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.
Hugh Version.pptx, 03 Aug 2026) is 8 slides: title, DLGNT ×4 (slides 2–5), PXA ×3 (slides 6–8). Every shape on every slide was searched. GBM appears exactly once in the entire deck — a single clause on slide 6, inside the PXA genetics block, listing where else BRAF-V600E turns up: "…but can also be found in gangliogliomas, EGBM, HGAP." There are no GBM slides, no eGBM slides, and no speaker notes anywhere in the file.
So this is not a citation-gap analysis of an existing section — it is a readiness analysis for a section that does not exist yet, and which the supplied 4-item reference list appears intended to seed. The question it answers: can a GBM/eGBM histology overview be written from these four references, to the same standard as the DLGNT and PXA sections?
The answer is: for GBM, largely yes. For eGBM, no. Three findings drive everything below.
The DLGNT and PXA slides both follow one fixed structure. Below, that structure is held constant and each section is tested against the four references in GBM references.docx. Covered writable now · Partial writable but thin or single-source · Gap not supported by any supplied reference.
| Deck section | GBM | eGBM | What the supplied references actually support |
|---|---|---|---|
| Description / definition | Covered | Gap | Louis 2021 defines GBM IDH-wildtype as an IDH-wt diffuse and astrocytic glioma in adults meeting one or more of: microvascular proliferation, necrosis, TERT promoter mutation, EGFR amplification, or +7/−10. That is a complete, citable definition. Nothing in the four references defines eGBM. |
| Subtypes | Covered | Gap | Guo 2023 supplies the clinically meaningful split: histological vs molecular GBM, 146 (76%) vs 45 (24%) of 191. Louis 2021 notes classic morphologic subtypes "are not listed in the classification" — see §4 for why this matters for eGBM. |
| Localization | Gap | Gap | Nothing usable. The only locational statement anywhere in the four is McNulty's incidental observation that 3 of 5 BRAF-mutant tumors were "found adjacent to the ventricular system" — an n=3 finding about a mutational subset, not a localization claim for GBM. |
| Imaging | Gap | Gap | Total gap — none of the four references contains any imaging content. Both the DLGNT and PXA slides carry a full MRI paragraph; neither can be written for GBM from this list. This is the single largest hole. |
| Epidemiology | Covered | Gap | Ostrom 2021 (CBTRUS) is strong and authoritative for the US: incidence 3.23/100,000 (Table 5); 49.1% of malignant CNS tumors and 14.3% of all CNS tumors (Executive Summary); median age 65 (Table 5); male 4.04 vs female 2.53 per 100,000 (≈1.6:1). Guo 2023 adds a single-center comparator (mean age 55.5±14.7; 52.4% male). No eGBM epidemiology in any of the four. |
| Genetic profile | Covered | Partial | Guo 2023 gives the real-world molecular frequencies: TERTp 136/182 (74.7%), EGFR amplification 46/80 (57.5%), +7/−10 73/77 (94.8%) (Table 1). For the BRAF angle the list is very weak — see the next row. |
| Genetic profile — BRAF / V600E |
Partial | Gap | McNulty 2021 is the only BRAF source, and it cannot carry this slide. n=91 specimens / 89 patients: 5 BRAF-mutant, of which only one is canonical p.V600E (VAF 50.2%); the rest are p.K483E, p.H574Q, p.G596R (class III) and a likely loss-of-function frameshift. Survival claim rests on three patients. And explicitly: "None were characterized as epithelioid GBM." |
| Grading | Covered | Partial | Louis 2021 — CNS WHO grade 4; the three molecular criteria permit grade 4 without histologic malignancy features. Applies to eGBM only insofar as eGBM is graded as GBM. |
| WHO diagnostic criteria | Covered | Gap | Louis 2021 is the correct and sufficient anchor for GBM. For eGBM it is the opposite of support — see §4. |
| Diagnostic checklist | Covered | Gap | Derivable from Louis 2021 + Guo 2023 for GBM (IDH status → histology → the three molecular criteria). Not derivable for eGBM from this list. |
All four supplied references were retrieved in full and checked. Identifiers as given in GBM references.docx were correct in every case — no PMID or DOI errors found.
| Reference | ID check | What it actually supports — and its limits |
|---|---|---|
| Louis 2021 Neuro Oncol 23(8):1231–51 PMID 34185076 |
Verified | The anchor for definition, grading and diagnostic criteria. Supports the GBM IDH-wt definition and the TERTp / EGFR-amp / +7−10 grade-4 route. Limit: it is a classification summary — it carries no epidemiology, no imaging, and no frequency data, and must not be cited for any of those. |
| McNulty 2021 Sci Rep 11:19999 PMID 34625582; PMC8501013 |
Verified | The weakest link, and the most consequential. 91 specimens / 89 patients referred for molecular testing at Washington University, Nov 2012 – Dec 2015. BRAF-mutant n=5; EGFR-mutant n=21 (23.1%). Only one canonical V600E. BRAF-mutant mean age 41.82 ± 15.94 y vs EGFR 53.19 ± 11.33 y. Survival: 48.01 ± 24.42 vs 16.13 ± 11.46 months (log-rank p=0.02) — but from only three BRAF patients. Explicitly: "None were characterized as epithelioid GBM." Also a referral-based cohort, so mutation frequencies are enriched and must not be read as population rates. |
| Ostrom 2021 (CBTRUS) Neuro Oncol 23(Suppl 2):iii1–iii105 PMID 34608945; PMC8491279 |
Verified | The strongest reference in the set. Population-based US registry, authoritative for incidence, age and sex. Limit: it is 2014–2018 data classified before WHO 2021, so its "glioblastoma" is the histologic entity — it does not map cleanly onto the molecular GBM definition Louis 2021 introduces. Say so if the two are cited on one slide. |
| Guo 2023 Front Oncol 13:1200815 PMID 37483487; PMC10358772 |
Verified | The best available bridge to WHO 2021 in this set — it applies the new criteria to a real cohort. 191 GBM (146 hist / 45 mol); mOS 12.6 mo overall, 11.4 (95%CI 10.7–15.8) hist vs 15.6 (95%CI 12.5–21.9) mol. Limits: single center (Peking Union Medical College Hospital), retrospective, and molecular denominators are incomplete (EGFR amp assessed in only 80 of 191). Reports no BRAF and no eGBM data. |
Epithelioid GBM is presumably in scope because it is BRAF-V600E–enriched, which is exactly the plixorafenib rationale. But writing an eGBM section requires confronting two facts that the supplied reference list does not surface, and that cut against treating eGBM as a clean, separate indication.
Louis 2021 states that classic morphologic subtypes "are not listed in the classification, but these classic variants are discussed in their respective chapters" — naming gliosarcoma and giant cell glioblastoma. Epithelioid glioblastoma is not mentioned in the classification summary at all. A slide that presents eGBM as a WHO entity alongside DLGNT and PXA — both of which are formally defined entities — would be making a categorical claim the WHO document does not support.
Korshunov 2018 (Brain Pathol; PMID 28990704) profiled 64 tumors originally diagnosed as eGBM by genome-wide methylation plus BRAF/TERT sequencing. They did not form a separate cluster. They distributed across the methylation groups of canonical PXA, IDH-wildtype GBM, and pediatric RTK1 GBM — leading the authors to argue against eGBM as a distinct tumor variant and to reduce the epithelioid appearance to a morphologic pattern.
Behling 2016 (Diagn Pathol 11:55; PMID 27350555) tested 969 CNS neoplasms: BRAF-V600E was found in 3 of 312 glioblastomas (0.96%) — and notably all three showed epithelioid features (Table 1; case detail in Table 2). That single observation carries the eGBM rationale far better than anything in the supplied list: V600E is rare in GBM overall, and concentrates in the epithelioid morphology. This reference, not McNulty, is the one the argument needs.
| Gap | What is needed |
|---|---|
| GBM imaging | An MRI-description source, to match the imaging paragraph both DLGNT and PXA slides carry. Nothing in the folder supports this. Highest-priority gap. |
| GBM localization | A source for anatomic distribution (supratentorial predominance, lobar distribution). Not present. |
| eGBM clinical series | A dedicated eGBM cohort for age, sex, location, survival. Candidate identified — "Clinicopathologic features and prognosis of epithelioid glioblastoma" (PMID 32782671) — but not retrieved: PubMed served a CAPTCHA and it was not bypassed. Flagged rather than summarized second-hand. |
| eGBM ↔ PXA boundary | A diagnostic-criteria source for separating anaplastic PXA from eGBM in practice, given §3.2. Directly relevant since both would appear in this deck. |