological vs molecular GBM distinction. |
| 6 · Note the route's prognostic weight | Real-world mOS 11.4 mo (95% CI 10.7–15.8) for histological vs 15.6 mo (95% CI 12.5–21.9) for molecular GBM In folder Guo 2023. Molecular-route tumors did better in that cohort. |
| 7 · Therapy-relevant markers | MGMT promoter methylation (temozolomide). For a RAF programme: test BRAF V600E — especially if the morphology is epithelioid (§B). |
B · Epithelioid glioblastoma (eGBM)
Read §B1 before using any of this. eGBM's status is genuinely unsettled, and the honest framing is "a morphologic pattern with V600E enrichment," not "a distinct entity." Everything below is written on that basis.
B1 · Description & nomenclature status
A glioblastoma composed predominantly of discohesive epithelioid cells — abundant eosinophilic cytoplasm, eccentric nuclei and prominent nucleoli, often described as melanoma-like or rhabdoid — with palisading and solid laminar necrosis, high mitotic activity and microvascular proliferation Added Wang 2020. Historically also called adenoid GBM or GBM with epithelioid metaplasia.
Classification status — three facts that must travel together:
- It entered the WHO CNS classification as a provisional/tentative pattern in 2016 Added Wang 2020.
- Under WHO 2021 morphologic subtypes are not listed in the classification, and epithelioid glioblastoma is not named in the classification summary at all In folder Louis 2021.
- Integrated molecular analysis of 64 tumors originally diagnosed as eGBM found they do not form a separate methylation cluster — they distribute across the groups of canonical PXA, IDH-wildtype GBM and paediatric RTK1 GBM, which argues against a distinct tumor variant and reduces the epithelioid appearance to a morphologic pattern Added Korshunov 2018.
B2 · Subtypes
None defined. Practically, eGBM resolves into the three molecular destinations in §B1(3) — and that resolution, not the morphology, is what predicts behaviour Added Korshunov 2018.
B3 · Localization
- All 33 tumors were located in the cerebral hemisphere in the largest single-institution series Added Wang 2020.
- 26/33 (78.79%) showed cortical invasion on MR — a superficial, cortex-involving pattern, and an independent adverse prognostic factor on multivariate Cox analysis (P = 0.000) Added Wang 2020.
- Superficial/cortical location is the practical contrast with conventional GBM's deeper white-matter-centred growth — and the feature it shares with PXA (§C).
B4 · Imaging
No dedicated eGBM imaging source was retrieved. What can be stated from the clinical series Added Wang 2020:
- Tumor size 1–6 cm, mean largest diameter 3.5 cm.
- Cortical invasion visible on MR in 78.79%.
- Otherwise the appearance follows conventional GBM (§A4) — necrosis and peripheral enhancement.
Gap: a dedicated eGBM neuroimaging source is not in the folder and was not retrieved. Do not assert an eGBM-specific imaging signature without one.
B5 · Epidemiology
| Measure | Value | Source & location |
|---|
| Incidence | Rare — no rate established | Added Wang 2020 — "nearly 40 cases" reported in the literature as of 2017 |
| Median age | 36 years (mean 38; range 9–67) | Added Wang 2020 (n=33) — ~29 years younger than conventional GBM |
| V600E-mutant age range | 10–50 years | Added Kleinschmidt-DeMasters 2013 |
| Sex | 16 M : 17 F — essentially equal | Added Wang 2020. Contrast conventional GBM's 1.6:1 male excess. |
| Presentation | Headache/dizziness 51.5%; incidental 18.2%; seizure 15.2%; vomiting 6.1% | Added Wang 2020 |
| Median overall survival | 10 months (range 6–31) | Added Wang 2020 — 25/33 died; median follow-up 11 mo |
| Survival at 1 / 3 / 5 years | 63.6% / 39.4% / 0% | Added Wang 2020 |
Two honest limits on the survival figures. (1) Median follow-up was only 11 months against a 10-month median OS, so the 3- and 5-year rates rest on few patients. (2) The paper's own treatment analysis is confounded by indication — it reports both that chemotherapy and radiotherapy "could improve the prognosis" and that they "could not prolong the overall survival time," with worse raw survival in the treated groups. Read that as sicker patients being treated, not as evidence against treatment; do not quote the treatment comparison.
B6 · Genetic profile
| Alteration | Frequency | Source & location |
|---|
| BRAF p.V600E | ~50% — literature consensus
7/13 (54%) in the defining series | Added Kleinschmidt-DeMasters 2013 (Am J Surg Pathol) — 13 eGBM. The number to quote. |
| BRAF V600E — outlier series | 33/33 (100%) | Added Wang 2020 — see the caveat below; do not quote 100%. |
| TERT mutation | 18/33 — 54.5% | Added Wang 2020. Independent adverse prognostic factor (P = 0.018) |
| MGMT promoter methylation | 15/33 — 45.4% | Added Wang 2020 |
| IDH1 mutation | 1/33 — 3% | Added Wang 2020 — essentially IDH-wildtype |
| EGFR amplification | 0/33 — none | Added Wang 2020. Sharp contrast with conventional GBM's 57.5% In folder Guo 2023 |
| 1p/19q co-deletion; PTEN mutation | None detected | Added Wang 2020 |
Do not quote the 100% V600E figure. Wang 2020 reports BRAF V600E in 33 of 33 cases. That is far outside the literature, where roughly half of epithelioid GBMs are V600E-mutant Added Kleinschmidt-DeMasters 2013 (7/13). A single-institution series in which every case is mutant strongly suggests the cohort was ascertained in a way that selected for the mutation. Use ~50% (7/13) as the citable frequency and treat Wang 2020 as the source for demographics, IHC, survival and the non-BRAF molecular findings.
Specificity of the association Added Kleinschmidt-DeMasters 2013: in the same study, 9 giant-cell GBMs and 2 rhabdoid GBMs were tested and none carried V600E. The mutation tracks the epithelioid morphology specifically, not high-grade morphology in general.
B7 · Immunohistochemistry (n=33)
Added Wang 2020:
- GFAP 30/33 — glial lineage confirmed
- CK 20/33 — cytokeratin positivity is a diagnostic trap; it drives the differential toward carcinoma/metastasis
- INI-1 retained 33/33 — excludes AT/RT and rhabdoid tumor, the key mimic
- ATRX retained 30/33
- HMB45 0/33, Desmin 0/33 — excludes melanoma, the other main mimic given the melanoma-like cytology
- S-100 6/33; p53 5/33; IDH-1 1/33
- Ki-67 >40% in all cases, >60% in 11 — uniformly high proliferation
B8 · Grading
CNS WHO grade 4, graded as glioblastoma. The epithelioid morphology does not carry its own grade.
B9 · Diagnostic checklist
| Step | What to establish |
|---|
| 1 · Morphology | Sheets of discohesive epithelioid / rhabdoid / melanoma-like cells, abundant eosinophilic cytoplasm, eccentric nuclei, prominent nucleoli, necrosis, high mitotic rate. |
| 2 · Confirm glial lineage | GFAP. Essential — the epithelioid cytology plus CK positivity (20/33) otherwise reads as metastatic carcinoma. |
| 3 · Exclude rhabdoid tumor | INI-1 must be retained (33/33 retained in eGBM). Loss indicates AT/RT. |
| 4 · Exclude melanoma | HMB45 and Desmin negative (0/33 in eGBM). |
| 5 · IDH / H3 | IDH-wildtype (1/33 mutant); exclude H3-altered midline glioma. |
| 6 · BRAF V600E | The step that matters for a RAF programme. ~50% positive. Test every epithelioid tumor. |
| 7 · Resolve against PXA | Do not skip — see §C. Add CDKN2A/B status and, where available, methylation profiling. A V600E-mutant superficial epithelioid tumor may be anaplastic PXA. |
| 8 · Prognostic adjuncts | TERT (adverse, P = 0.018); MGMT methylation; cortical invasion on MR (adverse, P = 0.000). |
C · The differential that this deck creates for itself: eGBM vs PXA
The deck already has a PXA section. Adding an eGBM section without addressing the boundary would put partially the same tumors in two places. This is not a pedantic concern — it is the direct consequence of Korshunov 2018, and it has program implications.
Tumors diagnosed as eGBM merge with the canonical PXA methylation group on unsupervised clustering Added Korshunov 2018. Both entities are superficial, cortex-involving, BRAF-V600E–enriched and occur in young patients. Slide 6 of the deck already concedes the overlap — CDKN2A/B deletion is "potentially molecularly defining for PXA (but can also be found in gangliogliomas, EGBM, HGAP)."
| Feature | eGBM | PXA (deck slide 6–7) | Conventional GBM |
|---|
| Median age | 36 y (9–67) | ≈26 y | 65 y |
| Sex | 16 M : 17 F (equal) | Equal in M and F | M:F ≈ 1.6:1 |
| Location | Cerebral hemisphere; cortical invasion 78.8% | Superficial, 98% supratentorial, temporal lobe | 85% supratentorial; frontal ~25%; deeper |
| BRAF V600E | ~50% | Characteristic — most common alteration | ~1% |
| CDKN2A/B homozygous deletion | Reported; overlaps PXA | ~87–95% | Not defining |
| EGFR amplification | 0/33 | Not characteristic | 57.5% |
| Grade | 4 | 2 or 3 | 4 |
| Median OS | 10 months | Substantially better | ~15 months |
Three consequences to decide deliberately.
(1) A patient labelled "eGBM" at one centre may be reported as anaplastic PXA at another, and vice versa — relevant to how a story board or a case narrative is framed.
(2) EGFR amplification is the cleanest discriminator in the data above: 57.5% in conventional GBM, 0/33 in eGBM. It separates eGBM from conventional GBM well, but does not separate eGBM from PXA.
(3) Any V600E frequency quoted for "eGBM" is a frequency for a morphologic pattern, not for a molecularly defined population — because the population itself is not molecularly defined.
D · References
D1 · Supplied in GBM/GBM references.docx
- In folder 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. PMID 34185076; PMC8328013. Used for: §A1, A2, A6, A7, A8, B1. Definition, grading and diagnostic criteria. Carries no epidemiology, imaging or frequency data.
- In folder Ostrom QT, Cioffi G, Waite K, Kruchko C, Barnholtz-Sloan JS. CBTRUS Statistical Report: Primary Brain and Other CNS Tumors Diagnosed in the United States in 2014–2018. Neuro Oncol. 2021;23(12 Suppl 2):iii1–iii105. PMID 34608945; PMC8491279. Used for: §A5. Incidence 3.23/100,000; 49.1% of malignant CNS tumors; median age 65; M 4.04 / F 2.53. Pre-WHO-2021 histologic entity.
- In folder Guo X, Gu L, Li Y, et al. Histological and molecular glioblastoma, IDH-wildtype: a real-world landscape using the 2021 WHO classification. Front Oncol. 2023;13:1200815. PMID 37483487; PMC10358772. Used for: §A2, A5, A6, A9. n=191 (146 hist / 45 mol); TERTp 74.7%, EGFR amp 57.5%, +7/−10 94.8%; mOS 12.6 mo. Single centre; incomplete molecular denominators.
- In folder McNulty SN, Schwetye KE, Ferguson C, et al. BRAF mutations may identify a clinically distinct subset of glioblastoma. Sci Rep. 2021;11:19999. PMID 34625582; PMC8501013. Used for: §A6 caveat only. n=5 BRAF-mutant, one canonical V600E; survival from n=3; referral cohort. States "None were characterized as epithelioid GBM" — not usable for §B.
D2 · Added — the folder list did not cover these sections
- Added Kleinschmidt-DeMasters BK, Aisner DL, Birks DK, Foreman NK. Epithelioid GBMs show a high percentage of BRAF V600E mutation. Am J Surg Pathol. 2013;37(5):685–698. PMID 23552385. Used for: §B5, B6. The defining eGBM/BRAF reference. V600E in 7/13 eGBM (~54%), mutant patients aged 10–50 y; 0/9 giant-cell and 0/2 rhabdoid GBM mutant.
- Added Wang S, He Q, Zhang Q, Guan B, Zhou X. Clinicopathologic features and prognosis of epithelioid glioblastoma. Int J Clin Exp Pathol. 2020;13(7):1529–1539. PMID 32782671. Used for: §B1, B3–B7. Largest single-institution series (n=33). Demographics, IHC panel, molecular profile, survival. Its 100% V600E rate is an outlier — not used as the frequency.
- Added Korshunov A, et al. Epithelioid glioblastomas stratify into established diagnostic subsets upon integrated molecular analysis. Brain Pathol. 2018. PMID 28990704. Used for: §B1, B2, §C. n=64; eGBM does not cluster separately — merges with canonical PXA, IDH-wt GBM and paediatric RTK1 GBM.
- Added Behling F, Barrantes-Freer A, Skardelly M, et al. Frequency of BRAF V600E mutations in 969 central nervous system neoplasms. Diagn Pathol. 2016;11:55. PMID 27350555. Used for: §A6. GBM V600E 3/312 (0.96%) — Table 1; all three showed epithelioid features.
- Added Grochans S, et al. / StatPearls: Glioblastoma Multiforme. StatPearls [Internet]. NCBI Bookshelf NBK558954. Bookshelf NBK558954. Used for: §A3, A4, A5, A6. Review-grade source — carries localization, imaging and 5-year survival, none of which any supplied reference covers. Replace with primary sources before external use.
Two provenance flags before this goes anywhere external. (1) Localization and imaging for conventional GBM (§A3, §A4) rest entirely on a review-grade source (StatPearls), because no supplied reference contains either. Substitute primary references before external use. (2) eGBM imaging (§B4) has no dedicated source at all and is stated only as far as the clinical series supports.
Method & limits. The deck was parsed programmatically (all 8 slides, every shape and notes slide); it contains no GBM or eGBM slides, so §A and §B are written to the structure the DLGNT and PXA slides use rather than against existing slide text. All four supplied references and all five added references were retrieved and read in full text, PMC or publisher PDF; every figure above is taken from the source with its table or section named. Korshunov 2018 was read via abstract and secondary description rather than full text, so no interior figure or table is cited from it. Percentages shown as "calculated" were derived from counts given in the source; all others are quoted as printed. Where sources conflict — notably the eGBM V600E frequency — both are shown and the recommended figure is stated with the reason.