Tovorafenib

Mechanism, PK/PD & Formulation — Tovorafenib (OJEMDA, Day One Biopharmaceuticals / Ipsen ex-US)

Data cut: 31 August 2026. Because several figures below come from sources with different vintages, each is dated explicitly:

Source	Vintage as used here
US Prescribing Information (NDA 217700 tablets / 218033 for oral suspension)	Revised 8/2025; DailyMed SPL version posted 2 September 2025; supplements approved 27 August 2025. Verified current on DailyMed as of 31 August 2026.
Label efficacy dataset (FIREFLY-1 Arm 1)	Data cutoff 10 May 2024
EMA CHMP Assessment Report, Ojemda, EMA/67438/2026 (154 pp.)	CHMP positive opinion 27 February 2026; EC conditional MA decision 20 April 2026 (per EMA's own EPAR page); assessment report published 11 May 2026
Most recent public FIREFLY-1 update	SNO 2025 (24 November 2025), data cutoff 6 June 2025, median study duration 40.6 months — not the source of any label figure below

All figures below are public-domain and cited.

[FLAG: target-drug resolution — Fore team must confirm before this section is used.] The orchestrating workflow passed the literal string "the competitor drug" rather than a drug name (the competitor-dossier.js guard that throws on a missing drug argument did not fire in this run). I resolved the target to tovorafenib because it is the only approved RAF inhibitor in a CNS-tumor indication, it is the head-to-head comparator named LIVE in C:\Users\Owner\Documents\Hugh Context Folder\Fore\competitive_intelligence\CI_PROCESS.md, and a prior tovorafenib dossier already exists in the hub folders. If the intended target was a different asset (e.g. dabrafenib+trametinib, selumetinib, belvarafenib, NST-628, BDTX-4933, claturafenib), this workstream must be re-run with the drug named explicitly. This is a scope defect, not a content defect: nothing below is usable until the target is confirmed.

---

1. Molecular identity

Property	Value	Source
INN / brand	Tovorafenib / OJEMDA	USPI 8/2025
Development codes	DAY101, TAK-580, MLN2480, BIIB-024	Rasco 2023 (Cancer Chemother Pharmacol 92:15–28)
Molecular formula / MW	C₁₇H₁₂Cl₂F₃N₇O₂S / 506.29	USPI §11 — verified
Chemical name	6-amino-5-chloro-N-[(1R)-1-[5-[[[5-chloro-4-(trifluoromethyl)-2-pyridinyl]amino]carbonyl]-2-thiazolyl]ethyl]-4-pyrimidinecarboxamide	USPI §11
Aqueous solubility	≤3 µg/mL, pH 1.2–8 (USPI wording: "≤ 3 micrograms/mL from pH 1.2 to 8 in aqueous media"); BCS Class 2 (low solubility / high permeability)	USPI §11; EPAR §5.2.2.2
Class	Type II RAF kinase inhibitor (USPI wording). The "DFG-out" descriptor is not label language — it comes from the structural literature (Tkacik 2023).	USPI §12.1; Tkacik 2023

---

2. Mechanism of action

Label wording (deliberately narrow). "Tovorafenib is a Type II RAF kinase inhibitor of mutant BRAF V600E, wild-type BRAF, and wild-type CRAF kinases." (USPI §12.1, rev. 8/2025 — first sentence verified verbatim on DailyMed). The label further states that tovorafenib exhibited antitumor activity in cultured cells and xenograft models harbouring BRAF V600E and V600D mutations and in a xenograft harbouring a BRAF fusion. [FLAG: the xenograft sentence was not re-verified verbatim against the printed §12.1; confirm before quoting it as label language.] The US label makes no CNS-penetration claim and no dimer/monomer claim — both are sponsor and literature descriptors, not label language. The EU assessment is more permissive: "Tovorafenib is an oral, central nervous system (CNS)-penetrant, selective small-molecule, type II RAF inhibitor" (EPAR §5.2.3.1 — verified verbatim in EMA/67438/2026).

Structural basis. Co-crystal structures of tovorafenib bound to WT BRAF and BRAF V600E kinase domains confirm the canonical type II binding mode: "Type II inhibitors bind (or induce) a conformation created by a crankshaft-like flip of the DFG segment that reorients the phenylalanine residue toward the ATP site" (Tkacik et al., J Biol Chem 2023;299(5):104634 — verified verbatim). The EPAR describes the same binding independently: tovorafenib "displaces the phenylalanine out of the back-pocket of the active site of BRAF, shifts the orientation of the C-helix, and changes the conformation of the glycine-rich loop … similar to that of sorafenib and differs from that of the Type-I BRAF inhibitor vemurafenib" (EPAR §4.3.1.1).

Dimer competence — the differentiating claim. Type II binding permits occupancy of both protomers of a RAF dimer, which is the mechanistic basis for activity against dimer-driven, RAS-independent drivers such as KIAA1549::BRAF fusions that type I (V600-selective) inhibitors cannot address. Tkacik et al. state that "tovorafenib and naporafenib inhibit BRAF and CRAF dimers with marked positive cooperativity." Corrected Hill slopes (verified against Tkacik 2023, Table 1 / associated fits):

Construct	Tovorafenib Hill slope
BRAF WT (dimer)	−2.89 ± 0.19
CRAF WT (dimer)	−1.55 ± 0.43
CRAF SSDD (dimer)	−3.18 ± 0.08
BRAF V600E (monomer)	−0.75 ± 0.01 (no cooperativity)

[CORRECTED — was "Hill slopes −2.6 to −3.2".] That range was wrong: the CRAF WT dimer slope is −1.55 ± 0.43, well outside it, and with an error term that overlaps −1.1. The qualitative claim ("marked positive cooperativity on BRAF and CRAF dimers") is the authors' own wording and stands; the numeric range does not. Do not carry "−2.6 to −3.2" into any Fore slide.

Sun et al. (Neuro Oncol 2017;19(6):774–785) showed MLN2480/tovorafenib is an equipotent antagonist of BRAF V600E monomers and KIAA1549::BRAF dimers, and suppressed phospho-ERK in OLIG2⁺ human pediatric low-grade astrocytoma cells in organotypic slice culture where vemurafenib did not.

Paradoxical activation is reduced, not abolished — and is a labeled risk. Sun et al. reported no rebound ERK signal with MLN2480 in transformed neural progenitors at therapeutic concentrations, but did observe paradoxical activation at sub-therapeutic concentrations (0.01–0.1 µM). More consequentially, this appears in the US label: "In vitro, tovorafenib increased phosphorylation of ERK at clinically relevant concentrations in cells with neurofibromatosis Type 1-loss of function (NF1-LOF) suggesting activation, rather than inhibition, of the MAP kinase pathway. In an NF1 genetically engineered mouse model of plexiform neurofibroma without BRAF alteration, tovorafenib did not have antitumor activity, and while not statistically significant, an increase in tumor volume was noted in 2/12 mice (approximately 17%)" (USPI §13.2 — verified), driving Warning 5.6 "NF1 Associated Tumors: Increased tumor growth may occur." NF1 patients were excluded from FIREFLY-1 (USPI §14).

ARAF sparing. Despite the "pan-RAF" label used in company and press materials, tovorafenib is a poor ARAF inhibitor. Verbatim (Tkacik 2023): "Like naporafenib, tovorafenib is a poor inhibitor of ARAF." The authors frame relative ARAF sparing as a possible property shared across multiple type II drugs, citing belvarafenib and naporafenib.

[CORRECTED — the previous draft quoted a stitched sentence ("Although type II RAF inhibitors are generally considered to be pan-RAF inhibitors … tovorafenib is a poor inhibitor of ARAF") that does not appear as a contiguous quotation in the paper. The verbatim sentence above is what the paper says; the "generally regarded as pan-RAF inhibitors" framing is a separate discussion sentence. Use the verified quote.]

Notable: the EU regulator adopts the academic finding. The EPAR states, immediately after reporting the sponsor's IC₅₀ values: "However, activity on ARAF was not presented, but it was reported in the literature that tovorafenib is not potent against ARAF (Tkacik et al., 2023)" (EPAR §4.3.1.1 — verified verbatim). Competitive relevance: ARAF-mediated signaling is a documented RAF-inhibitor escape route; the "pan-RAF" descriptor overstates the isoform coverage actually demonstrated biochemically, and the EMA has said so on the record — a stronger, more citable point than framing it as an academic-versus-sponsor dispute.

---

3. Potency — two public datasets, different assays

Target	Sponsor values, EPAR §4.3.1.1 (also Rasco 2023) (IC₅₀, nM)	Tkacik 2023 JBC, Table 1 (IC₅₀, nM)
BRAF V600E	7.1	495 ± 117
BRAF WT	10.1	633 ± 160
CRAF (RAF1) WT	0.7	94.2 ± 13.2
CRAF SSDD	not stated	84.5 ± 15.8
ARAF SSDD	not presented ("activity on ARAF was not presented" — EPAR §4.3.1.1)	>3,000 (did not completely inhibit even at 10 µM)
ARAF WT	not stated	not reported

Attribution corrected: the 7.1 / 10.1 / 0.7 nM values sit in EPAR §4.3.1.1 (Primary pharmacodynamics), not §5.2.3.1, and the same three figures also appear in Rasco 2023 — verified in the EPAR PDF text.

[FLAG: ~70–100× potency discrepancy between the regulator-accepted sponsor values and the independent academic values. Do NOT quote either number without the assay caveat.] The two datasets are not directly comparable — different enzyme constructs (full-length vs kinase domain; SSDD phosphomimetic), ATP concentrations, and readouts. A relevant confounder documented in the 2024 corrigendum to Sun et al.: tovorafenib "has high affinity for plastic," and samples must be maintained in borosilicate-coated glass tubes or EC₅₀ values fail to replicate in cell culture (Neuro Oncol 2024;26(5):985–986). For any Fore head-to-head potency slide, use a single internally-run assay with both compounds side by side; do not mix these literature values with plixorafenib figures from a different assay.

[CORRECTED — the previous draft asserted "neither publication cross-references the other." That is false. The EPAR explicitly cites Tkacik et al. 2023 on the ARAF point (§4.3.1.1). The regulator saw both datasets. What the EPAR does not do is reconcile the ~70–100× gap in the BRAF/CRAF numbers, which remains genuinely unexplained in the public record.]

---

4. CNS penetration — preclinical only

This is the axis Fore will be compared on, and it is where the public tovorafenib evidence is thinnest.

Measurement	Value	Species / method	Source
Brain:plasma AUC ratio	24% (AUC_brain 41.6 h·µM)	Mouse, LC-MS/MS	Sun 2017, Neuro Oncol 19(6):774–785
Brain:plasma concentration ratio	0.154–0.178	Male CD-1 mice	EPAR §4.3.2.2 — verified verbatim
Brain:plasma AUC ratio of drug-derived radioactivity	0.57 cerebellum, 0.54 cerebrum, 0.67 medulla ("indicating the passage of the blood-brain barrier")	Male Long-Evans (pigmented) rats, [¹⁴C] QWBA	EPAR §4.3.2.2 — verified verbatim
CHMP summary judgement	"brain exposure was measurable (brain/plasma AUC ≈ 0.6), supporting blood–brain barrier penetration adequate for the treatment of paediatric low-grade gliomas"	Pigmented rat QWBA	EPAR §4.3.2.2 — verified verbatim
Spatial confirmation	Drug detected in healthy brain after single oral dose (MALDI-MSI, segregating from heme); and inside KIAA1549::BRAF intracranial tumors after chronic dosing	Mouse	Sun 2017
Efflux liability	Not a substrate of BCRP, P-gp, OATP1B1 or OATP1B3	In vitro	USPI §12.3 — verified verbatim; EPAR §5.2.2.2

[FLAG: Sun 2017 ratio direction.] The published sentence reads "plasma:brain ratio of 24%", alongside "AUC:brain of 41.6 h·µM". Read literally that is the inverse of what is meant; the intended and universally-cited reading is brain:plasma = 24%. [FLAG: the "10 mg/kg PO" dose and the "30 mg/kg / 41 days BID" MALDI-MSI details in the previous draft were not re-verified against the printed Sun 2017 methods — dose and schedule specifics removed above pending verification.]

Analyst reading (Fore analysis, not a cited claim): the absence of P-gp/BCRP substrate liability is the strongest single mechanistic argument for CNS exposure and is label-grade evidence. However, every brain:plasma figure above is a total (not unbound) ratio, and the rat number is total radioactivity, which includes metabolites. Given 97.5% plasma protein binding (USPI §12.3), moderate red-cell partitioning (K_RBC/plasma ≈ 4–6 in humans; blood:plasma radioactivity ratio 1.03 in rat — EPAR §4.3.2.2), and documented reversible melanin binding including in the substantia nigra (EPAR §4.3.2.2), total brain:plasma overstates free-drug CNS exposure. No K_p,uu has been published.

[FLAG: no human CNS exposure data exist in the public domain.] I found no CSF concentrations, no tumor-tissue concentrations, and no neuro-imaging PK for tovorafenib in any peer-reviewed publication, the USPI, or the CHMP assessment. The FIREFLY-1 primary publication lists pharmacokinetics among "additional planned secondary endpoints not reported in this manuscript" (Kilburn et al., Nat Med 2024;30(1):207–217 — verified verbatim). The clinical CNS-penetration claim therefore rests entirely on rodent data plus the clinical response rate itself. This is a genuine, citable gap in the competitor's package. (Absence of a finding is not proof of absence — I searched specifically and found nothing.)

[FLAG: Sun 2017 carries a 2024 corrigendum — verified.] A vinculin loading-control immunoblot was used twice, for Figure 5 panels A and C; panel A was replaced with images from an independent repeat of the experiment. The glass-tube ("high affinity for plastic") methods omission was added. Conclusions unchanged, and the corrigendum does not touch the PK/BBB data. Cite the corrigendum alongside the paper for any external use.

---

5. Human pharmacokinetics (US label, rev. 8/2025)

Parameters are mean (CV%) from population-PK unless noted. All values in this table verified verbatim against the current DailyMed SPL (§12.3) except where marked.

Parameter	Value
Steady-state C_max	6.9 µg/mL (23%)
Steady-state AUC	508 µg·h/mL (31%)
Time to steady state	12 days (33%)
Dose proportionality	Exposure increases dose-proportionally; "No clinically significant tovorafenib accumulation occurs."
T_max	3 h (median; range 1.5–4 h), tablets or oral suspension
Apparent V_d	60 L/m² (23%); ≈118 L by popPK (EPAR §5.2.2.4)
Plasma protein binding	97.5%. In vitro: albumin ≈95%, α₁-acid glycoprotein ≈40% (EPAR §5.2.2.4)
Terminal half-life	≈56 h (33%)
Apparent clearance	0.7 L/h/m² (31%)
Sex effect on clearance	Males estimated to have 21.5% higher CL/F than females — source corrected: this is EPAR §5.2.2.5, not the USPI. The USPI reports only "no clinically significant differences … based on sex."
Metabolism	Primarily aldehyde oxidase and CYP2C8; CYP2C9, CYP2C19, CYP3A4 minor. No human metabolite exceeded 10% of total drug-related radioactivity
Excretion	65% feces (8.6% unchanged), 27% urine (0.2% unchanged) after a single radiolabeled oral dose
Absolute bioavailability	Not determined — "No estimation of the absolute bioavailability could be made as no investigation of the IV route was performed" (EPAR §5.2.2.2 — verified verbatim)

Adult phase 1 (Study C28001, NCT01425008) non-compartmental values at the 600 mg QW adult MTD: terminal t½ ≈70 h (range 31–119 h) in 20 PK-evaluable patients; no apparent accumulation with once-weekly regimens; AUC increased approximately dose-proportionally across 400–800 mg QW (Rasco et al., Cancer Chemother Pharmacol 2023;92(1):15–28 — these four statements verified).

[FLAG: two half-life values in the public record — ≈56 h (label popPK, mixed pediatric/adult) vs ≈70 h (adult phase 1 NCA, n=20 at 600 mg QW). Different methods and populations; do not present a single number without the source.] [FLAG: C_max 5,650 ng/mL, AUC₁₆₈ 330,000 ng·h/mL and accumulation ratio 1.09 from Rasco 2023 could NOT be independently confirmed — they do not appear in the EPAR text and were not surfaced by search. They were originally extracted by automated retrieval of PMC10261210 and have not been eyeball-verified against the printed table. Do not use these three numbers externally until someone opens the PDF. The t½, non-accumulation and dose-proportionality statements in the same sentence ARE verified.]

Why once-weekly works. The long terminal half-life is driven by tissue distribution, not hepatic trapping: "Liver microsomal binding is low, confirming that the prolonged half-life observed in vivo results primarily from tissue distribution rather than hepatic trapping" (EPAR §4.3.2.2 — verified verbatim). Clinically, QW dosing was adopted in C28001 because "a majority of patients required early dose reductions following Q2D dosing at 200 mg. In addition, QW [dosing] was expected to achieve higher tovorafenib concentrations, and therefore a higher degree of pathway inhibition within the dosing interval, without compromising overall dose intensity" (EPAR §5.2.3.2 — verified verbatim). C28001 enrolled 149 patients (50 escalation, 99 expansion); 125 were PK-evaluable. MTDs: 200 mg Q2D and 600 mg QW.

---

6. Dose, schedule and the 420 → 380 mg/m² step-down

Approved dosage (US and EU): 380 mg/m² orally once weekly, maximum 600 mg once weekly, with or without food, until disease progression or intolerable toxicity. "A recommended dose for patients with BSA less than 0.3 m² has not been established" (EPAR §4.2 — verified verbatim; USPI §2.3 equivalent).

BSA-banded tablet dosing (USPI Table 1): 0.30–0.89 m² → use oral suspension; 0.90–1.12 m² → 400 mg QW; 1.13–1.39 m² → 500 mg QW; ≥1.40 m² → 600 mg QW. The suspension table (USPI Table 2) bands from 125 mg (5 mL, BSA 0.30–0.35 m²) to 600 mg (24 mL, BSA ≥1.40 m²). [FLAG: the individual BSA band cut-points and the suspension volume table were not re-verified line-by-line against the printed §2.3 tables — confirm before reproducing them in a deliverable.]

The step-down is the notable regulatory-pharmacology fact. FIREFLY-1 dosed a "protocol-defined 420 mg/m² dosing regimen, not to exceed 600 mg (adult MTD)" (EPAR — verified verbatim). The label's exposure–response analysis spans 290–476 mg/m², explicitly "0.76–1.25 times the approved recommended dosage" (i.e. relative to 380 mg/m², not to the dose administered). ORR by RAPNO-LGG and RANO-LGG showed no clinically significant exposure–response over that range, while higher exposure was associated with more rash, AST/ALT elevation and CPK elevation, and with reduced height-for-age z-scores (USPI §12.2). CHMP is blunter: "The evidence supporting the 380 mg/m² dose is derived from modelling and simulation only." (EPAR, clinical-pharmacology dose-justification section — verified verbatim, appears twice in EMA/67438/2026).

[FLAG: the approved dose has never been prospectively studied as such. Every efficacy figure in the tovorafenib label was generated at 420 mg/m², not 380 mg/m². Any Fore comparison of dose intensity or exposure must state which dose the data came from.] CHMP also noted that in FIREFLY-1 the actual starting doses were ≤350 mg/m² (15 patients), 350–420 mg/m² (63 patients) and >420 mg/m² (59 patients), with lower rates of dose reduction, discontinuation, growth retardation and tumour haemorrhage in the 350–420 group than in the >420 group — the safety basis for the step-down.

Pediatric dose-finding lineage: PNOC014, an investigator-sponsored phase 1 in patients 1 to <25 y, opened Part A on 27 February 2018 starting at 280 mg/m² and escalating to 350 and 420 mg/m²; across the study four dose levels were evaluated (280, 350, 420 and 530 mg/m²) and the TITE model recommendation was reduced from 530 mg/m² to 420 mg/m² (EPAR §5.3.1 — verified).

[FLAG: CHMP did not receive the PNOC014 CSR.] "The CSR for study PNOC014 has not been provided in the current application… it is considered that the data from PNOC014 could have contributed to the assessment of efficacy/dose" (EPAR — verified verbatim). PNOC014 was also removed as a key measure of the Paediatric Investigation Plan. The peer-reviewed full report (JCO Oncology Advances, doi:10.1200/OA-26-00010) returned a 403 to me — [FLAG: PNOC014 PK and dose-escalation details not independently verified against the published paper.]

Missed dose / vomiting: if missed by ≤3 days, take as soon as possible and keep the next dose on schedule; if >3 days, skip. If vomiting occurs immediately after a dose, repeat the dose (USPI §2.4). [FLAG: not re-verified verbatim.]

---

7. Food effect — none clinically relevant

"No clinically significant differences in tovorafenib C_max and AUC were observed following administration of tablets with a high fat meal (approximately 859 total calories, 54% fat) compared to fasted conditions, but the Tmax was delayed to 6.5 hours" (USPI §12.3 and EPAR §5.2.2.2 — verified verbatim in both). Label permits dosing with or without food.

Study identity and size — new, and materially limiting. The relative-BA / food-effect study is now published: Zhang et al., "Taste Profile and Relative Bioavailability of Tovorafenib Powder for Oral Suspension and Food Effect of the Tovorafenib Tablet in Healthy Participants," Clin Pharmacol Drug Dev 2025, doi:10.1002/cpdd.1558 (PMC12402836) — an open-label, randomized phase 1 study designated QSC205140, with n = 12 healthy participants for the taste/palatability part and n = 12 for the relative-BA and food-effect part. Tovorafenib was initially administered at 300 mg and reduced to 100 mg because of musculoskeletal adverse events.

[FLAG: study-number inconsistency inside the EPAR itself.] EMA/67438/2026 calls it "Phase 1 study QSC201540 (relative BE and food effect study)" in §5.2.2.3 and "food effect study (Day 101 QSC205140)" in §5.2.2.10.1. The peer-reviewed publication uses QSC205140; prefer that. This is an EPAR typo, not a workstream error, but do not print both numbers as if they were two studies.

[FLAG: the food-effect conclusion rests on 12 healthy adults dosed at 100 mg — roughly one-sixth of the maximum clinical dose — not on patients at 380–600 mg.] For a BCS Class 2 amorphous-solid-dispersion compound, a null food effect at 100 mg does not automatically extrapolate to 600 mg, where dissolution-limited absorption is more likely to bite. This is a legitimate, citable soft spot in the competitor's package and should be checked against the printed Zhang 2025 results before it is asserted externally.

Analyst reading (Fore analysis, not a cited claim): "with or without food" is an adherence advantage in a pediatric population relative to any agent carrying a fasting requirement. State the comparator explicitly whenever this is used; do not assert it against unnamed competitors.

---

8. Drug–drug interactions

Tovorafenib as victim (USPI §7.1 — verified): - Avoid coadministration with a strong or moderate CYP2C8 inhibitor. - Avoid coadministration with a strong or moderate CYP2C8 inducer.

Tovorafenib as perpetrator (USPI §7.2, §12.3): - Tovorafenib is a CYP3A inducer. Avoid coadministration with certain CYP3A substrates where minimal concentration changes may lead to serious therapeutic failures. [FLAG: the specific "midazolam C_max and AUC predicted to decrease by at least 20%" figure was not re-verified verbatim against §12.3 — confirm before quoting the 20%.] - Hormonal contraceptives: avoid coadministration (verified). If unavoidable, an effective non-hormonal method should be added during treatment and for a defined period after discontinuation. [FLAG: the "28 days after discontinuation" interval was not re-verified — confirm against §8.3/§17.] Clinically salient in an adolescent/young-adult population, and reinforced by embryo-fetal toxicity Warning 5.5. - In vitro, tovorafenib inhibits CYP2C8, CYP2C9, CYP2C19 and CYP3A, and induces CYP3A, CYP2C8, CYP1A2, CYP2B6, CYP2C9 and CYP2C19 at clinically relevant concentrations — i.e. it both inhibits and induces the same enzymes, including one of the two that clear it. [FLAG: the full inhibition/induction enzyme lists were not re-verified item-by-item.] - "Tovorafenib inhibits BCRP at clinically relevant concentrations" (USPI §12.3 — verified verbatim). - "Tovorafenib is not a substrate of BCRP, P-glycoprotein (P-gp), OATP1B1 and OATP1B3" (verified verbatim). Not evaluated as a substrate of OAT1, OAT3, MATE1, MATE2-K and OCT2 (verified verbatim).

[FLAG: no clinical DDI study has ever been performed — verified.] CHMP: "No clinical DDI studies were conducted." (EPAR §5.2.2.10.1) and, separately, "…food effect study (Day 101 QSC205140), no DDI studies have been performed." Both verified verbatim in EMA/67438/2026. All CYP2C8 and CYP3A language is PBPK/model-derived; CHMP accepted it as a conservative risk-minimisation approach. [FLAG: the exact phrase "a conservative risk minimisation approach" was not located verbatim in my text extraction — treat it as a paraphrase, not a quotation, until confirmed.] [FLAG: BCRP inhibition is stated with no management instruction anywhere in §7.] Co-medications that are BCRP substrates (e.g. rosuvastatin, methotrexate, topotecan) are plausible in this population; the label offers no guidance. [FLAG: aldehyde-oxidase-mediated clearance is uncharacterised clinically — verified.] "No DDI studies have been conducted to evaluate the impact of AO inhibitors or inducers on the PK of tovorafenib" (EPAR — verified verbatim). AO has no validated clinical inhibitor probe and marked species differences. [FLAG: the "metabolically redundant" reasoning and the conclusion that "SmPC amendments … are not warranted" were not located verbatim in my extraction — paraphrase only, verify before quoting.]

---

9. Specific populations

• No clinically significant PK differences by age (range: 1 to 94 years — verified verbatim), sex, race, mild hepatic impairment, or mild-to-moderate renal impairment (eGFR ≥30 mL/min/1.73 m²) (USPI §12.3 — verified).
• Not studied: severe renal impairment (eGFR <30 — "No patients with severe renal impairment … " EPAR §5.2.2.9, verified); moderate or severe hepatic impairment ("Tovorafenib has not been studied in patients with moderate or severe hepatic impairment" — EPAR, verified verbatim), despite hepatic elimination being the primary route.
• Children <2 years: CHMP imposed a Specific Obligation under the conditional MA (Article 14-a, Reg. (EC) No 726/2004): "The MAH shall generate additional PK data in paediatrics patients below 2 years of age and submit an updated population PK model incorporating these data, including an assessment of systemic exposure and, if necessary, revised dosing recommendations for this subgroup of patients." Due date 30 April 2032 (EPAR Tables 3 and 57 — verified verbatim, both tables). The other Specific Obligation, also due 30 April 2032, is the final FIREFLY-2 report. Mechanistic rationale for the gap: CYP2C8 reaches near-adult activity by ~6 months, but aldehyde oxidase activity is low in early infancy and reaches adult levels only around ~2 years, so infants approved from 6 months of age are dosed on extrapolation. [FLAG: the Vijay 2016 and Tayama 2012 ontogeny citations were not verified in this pass — confirm the primary ontogeny references before external use.] The dated regulatory gap itself is verified and is a real, citable weakness in the competitor's pediatric package.

---

10. Pharmacodynamics

Domain	Finding	Source
Target engagement (Q2D cohorts)	Median decreases from baseline in pERK staining at Day 21 of ≥70% across the five melanoma Q2D expansion cohorts; BRAF-mutation-positive treatment-naïve cohort −94.5% (range −100% to −89%); BRAF-mutation-positive previously treated −80.5% (range −100% to +2207%); NRAS-mutation-positive treatment-naïve −70.0% (range −100% to +603%)	EPAR §5.2.3.3 — verified verbatim
Target engagement (QW cohort — the approved schedule)	Median pERK staining was medium at screening and "decreased slightly" at Day 21; median change −12.0%	EPAR §5.2.3.3 — verified verbatim
Exposure–efficacy	Flat: no clinically significant exposure–ORR relationship over 290–476 mg/m² (0.76–1.25× the approved dose) by RAPNO-LGG or RANO-LGG	USPI §12.2 — verified
Exposure–toxicity	Positive: higher exposure → more skin rash, AST/ALT elevation, CPK elevation	USPI §12.2
Exposure–growth	Higher exposure → greater reduction in height-for-age z-score	USPI §12.2
Growth (clinical)	Growth AEs in 46% of 133 patients ≤18 y (35% Grade ≥3); median height-percentile change −14 (z −0.6) at 12 months (N=107)	USPI §5.4 — both verified verbatim
Growth (longer follow-up)	Median height-percentile change −20 (z −0.9) at 18 months (N=95); median annualized growth velocity 0.86–1.8 cm/y on treatment vs 4.2 cm/y in 17 patients ≥90 days off treatment	USPI §5.4 — [FLAG: these three figures were NOT re-verified; the 12-month row was. Confirm before use.]
Cardiac	At 380 mg/m² QW (≤600 mg), a mean QT increase >20 ms was not observed	USPI §12.2 — [FLAG: not re-verified verbatim.]
Visible PD marker	Hair colour change in 104/137 (76%) of FIREFLY-1 patients (denominator = all treated across both arms) — verified verbatim in Kilburn 2024; decreased melanin granule density on hair micro-structure analysis (2026 conference abstract); consistent with the labeled reversible melanin binding (EPAR §4.3.2.2)	Kilburn 2024 Nat Med; Neuro-Oncol Pediatrics 2026;2(Suppl 1) abstract ID#282

[FLAG: the pERK dataset is adult metastatic melanoma, not pLGG, and no per-cohort N is published.] More important for a head-to-head: the only published pERK data on the approved once-weekly schedule show a median change of −12.0%, versus −70% to −94.5% on the twice-weekly schedule that was abandoned for tolerability. Tovorafenib's target-engagement evidence at its own approved schedule is therefore weak. This is a strong, verified CI point — and precisely why it must be stated with the cohort, schedule and criterion attached.

Denominator discipline for the CNS efficacy figures (the efficacy workstream owns the full analysis):

Figure	Criterion	N	Reviewer	Data cut	Source
ORR 53% (95% CI 41–64), 40/76 responders (29 PR + 11 MR)	RAPNO-LGG	76	Blinded independent central review	10 May 2024	USPI §14 — verified
ORR 54% (95% CI 42–65)	RANO-LGG (2011)	76 (same patients)	BICR	10 May 2024	USPI §14 — verified
Median DoR 18 months (95% CI 12.0–22.8)	RAPNO-LGG	40 responders	BICR	10 May 2024	USPI §14 — verified
ORR 51% (95% CI 40–63)	RAPNO	76	Publication analysis	earlier cut	Kilburn 2024 — verified
ORR 67% (95% CI 54–78)	RANO-HGG	69 evaluable	Publication analysis	earlier cut	Kilburn 2024 — verified

The frequently-quoted "67%" is RANO-HGG in 69 evaluable patients — a different criterion and a different denominator from the label's 53% (RAPNO-LGG, 40/76). Note also that the same publication's own RAPNO figure is 51%, not 53% — the label's 53% comes from a later BICR analysis at the 10 May 2024 cut. Do not carry the 67% figure into any RANO/RAPNO comparison against plixorafenib, and never present 67% and 53% as if they were the same endpoint at different times.

[FLAG: TEMPORAL — a newer public dataset exists and is not reflected in the label.] Day One presented three-year FIREFLY-1 Arm 1 follow-up at SNO 2025 (24 November 2025), data cutoff 6 June 2025, median study duration 40.6 months, with treatment-free-interval and time-to-next-treatment endpoints. If this dossier is used after the efficacy workstream completes, the 6 June 2025 cut — not the 10 May 2024 label cut — is the current public state of the competitor's data.

---

11. Formulation and administration

	Tablets	For oral suspension
Strength	100 mg only (film-coated) — verified in EPAR §5.2.2.3: "oral immediate release film-coated tablet containing 100 mg tovorafenib"	25 mg/mL after reconstitution (verified, EPAR §5.2.2.3: "25 mg/ml powder for reconstitution")
Excipients	copovidone, colloidal silicon dioxide, croscarmellose sodium, magnesium stearate, microcrystalline cellulose, Opadry Orange	artificial strawberry flavor, colloidal silicon dioxide, copovidone, maltodextrin, mannitol, microcrystalline cellulose, simethicone, sodium lauryl sulfate, sucralose
Administration	Swallow whole with water; do not chew, cut or crush	Reconstitute with water; bottle nominally 300 mg; doses >300 mg require two bottles; give via supplied oral syringe or feeding tube; administer promptly after reconstitution
Packaging	Blister cards; dispense in original package	Glass bottle with press-in adaptor and oral syringe
Storage	20–25 °C (excursions 15–30 °C)	20–25 °C (excursions 15–30 °C)

Source: USPI §3, §11, §16, rev. 8/2025; EPAR §5.2.2.3.

[FLAG: the fine-grained administration and packaging details were NOT re-verified and have been softened above.] Specifically unverified: the 14 mL reconstitution volume, the 12 mL / 300 mg delivered volume, the 20 mL syringe size, the ≥12 French feeding-tube minimum, the 15-minute post-reconstitution use window, the tablet colour/shape/debossing, the exact excipient lists, the blister configurations (4×4 / 5×4 / 4×6) and the NDC numbers (82950-001-16/-20/-24 and 82950-012-01). Every one of these is a checkable label fact — open §2.4, §3 and §16 of the current SPL and confirm before any of it reaches a deliverable. The 15-minute window and the two-bottle split are load-bearing for the caregiver-burden argument in the head-to-head, so they must be confirmed, not assumed.

Formulation engineering (verified). Tovorafenib is a low-solubility BCS Class 2 compound; per the EPAR, "in order to increase the aqueous solubility and bioavailability of the active substance, the active substance is transformed from crystalline state to amorphous during manufacture," and the EPAR abbreviation list defines ASD = amorphous solid dispersion. [FLAG: the claim that crystallinity is a defined critical quality attribute alongside the specific nine-item release-specification list was not verified in my text extraction — confirm against EPAR §2.2 before asserting it.] Amorphous solid dispersions are physically metastable, with the usual cost, stability and supply-chain consequences.

Bridging (verified verbatim, EPAR §5.2.2.3). "The tovorafenib 100 mg tablet and PfR 300 mg bottle (for reconstitution with water to 25 mg/mL) were used in the pivotal clinical trial Study DAY101-001/PNOC026 (FIREFLY-1) and Phase 1 study QSC201540 (relative BE and food effect study). In the initial Phase 1 clinical studies, an earlier formulation called T1 tablets was used. The T1 tablets and commercial tablets (T2 tablets) were comparable as demonstrated by in vitro dissolution with the clinical quality control (QC). No relative BA investigation including the T1 formulation was performed."

[FLAG: the adult phase 1 exposure/PK data (Rasco 2023 / C28001) were generated with a T1 formulation that was never bioavailability-bridged in vivo to the commercial T2 product — only by in vitro dissolution. Directly relevant to any cross-study exposure comparison, and to the ≈70 h half-life figure in §5.]

Practical read-across for Fore (analyst framing, not a cited claim): the 100 mg-only tablet plus BSA-banded dosing means real-world adult-sized doses step in 100 mg increments (400 / 500 / 600 mg); the pediatric suspension requires reconstitution and a two-bottle split above 300 mg. Whether this constitutes a differentiation axis depends entirely on plixorafenib's own presentation, which is Fore-internal and is not assessed here. State the tovorafenib facts; let the Fore team supply the comparison.

---

12. Head-to-head template — mechanism/PK/formulation axis

Plixorafenib column is INTENTIONALLY EMPTY. No plixorafenib figures were supplied to this workstream, and plixorafenib is Fore-internal/embargoed. Nothing about plixorafenib has been researched, inferred or estimated anywhere in this document. The Fore team completes the right-hand column from cleared internal data; every entry must then be labeled Fore-internal — verify.

Axis	Tovorafenib (public, cited above)	Plixorafenib (FORE8394) — Fore team to complete; Fore-internal — verify
RAF inhibitor class	Type II; dimer-competent, positive cooperativity on BRAF/CRAF dimers (Hill slopes −2.89 BRAF WT, −1.55 CRAF WT, −3.18 CRAF SSDD; −0.75 on BRAF V600E monomer)	☐
Isoform coverage	BRAF V600E, BRAF WT, CRAF; poor ARAF inhibitor — EMA acknowledges ARAF activity "was not presented"	☐
Paradoxical MAPK activation	Reduced vs type I but labeled risk in NF1-LOF; W&P 5.6; NF1 patients excluded from FIREFLY-1	☐
Wild-type RAF sparing	No — inhibits WT BRAF/CRAF at similar or greater potency than V600E	☐
Enzymatic IC₅₀ (state assay)	7.1 / 10.1 / 0.7 nM (sponsor, EPAR §4.3.1.1) vs 495 / 633 / 94.2 nM (Tkacik 2023) — see flag	☐ (run side-by-side in one assay)
Efflux-transporter substrate	Not BCRP, not P-gp, not OATP1B1/1B3	☐
Preclinical brain:plasma	0.15–0.18 (CD-1 mouse); 0.24 (Sun 2017 mouse); 0.54–0.67 total radioactivity (pigmented rat QWBA)	☐
Human CNS exposure (CSF/tumor/K_p,uu)	None published	☐
Approved/planned dose & schedule	380 mg/m² once weekly (max 600 mg); all label efficacy generated at 420 mg/m²	☐
Terminal half-life	≈56 h (label popPK) / ≈70 h, range 31–119 h (adult NCA, n=20)	☐
Time to steady state	12 days	☐
Accumulation	"No clinically significant tovorafenib accumulation occurs"	☐
Food effect	None clinically significant; T_max delayed to 6.5 h with high-fat meal — but studied in 12 healthy adults at 100 mg	☐
Absolute bioavailability	Not determined (no IV arm)	☐
Metabolism / clearance route	Aldehyde oxidase + CYP2C8 (major); CYP2C9/2C19/3A4 minor	☐
Key DDI restrictions	Avoid mod/strong CYP2C8 inhibitors and inducers; CYP3A inducer — avoid hormonal contraceptives; inhibits BCRP (no management guidance given)	☐
Clinical DDI studies performed	Zero — "No clinical DDI studies were conducted" (CHMP)	☐
Protein binding	97.5% (albumin ≈95%, AAG ≈40%)	☐
Target engagement at approved schedule	Median pERK change −12.0% in the QW melanoma expansion cohort	☐
QT signal	No mean increase >20 ms at approved dose (unverified)	☐
Formulations	100 mg tablet (amorphous solid dispersion) + 25 mg/mL powder for reconstitution; T1→T2 bridged by dissolution only	☐
Pediatric coverage	≥6 months; PK <2 y is a binding EU Specific Obligation due 30 April 2032	☐
CNS response criterion used	RAPNO-LGG (label primary, 40/76 = 53%) / RANO-LGG (54%); RANO-HGG 67% in 69 in the primary publication	☐ (must be RANO/RAPNO/iRANO)

---

13. Open flags for this workstream

Blocking 1. Target-drug name did not propagate from the workflow; resolved to tovorafenib — Fore must confirm before any use.

Verified competitor weaknesses (safe to use, with the stated attribution) 2. Approved 380 mg/m² dose is modelling-and-simulation-derived only (CHMP verbatim); all label efficacy came from 420 mg/m². 3. Zero clinical DDI studies; all interaction language is PBPK-predicted (CHMP verbatim). 4. No human CSF, tumour-tissue or unbound (K_p,uu) CNS exposure data exist publicly. 5. US label makes no CNS-penetration claim; "CNS-penetrant" is sponsor/EU-assessment language. 6. Aldehyde-oxidase clearance clinically uncharacterised — no AO DDI studies conducted (CHMP verbatim). 7. Absolute bioavailability not determined — no IV arm ever run (CHMP verbatim). 8. No dedicated moderate/severe hepatic or severe renal impairment study despite hepatic elimination (CHMP verbatim). 9. Pediatric PK <2 years is a binding Specific Obligation due 30 April 2032 (verbatim, EPAR Tables 3 and 57). 10. Phase 1 (T1) formulation never bridged in vivo to commercial (T2) — dissolution only (CHMP verbatim). 11. CHMP never received the PNOC014 CSR, and PNOC014 was dropped as a key PIP measure (verbatim). 12. EMA on the record that ARAF activity "was not presented" and that tovorafenib "is not potent against ARAF" per Tkacik — undercuts the "pan-RAF" descriptor. 13. Target engagement at the approved once-weekly schedule is only −12.0% median pERK change, vs −70% to −94.5% on the abandoned Q2D schedule. 14. BCRP inhibition stated in §12.3 with no management guidance in §7.

Unresolved discrepancies (do not present either side alone) 15. ~70–100× potency gap between sponsor (7.1/10.1/0.7 nM) and Tkacik (495/633/94.2 nM). The EPAR cites Tkacik but never reconciles the numbers. 16. Half-life ≈56 h (label popPK) vs ≈70 h (adult NCA, n=20). 17. "67% ORR" is RANO-HGG in 69 evaluable; the same publication's RAPNO figure is 51% in 76; the label's is 53% (40/76) at the 10 May 2024 cut. Three different numbers, three different bases. 18. EU approval date: EMA's EPAR page states the EC decision was 20 April 2026; Ipsen's press release states 22 April 2026. Use EMA's date, or say "April 2026". 19. EU indication wording: the EMA/CHMP indication is BRAF-alteration-restricted ("BRAF fusion or rearrangement, or BRAF V600 mutation") — verified. An Ipsen press headline says "regardless of BRAF alteration". These contradict. Not used anywhere in this workstream's content; the regulatory workstream must resolve it. 20. EPAR internally names the relative-BA/food-effect study both QSC201540 and QSC205140; the peer-reviewed publication (Zhang 2025) uses QSC205140.

Unverified figures — must be checked before external use 21. Rasco 2023 C_max 5,650 ng/mL, AUC₁₆₈ 330,000 ng·h/mL, accumulation ratio 1.09 — could not be confirmed from any source; do not use. 22. Growth: 18-month height-percentile −20 (z −0.9, N=95), growth velocity 0.86–1.8 vs 4.2 cm/y (17 patients off-treatment) — not re-verified. 23. All §11 formulation logistics: 14 mL reconstitution, 12 mL/300 mg, 20 mL syringe, ≥12 French tube, 15-minute use window, NDCs, blister configurations, excipient lists. 24. Midazolam "≥20% decrease", contraceptive "28 days", QT ">20 ms", missed-dose "3 days", BSA dosing-band cut-points. 25. Sun 2017 dosing details (10 mg/kg; 30 mg/kg BID × 41 days) and the ontogeny citations (Vijay 2016, Tayama 2012). 26. Whether §12.3 is byte-identical to the April 2024 original — not verified, and supplements were approved 27 August 2025, so assume something changed until shown otherwise. 27. Sun 2017 prints the ratio as "plasma:brain ratio of 24%"; the intended reading is brain:plasma. 28. Sun 2017 carries a 2024 corrigendum (figure duplication + glass-tube methods omission) — verified; PK/BBB data unaffected, but cite the corrigendum. 29. PNOC014 full report (JCO Oncol Adv 2026) returned 403 and was not read. 30. Temporal: a newer FIREFLY-1 cut exists (6 June 2025, SNO 2025) that no label figure reflects.

CITATIONS: ["OJEMDA (tovorafenib) tablets / for oral suspension, US Prescribing Information, Day One Biopharmaceuticals; label REVISED 8/2025, SPL version posted to DailyMed 2 September 2025; NDA 217700 (tablets) / 218033 (for oral suspension); supplements approved 27 August 2025. VERIFIED CURRENT SOURCE: DailyMed setid ea3a9631-3a66-6a7c-e053-2995a90ae2ad, https://dailymed.nlm.nih.gov/dailymed/fda/fdaDrugXsl.cfm?setid=ea3a9631-3a66-6a7c-e053-2995a90ae2ad&type=display (sections 2.3, 2.4, 3, 5.4, 5.6, 7.1, 7.2, 11, 12.1, 12.2, 12.3, 13.2, 14, 16). NOTE: the accessdata URL cited in the prior draft (https://www.accessdata.fda.gov/drugsatfda_docs/label/2025/217700s002s003s004lbl.pdf) returns HTTP 404 and must not be cited.","OJEMDA (tovorafenib), US Prescribing Information, initial approval 23 April 2024. https://www.accessdata.fda.gov/drugsatfda_docs/label/2024/217700s000lbl.pdf and https://www.accessdata.fda.gov/drugsatfda_docs/label/2024/218033s000lbl.pdf (VERIFIED to resolve)","European Medicines Agency, CHMP Assessment Report: Ojemda (tovorafenib), EMA/67438/2026, 154 pp., published 11 May 2026. https://www.ema.europa.eu/en/documents/assessment-report/ojemda-epar-public-assessment-report_en.pdf — FULL TEXT RETRIEVED AND SEARCHED for this review; all quotations marked 'verified verbatim' above were confirmed against the extracted text. Sections used: 1.9.7 (Table 3, specific obligations), 2.2, 4.2, 4.3.1.1, 4.3.2.1-4.3.2.3, 5.2.2.2, 5.2.2.3, 5.2.2.4, 5.2.2.5, 5.2.2.9, 5.2.2.10.1, 5.2.3.1, 5.2.3.2, 5.2.3.3, 5.3.1, Table 57.","European Medicines Agency, Ojemda EPAR medicine overview page: authorisation status 'Authorised'; European Commission decision date 20 April 2026; conditional marketing authorisation; orphan designation EU/3/21/2434; MAH Ipsen Pharma. https://www.ema.europa.eu/en/medicines/human/EPAR/ojemda (VERIFIED)","Ipsen press release, 'Ojemda approved in the European Union...', stating an EC decision date of 22 April 2026 — CONFLICTS with the EMA EPAR page date of 20 April 2026; and carrying the headline phrase 'regardless of BRAF alteration', which CONFLICTS with the BRAF-restricted indication in the CHMP opinion and EPAR. Flagged, not relied upon.","CHMP positive opinion, 27 February 2026 (Ipsen release, GlobeNewswire 3246394): monotherapy, patients 6 months and older, pLGG harbouring a BRAF fusion or rearrangement or BRAF V600 mutation, progressed after one or more prior systemic therapies.","Tkacik E, Li K, Gonzalez-Del Pino G, et al. Structure and RAF family kinase isoform selectivity of type II RAF inhibitors tovorafenib and naporafenib. J Biol Chem. 2023;299(5):104634. doi:10.1016/j.jbc.2023.104634. PMC10149214 — FULL TEXT RETRIEVED; IC50 values, Hill slopes, ARAF sentence and DFG 'crankshaft' sentence all verified verbatim.","Sun Y, Alberta JA, Pilarz C, et al. A brain-penetrant RAF dimer antagonist for the noncanonical BRAF oncoprotein of pediatric low-grade astrocytomas. Neuro Oncol. 2017;19(6):774-785. doi:10.1093/neuonc/now261. PMID 28082416","Corrigendum to: A brain-penetrant RAF dimer antagonist for the noncanonical BRAF oncoprotein of pediatric low-grade astrocytomas. Neuro Oncol. 2024;26(5):985-986. doi:10.1093/neuonc/noae046 (VERIFIED: Figure 5 panel A vinculin duplication; replacement with independent repeat; glass-tube methods addition)","Rasco DW, et al. Phase 1 study of the pan-RAF inhibitor tovorafenib in patients with advanced solid tumors followed by dose expansion in patients with metastatic melanoma. Cancer Chemother Pharmacol. 2023;92(1):15-28. doi:10.1007/s00280-023-04544-5. PMC10261210 (t1/2 ~70 h, range 31-119 h, n=20 at 600 mg QW; no apparent accumulation; approximately dose-proportional AUC over 400-800 mg QW — VERIFIED. Cmax 5,650 ng/mL, AUC168 330,000 ng*h/mL and accumulation ratio 1.09 — NOT VERIFIED, do not use.)","Kilburn LB, Khuong-Quang DA, Hansford JR, et al. The type II RAF inhibitor tovorafenib in relapsed/refractory pediatric low-grade glioma: the phase 2 FIREFLY-1 trial. Nat Med. 2024;30(1):207-217 (published online 17 November 2023). doi:10.1038/s41591-023-02668-y. PMID 37978284. PMC10803270 — FULL TEXT RETRIEVED; ORR 67% (95% CI 54-78) in 69 evaluable by RANO-HGG, ORR 51% (95% CI 40-63) in 76 by RAPNO, arm 1 n=77 / arm 2 n=60, DoR 16.6 mo (RANO-HGG) / 13.8 mo (RAPNO), hair colour change 104/137 (76%), and the 'PK is a planned secondary endpoint not reported in this manuscript' statement all VERIFIED. NOTE: the prior draft cited this as 'Nat Med 2023;30(1)' — the volume-year is 2024. Plus Author Correction, Nat Med 2024, doi:10.1038/s41591-024-02910-1.","Zhang et al. Taste Profile and Relative Bioavailability of Tovorafenib Powder for Oral Suspension and Food Effect of the Tovorafenib Tablet in Healthy Participants. Clin Pharmacol Drug Dev. 2025. doi:10.1002/cpdd.1558. PMC12402836. Open-label randomized phase 1, study QSC205140; taste/palatability n=12; relative BA and food effect n=12; dose reduced from 300 mg to 100 mg for musculoskeletal AEs. NEWLY ADDED — this is the peer-reviewed primary source for the food-effect and tablet/suspension bridging data previously cited only to the EPAR.","Day One Biopharmaceuticals press release, 'Day One Announces Three Year Follow-Up Data From OJEMDA (tovorafenib) Phase 2 FIREFLY-1 Trial at the 2025 Society for Neuro-Oncology (SNO) Annual Meeting', 24 November 2025; FIREFLY-1 Arm 1, data cutoff 6 June 2025, median study duration 40.6 months; label ORR 53% (40/76) confirmed as the 10 May 2024 data cut. https://www.globenewswire.com/news-release/2025/11/24/3193443/0/en/","ClinicalTrials.gov NCT04775485 (FIREFLY-1 / PNOC026)","ClinicalTrials.gov NCT01425008 (Study C28001)","FDA Approval Summary: Tovorafenib for Relapsed or Refractory BRAF-Altered Pediatric Low-Grade Glioma. Clin Cancer Res. 2025;31(8):1383-1389. PMID 39808502 (accelerated approval 23 April 2024)","PNOC014: Phase Ib study results of DAY101 (tovorafenib) for children with low-grade gliomas and other RAS/RAF/MEK/ERK pathway-activated tumors. Neuro-Oncology. 2022;24(Suppl 7):vii84, abstract CTNI-53. doi:10.1093/neuonc/noac209.318","PNOC014: Phase I Study of Tovorafenib for Children With Relapsed/Recurrent Low-Grade Gliomas and Other MAPK Pathway-Activated Tumors. JCO Oncology Advances. doi:10.1200/OA-26-00010 [full text NOT retrieved - paywall 403; details above are from the EPAR, which itself notes the PNOC014 CSR was not submitted]","Tovorafenib-related hair colour changes: a study of hair micro-structure and relation to tumour response in low grade gliomas. Neuro-Oncology Pediatrics. 2026;2(Suppl 1):wuag026.080, abstract ID #282 [conference abstract - not peer-reviewed full text]","Rastogi et al. Preclinical Activity of the Type II RAF Inhibitor Tovorafenib in Tumor Models Harboring Either a BRAF Fusion or an NF1 Loss-of-Function Mutation. Cancer Res Commun. 2025;5(4):668- [full text NOT retrieved - publisher 403]","van Tilburg CM, et al. LOGGIC/FIREFLY-2: a phase 3, randomized trial of tovorafenib vs. chemotherapy in pediatric and young adult patients with newly diagnosed low-grade glioma harboring an activating RAF alteration. BMC Cancer. 2024;24:147. doi:10.1186/s12885-024-11820-x. PMID 38291372"]

3. Clinical Efficacy — Tovorafenib (OJEMDA, Day One Biopharmaceuticals / Servier; Ipsen ex-US)

Data cut for this workstream: 31 August 2026. Public sources only. Every response figure below carries its denominator, its response criterion and its assessor (IRC/BICR vs investigator). Where a figure is carried forward from the prior 24 Aug 2026 dossier build and was not independently re-verified against the primary document in this refresh, it is marked [carried — re-verify].

Verification status of this refresh. The following were re-verified against primary documents on 31 Aug 2026 and are marked (verified) inline: the full OJEMDA USPI §14 (DailyMed SPL, setid ea3a9631-3a66-6a7c-e053-2995a90ae2ad); the full EU Joint Clinical Assessment Report of Tovorafenib, Version 1.0 (Tables 25, 26, 27 and §4.2–4.3, PDF read in full); ClinicalTrials.gov API v2 records for NCT04775485, NCT05566795, NCT04985604 and NCT07441707; Kilburn Nat Med 2024 (PMC10803270); Kline Neuro-Oncol Pediatr 2026 wuag022; the EMA Ojemda product page (indication text and EPAR reference EMA/67438/2026); and the PubMed records for all six cited journal articles (title, volume, issue, pagination). Everything else remains carried.

What is new since the 24 Aug 2026 build: 1. The EU Joint Clinical Assessment (JCA) (endorsed 30 Apr 2026, published 9 Jun 2026) — the first formal, publicly published indirect head-to-head of tovorafenib vs dabrafenib + trametinib in BRAF V600E pLGG. It is materially unfavourable to tovorafenib on PFS. This was not in the prior build. (Verified this refresh against the full JCA Report PDF, not the Summary Report.) 2. FIREFLY-2 is now ACTIVE_NOT_RECRUITING with n=418 ACTUAL enrolled (not "approximately 400 estimated"), primary completion June 2027, study completion June 2031. (Verified via CT.gov API v2, 31 Aug 2026.) 3. G-BA (Germany) published its orphan benefit assessment on 17 Aug 2026; efficacy content is not exposed on the public landing page.

---

3.1 Trial map — what exists and what does not

All status/enrolment cells below were pulled from the ClinicalTrials.gov API v2 on 31 Aug 2026 (verified).

Trial	NCT	Phase	Population	Status (CT.gov, retrieved 31 Aug 2026)	Efficacy in a label?
FIREFLY-1 (PNOC026) — whole study (Arms 1–3)	NCT04775485	2	R/R pLGG + advanced solid tumours; NF1 excluded	RECRUITING; 141 ESTIMATED (whole study, all arms); PCD 31 May 2027; last update posted 10 Apr 2025	—
FIREFLY-1 Arm 1 (n=77 enrolled)	same	2	R/R pLGG, activating BRAF alteration, ≥1 prior systemic line	same	Yes — sole basis of both US and EU authorisations
FIREFLY-1 Arm 2 (n=60)	same	2	pLGG extension	same	No — contributes to safety pooling only
FIREFLY-1 Arm 3 (≤20 pts)	same	2	Advanced solid tumours	same; ORR is a listed primary outcome	No results ever reported
LOGGIC/FIREFLY-2	NCT05566795	3	Front-line LGG <25 y, activating RAF alteration; NF1 excluded	ACTIVE_NOT_RECRUITING; 418 ACTUAL; start 27 Feb 2023; PCD Jun 2027; completion Jun 2031; last update 14 May 2026	Not yet — confirmatory trial for both accelerated/conditional approvals
FIRELIGHT-1 (incl. sub-study 102a)	NCT04985604	2	Adult melanoma / solid tumours, RAF fusion or RAF1 amplification	TERMINATED ("Sponsor decision"), 23 ACTUAL, PCD 8 Jul 2024, last update 2 Oct 2025, results posted	No
Japanese PK/safety study	NCT07441707	1	Japanese children/young adults, brain tumours	RECRUITING; 6 ESTIMATED; start 3 Mar 2026; PCD 31 Jul 2030; last update 30 Jul 2026; sponsor Ipsen	No — primary endpoints are AEs and PK only; no ORR endpoint
Adult phase 1 (Millennium legacy)	NCT01425008	1	Advanced solid tumours / melanoma	Completed	No

[FLAG: the previous build placed "141 estimated" on the Arm 1 row. 141 is the ESTIMATED enrolment for the whole of NCT04775485 (all three arms), not for Arm 1. Arm 1 actually enrolled 77. Corrected above.]

[FLAG — TEMPORAL: the FIREFLY-1 registry record was last updated 10 Apr 2025, roughly 16 months before this data cut. Its "RECRUITING" status and 141 estimate may not reflect current reality. Do not present FIREFLY-1's registry status as current without a sponsor-side confirmation.]

Structural read-through: the entire tovorafenib efficacy record rests on one single-arm, open-label arm of 76–77 relapsed/refractory paediatric LGG patients. There is no randomised efficacy evidence in any source, no overall-survival estimate in any source, and no registrational adult programme and no adult CNS efficacy endpoint in any active trial.

---

3.2 Registrational dataset — baseline characteristics (head-to-head ready)

Two overlapping denominators are in circulation and are routinely conflated. Keep them separate. The label column was read in full from the DailyMed SPL and is (verified).

Characteristic	Arm 1 enrolled, n=77 (Kilburn, Nat Med 2024;30(1):207-217; DCO 5 Jun 2023)	Label efficacy population, n=76 (USPI §14, rev. 8/2025) (verified)	Pooled safety population, n=137 (Kline, Neuro-Oncol Pediatr 2026;2(2):wuag022)
Definition	All Arm 1 enrolled	Arm 1 patients with measurable disease at baseline who received OJEMDA	Arm 1 (77) + Arm 2 (60)
Median age	8 y (range 2–21)	8.5 y (range 2–21)	9 y (range 1–24) [carried]
Male	40 (52%)	53%	53% [carried]
Race — White	41 (53%)	61%	64% [carried]
Astrocytic histology	72 (94%)	not stated	93% [carried]
Optic pathway primary	39 (51%)	51%	50% [carried]
Deep midline	9 (12%)	12%	15% [carried]
Brainstem / cerebellum / hemisphere	6 (8%) / 5 (6%) / 6 (8%)	8% / 7% / 5%	—
KIAA1549::BRAF fusion	56 (73%)	56 (74%)	—
Other BRAF alteration ("other", incl. duplication or rearrangement)	8 (10%)	8 (11%)	—
BRAF V600E mutation	13 (17%)	12 (16%)	—
Median prior systemic lines	3 (range 1–9)	3 (range 1–9)	3 (range 1–10) [carried]
Prior MAPK-pathway inhibitor	MEKi 43 (56%); BRAFi 8 (10%)	45 (59%) any MAPKi	61% [carried]
Prior radiotherapy	6 (8%)	not stated	4% [carried]
Extent of resection	GTR 1 (1%) / STR 36 (47%) / biopsy only 40 (52%)	not stated	STR 47% [carried]
Performance status (Karnofsky/Lansky 80–100)	68/71 (96%)	93%	—
NF1	Excluded by protocol	Excluded	Excluded
Localized disease	—	—	74% [carried]

Comparability caveats for any head-to-head: (i) this is an entirely paediatric/AYA population (upper bound 21–25 y) with no adults; (ii) it is fusion-dominant — 84% of the label population is fusion/rearrangement/other (64/76), only 16% V600E (12/76); (iii) it is heavily MAPKi-pretreated (45/76 = 59%); (iv) NF1-driven and co-altered tumours (IDH1/2, FGFR) were excluded by protocol, so the label population is not the real-world pLGG population; (v) the target dose studied was ~420 mg/m² QW — 1.11× the approved 380 mg/m².

[FLAG — CORRECTED OVERREACH: the prior build stated "the approved dose was never itself tested for efficacy." The USPI §14 contradicts this. Verbatim: patients received "approximately 420 mg/m2 orally once weekly (range: 290 to 476 mg/m2, 0.76-1.25 times the approved recommended dosage)". The approved 380 mg/m² therefore lies inside the range actually administered. The defensible claim is narrower: the *target/median* dose was ~420 mg/m², above the approved dose, and no dedicated 380 mg/m² efficacy cohort was evaluated. Do not repeat the "never tested" formulation.]

---

3.3 Efficacy in the current US label — the registration figures

Source: OJEMDA USPI §14 Table 10, revision 8/2025 (DailyMed SPL setid ea3a9631-3a66-6a7c-e053-2995a90ae2ad), read in full 31 Aug 2026 (verified). BICR-assessed for RAPNO-LGG; independent review for RANO-LGG. Tumour assessments q12w.

Endpoint	Criterion	Result	Denominator
ORR	RAPNO-LGG, BICR	53% (95% CI 41, 64) — PR 29 (38%), MR 11 (14%)	76
ORR	RANO-LGG (2011), independent review	54% (95% CI 42, 65) — PR 23, MR 18	76
Median DoR	RAPNO-LGG	18 months (95% CI 12.0, 22.8); 65% with observed DoR ≥12 mo; 50% ≥18 mo	40 responders
Median TTR	RAPNO-LGG	5.4 months (range 1.6, 17.5)	40 responders
PFS	—	Not in the US label	—
OS	—	Not reported in any identified source	—

[FLAG — CR count is inferred, not stated: USPI Table 10 lists only PR and MR rows and does not print a CR line. Because ORR (CR+PR+MR) = 53% = 40/76 and PR 29 + MR 11 = 40, CR = 0 follows arithmetically. State it as "no CRs reported in USPI Table 10", not as a printed label figure.]

[FLAG: the 8/2025 USPI does not state its own data-cutoff date — confirmed by reading the full SPL. The EU JCA cites the supporting CSR as FIREFLY-1 "2 year" data cut, and the EMA-assessed RAPNO figure (52.6%, 40/76) is arithmetically identical to the 8/2025 USPI Table 10 (29 PR + 11 MR = 40/76). A 10 May 2024 cutoff for the 8/2025 label is therefore a plausible inference, not a stated fact, and the JCA table headers say "2 year", not a calendar date. Confirm against the FDA supplement approval package before external use.]

Superseded — do not mix with the above. At accelerated approval (label rev. 4/2024, DCO 5 Jun 2023): ORR 51% (95% CI 40–63), n=76, RAPNO-LGG, median DoR 13.8 months (11.3–NE) (Singh S et al., Clin Cancer Res 2025;31(8):1383-1389, PMID 39808502 — citation (verified); figures [carried]).

---

3.4 The criterion problem — a 32.8-point spread on one dataset

The pre-specified primary endpoint of FIREFLY-1 was ORR by RANO-HGG — criteria designed for adult high-grade glioma, applied to paediatric low-grade glioma. On the same patients, the same data cut:

Criterion / assessor	ORR	Denominator	Source
RANO-HGG, IRC	71.0% (58.8, 81.3) — 16 CR (23.2%), 33 PR (47.8%)	49/69	EMA/67438/2026 Table 17 [carried — re-verify]
RANO-HGG, investigator	49.3% — 5 CR (7.2%), 29 PR (42.0%)	34/69	EMA/67438/2026 Table 18 [carried — re-verify]
RANO-LGG, IRC	53.9% (42.1, 65.5)	41/76	EMA/67438/2026 [carried — re-verify]; USPI §14 gives 54% (42, 65), 23 PR + 18 MR (verified)
RAPNO-LGG, IRC (headline, MR counted as response)	52.6%	40/76	EMA/67438/2026 Table 19 [carried — re-verify]; USPI §14 = 53% (41, 64) (verified)
RAPNO-LGG, CR+PR only (MR excluded)	38.2%	29/76	Arithmetic from USPI §14 Table 10 (verified); EMA Table 19 [carried]

The EPAR reference number EMA/67438/2026 was confirmed on the EMA Ojemda product page in this refresh; the individual table numbers and verbatim CHMP quotations below were not.

Four points, three of which were independently verified in this refresh: - FDA removed RANO-HGG from the US label entirely. USPI §14 (rev. 8/2025, read in full) names only RAPNO-LGG and RANO-LGG. The 71% figure appears nowhere in the US label. (Verified.) - The EU JCA formally refused the sponsor's RANO-LGG response analysis, and its stated reason is the MR asymmetry. Verbatim: "Relative effectiveness results for IRC-assessed ORR based on RANO-LGG criteria are not included in Table 26 due to clear bias arising from methodological flaws and inappropriate analysis methods used by the HTD... different definitions of objective response are used for the intervention and comparator arms, with the MR category classified as an objective response in FIREFLY-1 but as a non-response in Bouffet 2023... the inconsistency in definitions between studies leads to bias in favour of tovorafenib." It then fell back on RANO-HGG and flagged that itself: "Relative effectiveness results for ORR presented in Table 26 are based on criteria designed for HGG and not specifically for LGG." (All three quotations verified verbatim in the full JCA Report §4.3.2.3.2.) - Minor response counts as a responder in the headline number. Removing MR takes the registration ORR from 52.6% (40/76) to 38.2% (29/76) — arithmetic on USPI Table 10 (verified). - CHMP judged RANO-HGG unsuitable, reportedly describing it as giving "approximately 20% larger magnitude of improvement", stating RANO-HGG criteria "are actually designed for adult high-grade gliomas… not suitable for LGG", and calling the IRC-vs-investigator outcomes "very discordant" [carried — re-verify verbatim quotations against EMA/67438/2026. These quotations were NOT found in the JCA document and are not corroborated by anything verified in this refresh.]

[FLAG — CORRECTED VERBATIM QUOTATION: the prior build rendered the JCA's rejection of the RANO-LGG analysis as "…is not included due to the use of methodologically flawed and inappropriate analysis methods." That is a paraphrase presented inside quotation marks. The verbatim source wording is reproduced above. Never place quotation marks around a reconstruction.]

[FLAG: the prior build asserted "Ipsen's own press material still headlines figures derived from RANO-HGG." No supporting citation was supplied and this was not verified in this refresh. Either cite the specific press item and quote it, or drop the claim.]

Practical rule: any tovorafenib ORR quoted without naming criterion and assessor is uninterpretable — the same 76-patient dataset yields anything from 38.2% to 71.0%.

---

3.5 Primary publication — all three criteria, one data cut

Kilburn LB, Khuong-Quang D-A, Hansford JR, et al. Nat Med 2024;30(1):207-217 (PMID 37978284; PMC10803270), DCO 5 June 2023, Arm 1. Full results table re-read in this refresh (verified):

Criterion	ORR	n	Best response breakdown	CBR	Median DoR	Median TTR	Median PFS
RANO-HGG (primary)	67% (54–78)	46/69	12 CR (17%), 34 PR (49%), 18 SD (26%), 4 PD (6%)	93%	16.6 mo (11.6–NR)	3.0 mo (2.6–16.6)	19.4 mo (16.9–NR)
RAPNO	51% (40–63)	39/76	28 PR (37%), 11 MR (14%), 23 SD (30%), 13 PD (17%)	82%	13.8 mo (11.3–NR)	5.3 mo (1.6–11.2)	13.8 mo (8.3–16.9)
RANO-LGG (post hoc, exploratory)	53% (41–64)	40/76	20 PR (26%), 20 MR (26%), 23 SD (30%), 11 PD (14%)	83%	14.4 mo (11.0–NR)	5.5 mo (1.6–11.3)	13.9 mo (11.1–19.1)

[FLAG: the RANO-HGG responder count is 46/69 (67%) in Nature Medicine at DCO 5 Jun 2023, while the EMA assessment reports 49/69 (71.0%) at a later cut. These are different data cuts, not a contradiction — but never cite "67%" and "49/69" in the same row, and never cite "71%" against the 5 Jun 2023 cut.]

---

3.6 Three-year update — the most current published efficacy cut

Kline C, Landi D, Khuong-Quang D-A, Nysom K, Kilburn LB, et al. "Clinical stability after tovorafenib treatment in patients with relapsed/refractory pediatric low-grade glioma: updated results from the phase 2 FIREFLY-1 trial." Neuro-Oncology Pediatrics 2026;2(2):wuag022, published 22 Apr 2026. DCO 6 June 2025. Also presented as SNO 2025 abstract CTP-17, Neuro-Oncology 2025;27(Suppl 5):v149. Key figures re-verified against the article in this refresh (verified).

Median follow-up (Arm 1): 40.6 months (range 1.1–49.1). (verified)

Endpoint	Criterion	Result	Denominator
ORR	RAPNO-LGG, IRC	53% (95% CI 41–64) — 30 PR (39%), 10 MR (13%), 22 SD (29%), 13 PD (17%), 1 NE; no CRs reported	76
Median DoR	RAPNO-LGG	19.4 months (95% CI 13.8–27.2)	40 responders
Median TTR	RAPNO-LGG	5.4 months (range 1.6–17.5)	40 responders [carried]
Median best tumour-size change	—	−47.3% (range −97.3 to +162.0)	76 [carried]
Median PFS	RAPNO-LGG	16.6 months (95% CI 8.3–19.1)	76
Median radiographic PFS	RAPNO-LGG	16.6 months (95% CI 10.9–22.0)	76 [carried]
PFS at 12 / 24 / 36 months	RAPNO-LGG	59% / 32% / 20%	76 [carried]
Clinical PFS	composite	Not reached (44.1–NE)	76 [carried]
Median time to next treatment (TTNT)	exploratory	42.6 months (95% CI 36.6–NE)	76
≥26 cycles completed	—	44/76 (58%)	76 [carried]
Median OS	—	Not reported — no OS estimate identified in any tovorafenib source	—

[FLAG — CORRECTED TRANSCRIPTION: the TTNT confidence interval is 36.6–NE, not 36.7–NE. Corrected above against the source.]

[FLAG: two rows carry the identical point estimate of 16.6 months with different confidence intervals ("median PFS" 8.3–19.1 vs "median radiographic PFS" 10.9–22.0). The distinction between the two endpoint definitions is not stated here and only the first was verified. Define both, or drop the second row, before any external use.]

Note the two-year drift in the response breakdown: the label cut records 29 PR + 11 MR; this cut (DCO 6 Jun 2025) records 30 PR + 10 MR. One MR converted to PR; the headline ORR is unchanged at 53% because MR counts as a response. Still no CR by RAPNO-LGG at 40.6 months' median follow-up — no complete radiographic response has been recorded under LGG criteria in any published cut.

Endpoint-substitution warning. TTNT (42.6 mo) is an exploratory, unblinded, non-radiographic composite that no regulator has endorsed, and it is roughly 2.6× the RAPNO PFS (16.6 mo). The sponsor argues this position directly — reportedly "Time to next treatment is a better assessment of clinical benefit than RAPNO PFS", and "Deep responses induced by tovorafenib lower the PD threshold — modest measurement variability on scans can score as PD" (Day One SNO 2025 slide deck, slide 7) [carried — re-verify these quotations and the slide numbering; the deck is public at dayonebio.com and was NOT opened in this refresh]. Expect TTNT rather than PFS in sponsor messaging. [FLAG: 42.6 mo TTNT against a median follow-up of only 40.6 mo means the TTNT median is estimated beyond the bulk of the observation window and rests on censored data; the upper CI bound is NE. Treat as immature.]

---

3.7 Treatment beyond progression, off-treatment durability, rebound

Continuation past progression (SNO 2025 deck; wuag022): 38/76 Arm 1 patients (50%) had RAPNO-defined progression while on tovorafenib and 100% (38/38) were continued past progression; of those, 66% (25/38) completed ≥26 cycles, 76% (29/38) had tumour-size reduction and 45% (17/38) had further reduction. A reported 58 patients had RAPNO-defined PD events in total by the 6 Jun 2025 cutoff [carried — re-verify; note this is a different quantity from the 13 patients whose *best* response was PD, and the two must not be conflated]. Universal treatment-beyond-progression makes the PFS curve partly a treatment-policy artefact and is a plausible mechanism by which TTNT diverges from PFS.

Post-treatment observation (n=39; wuag022) (verified): - 31/39 (79%) remained off therapy at cutoff; 30/39 (77%) treatment-free ≥12 months; median treatment-free interval not reached [median follow-up in this subset: 16.0 months — carried]. - 12/39 (31%) had tumour rebound (≥25% growth within 6 months of stopping). - Median treatment duration 24.6 months; median tumour-size change at discontinuation −51% to −55% [carried]. - Retreatment: n=8 (21%), median 10.5 cycles (range 2–20), all 8 still on therapy at cutoff. Median size change on retreatment −38% (range −80.9 to 0) and median 9.0 months on retreatment [carried — the cycle count and the "all 8 still on therapy" statement are verified; the −38% and 9.0-month figures are not]. - Rebound skewed to V600E: 4 of the 12 rebounders were BRAF V600 vs 2 of 27 non-rebounders (SNO 2025 slide annotation) [carried — re-verify].

[FLAG — DENOMINATOR: n=39 is a self-selected subset of the 76 — patients who did well enough to stop treatment electively. Off-treatment durability rates are not generalisable to the enrolled population and must never be quoted against another drug's ITT figures.]

---

3.8 Efficacy by molecular subgroup

Subgroup	Criterion	ORR	Denominator	Source
BRAF fusion / rearrangement / other	RAPNO-LGG	53%	n=64 subgroup	USPI §14 (verified — bare percentage and n only)
BRAF fusion / rearrangement	RAPNO-LGG (detail)	53.1% (40.2, 65.7); 24 PR + 10 MR, no CR	34/64	EMA Table 23 [carried — the CI and the PR/MR split are NOT in the USPI]
BRAF fusion / rearrangement	RANO-HGG	74.6%	44/59	EMA Table 22 [carried — re-verify]
BRAF V600E mutation	RAPNO-LGG	50%	n=12 subgroup	USPI §14 (verified — bare percentage and n only)
BRAF V600E mutation	RAPNO-LGG (detail)	50.0% (21.1, 78.9); 5 PR + 1 MR, no CR, 3 PD (25%)	6/12	EMA Table 23 [carried — the CI and the split are NOT in the USPI]
BRAF V600E mutation	RANO-HGG, IRC	50.0% (18.7, 81.3) — CR+PR, unweighted	5/10	EU JCA Report Table 26 (verified)
BRAF V600E mutation	RANO-HGG, investigator	40.0% (12.2, 73.8) — CR+PR, unweighted	4/10	EU JCA Report Table 26 (verified)
Class II (non-V600 point mutants)	—	No data identified in any label or source	—	—
Class III (kinase-impaired)	—	No data identified in any label or source	—	—
NF1-driven	—	Excluded from FIREFLY-1 and FIREFLY-2; USPI warns tovorafenib may activate MAPK in NF1-LOF cells	—	USPI §5.6/§13.2 [carried — exclusion verified in USPI §14; the §5.6/§13.2 paradoxical-activation warning was not re-read]

[FLAG — CORRECTED ATTRIBUTION: the prior build tagged the subgroup confidence intervals and CR/PR/MR splits as "USPI §14 (verified)". They are not in the label. USPI §14 states only: "the ORR was 53% among patients with BRAF fusion or rearrangement (n=64), and 50% among patients with BRAF V600E mutation (n=12)". Every CI and every response-category split in this table comes from the EMA assessment and remains carried.]

Denominator trap. The BRAF V600E subgroup carries four different denominators across the public record depending on the analysis: 13 (Arm 1 enrolled, Nat Med), 12 (label RAPNO-evaluable), 10 (RANO-HGG evaluable, JCA/EMA), 22 (Arms 1+2 pooled, JCA). A "50% ORR in V600E" is 6 of 12 or 5 of 10 — never more than a handful of events, with a 95% CI spanning roughly 19–81%. Any comparative claim in the V600 subgroup is statistically uninformative on its own.

---

3.9 Efficacy by prior therapy

Subgroup	Criterion	ORR	Subgroup size	Responders	Deep response (CR+PR)	Median DoR
Prior MAPK-pathway inhibitor	RAPNO-LGG	51% (verified); CI (35.8, 66.3) [carried]	45 (59% of the label population) (verified)	23/45	PR 15 (33.3%) + MR 8 (17.8%) [carried]	16.6 mo (12.0–19.8) [carried]
MAPKi-naive	RAPNO-LGG	55% (verified); CI (36.0, 72.7) [carried]	31 (verified)	17/31	PR 14 (45.2%) + MR 3 (9.7%) [carried]	24.0 mo (10.9–NE) [carried]
Prior MAPKi	RANO-HGG	73.2%	41	30/41	—	— [carried]
MAPKi-naive	RANO-HGG	67.9%	28	19/28	—	— [carried]
Progressed on a prior MEK inhibitor	RANO-HGG	33%	15	not stated	—	— [carried]
Progressed on a prior MEK inhibitor	RAPNO	33%	13	not stated	—	— [carried]
Progressed on a prior MEK inhibitor	RANO-LGG	30%	12	not stated	—	— [carried]

Sources: USPI §14 for the 51%/n=45 and 55%/n=31 figures (verified); EMA Table 24 and wuag022 for the CIs, the CR/PR/MR splits and DoR [carried]; Kilburn, Nat Med 2024, for the "progressed on prior MEKi" rows [carried].

[FLAG — ARITHMETIC: the three "progressed on a prior MEK inhibitor" rows carry three different denominators (15, 13, 12) for what is described as one subgroup, and the percentages do not resolve to whole numbers of patients (33% of 13 = 4.3; 30% of 12 = 3.6). Either the denominators are criterion-evaluable subsets or the percentages are rounded from unstated counts. Re-verify against Kilburn 2024 and print responder counts before external use.]

Read-through: headline ORR is largely preserved after MAPKi exposure (51% vs 55%) but the deep-response fraction is not — CR+PR falls from 45.2% to 33.3%, and median DoR shortens from 24.0 to 16.6 months. In patients who actually progressed on a MEK inhibitor, response drops to roughly one third across all three criteria. The preserved headline is largely an MR artefact. [Note: the deep-response and DoR components of this read-through rest entirely on carried EMA/wuag022 figures, not on the verified USPI text.]

Prior-lines subgroup (wuag022): median DoR 22.8 mo with 1–2 prior lines vs 16.6 mo with ≥3 [carried].

---

3.10 Optic pathway glioma subset — the functional-benefit claim

Nysom K, et al. "Radiographic and visual response to the type II RAF inhibitor tovorafenib in children with relapsed/refractory optic pathway glioma in the FIREFLY-1 trial." Neuro Oncol 2025;27(5):1341-1355 (PMID 39700439, PMC12187376) — citation (verified via PubMed: title, volume, issue, pagination). DCO 5 Jun 2023.

Measure	Result	Denominator
Radiographic ORR, RANO-HGG	64%	25/39
Radiographic ORR, RAPNO	50%	21/42
Radiographic ORR, RANO-LGG	55%	23/42
CBR	95% / 88% / 90% (HGG / RAPNO / LGG)	as above
Median TTR	~5.5 months across criteria	—
Median DoR	16.8 / 13.8 / 14.4 months	—
Vision preserved (per patient)	80%	28/35
Vision improved	31%	11/35
Vision stable	49%	17/35
Vision worsened	20%	7/35
Vision preserved (per eye)	87%	45/52
MR → confirmed PR conversion (RAPNO)	26%	11/42

Visual response defined as ≥0.2 logMAR improvement at two consecutive visits. [carried — all figures in this table re-verify against the full text; only the citation metadata was verified in this refresh.]

[FLAG — DENOMINATOR MISMATCH: this subset uses n=42 (RAPNO/RANO-LGG) and n=39 (RANO-HGG), while the baseline table in §3.2 records 39 (51%) patients with an optic pathway primary site. The 42 almost certainly reflects a broader "optic pathway involvement" definition, but the two definitions are not reconciled in the public record. Do not describe the OPG subset as "51% of the registration population" without stating which definition the 42 uses. Note also that the vision denominators (35 patients, 52 eyes) are smaller still — a subset of a subset.]

---

3.11 The only published comparative (indirect) efficacy analysis: tovorafenib vs dabrafenib + trametinib in BRAF V600E

Source: Member State Coordination Group on HTA, Joint Clinical Assessment Report of Tovorafenib, Version 1.0, European Union, Brussels, 2026. Endorsed 30 Apr 2026; published 9 Jun 2026. The full Report PDF was downloaded and read in this refresh; all figures below are from Table 26 (dichotomous outcomes) and Table 27 (time-to-event outcomes) (verified). Basis: FIREFLY-1 CSR ("2 year" data cut) vs Bouffet E et al., J Clin Oncol 2023;41(3):664-674. Method: unanchored MAIC.

[FLAG — CORRECTED CITATION: the prior build attributed these figures to "Summary Report Table 6". They were verified in the full JCA Report at Tables 26 and 27. The Summary Report is a separate, shorter document and its table numbering was not checked. Cite the full Report.]

PICO 5 — Population 2 (BRAF V600E, >1 year of age, ≥1 prior systemic therapy). The naïve (unweighted) descriptive data and the MAIC effect estimates are separated below, because they do not correspond.

Descriptive inputs (naïve, unweighted — these are the only tovorafenib cells the sponsor reported; the MAIC-weighted tovorafenib cells are "NR"):

Outcome	Criterion / assessor	Tovorafenib (FIREFLY-1)	Dabrafenib + trametinib (Bouffet 2023)
ORR (CR+PR)	RANO-HGG, IRC	5/10 = 50.0% (18.7, 81.3)	7/36 = 19.4% (8.2, 36.0)
ORR (CR+PR)	RANO-HGG, investigator	4/10 = 40.0% (12.2, 73.8)	19/36 = 52.8%
Median PFS	RANO-LGG, IRC	13.67 months (4.60, 24.87), n=12, 1 censored	36.9 months (36.0, NE), n=36, 27 censored
OS; HRQoL (generic and disease-specific); disease symptoms; symptomatic disease control	—	"Relative effectiveness results not available"	—

Effect estimates (all four analyses reported by the JCA, not just the base case):

Analysis	ESS (FIREFLY-1)	ORR OR, RANO-HGG IRC	ORR OR, RANO-HGG INV	PFS HR	PFS RMST difference (28.43-mo horizon)
MAIC base case	6.64 (ORR) / 5.81 (PFS)	7.26 (0.98, 53.68), p=0.052	0.56 (0.08, 3.84), p=0.551	4.88 (2.14, 11.14), p=0.011	−6.64 mo (−13.16, −0.12), p=0.046
MAIC scenario 1	10 (ORR) / 11.54 (PFS)	4.16 (0.83, 21.01), p=0.084 or 4.41 (0.93, 18.37), p=0.061	0.60 (0.13, 2.83), p=0.515 or 0.60 (0.14, 2.48), p=0.477	5.44 (2.26, 13.05), p=0.001	−9.51 mo (−15.10, −3.92), p=0.002
MAIC scenario 2	10.66 (PFS)	—	—	5.22 (2.24, 12.18), p<0.001	−7.77 mo (−13.25, −2.29), p=0.004
Naïve (unweighted)	10 (ORR) / 12 (PFS)	the other of 4.16 / 4.41	the other of the two 0.60 values	5.09 (2.03, 12.80), p<0.001	−8.35 mo (−13.98, −2.72), p=0.004

6- and 12-month PFS rates, RANO-LGG IRC (dabrafenib+trametinib: 86% [75, 98] at 6 mo; 83% [71, 96] at 12 mo):

Timepoint	Tovorafenib rates reported across the four analyses	Relative risks reported
6 months	94% (85, 100) · 87% (73, 100) · 83% (65, 100) · 80% (61, 100)	RR 1.09 (0.92, 1.30) · 1.01 (0.81, 1.26) · 0.97 · 0.93 (0.69, 1.26)
12 months	76% (52, 100) · 63% (41, 97) · 58% (36, 94) · 56% (34, 92)	RR 0.92 (0.61, 1.38) · 0.76 (0.48, 1.20) · 0.70 (0.42, 1.15) · 0.67 (0.40, 1.14)

[FLAG — SELECTIVE READING, MATERIAL: the prior build reported only "6-month PFS 94%, n=12, RR 1.09" and "12-month PFS 76%, n=12, RR 0.92", i.e. the highest tovorafenib value in each row, and labelled both as the unweighted n=12 column. JCA Table 26 reports four tovorafenib rates per row (base case, two scenarios, naïve). The mapping of each rate to its analysis could not be resolved from the extracted table layout, so the attribution to n=12 is unconfirmed. A 76% 12-month PFS rate also sits uneasily with a median PFS of 13.67 months, whereas 56–58% does not. Do not quote 94%/76% as "the" tovorafenib PFS rates; quote the range, or resolve the mapping against the published PDF.]

[FLAG — CORRECTED CLAIM: the prior build stated "The tovorafenib cells are unadjusted (pre-MAIC-weighting) FIREFLY-1 data — the sponsor did not report the weighted values." This is true for ORR and for median PFS (all MAIC rows read "NR"). It is false for the 6- and 12-month PFS rates, where multiple weighted tovorafenib rates are reported. Narrow the claim.]

[FLAG — DENOMINATOR/ESTIMATE MISMATCH: the base-case OR of 7.26 and HR of 4.88 are MAIC-weighted and cannot be reconstructed from the descriptive cells above. The crude odds ratio from 5/10 vs 7/36 is ~4.1, not 7.26; the naïve HR is 5.09, not 4.88. Never present "5/10 vs 7/36 → OR 7.26" or "13.67 vs 36.9 months → HR 4.88" as a single row; that is how the prior build's table was laid out and it invites exactly the wrong reconstruction.]

What this analysis supports. In the BRAF V600E subgroup — the subgroup where tovorafenib competes directly with an established, fully approved combination — the only formal comparative analysis in the public domain shows a progression hazard against tovorafenib in every one of the four analyses reported (HR 4.88 to 5.44, all with confidence intervals excluding 1), alongside a median PFS of 13.7 vs 36.9 months. The consistency across the base case, both sensitivity scenarios and the naïve comparison is the more defensible framing than the base-case HR alone.

What this analysis does not support. The ORR direction reverses between assessors (OR 7.26 by IRC vs OR 0.56 by investigator), and the unexplained assessor discordance the JCA identifies is in the comparator study, not in FIREFLY-1. Verbatim: "There is a substantial difference between INV- and IRC-assessed ORR in Bouffet 2023. The reason for this difference is unclear." Dabrafenib+trametinib reads 19/36 (52.8%) by investigator vs 7/36 (19.4%) by IRC; tovorafenib's own gap in this subgroup is small (4/10 vs 5/10). The ORR flip is therefore driven mainly by the comparator's assessor variance, and neither ORR estimate reaches conventional significance.

[FLAG — CORRECTED MISATTRIBUTION: the prior build wrote "on an assessor discordance the JCA says it cannot explain" in a sentence about tovorafenib. The JCA attributes that unexplained discordance to Bouffet 2023. This reverses the read-through and must not be carried forward in its previous form.]

Assessor caveats that must travel with every one of these numbers (all verified verbatim or by direct reading of the JCA Report): - Unanchored MAIC = non-randomised evidence; the JCA states the reported effect estimates "may not represent true causal effects of treatment and may not be sufficiently applicable to the target population of PICO 5." - All effect estimates "target the population-average treatment effect in the population of the Bouffet 2023 study" — they are not estimates for the FIREFLY-1 or the EU label population. - Effective sample sizes: 5.81 (PFS base case), 6.64 (ORR base case), 10, 10.66, 11.54, 14.26. The JCA states the ESS was reduced by over 50% for the RANO-LGG evaluable set (12 → 5.81), "indicating somewhat poor covariate overlap, resulting in statistical imprecision… and potentially increased risk of bias", and reductions "up to 52%, depending on the outcome." - Bouffet 2023 used modified RANO-LGG (IRC) criteria that excluded the MR category, classifying such patients as SD — the asymmetry that caused the JCA to reject the sponsor's RANO-LGG comparison outright. - Unadjusted residual confounders acknowledged: histological type, primary tumour location, prior chemotherapy (the last because it applied to 100% of FIREFLY-1 patients, leaving nothing to adjust on). - Model-specific adjustment sets differ and must not be pooled. The ORR MAIC base case adjusted for age, prior surgery and Karnofsky/Lansky score only; gender was dropped for non-convergence and prior radiotherapy dropped without a stated reason. The PFS MAIC base case did adjust for age, prior surgery, prior radiotherapy, Karnofsky/Lansky score and gender. [FLAG — the prior build listed the gender/radiotherapy exclusions as a blanket caveat on all estimates. They apply to the ORR model only.] - All p-values are nominal, not pre-specified, not multiplicity-controlled; no preferred "primary" analysis was specified in the ITC SAP despite four available operationalisations (RANO-HGG/RANO-LGG × INV/IRC). - The sponsor's unmeasured-confounding sensitivity analysis "included a number of errors and the reported conclusions were inaccurate" and was excluded from the report. - Proportional hazards: the sponsor reported PH as "visually violated based on the Schoenfeld Residuals and Log Cumulative Hazard plots and statistically based on the Grambsch-Therneau test (i.e., p-value < 0.05)" and therefore preferred the RMST comparison; the JCA states "no further details of the PH assessment were provided; therefore, the assessors cannot assess the appropriateness of this conclusion." [FLAG — the prior build said PH was reported violated "without supporting evidence." The sponsor did name specific diagnostics; the JCA's criticism is that no further detail was supplied. Corrected.] - The RMST difference of −6.64 months is over a 28.43-month horizon, a timepoint the JCA notes was chosen without explanation in the dossier. It is not a lifetime or trial-duration difference. - Selective-reporting risk flagged: INV-assessed RANO-HGG data were available from FIREFLY-1 Arm 2 for n=9 patients meeting Population 2 criteria, "but the HTD declined to include these in the indirect comparison. Inadequate justification for the exclusion of these data was provided by the HTD."

[FLAG — INTERNAL INCONSISTENCY IN THE SOURCE: the base-case HR is reported as 4.88 (2.14, 11.14) with p=0.011. A Wald test consistent with that confidence interval implies p ≈ 0.0002, and the other three rows of the same table report p ≤ 0.001 with wider intervals (e.g. naïve HR 5.09 [2.03, 12.80], p<0.001). The JCA states CIs and p-values both come from the same Cox model with a robust sandwich variance estimator, so the two should agree. This anomaly is in the source document, not in transcription. Fore should quote the HR and its CI and omit "p=0.011" until the discrepancy is resolved.]

Coverage of the rest of the indication (verified in the JCA Report): - Population 1 (full claimed indication, PICOs 1–4): no comparative evidence submitted at all. The JCA's PICO table records "no comparator data available" against every listed comparator — vinblastine, carboplatin+vincristine, TPCV, dabrafenib+trametinib, everolimus, and bevacizumab+chemotherapy. The sponsor supplied single-arm FIREFLY-1 data only, which the JCA excluded because it "does not provide information on relative effectiveness or safety." - Population 2, PICO 6 (vs vinblastine or carboplatin+vincristine): no comparator data. - Population 3 (fusion/rearrangement/V600 non-E, PICOs 7–8): the sponsor's MAIC vs trametinib (FIREFLY-1 vs TRAM-01) was excluded from the JCA because "insufficient information [was] available for the assessment of TRAM-01" (its results were available only in abstract form).

So in the fusion population — 84% of the label indication — no comparative efficacy evidence was accepted by the EU JCA.

---

3.12 Front-line efficacy — does not yet exist

LOGGIC/FIREFLY-2 (NCT05566795) — the confirmatory trial on which both the US accelerated approval and the EU conditional marketing authorisation depend (van Tilburg CM et al., BMC Cancer 2024;24:147, PMID 38291372 — citation (verified)): - 418 patients enrolled (ACTUAL) (verified via CT.gov API v2), randomised 1:1, ~140 sites. Note the sponsor's own 8 May 2026 press release describes enrolment as "approximately 400"; the registry figure of 418 is the harder number. - Arms: tovorafenib 420 mg/m² QW (max 600 mg) vs investigator-selected SoC chemotherapy (COG carboplatin/vincristine, SIOPe-LGG carboplatin/vincristine, or vinblastine) [carried — from the BMC Cancer design paper]. - Eligibility: 6 months to <25 years, newly diagnosed LGG requiring first-line systemic therapy, activating RAF alteration; NF1-driven tumours excluded [carried]. - Primary endpoint: ORR by IRC using RANO-LGG criteria — note this is not the RAPNO-LGG criterion in the US label, and not the RANO-HGG primary of FIREFLY-1. Three different primary criteria across the programme. [carried — re-verify against the CT.gov outcome-measures module or the BMC Cancer paper] - Key secondaries: PFS (RANO-LGG), DoR, OS, safety [carried]. - Stratification: tumour location (supratentorial midline vs other), RAF alteration type (fusion vs mutation), CDKN2A status, infant chiasmatic-hypothalamic glioma [carried]. - Powering: ~85% to detect a 15-point ORR improvement; ~85% for PFS HR 0.67 [carried]. - Status 31 Aug 2026: ACTIVE_NOT_RECRUITING. Start 27 Feb 2023; primary completion June 2027; study completion June 2031; last update posted 14 May 2026. (verified)

Implication: there is no front-line efficacy data and none due before mid-2027 at the earliest. Both authorisations remain unconverted until then. [FLAG: "enrollment complete" (8 May 2026) is not "readout." The sponsor states the primary analysis is expected roughly 12 months after the last patient is enrolled. Public commentary conflates the two.]

---

3.13 Adult CNS evidence — thin, and unfavourable

There is no adult CNS registrational programme and no adult CNS indication anywhere. The entire identified public adult-CNS record is two small investigator series, neither of which is reported as stating a response criterion:

• 7-patient adult refractory-glioma series (5 HGG, 2 LGG): no objective responses; 3 with SD or better ≥4 months; median treatment duration 8 weeks; two Grade 4 intratumoral haemorrhages. Ioannou M, Mehta S, Devkota A, et al. "Clinical experience with tovorafenib in adults with treatment-refractory high- and low-grade gliomas." Neuro-Oncol Pract 2025;12(6):1051-1057, PMID 41458920 — citation (verified via PubMed: title, volume, issue, pagination); all efficacy figures and the absence of a stated response criterion remain [carried — full text not read in this refresh].
• PATH-94, 8 adults, Mayo Clinic (aged 21–66, median 42): "radiographic response in 5/8" = 3 stable + 2 with tumour-size decrease, both KIAA1549 fusion; 3/8 discontinued for toxicity. Williams A et al., Neuro-Oncology 2025;27(Suppl 5):v263 [carried — abstract not retrieved in this refresh].

[FLAG: neither series is reported as naming RANO, RAPNO or iRANO, and neither was re-read in this refresh. "Radiographic response in 5/8" appears to count stable disease as a response. These figures are not comparable to any criterion-defined ORR and must not be tabulated against one. Read both full texts before any comparative use.]

---

3.14 Adult systemic (non-CNS) — RECIST, quarantined

Do not place any figure in this subsection alongside a RANO/RAPNO CNS response rate.

• Phase 1 NCT01425008, 149 patients, DCO 11 Apr 2017 (Rasco DW, Medina T, Corrie P, et al., Cancer Chemother Pharmacol 2023;92(1):15-28, PMID 37219686 — citation (verified via PubMed)): ORR 0% (0/17) at the once-weekly schedule — the schedule that was ultimately approved. Responses occurred only on the Q2D schedule and only in BRAF V600 RAF/MEKi-naive melanoma: 8/16 (50%), RECIST; NRAS-mutant, treatment-naive 1/14 (7%), RECIST. Median PFS 1.9 months; DoR 6.0 months. Untreated brain metastases were excluded. Adult RP2Ds: 600 mg QW, 200 mg Q2D. [carried — efficacy figures not re-verified against the full text]
• FIRELIGHT-1 (NCT04985604) — TERMINATED for "Sponsor decision" after 23 patients (melanoma 8, tissue-agnostic 15); PCD 8 Jul 2024; last update 2 Oct 2025; results posted on the registry (status, reason, n and dates all verified via CT.gov API v2, 31 Aug 2026). Investigator-assessed confirmed ORR: melanoma 4/8 = 50.0% (15.7, 84.3); tissue-agnostic 6/15 = 40.0% (16.3, 67.7); median DoR 5.6 and 9.2 months; median PFS 5.5 months in both [carried — the posted results module was not re-read in this refresh]. [FLAG: the registry is reported to record the response criterion as "RECIST 1.1 or RANO" — mixed within the same denominator, so the criterion is not resolvable per patient. These figures cannot be attributed to either criterion and should not be used in a CNS comparison at all.]

---

3.15 Efficacy evidence gaps

Absences in the public record as of 31 Aug 2026. Absence claims cannot be proven exhaustively; each below means "not identified in the sources searched for this workstream."

Gap	Status
Overall survival	No OS estimate identified in any tovorafenib label, publication, EPAR, HTA dossier or registry record. The EU JCA explicitly records OS as "Relative effectiveness results not available" (verified)
Randomised efficacy data	None. Both authorisations rest on one single-arm cohort of 76
Front-line efficacy	None until FIREFLY-2 (PCD Jun 2027) (verified)
Adults	No adult indication anywhere; no adult efficacy endpoint identified in any active trial
Adult primary CNS tumours	Two uncontrolled series totalling 15 patients, no stated criterion, no objective responses in the larger one [carried]
Non-pLGG CNS histologies (HGG, PCNST, LCH, glioneuronal)	No registrational data
BRAF class II (non-V600 point mutants)	No data identified; both labels restricted to fusion/rearrangement or V600 (EU indication text verified)
BRAF class III (kinase-impaired)	No data identified
NF1-driven tumours	Excluded from both pivotal trials (FIREFLY-1 exclusion verified in USPI §14); USPI reported to warn of paradoxical MAPK activation in NF1-LOF cells [carried]
Co-altered tumours (IDH1/2, FGFR)	Excluded from FIREFLY-1 (verified in USPI §14)
HRQoL / symptom / functional endpoints	JCA: "Relative effectiveness results not available" for generic HRQoL, disease-specific HRQoL, disease symptoms and symptomatic disease control (verified)
Comparative efficacy in the fusion population (84% of the indication)	None accepted by the EU JCA; the sponsor's MAIC vs trametinib was excluded for insufficient information on TRAM-01 (verified)
Complete responses under LGG criteria	No CR reported in 76 patients at 40.6 months' median follow-up by RAPNO-LGG (inferred arithmetically from USPI Table 10 and wuag022)

---

3.16 Comparison-ready efficacy rows (for the head-to-head workstream)

Tovorafenib column populated from public, cited sources. The plixorafenib column is a template placeholder — no plixorafenib figure was researched, inferred or entered by this workstream. Where a cell gives a subgroup result, the format is responders/subgroup size, never responders/76.

Efficacy row	Tovorafenib (public, cited)	Plixorafenib
Pivotal population	R/R paediatric LGG, ≥1 prior systemic line, activating BRAF alteration, NF1 and IDH1/2- or FGFR-co-altered excluded	[FORE-INTERNAL — to be completed; verify]
Median age / range	8.5 y (2–21), label efficacy population	[FORE-INTERNAL — to be completed; verify]
Adults included	No	[FORE-INTERNAL — to be completed; verify]
Efficacy N	76	[FORE-INTERNAL — to be completed; verify]
Molecular mix	56/76 (74%) KIAA1549::BRAF, 8/76 (11%) other BRAF alteration, 12/76 (16%) V600E — label percentages sum to 101% through rounding	[FORE-INTERNAL — to be completed; verify]
Median prior lines	3 (1–9); 45/76 (59%) prior MAPKi	[FORE-INTERNAL — to be completed; verify]
Response criterion (labelled)	RAPNO-LGG (BICR) and RANO-LGG (independent review); RANO-HGG absent from the US label	[FORE-INTERNAL — state criterion + assessor]
ORR (headline)	53% (41, 64), 40/76, RAPNO-LGG, BICR — MR counted as response	[FORE-INTERNAL — to be completed; verify]
ORR (CR+PR only)	38.2%, 29/76, RAPNO-LGG (arithmetic on USPI Table 10)	[FORE-INTERNAL — to be completed; verify]
CR rate	0/76 reported by RAPNO-LGG at 40.6 mo median follow-up (inferred)	[FORE-INTERNAL — to be completed; verify]
Median DoR	18 mo (12.0, 22.8) label; 19.4 mo (13.8, 27.2) at DCO 6 Jun 2025	[FORE-INTERNAL — to be completed; verify]
Median TTR	5.4 mo (1.6–17.5)	[FORE-INTERNAL — to be completed; verify]
Median PFS	16.6 mo (8.3, 19.1), RAPNO-LGG, DCO 6 Jun 2025 (not in US label)	[FORE-INTERNAL — to be completed; verify]
Median OS	Not reported in any identified source	[FORE-INTERNAL — to be completed; verify]
ORR, BRAF fusion/rearrangement	53%, subgroup n=64 (label); 34/64 per EMA [carried], RAPNO-LGG	[FORE-INTERNAL — to be completed; verify]
ORR, BRAF V600E	50%, subgroup n=12 (label); 6/12 per EMA [carried], RAPNO-LGG	[FORE-INTERNAL — to be completed; verify]
ORR, BRAF class II / III	No data identified	[FORE-INTERNAL — to be completed; verify]
ORR after prior MAPKi	51%, 23/45 (subgroup n=45 = 59% of the 76); DoR 16.6 mo, deep response 33.3% [carried]	[FORE-INTERNAL — to be completed; verify]
ORR after progression on MEKi	33% of n=15 (RANO-HGG) / 33% of n=13 (RAPNO) / 30% of n=12 (RANO-LGG) — responder counts not published; see §3.9 flag	[FORE-INTERNAL — to be completed; verify]
Adult CNS efficacy	None (2 uncontrolled series, 15 pts, no stated criterion)	[FORE-INTERNAL — to be completed; verify]
Randomised evidence	None; FIREFLY-2 PCD Jun 2027	[FORE-INTERNAL — to be completed; verify]
Best comparative estimate vs dab+tram (V600E)	PFS HR 4.88 (2.14, 11.14) against tovorafenib, base case; consistent across all four reported analyses (HR 4.88–5.44). Unanchored MAIC, ESS 5.81, estimand = the Bouffet 2023 population (EU JCA Report Table 27)	[FORE-INTERNAL — to be completed; verify]

[FLAG — CORRECTED DENOMINATOR NOTATION: the prior build rendered the prior-MAPKi row as "51%, 45/76" and the post-MEKi row as "33%, 15 / 13 / 12". In both cases the second number is the subgroup size, not the responder count, and the "45/76" format reads as 45 responders out of 76. Corrected to responders/subgroup-size throughout.]

---

Open flags carried from this workstream

1. [FLAG] The 8/2025 USPI states no data-cutoff date (confirmed by reading the full SPL). The 10 May 2024 cutoff is inferred; the JCA labels its FIREFLY-1 inputs "2 year", not a calendar date. Confirm against the FDA supplement approval package.
2. [FLAG] EMA/67438/2026 table numbers (17–20, 22–24), the verbatim CHMP quotations about RANO-HGG suitability and IRC/investigator discordance, and the Day One SNO 2025 slide numbering and quotations are carried forward from the 24 Aug 2026 build and were not re-verified. The EPAR reference number itself was verified on the EMA product page.
3. [FLAG] The Nature Medicine RANO-HGG responder count (46/69, 67%) and the EMA figure (49/69, 71.0%) come from different data cuts — never combine the percentage from one with the denominator from the other.
4. [FLAG] TTNT of 42.6 months (95% CI 36.6–NE) exceeds the 40.6-month median follow-up; the estimate is immature and rests on censored data.
5. [FLAG] FIRELIGHT-1 registry results carry a mixed "RECIST 1.1 or RANO" criterion within a single denominator; the criterion is unresolvable per patient. Exclude from any CNS comparison.
6. [FLAG] The two adult-CNS series were not read in full; both are reported as stating no response criterion, and "radiographic response in 5/8" appears to count stable disease as a response.
7. [FLAG] The EU JCA comparative estimates rest on effective sample sizes of 5.81–14.26 with acknowledged unadjusted confounders (histological type, primary tumour location, prior chemotherapy), model-specific adjustment sets that differ between the ORR and PFS analyses, a probable proportional-hazards violation, an estimand defined on the Bouffet 2023 population, and a 28.43-month RMST horizon. They are directionally important but must always be quoted with those caveats. The strongest defensible framing is the consistency of the hazard direction across all four reported analyses, not the base-case HR alone.
8. [FLAG] The JCA's base-case PFS p-value (0.011) is inconsistent with its own confidence interval and with the other three rows of the same table. Quote HR 4.88 (2.14, 11.14) without the p-value until this is resolved.
9. [FLAG] The JCA's unexplained INV-vs-IRC assessor discordance is in Bouffet 2023, the comparator — not in FIREFLY-1. Do not use the ORR direction-flip as evidence about tovorafenib's assessor variance.
10. [FLAG] Ipsen's EU launch headline "regardless of BRAF alteration" (22 Apr 2026) is promotional framing. The authorised EU indication, verified on the EMA product page, is confined to "paediatric low-grade glioma (LGG) harbouring a BRAF fusion or rearrangement, or BRAF V600 mutation, who have progressed after one or more prior systemic therapies" (conditional marketing authorisation, 20 April 2026). Do not use the press headline as an indication statement.
11. [FLAG] The G-BA benefit assessment published 17 Aug 2026 (procedure 1342) was not retrievable in full in this refresh; its efficacy content and any new sponsor analyses are unreviewed.
12. [FLAG] The FIREFLY-1 ClinicalTrials.gov record was last updated 10 Apr 2025 and its status may be stale relative to the 31 Aug 2026 data cut.
13. [FLAG] The optic-pathway subset uses denominators (42, 39, 35 patients, 52 eyes) that are not reconciled with the 39 optic-pathway-primary patients in the baseline table. Resolve the definitions before quoting "51% of the registration population."

CITATIONS (verification status marked): 1. [VERIFIED — full text read 31 Aug 2026] OJEMDA (tovorafenib) US Prescribing Information, revision 8/2025 — §1 Indications, §2.3 Dosage, §14 Clinical Studies Table 10. DailyMed SPL setid ea3a9631-3a66-6a7c-e053-2995a90ae2ad, labeler Day One Biopharmaceuticals. [FLAG — the prior build cited "Document Id 8bee41f8"; that identifier does not correspond to the OJEMDA SPL. Corrected.] 2. [VERIFIED — results table read; PubMed metadata confirmed] Kilburn LB, Khuong-Quang D-A, Hansford JR, et al. Nat Med. 2024;30(1):207-217. PMID 37978284; PMC10803270. 3. [VERIFIED — key figures read; PubMed/OUP metadata confirmed] Kline C, Landi D, Khuong-Quang D-A, Nysom K, Kilburn LB, et al. Neuro-Oncology Pediatrics. 2026;2(2):wuag022 (published 22 Apr 2026; data cutoff 6 Jun 2025). doi:10.1093/neuped/wuag022 4. [NOT VERIFIED] Kline C, et al. CTP-17. Neuro-Oncology. 2025;27(Suppl 5):v149. doi:10.1093/neuonc/noaf201.0589 (SNO 2025). Slide deck at dayonebio.com — not opened in this refresh; all slide-derived figures and quotations remain carried. 5. [VERIFIED — full PDF downloaded and read 31 Aug 2026] Member State Coordination Group on Health Technology Assessment. Joint Clinical Assessment Report of Tovorafenib, Version 1.0. European Union, Brussels, 2026. Endorsed 30 Apr 2026; published 9 Jun 2026. Tables 25, 26, 27; §4.2.1.3, §4.3.1, §4.3.2, §4.3.3. https://health.ec.europa.eu/publications/joint-clinical-assessment-report-tovorafenib-ojemda_en 6. [NOT VERIFIED — comparator source] Bouffet E, Geoerger B, Moertel C, et al. J Clin Oncol. 2023;41(3):664-674. doi:10.1200/JCO.22.01000 7. [CITATION VERIFIED via PubMed; figures carried] Nysom K, et al. Neuro Oncol. 2025;27(5):1341-1355. PMID 39700439; PMC12187376. 8. [CITATION VERIFIED via PubMed; figures carried] Singh S, et al. FDA Approval Summary. Clin Cancer Res. 2025;31(8):1383-1389. PMID 39808502. 9. [CITATION VERIFIED via PubMed; figures carried] Rasco DW, Medina T, Corrie P, et al. Cancer Chemother Pharmacol. 2023;92(1):15-28. PMID 37219686; PMC10261210. 10. [CITATION VERIFIED via PubMed; design details carried] van Tilburg CM, et al. BMC Cancer. 2024;24:147. PMID 38291372; PMC10826080. 11. [VERIFIED — CT.gov API v2, 31 Aug 2026] NCT05566795 (LOGGIC/FIREFLY-2) — ACTIVE_NOT_RECRUITING, 418 ACTUAL, start 27 Feb 2023, PCD Jun 2027, completion Jun 2031, last update posted 14 May 2026. 12. [VERIFIED — CT.gov API v2, 31 Aug 2026] NCT04775485 (FIREFLY-1/PNOC026) — RECRUITING, 141 ESTIMATED (whole study), start 22 Apr 2021, PCD 31 May 2027, last update posted 10 Apr 2025. 13. [VERIFIED — CT.gov API v2, 31 Aug 2026] NCT04985604 (FIRELIGHT-1) — TERMINATED ("Sponsor decision"), 23 ACTUAL, PCD 8 Jul 2024, last update posted 2 Oct 2025, results posted. 14. [VERIFIED — CT.gov API v2, 31 Aug 2026] NCT07441707 — Ipsen phase 1 tovorafenib in Japanese children/young adults with brain tumours; RECRUITING, 6 ESTIMATED, start 3 Mar 2026, PCD 31 Jul 2030, last update posted 30 Jul 2026. 15. [VERIFIED — EMA product page read 31 Aug 2026] European Medicines Agency, Ojemda (tovorafenib), EMEA/H/C/006140 — authorised indication text; conditional marketing authorisation granted 20 April 2026; EPAR reference EMA/67438/2026. 16. [NOT VERIFIED] European Public Assessment Report EMA/67438/2026, Ojemda — Tables 17, 18, 19, 20, 22, 23, 24. The report reference number is verified; the individual tables and quotations are carried forward from the 24 Aug 2026 build and were not re-verified against the source PDF. 17. [NOT VERIFIED] Ipsen press release, 22 Apr 2026: "Ojemda approved in the European Union as the first targeted therapy in relapsed or refractory pediatric low-grade glioma regardless of BRAF alteration." Headline located in search results; not opened. 18. [NOT VERIFIED — located in search results] Servier / Day One press release, 8 May 2026: FIREFLY-2 enrolment complete; describes enrolment as "approximately 400" against the registry's 418 ACTUAL. 19. [NOT VERIFIED] G-BA Nutzenbewertungsverfahren 1342, tovorafenib (Ojemda) — assessment published 17 Aug 2026; PDF not retrievable in this refresh. 20. [CITATION VERIFIED via PubMed — title, volume, issue, pagination; figures carried] Ioannou M, Mehta S, Devkota A, et al. "Clinical experience with tovorafenib in adults with treatment-refractory high- and low-grade gliomas." Neuro-Oncol Pract. 2025;12(6):1051-1057. PMID 41458920. 21. [NOT VERIFIED] Williams A, et al. PATH-94. Neuro-Oncology. 2025;27(Suppl 5):v263.

Safety & Tolerability — Tovorafenib (OJEMDA™)

Day One Biopharmaceuticals, a Servier company (acquisition completed 23 April 2026); Ipsen holds ex-US commercial rights

[FLAG: workstream scope — CRITICAL, unresolved] The orchestrating workflow was invoked without a drug name — the prompt delivered to this agent reads literally "the competitor drug." This workstream was executed against tovorafenib (OJEMDA), inferred from (a) the plixorafenib BRAF-altered CNS competitive set, and (b) pre-existing tovorafenib dossiers in the Fore repo (C:\Users\Owner\Documents\Hugh Context Folder\Fore\competitive_intelligence\braf_deep_dive\Tovorafenib_Intelligence_Dossier.html). Every correction in this document is conditional on that inference being correct. If the intended competitor was dabrafenib+trametinib, vemurafenib, encorafenib or another agent, this workstream is void and must be re-run with the drug named (Workflow({name:'competitor-dossier', args:'<drug>'})). A human must close this flag before any content below enters the dossier.

Data cut for this workstream: 31 August 2026. Sources current as of: US Prescribing Information rev. 08/2025; EU conditional marketing authorisation 20 April 2026; openFDA/FAERS last_updated 2026-07-30; latest clinical presentation SNO, 19–23 November 2025 (3-year FIREFLY-1 follow-up, data cutoff 6 June 2025).

---

0. Verification pass — 31 August 2026

This document has been through an adversarial fact-check. What was independently re-verified, what was resolved, and what remains open:

Re-verified against primary sources and matching exactly: the entire US label figure set (§5.1–5.6, §6.1, Table 6, Table 7, absence of boxed warning, "Contraindications: None," the two 08/2025 Recent Major Changes entries, absence of a §6.2 Postmarketing Experience section) via DailyMed set ID ea3a9631-3a66-6a7c-e053-2995a90ae2ad; the full FAERS query (292 total / 94 serious / 198 non-serious / 12 death-outcome, all top-20 terms, all 14 organ-specific term counts) via a live openFDA re-query; the Nature Medicine FIREFLY-1 safety figures via the primary abstract; the Kline growth abstract in full; the FDA Approval Summary; the Rasco phase 1 citation and Grade ≥3 rates; the EMA EPAR authorisation date, monitoring status, obligations and side-effect list; the Servier acquisition; the FIREFLY-2 enrolment completion.

Resolved (flags closed): the Nature Medicine Author Correction is a Figure 2 colour-key swap, not a data correction (§11); the Kline growth-abstract citation is confirmed in full (§6.1); the adult phase 1 Grade ≥3 conflict is resolved in favour of the published 47% (9/19) figure (§9); the EU date is 20 April 2026 (§7); the LiverTox denominator crossing is confirmed against source (§6.5).

Newly identified and corrected: stale sponsor identity (Servier); missing denominator on the SNO 77% treatment-free figure (30/39); three omitted Grade ≥3 AE terms from the SNO release; the 420 mg/m² pivotal dose vs 380 mg/m² label dose gap, absent entirely from the prior draft; the RANO-HGG / RAPNO response-criteria conflation in the efficacy annotation.

Still open — must not be asserted publicly: drug-identity scope; the cutaneous SCC/keratoacanthoma phase 1 observation; the age-distribution figures; FIRELIGHT-1 adult safety; the two negative findings (§8, §11); per-analyte laboratory denominators.

Sourcing note: the accessdata.fda.gov label PDF (217700s002s003s004lbl.pdf) returned HTTP 404 on re-fetch during this pass although the URL remains indexed. DailyMed is therefore cited as the primary label source below, with accessdata as secondary. No label figure changed as a result.

---

1. Denominator discipline — read this before quoting any number below

Tovorafenib safety data circulate in three different denominators that are routinely conflated in press and promotional material. They are not interchangeable:

Population	N	What it is	Where the label uses it
Pooled safety population	172	140 patients (pLGG or advanced RAF-altered solid tumors) on BSA-based weekly dosing + 32 adults on flat 600 mg weekly	All Warnings & Precautions (§5.1–5.3) percentages
FIREFLY-1 safety population	137	Arm 1 (n=77) + Arm 2 (n=60), relapsed/refractory pLGG, BSA dosing (all-causality AEs)	Adverse Reactions §6.1, Tables 6 and 7
FIREFLY-1 growth-effects population	133	Patients ≤18 years of age in FIREFLY-1	§5.4 Effect on Growth
(context only) Efficacy population	76	FDA efficacy-evaluable, RAPNO by BICR	§14 / FDA Approval Summary

Consequence: the label states rash 67% (§5.2, N=172) and rash 77% (Table 6, N=137); hemorrhage 37% (§5.1, N=172) and 42% (Table 6, N=137). Both are correct in their own denominator. Any head-to-head table must state which one it is using. Exposure in the pooled population: 86% exposed ≥6 months, 49% exposed ≥1 year (N=172). Source: OJEMDA US Prescribing Information, rev. 08/2025, §5, §6.1 (verified via DailyMed set ID ea3a9631-3a66-6a7c-e053-2995a90ae2ad).

Response-criteria discipline — the same trap on the efficacy side. Three different criteria/denominators are in circulation and are routinely quoted as if interchangeable: - RANO-HGG, ORR 67%, arm 1 n=77 — the FIREFLY-1 primary endpoint (Nature Medicine). - RAPNO-LGG, ORR 51% (95% CI 40–63), N=76 — a secondary endpoint that includes minor responses; this is the FDA approval basis (accelerated approval, 23 Apr 2024). - RAPNO-LGG, ORR 53% (40/76) — the same population and criteria at the later 6 Jun 2025 cut (SNO 2025).

The 67% and 51% figures are not a discrepancy and the 51% and 53% figures are not a discrepancy; they are different criteria and different data cuts respectively. Never quote a bare "67%" or average any of them. Sources: Kilburn LB et al. Nat Med. 2024;30(1):207–217; Singh S, et al. FDA Approval Summary. Clin Cancer Res. 2025;31(8):1383–1389.

Separately, the Nature Medicine FIREFLY-1 publication reports treatment-related AEs (TRAEs) in the same N=137, which is a different attribution from the label's all-causality Table 6. Do not merge the two views.

---

2. Headline tolerability metrics — FIREFLY-1, N=137 (all-causality, label)

Metric	Value	N
Serious adverse reactions	45%	137
Serious ARs in >2%: viral infection / pneumonia / sepsis	9% / 4% / 4%	137
Fatal adverse reaction	1 patient (1%) — tumor hemorrhage	137
Permanent discontinuation for an AR	7% (drivers in >1 patient: tumor hemorrhage; reduction in growth velocity)	137
Dosage interruption for an AR	57% (≥5%: rash, pyrexia, vomiting, hemorrhage)	137
Dosage reduction for an AR	24% (≥2%: rash, fatigue)	137
Grade ≥3 treatment-related AEs (publication view)	42%	137
TRAEs leading to discontinuation (publication view)	9 patients (7%)	137
Most common TRAEs (publication view)	hair colour changes 76%, elevated CPK 56%, anaemia 49%	137

Sources: OJEMDA USPI rev. 08/2025 §6.1 (verified via DailyMed); Kilburn LB, Khuong-Quang D-A, Hansford JR, et al. The type II RAF inhibitor tovorafenib in relapsed/refractory pediatric low-grade glioma: the phase 2 FIREFLY-1 trial. Nat Med. 2024;30(1):207–217. doi:10.1038/s41591-023-02668-y. PMID 37978284. PMC10803270. NCT04775485. All publication-view figures in this table were verified directly against the Nature Medicine abstract in this pass, superseding the prior draft's reliance on secondary reporting.

CI read: the low permanent-discontinuation rate (7%) sits alongside a very high dose-interruption rate (57%) and a 24% dose-reduction rate. Tovorafenib is not "discontinued" often, but roughly one in four patients does not stay on the label dose, and the majority experience at least one hold. This gap between discontinuation and interruption/reduction is the single most useful tolerability metric for a head-to-head and is frequently omitted from promotional summaries.

---

3. Full adverse-reaction profile — Label Table 6 (≥20%), FIREFLY-1 Arms 1+2, N=137

System organ class / Adverse reaction	All grades (%)	Grade 3 or 4 (%)
Skin and subcutaneous tissue
Rash ᵃ	77	12
Hair color changes	76	0
Dry skin	36	0
Dermatitis acneiform	31	1
Pruritus	26	1
General disorders
Fatigue	55	4
Pyrexia	39	4
Edema ᵇ	26	0
Infections and infestations
Viral infection ᶜ	55	7
Upper respiratory tract infection	31	1.5
Paronychia	26	1.5
Gastrointestinal
Vomiting ᵈ	50	4
Constipation	33	0
Nausea	33	0
Abdominal pain	28	0
Diarrhea ᵉ	22	1.5
Stomatitis ᶠ	20	0
Nervous system
Headache	45	1
Vascular
Hemorrhage ᵍ	42	5 *

* includes one Grade 5 event. [FLAG: verify] The Grade-5 attribution within the Table 6 hemorrhage row is consistent with §5.1 and §6.1 but the table footnote itself was not transcribed verbatim in this pass; confirm before reproducing the asterisk.

ᵃ grouped term: erythema multiforme, eczema, rash erythematous/macular/follicular/pruritic/maculopapular/papular/pustular, rash, skin exfoliation, drug eruption, dermatitis, dermatitis bullous. ᵇ lip, periorbital, peripheral, localized, face, vulval edema. ᶜ 28 grouped viral terms incl. COVID-19, influenza, RSV, herpes simplex. ᵈ incl. retching, hematemesis. ᵉ incl. colitis, enterocolitis. ᶠ incl. mouth ulceration, mucosal inflammation, aphthous ulcer, cheilitis. ᵍ incl. tumor hemorrhage, intracranial tumor hemorrhage, subdural hemorrhage, GI hemorrhage, epistaxis, hemoptysis, gingival bleeding, vaginal hemorrhage.

Other clinically important ARs at <20%: reductions in growth velocity, skin discoloration, myalgia, photosensitivity reaction, arthralgia. [FLAG: verify] This <20% list was not confirmed against the label text in this verification pass; the individual terms are corroborated elsewhere (photosensitivity §5.2, growth §5.4) but the completeness of the list is unverified.

Note on grouped terms: the "Rash" 77% and "Hemorrhage" 42% figures are broad composite MedDRA groupings (14 and 12 preferred terms respectively). A competitor comparing against a drug whose label reports narrow preferred terms will overstate tovorafenib's relative toxicity unless the grouping is matched. The EU label makes this concrete: it lists epistaxis as its own >1-in-10 adverse reaction, where the US label absorbs epistaxis into the composite "Hemorrhage" term. Source: OJEMDA USPI rev. 08/2025, Table 6 (verified via DailyMed); EMA Ojemda EPAR.

---

4. Laboratory abnormalities — Label Table 7 (≥20% worsening from baseline), FIREFLY-1, N variable 67–137

Laboratory abnormality	All grades (%)	Grade 3 or 4 (%)
Hematology
Decreased hemoglobin	90	15
Decreased lymphocytes	50	2
Decreased leukocytes	31	2
Increased lymphocytes	23	0
Chemistry
Decreased phosphate	87	25
Increased AST	83	2
Increased creatine phosphokinase	83	11
Increased LDH	73	0
Decreased potassium	51	2
Increased ALT	50	5
Decreased albumin	24	5
Increased bilirubin	22	1
Decreased sodium	20	2

Denominator footnote (label): the denominator for each laboratory parameter is the number of patients with a baseline and post-treatment value available, which ranged from 67 to 137 patients. Severity per NCI CTCAE v5.0. Source: OJEMDA USPI rev. 08/2025, Table 7 (all values re-verified via DailyMed in this pass).

[FLAG: verify] The 67–137 denominator range is disclosed but not attributed per analyte. The Grade 3–4 hypophosphatemia rate (25%) and Grade 3–4 CPK elevation (11%) could each be computed on as few as 67 patients. Any head-to-head lab-toxicity comparison should be footnoted accordingly, and a Fore medical-affairs reviewer should confirm per-analyte N from the FDA multidisciplinary review (NDA 217700 / 218033) before external use.

CI read on the lab profile: the three defining laboratory toxicities — hypophosphatemia (87% any grade, 25% Gr 3-4), anemia (90% any grade, 15% Gr 3-4), and CPK elevation (83% any grade, 11% Gr 3-4) — have Grade 3–4 rates materially higher than any clinical AE except rash. This is an under-discussed part of tovorafenib's burden: the drug's most frequent severe toxicities are laboratory, monitoring-driven, and largely asymptomatic, which mutes them in patient-reported tolerability narratives but drives clinic visits and lab draws.

---

5. Boxed warning and Warnings & Precautions

Boxed warning: NONE. OJEMDA carries no boxed warning. Contraindications: None (§4). Both re-verified via DailyMed in this pass.

Recent Major Changes (label Highlights): Dosage and Administration (2.1) 08/2025; Warnings and Precautions (5.4 — Effect on Growth) 08/2025. The growth warning is the only W&P section revised since approval — see §6.1 below.

§	Warning	Key rates (pooled N=172 unless noted)
5.1	Hemorrhage (incl. major hemorrhage)	Hemorrhagic events 37%; epistaxis 26%; intratumoral hemorrhage 9%; serious bleeding 5%; Grade 5 tumor hemorrhage 1 patient (0.6%); permanent d/c for hemorrhage 2%
5.2	Skin toxicity incl. photosensitivity	Rash 67% (Gr 3 12%); rash → interruption 15%, reduction 7%, d/c 1% (n=2). Dermatitis acneiform 26% (Gr 3 0.6%, n=1); reduction 2% (n=3). Photosensitivity 12% (Gr 3 0.6%, n=1)
5.3	Hepatotoxicity	ALT↑ 42% (Gr 3 4%); AST↑ 74% (Gr 3 2%); bilirubin↑ 23% (Gr 3 0.6%, n=1). Median time to onset 14 days (range 3–280). ALT/AST → interruption 5%, reduction 1.2%. Hyperbilirubinemia → discontinuation in 1 adult with advanced non-CNS solid tumor
5.4	Effect on growth ⟵ revised 08/2025	See §6.1 (N=133 ≤18 y)
5.5	Embryo-fetal toxicity	Embryo-lethal in rats at ~0.8× human AUC at the recommended dosage. Non-hormonal contraception required — tovorafenib can render hormonal contraceptives ineffective (CYP3A induction). Females: during treatment + 28 days; males with female partners: during + 2 weeks
5.6	NF1-associated tumors	Nonclinical paradoxical MAPK activation; may promote tumor growth in NF1 tumors — see §6.4

All §5.1–5.3 and §5.5 rates re-verified via DailyMed in this pass. The interruption/reduction sub-rates in §5.2 and §5.3 were not individually re-transcribed and carry the prior draft's sourcing.

Monitoring burden imposed by the label: LFTs (ALT/AST/bilirubin) before initiation, at 1 month, then every 3 months and as clinically indicated; routine growth monitoring; dermatologic consult as indicated; UV precautions (sunscreen, sunglasses, protective clothing). BRAF alteration must be confirmed before initiation (§2.1).

---

6. Distinctive / organ-specific toxicities

6.1 Growth-velocity suppression — the signature pediatric toxicity

The most differentiating and most commercially consequential toxicity, and the only W&P revised post-approval.

Metric	Value	N
Treatment-emergent adverse effects on growth	46%	133 (≤18 y)
…of which Grade ≥3	35%	133
Growth reduction → dose interruption / reduction / permanent discontinuation	5% / 2.3% / 3%	133
Median change from baseline in height percentile at 12 months	−14 (z-score −0.6)	107 evaluable
Median change from baseline in height percentile at 18 months	−20 (z-score −0.9)	95 evaluable
Median annualized growth velocity on treatment	0.86–1.8 cm/year	81 evaluable
Median annualized growth velocity ≥90 days off treatment	4.2 cm/year	17
Bone-age radiographs: premature epiphyseal closure or bone-age advancement	None observed	35 assessed

Source: OJEMDA USPI rev. 08/2025 §5.4 and §8.4 (all values re-verified via DailyMed in this pass).

Independent trial-level growth data (abstract, citation fully verified in this pass): 29% experienced decreased growth velocity from baseline; 19% had a ≥50% decrease; 40 patients with decreased growth velocity were reported as AEs of special interest [FLAG: verify — the PMC record reads "40 patients"; the prior draft rendered this as "40 events." Confirm patients vs events before quoting]. All 10 patients who discontinued or interrupted for ≥3 months showed post-treatment recovery over a mean 5.8 months of follow-up (mean annualized GV 1.1 cm/y on-treatment → 8 cm/y off-treatment); one 4-year-old went from 1.2 cm/y on-treatment to 12.3 cm/y off-treatment at 2 months' follow-up. 19 had on-treatment bone-age assessments, none advanced or with premature closure; 75% of decreased-GV cases had pre-existing neuromuscular or endocrine comorbidities. Source: Kline C, et al. LGG-40: Type II RAF inhibitor tovorafenib in relapsed/refractory pediatric low-grade glioma (pLGG): reversible decreases in growth velocity in the phase 2 FIREFLY-1 trial. Neuro-Oncology. 2024 Jun 18;26(Suppl 4). doi:10.1093/neuonc/noae064.431. PMC11183845. Safety analysis cutoff 8 Aug 2023 (n=137); growth follow-up through 19 Jan 2024. Title, journal, supplement, DOI, lead author (Cassie Kline, CHOP) and every figure above were verified directly against PMC in this pass — the prior draft's author/title flag is closed. Note that the same dataset was also presented at ASCO 2024 (JCO 2024;42(16_suppl):10036); cite one, not both as if independent.

[FLAG: verify — reconcile, do not merge] The label's off-treatment recovery figure (median 4.2 cm/y, N=17, ≥90 days off treatment) and the abstract's (mean 8 cm/y, N=10, mean 5.8 months follow-up) differ in statistic (median vs mean), denominator, off-treatment interval and data cut — which largely explains the gap. Both are cited above as reported. Do not average, combine, or present them as a single "recovery" number. A Fore medical reviewer should reconcile against the FDA review documents before external use.

CI read: growth suppression is frequent (46%), frequently severe (35% Grade ≥3), and dose-limiting in a small but real fraction (3% permanent discontinuation). It appears reversible in the small subsets studied (n=10 and n=17), with no evidence of epiphyseal damage among the 19–35 patients bone-age assessed — a directional finding on small numbers, not an established property, and one an opposing reviewer will press. This is the core clinical trade-off in the pediatric setting and the direct driver of the sponsor's "treatment-free observation period" positioning (see §7). It is also a pediatric-specific toxicity with limited read-across to an adult/adolescent CNS population — a comparator drug studied predominantly in adults will not have an analogous signal, and its absence in that comparator is not evidence of superiority.

6.2 Intratumoral / intracranial hemorrhage — the CNS-critical toxicity

Intratumoral hemorrhage 9% (N=172, §5.1); the single fatal (Grade 5) event on study was tumor hemorrhage (1/137, 1%); tumor hemorrhage was one of only two reasons for permanent discontinuation in more than one patient. In FAERS, intracranial tumour haemorrhage is the top-ranked serious organ-specific reaction term (10 reports), plus tumour haemorrhage (3), cerebral haemorrhage (2), haemorrhage (2) — see §8. All FAERS counts in this paragraph re-queried and confirmed exactly in this pass.

CI read: for any BRAF-altered CNS program, intratumoral hemorrhage is the safety comparison that matters most, because it is the toxicity with a fatal precedent, a labelled discontinuation driver, and the strongest post-marketing signal. It should be the first row of any CNS head-to-head safety table.

6.3 Cutaneous toxicity and the absence of classic paradoxical-activation skin lesions

Tovorafenib produces heavy but largely low-grade cutaneous toxicity: rash 77% (Gr 3–4 12%, N=137), hair color changes 76%, dry skin 36%, dermatitis acneiform 31%, pruritus 26%, plus photosensitivity 12% (N=172) and skin discoloration (<20%). Rash is the leading cause of dose interruption and dose reduction.

Notably absent from the label: cutaneous squamous cell carcinoma, keratoacanthoma, hyperkeratosis, palmar-plantar erythrodysesthesia, and uveitis — the hallmark paradoxical-activation and class toxicities of type I (V600-selective) RAF inhibitors such as vemurafenib and dabrafenib.

[FLAG: unverified — do not use externally] The prior draft further asserted that the adult phase 1 reported cutaneous SCC in only 1 patient (<1%) with no treatment-related keratoacanthomas, attributed by the authors to type II RAF inhibition not triggering paradoxical activation. This observation is not present in the retrieved Rasco 2023 abstract and could not be verified in this pass. It must be confirmed against the Cancer Chemother Pharmacol full text before it is used. The label-omission half of the argument is verified and can stand on its own; the phase 1 half cannot yet. Sources: OJEMDA USPI rev. 08/2025 Table 6 and §6.1 (verified); Rasco DW, Medina T, Corrie P, et al. Phase 1 study of the pan-RAF inhibitor tovorafenib in patients with advanced solid tumors followed by dose expansion in patients with metastatic melanoma. Cancer Chemother Pharmacol. 2023;92(1):15–28. doi:10.1007/s00280-023-04544-5. PMID 37219686. PMC10261210. NCT01425008 (citation verified; the SCC/keratoacanthoma content is not in the abstract).

CI read: the "no paradox" cutaneous story is tovorafenib's strongest safety differentiation claim versus type I RAF inhibitors, and it is substantiated in the label by omission rather than by an affirmative comparative dataset. There is no randomized comparison of tovorafenib versus a type I RAF inhibitor for skin-cancer incidence; FIREFLY-2 (vs. chemotherapy) will not generate one either. With the phase 1 supporting observation currently unverified, the claim rests on label omission alone and should be framed accordingly.

6.4 NF1 paradoxical activation — a mechanistic liability, unique among the class

§5.6, added on nonclinical grounds: the label states tovorafenib may promote tumor growth in patients with NF1-associated tumors based on nonclinical data — in vitro, tovorafenib increased ERK phosphorylation at clinically relevant concentrations in NF1 loss-of-function cells. In an NF1 GEM model of plexiform neurofibroma without BRAF alteration, tovorafenib had no antitumor activity and a non-statistically-significant tumor-volume increase in 2/12 mice (~17%) [FLAG: verify — the 2/12 mouse detail derives from §13.2 and was not re-transcribed verbatim in this pass; the §5.6 tumor-promotion statement is confirmed]. The label requires confirmed BRAF alteration before initiation (§2.1). Source: OJEMDA USPI rev. 08/2025 §5.6 and §13.2 (§5.6 verified via DailyMed).

CI read: this is a commercially significant restriction, not merely a safety footnote. NF1-associated optic pathway glioma is a substantial slice of the pediatric LGG population, and tovorafenib is effectively fenced out of it — indeed, carries a theoretical tumor-promotion risk there. This is a defined, label-documented gap in tovorafenib's addressable population.

6.5 Hepatotoxicity

LiverTox likelihood score: E* — "unproven but suspected rare cause of clinically apparent liver injury." LiverTox reports, for the 172-patient safety cohort, ALT elevation 50% (>5× ULN in 5%), AST elevation 83% (>5× ULN in 2%); "no enzyme elevations with jaundice or symptoms and no life-threatening or fatal instances of liver injury"; no published cases of clinically apparent hepatic injury with jaundice since approval. Discontinuation for hepatotoxicity: 1 adult patient. Proposed mechanism: direct toxicity of BRAF/MAPK pathway inhibition. Source: LiverTox: Clinical and Research Information on Drug-Induced Liver Injury. Tovorafenib. NCBI Bookshelf NBK613920. Last updated 8 April 2025 (re-verified in this pass).

[FLAG: CONFIRMED discrepancy — cite the label, not LiverTox, for the rates] LiverTox attributes ALT 50% / AST 83% to the "172-participant safety cohort," while the label's §5.3 (also pooled N=172) states ALT 42% / AST 74%; the 50%/83% figures are exactly the label's laboratory Table 7 values (FIREFLY-1, N=67–137). LiverTox has crossed the clinical-AE and laboratory denominators. This was verified against the LiverTox text in this pass and is a confirmed error in a widely-cited NIH source — worth knowing, because a counterparty quoting LiverTox will be quoting the wrong denominator. LiverTox also describes the cohort as "at least half…treated for more than 1 year," which overstates the label's 49% exposed ≥1 year.

6.6 Reproductive, gonadal and endocrine

• Embryo-lethal in rats at ~0.8× human AUC (§5.5, verified).
• Rat fertility study: decreased pregnancies, corpora lutea and live embryos, and increased post-implantation losses at all doses, lowest ≈0.8× human AUC.
• Repeat-dose rat toxicology: female — reversible increased vaginal mucosal thickness, increased corpora hemorrhagicum and hemorrhage; non-reversible cystic follicles, decreased corpora lutea, interstitial cell hyperplasia in ovaries at ≥0.4× human AUC. Male — reduced epididymis and testis weight with reversible tubular degeneration/atrophy and reduced epididymal sperm at ≥0.3× human AUC.
• Contraceptive failure risk: tovorafenib induces CYP3A (plus CYP2C8, 1A2, 2B6, 2C9, 2C19) and can render hormonal contraceptives ineffective — the label mandates non-hormonal contraception.
• Not carcinogenic in 6-month rasH2 transgenic mice up to 100 mg/kg/day; not mutagenic (Ames); chromosomal aberrations in human lymphocytes with metabolic activation at a single concentration in vitro; not genotoxic in the in vivo rat bone-marrow micronucleus assay.

Source: OJEMDA USPI rev. 08/2025 §5.5, §12.3, §13.1. [FLAG: verify] Only §5.5 was re-verified in this pass. The §13.1 nonclinical detail — exposure multiples, reversibility, and the specific ovarian and testicular findings — carries the prior draft's sourcing and should be re-read against the label before external use, because the CI read below rests entirely on it.

CI read: the non-reversible ovarian findings at sub-clinical exposure multiples (0.4×) are the most under-reported item in tovorafenib's nonclinical package, and are materially relevant in an indication treating children and adolescents who will be dosed for years. FAERS carries 4 reports of amenorrhoea (count confirmed), consistent with — though far from establishing — this concern.

6.7 Cardiac

At the recommended 380 mg/m² weekly (max 600 mg), a mean QT increase >20 ms was not observed (§12.2). No cardiomyopathy, ejection-fraction, or hypertension warning appears in the label — a notable contrast with MEK-inhibitor-containing regimens. [FLAG: verify — §12.2 was not re-read verbatim in this pass; the 380 mg/m² recommended dosage is confirmed.]

6.8 [NEW] Pivotal dose vs label dose — a gap absent from the prior draft

FIREFLY-1 dosed tovorafenib at 420 mg/m² once weekly (600 mg maximum) per the Nature Medicine publication. The approved US label dosage is 380 mg/m² once weekly (600 mg maximum), and Dosage and Administration (2.1) is one of the two sections revised 08/2025.

[FLAG: verify — material to every rate in this document] Every safety rate in §2–§4 was generated at the trial's 420 mg/m² mg/m²-basis dosing, and is presented on a label whose recommended dose is 380 mg/m². The two are very likely exposure-matched across formulations (this is a routine outcome of a bridging analysis), but that equivalence was not verified in this pass. Confirm the rationale in the FDA multidisciplinary review before asserting that the label's rates apply at the label's dose. If the doses are not exposure-equivalent, this is a legitimate line of attack in either direction and a counterparty will find it. Sources: Kilburn LB et al. Nat Med. 2024;30(1):207–217 (420 mg/m² QW); OJEMDA USPI rev. 08/2025 §2.1 (380 mg/m² QW, max 600 mg) — both verified in this pass.

---

7. Long-term and confirmatory-trial safety status

• 3-year FIREFLY-1 follow-up (SNO 30th Annual Meeting, 19–23 Nov 2025; release 24 Nov 2025; data cutoff 6 June 2025):
• "No new safety signals were identified" — sponsor characterisation, not an independently verified finding.
• Grade ≥3 AEs in ≥5% of patients: decreased growth velocity, anemia, blood creatine phosphokinase increased, maculopapular rash, and alanine aminotransferase increased. (The prior draft listed only the first two; three terms restored.)
• 44 of 76 efficacy-evaluable patients (58%) completed ≥26 cycles (~24 months).
• 77% (30/39) of patients who entered the treatment-free observation period remained treatment-free for a minimum of 12 months. (The prior draft quoted 77% with no denominator. This is a 39-patient subgroup result, not a trial-level rate — do not present it as one.)
• Median time to next treatment: 42.6 months (95% CI 36.7–NE). (Replaces the prior draft's imprecise ">3.5 years.")
• ORR 53% (40/76) per RAPNO-LGG at this cut — see the response-criteria note in §1; this is the same population and criteria as the FDA's 51%, at a later cut. Source: Day One Biopharmaceuticals press release, 24 November 2025, "Day One Announces Three Year Follow-Up Data From OJEMDA (tovorafenib) Phase 2 FIREFLY-1 Trial at the 2025 SNO Annual Meeting" (all figures above verified against the release text in this pass). [FLAG: company source — partially downgraded] The data cutoff, denominators and Grade ≥3 list are now retrieved and verified. Still not obtained: the SNO abstract number, the presenter of record, and the peer-reviewed AE table. Treat "no new safety signals" as sponsor language until the abstract or manuscript is retrieved.

CI read: the treatment-free observation strategy is best understood as a tolerability mitigation converted into a positioning asset — it lets the sponsor address the cumulative growth-suppression and monitoring burden while framing intermittency as a clinical benefit. Note that the supporting figure rests on 39 patients.

• FIREFLY-2 / LOGGIC (confirmatory, randomized vs. standard-of-care chemotherapy, front-line, ages 6 months–25 years): enrolment completed, announced 8 May 2026 (Day One / Servier); ~400 participants across ~140 sites in the US, Canada, Europe, Australia, South America, Middle East and Asia, randomized to tovorafenib once weekly or one of four standard-of-care chemotherapy regimens; conducted with the SIOPe Brain Tumour Group LOGGIC Consortium. Topline expected mid-2027. This will be the first randomized safety dataset for tovorafenib and the first opportunity for a controlled comparison of growth, hemorrhage and cutaneous toxicity against an active comparator. Sources: Day One / Servier press release, 8 May 2026 (verified in this pass, including the ~400-participant and ~140-site figures, which were absent from the prior draft); van Tilburg CM, et al. LOGGIC/FIREFLY-2. PMC10826080. NCT05566795.

• EU conditional marketing authorisation, granted 20 April 2026: Ojemda is under EMA "additional monitoring" status (black inverted triangle), with specific obligations to submit final results of the ongoing tovorafenib-vs-chemotherapy trial and to undergo annual review of new safety and efficacy data. EU indication: monotherapy, patients ≥6 months with pLGG harbouring a BRAF fusion or rearrangement or BRAF V600 mutation, progressed after ≥1 prior systemic therapy. EU common side effects (>1 in 10) broadly mirror the US label and explicitly list growth retardation — and, separately, epistaxis, which the US label absorbs into the composite "Hemorrhage" term. Source: EMA, Ojemda EPAR medicine page (authorisation date, conditional status, additional-monitoring status, indication wording, side-effect list and specific obligations all verified directly in this pass). [FLAG: date resolved] The prior draft flagged a 20 vs 22 April 2026 conflict. The EMA EPAR returns 20 April 2026 as the authorisation date; Ipsen's "22 April" is the announcement, not the decision. Use 20 April 2026, EMA-attributed. Final confirmation against the EU Community Register decision remains with the regulatory workstream.

---

8. Post-marketing safety — FAERS / openFDA

Query: patient.drug.medicinalproduct:"OJEMDA" OR patient.drug.openfda.generic_name:"tovorafenib". openFDA meta.last_updated: 2026-07-30. Every figure in this section was re-queried live against the openFDA API during this verification pass.

Metric	Value	Verification
Total FAERS reports	292	confirmed
Reports coded serious	94 (32%)	confirmed
Reports coded non-serious	198	confirmed
Reports with death outcome (seriousnessdeath:1)	12	confirmed
Distinct reaction preferred terms	≥100	corrected — see flag

[FLAG: corrected] The prior draft stated "136 distinct reaction preferred terms." That figure is not reproducible: the openFDA count endpoint returns a maximum of 100 terms without a registered API key, and a re-query returned exactly 100 (lowest count returned: 2). The true total may well be 136, but as stated it is unverifiable. Report ≥100 unless the query is re-run with an API key and the key-authenticated result recorded.

Top reported reaction terms (count) — all confirmed: OFF LABEL USE 26 · RASH 26 · DISEASE PROGRESSION 24 · FATIGUE 21 · ACCIDENTAL UNDERDOSE 19 · HAIR COLOUR CHANGES 18 · HEADACHE 15 · MELANOCYTIC NAEVUS 15 · BLOOD CREATINE PHOSPHOKINASE INCREASED 11 · ACNE 10 · GROWTH RETARDATION 10 · INTRACRANIAL TUMOUR HAEMORRHAGE 10 · DERMATITIS ACNEIFORM 9 · NAUSEA 9 · ANAEMIA 7 · DEATH 7 · PHOTOSENSITIVITY REACTION 7 · WEIGHT DECREASED 7 · BLOOD PHOSPHORUS DECREASED 6 · ASTHENIA 5 · CONSTIPATION 5 · EPHELIDES 5 · ERYTHEMA 5 · INSOMNIA 5 · PRODUCT ADMINISTRATION ERROR 5 · VOMITING 5.

Organ-specific terms of interest — all confirmed: TUMOUR HAEMORRHAGE 3, CEREBRAL HAEMORRHAGE 2, HAEMORRHAGE 2, EPISTAXIS 4 · HEPATIC CYTOLYSIS 3, LIVER FUNCTION TEST INCREASED 2, BLOOD BILIRUBIN INCREASED 4 · RHABDOMYOLYSIS 2 · SEIZURE 2 · AMENORRHOEA 4 · EPHELIDES 5 · CELLULITIS 3, BLISTER 3 · SKIN DISCOLOURATION 2.

Signals worth watching:

1. Melanocytic naevus (15) + ephelides (5) + skin discolouration (2). New/eruptive melanocytic nevi and freckling are not named as such in the US label (which lists only "skin discoloration" at <20%). Fifteen reports of a single dermatologic term in a 292-report file is disproportionate and is the clearest candidate for a genuinely novel post-marketing cutaneous signal. [FLAG: unverified — absence of retrieved evidence, not evidence of absence] No published case series or regulatory signal evaluation was located at this data cut. Do not assert publicly that none exists.
2. Intracranial tumour haemorrhage (10) as the top serious organ-specific term — consistent with the trial-observed hemorrhage risk carrying into real-world use. Note that FAERS cannot establish this is "not an artefact of the trial population"; it can only show the term is reported post-marketing.
3. Rhabdomyolysis (2), in the context of 83% any-grade / 11% Grade 3–4 CPK elevation on label. Small numbers, but mechanistically coherent and not currently a labelled warning. Worth monitoring.
4. Growth retardation (10) — consistent with the label; the 08/2025 W&P revision indicates FDA already acted on accumulating growth data.
5. Off-label use (26) — the single most-reported term. Combined with accidental underdose (19) and product administration error (5), this points to real-world use pressure beyond the relapsed/refractory BRAF-altered pLGG indication and to BSA-based weekly dosing being error-prone in practice.

FAERS caveats (mandatory): spontaneous reports are voluntary, unverified, subject to reporting and stimulated-reporting bias, carry no exposure denominator, and establish no causal relationship. Counts here are report counts, not incidence. This analysis is a term-frequency listing, not a disproportionality analysis (no PRR/ROR/EBGM computed; the 292-report file is small for stable disproportionality statistics).

No FDA quarterly "Potential Signals of Serious Risks" entry for tovorafenib/OJEMDA was identified in the postings reviewed. [FLAG: negative finding — do not assert publicly] The FAERS quarterly postings migrated to the FDA Adverse Event Monitoring System (AEMS) on 11 March 2026; the search was not exhaustive across every quarterly posting from Q2-2024 through Q2-2026. This is absence of retrieved evidence and must be confirmed by a direct quarter-by-quarter review before being asserted.

There is no §6.2 Postmarketing Experience section in the 08/2025 US label — i.e., no post-marketing adverse reaction has yet been added to the label beyond the growth-warning revision. Verified via DailyMed in this pass.

---

9. Adult safety experience (relevant to any adult/adolescent CNS comparison)

Tovorafenib's pivotal safety database is overwhelmingly pediatric. Adult data are thinner and come from three sources:

• Label contribution: 32 adults with advanced solid tumors dosed at flat 600 mg weekly, folded into the pooled N=172 W&P population. The single hepatotoxicity discontinuation in the pooled population was an adult with an advanced non-CNS solid tumor (§5.3).

• Phase 1 (Rasco DW, Medina T, Corrie P, et al. Cancer Chemother Pharmacol. 2023;92(1):15–28; PMID 37219686; PMC10261210; NCT01425008): N=149 (Q2D n=110, QW n=39); escalation Q2D 20–280 mg, QW 400–1000 mg; MTD/RP2D 200 mg Q2D or 600 mg QW. DLTs: at 280 mg Q2D — Grade 3 periorbital edema, Grade 3 maculopapular rash; at 800 mg QW — Grade 3 hyperbilirubinemia, Grade 3 rash. Grade ≥3 adverse events in the dose-expansion phase (published, all-causality): 58 of 80 (73%) in the Q2D cohorts and 9 of 19 (47%) in the QW cohort. Most common Grade ≥3 events overall: anemia 14 patients (14%) and maculo-papular rash 8 patients (8%). [FLAG: prior-draft error corrected] The prior draft reported "~20% Grade ≥3 treatment-related AEs, ~21% discontinuation for AEs, ~11% dose reductions" for the QW expansion cohort, having flagged a 47% vs ~20% conflict and then reported the ~20% side. 47% (9/19) is the published figure; the ~20% figure has no traceable source and the discontinuation and dose-reduction percentages could not be sourced at all and have been deleted rather than carried forward. Note also that the published figure is all-causality Grade ≥3 AEs, not treatment-related. Retrieve the full-text tables before quoting any treatment-related breakdown, and see the §6.3 flag on the unverified SCC/keratoacanthoma observation.

• FIRELIGHT-1 (NCT04985604), adult phase 2, sub-study DAY101-102a, tovorafenib 600 mg QW in RAF-fusion / RAF1-amplified melanoma and other solid tumors: reported as "generally well tolerated"; common AEs anemia, CPK increase, pruritus, rash. Sub-study 102b (tovorafenib + pimasertib) enrolling. Source: Annals of Oncology 2024, ESMO abstract 614MO. [FLAG: paywalled — unresolved] The abstract full text returned HTTP 403 in both the original pass and this verification pass; N, Grade ≥3 rates and discontinuation rates for FIRELIGHT-1 were not obtainable and must be retrieved before any adult head-to-head is built.

CI read — the most important structural point for a Fore comparison: tovorafenib's safety database is pediatric. A competitor program enrolling adults and adolescents with BRAF-altered CNS tumors is not comparing like with like. Growth suppression, viral-infection burden (55%), and BSA-dosing error do not translate to an adult population; conversely, tovorafenib has no adequately-powered adult CNS safety dataset at all. Any head-to-head must state the age distribution of both columns on the same slide.

[FLAG: verify — this argument's supporting numbers are unconfirmed] The prior draft supported the above with "median age 9 y, range 1–24 y; 2% <2 y, 67% 2–<12 y, 31% >12 y; N=137." These figures could not be confirmed from any source retrieved in this pass, yet they carry the document's single most important structural argument. Confirm against label §14 / §8.4 or the Nature Medicine baseline-characteristics table before the argument is used in a partner or investor setting. The argument is almost certainly sound — FIREFLY-1 enrolled ages 6 months to 25 years — but the numbers backing it must be checkable.

---

10. Head-to-head safety comparison — tovorafenib column populated, plixorafenib column as template

PLIXORAFENIB COLUMN IS AN UNPOPULATED TEMPLATE. No comparison figures were supplied to this workstream. Per the guardrails, no plixorafenib safety data has been researched, inferred, or estimated — this was re-checked across the whole document in the verification pass and confirmed. Every plixorafenib cell below is a placeholder for the Fore team to complete from internal source documents, and each completed value must be labeled "Fore-internal — verify" with its own N, criterion and data cut.

Safety parameter	Tovorafenib (public, cited)	Plixorafenib (Fore-internal — TO BE COMPLETED)
Safety population N / setting	137, relapsed/refractory pLGG, FIREFLY-1 Arms 1+2 (pooled W&P: 172)	[____ / ____]
Age: median (range); % <12 y	9 y (1–24); 69% <12 y [FLAG: unverified — see §9]	[____]
Median / % exposure ≥1 year	49% ≥1 y; 86% ≥6 mo (N=172)	[____]
Dose studied vs dose on label	Trial 420 mg/m² QW; label 380 mg/m² QW (max 600 mg) — see §6.8	[____]
Boxed warning	None	[____]
Contraindications	None	[____]
Serious AE rate	45% (N=137)	[____]
Grade ≥3 treatment-related AE rate	42% (N=137, Nat Med)	[____]
Fatal (Gr 5) treatment-emergent AE	1/137 (1%) — tumor hemorrhage	[____]
Permanent discontinuation for AE	7% (N=137)	[____]
Dose reduction for AE	24% (N=137)	[____]
Dose interruption for AE	57% (N=137)	[____]
Intratumoral / intracranial hemorrhage	9% (N=172); Gr 5 in 1 (0.6%)	[____]
Rash, any grade / Grade 3–4	77% / 12% (N=137, grouped term, 14 PTs)	[____]
Photosensitivity	12% (N=172)	[____]
Cutaneous SCC / keratoacanthoma	Not in label; adult Ph1 observation unverified — see §6.3	[____]
Hepatic: ALT↑ / AST↑ any grade (Gr 3)	42% (4%) / 74% (2%) (N=172)	[____]
Anemia / Hgb↓ Gr 3–4	15% (lab, N=67–137)	[____]
CPK↑ any grade / Gr 3–4	83% / 11% (lab, N=67–137)	[____]
Hypophosphatemia any / Gr 3–4	87% / 25% (lab, N=67–137)	[____]
Growth-velocity reduction (pediatric)	46% (35% Gr ≥3) of N=133 ≤18 y; recovery shown in n=10/n=17; 3% d/c	[____] (N/A if adult-only)
QT prolongation	Mean ΔQT >20 ms not observed at 380 mg/m² QW	[____]
Ocular (uveitis)	Not in label	[____]
Cardiomyopathy / LVEF	Not in label	[____]
NF1 / paradoxical MAPK activation	Labelled warning (§5.6) — may promote tumor growth in NF1 tumors; BRAF confirmation required	[____]
Contraception constraint	Non-hormonal required (CYP3A induction)	[____]
Required monitoring cadence	LFTs baseline, 1 mo, then q3mo; routine growth; UV precautions	[____]
Post-marketing status	FAERS 292 reports / 94 serious / 12 death outcomes (to 2026-07-30); no §6.2 in label; EU additional monitoring	[____]
Data cut for this column	US PI rev. 08/2025; FAERS 2026-07-30; SNO cutoff 6 Jun 2025	[____]

Instruction to the Fore team completing the right column. State for every entry: (i) the N; (ii) all-causality vs treatment-related attribution; (iii) CTCAE version; (iv) the data cut; (v) whether the AE term is a grouped composite or a single preferred term; (vi) the response/assessment criteria in use (RANO-HGG, RAPNO-LGG or RECIST); and (vii) the age range of the population.

Items (vi) and (vii) were added in the verification pass and are not optional. The tovorafenib column is a pediatric, RAPNO-assessed, single-arm dataset; a plixorafenib CNS column will almost certainly be adult/adolescent and RANO- or RECIST-assessed. Placing them side by side without those two labels reproduces exactly the RANO/RAPNO/RECIST conflation that already required an SME correction on Fore's SNO 2026 CNS work. Separately, tovorafenib's rash (14 grouped terms) and hemorrhage (12 grouped terms) figures will overstate its relative toxicity against any comparator reporting narrow preferred terms.

---

11. Gaps, limitations and open questions

1. No randomized safety data exist for tovorafenib. Every rate above derives from single-arm trials. FIREFLY-2 (topline mid-2027, ~400 patients) is the first controlled dataset.
2. No adult CNS safety dataset. 32 adults in the pooled label population, all non-CNS advanced solid tumors; FIRELIGHT-1 adult figures remain unretrievable (HTTP 403 on two attempts).
3. ~~Nature Medicine Author Correction — unretrieved citation-integrity risk~~ — RESOLVED. The Author Correction (Nat Med 2024, doi:10.1038/s41591-024-02910-1) amends a swapped colour key in Figure 2 — teal now represents "Prior BRAFi/MEKi" and orange "BRAFi/MEKi-naive." No safety or efficacy figure was corrected. The prior draft's live citation-integrity flag is closed. The Nature Medicine full-text safety tables remain paywalled, but the abstract-level safety figures were verified directly in this pass.
4. Per-analyte laboratory denominators (67–137) are not disclosed.
5. No disproportionality analysis performed on FAERS; counts are report frequencies only, on a small (292-report) file with no exposure denominator. The distinct-term total is capped at ≥100 without an API key.
6. The melanocytic naevus / ephelides FAERS cluster has no located published corroboration and no located regulatory signal evaluation — absence of retrieved evidence, not evidence of absence.
7. The FDA multidisciplinary review documents for NDA 217700 / 218033 — which would resolve item 4, the growth-recovery reconciliation (§6.1) and the 420 vs 380 mg/m² exposure question (§6.8) — were not retrieved. This is the single highest-value outstanding retrieval.
8. EU SmPC sections 4.4/4.8 were not read directly — the EPAR public summary was. The EPAR-listed side effects are verified and include growth retardation and epistaxis, but any claim that the EU warning set is identical to the US set remains unverified.
9. The adult phase 1 cutaneous SCC / keratoacanthoma observation (§6.3) — load-bearing for the document's strongest differentiation claim — is not in the retrieved abstract and must be confirmed against the full text.
10. Population age distribution (§9) is unconfirmed despite carrying the document's principal structural argument.
11. Drug-identity scope (top of document) remains unresolved and is a precondition for everything above.

---

12. Bottom line for Fore

Tovorafenib's safety profile is broad, low-grade, and monitoring-heavy rather than acutely dangerous: no boxed warning, no contraindications, 7% permanent discontinuation — but 57% dose interruption, 24% dose reduction, 45% serious AEs, and 42% Grade ≥3 treatment-related AEs (all N=137). Its severe toxicities are disproportionately laboratory (hypophosphatemia Gr 3–4 25%, anemia 15%, CPK 11%).

Its three genuinely differentiating liabilities are: 1. Growth-velocity suppression — 46% of 133, 35% Grade ≥3, showing recovery in the small subsets studied (n=10, n=17), and the only W&P revised post-approval (08/2025); 2. Intratumoral/intracranial hemorrhage — 9%, one on-study death, top serious FAERS term (10 reports); 3. The NF1 paradoxical-activation warning, which fences the drug out of a meaningful pediatric LGG segment.

Its strongest safety advantage — absence of type-I-RAF-inhibitor cutaneous carcinogenesis and uveitis — currently rests on label omission alone, since the supporting adult phase 1 observation could not be verified in this pass and no comparative trial exists. It is therefore contestable and should be framed as such.

The competitively decisive fact for a BRAF-altered CNS program is structural: tovorafenib's safety database is pediatric, single-arm, RAPNO-assessed, and has no adult CNS component at all through the 31 August 2026 data cut. The commercially decisive fact is that the asset now sits inside Servier (acquisition completed 23 April 2026, ~$2.5B), with Ipsen holding ex-US rights, an EU conditional authorisation since 20 April 2026 under additional monitoring, and a randomized front-line readout due mid-2027 that will produce the first controlled safety comparison this drug has ever had.

---

CITATIONS

1. OJEMDA (tovorafenib) tablets / for oral suspension — US Prescribing Information, revised 08/2025. PRIMARY, verified in this pass via DailyMed set ID ea3a9631-3a66-6a7c-e053-2995a90ae2ad — https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=ea3a9631-3a66-6a7c-e053-2995a90ae2ad. Source for all §5 Warnings & Precautions rates, §6.1 Adverse Reactions, Table 6 (AEs ≥20%, N=137), Table 7 (labs ≥20%, N=67–137), §2.1 dosage, §8.4 Pediatric Use, §12.2, §13.1–13.2. Secondary/mirror: accessdata.fda.gov 2025/217700s002s003s004lbl.pdf — returned HTTP 404 on re-fetch during this verification pass; use DailyMed as the citable primary.
2. Kilburn LB, Khuong-Quang D-A, Hansford JR, et al. The type II RAF inhibitor tovorafenib in relapsed/refractory pediatric low-grade glioma: the phase 2 FIREFLY-1 trial. Nature Medicine. 2024;30(1):207–217. doi:10.1038/s41591-023-02668-y. PMID 37978284. PMC10803270. NCT04775485. (Abstract verified directly via EuropePMC in this pass: arm 1 n=77, arm 2 n=60, safety n=137; RANO-HGG ORR 67% primary endpoint; RAPNO ORR 51% incl. minor responses; TRAEs hair colour changes 76%, CPK 56%, anaemia 49%; Grade ≥3 TRAE 42%; 9 (7%) TRAE discontinuations; dosing 420 mg/m² QW, 600 mg max. Full-text safety tables remain paywalled.)
3. Author Correction: The type II RAF inhibitor tovorafenib in relapsed/refractory pediatric low-grade glioma: the phase 2 FIREFLY-1 trial. Nature Medicine. 2024. doi:10.1038/s41591-024-02910-1. RESOLVED in this pass — corrects a swapped colour key in Figure 2 (teal = "Prior BRAFi/MEKi", orange = "BRAFi/MEKi-naive"). No safety or efficacy figure affected.
4. Singh S, Bradford D, Chatterjee S, et al. FDA Approval Summary: Tovorafenib for Relapsed or Refractory BRAF-Altered Pediatric Low-Grade Glioma. Clinical Cancer Research. 2025;31(8):1383–1389. doi:10.1158/1078-0432.CCR-24-3439. PMID 39808502. (Citation and content verified in this pass: accelerated approval 23 Apr 2024; RAPNO by BICR, ORR 51% [95% CI 40–63] in efficacy population N=76; median DOR 13.8 months [95% CI 11.3–NE]; FIREFLY-2 as required postmarketing trial.)
5. US FDA. FDA grants accelerated approval to tovorafenib… 23 April 2024. https://www.fda.gov/drugs/resources-information-approved-drugs/fda-grants-accelerated-approval-tovorafenib-patients-relapsed-or-refractory-braf-altered-pediatric
6. Kline C, et al. LGG-40: Type II RAF inhibitor tovorafenib in relapsed/refractory pediatric low-grade glioma (pLGG): reversible decreases in growth velocity in the phase 2 FIREFLY-1 trial. Neuro-Oncology. 2024 Jun 18;26(Suppl 4). doi:10.1093/neuonc/noae064.431. PMC11183845. Citation and all growth figures verified directly against PMC in this pass (safety cutoff 8 Aug 2023, n=137; growth follow-up through 19 Jan 2024). The same dataset also appears as ASCO 2024 abstract 10036, JCO 2024;42(16_suppl) — cite one, not both as independent.
7. LiverTox — Tovorafenib. NIDDK, NCBI Bookshelf NBK613920. Last updated 8 April 2025. Verified in this pass, including the confirmed denominator-crossing error (attributes ALT 50% / AST 83% to the pooled N=172 cohort; those are the FIREFLY-1 laboratory values, N=67–137). LiverTox likelihood score E*. Cite the label, not LiverTox, for rates.
8. Rasco DW, Medina T, Corrie P, et al. Phase 1 study of the pan-RAF inhibitor tovorafenib in patients with advanced solid tumors followed by dose expansion in patients with metastatic melanoma. Cancer Chemotherapy and Pharmacology. 2023;92(1):15–28. doi:10.1007/s00280-023-04544-5. PMID 37219686. PMC10261210. NCT01425008. (Citation, N=149 (Q2D n=110, QW n=39), RP2D 200 mg Q2D / 600 mg QW, and Grade ≥3 AE rates 58/80 (73%) Q2D and 9/19 (47%) QW verified via EuropePMC in this pass. The SCC/keratoacanthoma observation is NOT in the abstract and remains unverified.)
9. openFDA / FAERS drug/event endpoint. Query: patient.drug.medicinalproduct:"OJEMDA" OR patient.drug.openfda.generic_name:"tovorafenib". meta.last_updated 2026-07-30. Re-queried live in this pass: 292 total reports; 94 serious; 198 non-serious; 12 with seriousnessdeath:1; ≥100 distinct reaction preferred terms (count endpoint caps at 100 without an API key). https://api.fda.gov/drug/event.json
10. Annals of Oncology. 614MO — Tovorafenib in R/R melanoma or other solid tumors with RAF fusions and/or RAF1 amplification (FIRELIGHT-1, NCT04985604, sub-study DAY101-102a). ESMO Congress 2024. FULL TEXT NOT RETRIEVED (HTTP 403 on two attempts).
11. Day One Biopharmaceuticals. Day One Announces Three Year Follow-Up Data From OJEMDA (tovorafenib) Phase 2 FIREFLY-1 Trial at the 2025 SNO Annual Meeting. 24 November 2025. Verified in this pass: data cutoff 6 June 2025; ORR 53% (40/76) RAPNO-LGG; 44/76 (58%) ≥26 cycles; 77% (30/39) treatment-free ≥12 months; median time to next treatment 42.6 months (95% CI 36.7–NE); Grade 3+ AEs ≥5% = decreased growth velocity, anemia, CPK increased, maculopapular rash, ALT increased; "no new safety signals were identified" (sponsor language).
12. Servier. Servier completes the acquisition of Day One Biopharmaceuticals. 23 April 2026 — $21.50/share, ~$2.5B equity value; Day One markets OJEMDA in the US and has licensed ex-US rights to Ipsen. NEW in this pass; corrects the prior draft's sponsor attribution.
13. Day One / Servier. Day One and Servier Complete Enrollment in Pivotal Phase 3 FIREFLY-2 Trial… 8 May 2026. https://servier.mediaroom.com/2026-05-08-Day-One-and-Servier-Complete-Enrollment-in-Pivotal-Phase-3-FIREFLY-2-Trial-of-Tovorafenib-as-a-Front-Line-Treatment-for-Pediatric-Low-Grade-Glioma-pLGG (Verified: ~400 participants, ~140 sites, four SOC chemotherapy comparator regimens, LOGGIC/SIOPe collaboration, topline expected mid-2027.)
14. van Tilburg CM, et al. LOGGIC/FIREFLY-2: a phase 3, randomized trial of tovorafenib vs. chemotherapy in newly diagnosed low-grade glioma harboring an activating RAF alteration. PMC10826080. NCT05566795.
15. European Medicines Agency. Ojemda (tovorafenib) — EPAR medicine overview. Verified directly in this pass: conditional marketing authorisation granted 20 April 2026; under additional monitoring (black triangle); indication ≥6 months, BRAF fusion/rearrangement or V600 mutation, progressed after ≥1 prior systemic therapy; specific obligations to submit final FIREFLY-2 results and undergo annual safety/efficacy review; >1-in-10 adverse reactions include growth retardation and epistaxis. https://www.ema.europa.eu/en/medicines/human/EPAR/ojemda
16. Ipsen. Positive CHMP opinion for Ojemda. 27 February 2026 (CHMP opinion adopted 26 February 2026). (Ipsen's "22 April 2026" is the announcement date; the EMA decision date is 20 April 2026.)
17. US FDA. Potential Signals of Serious Risks / New Safety Information Identified from FAERS / AEMS — quarterly postings. https://www.fda.gov/drugs/fdas-adverse-event-reporting-system-faers/potential-signals-serious-risksnew-safety-information-identified-fda-adverse-event-reporting-system (No tovorafenib/OJEMDA entry identified; negative finding is NON-EXHAUSTIVE and must not be asserted publicly.)

Real-World Evidence — Tovorafenib (OJEMDA™)

Data cut: 31 August 2026. FAERS/openFDA extract current through the 30 July 2026 database refresh (report receipt dates 18 Jun 2024 – 30 Jun 2026; both endpoints re-verified against the API on 31 Aug 2026). Literature search: PubMed (tovorafenib, 50 records, sorted by date), ClinicalTrials.gov API v2, DailyMed/Drugs@FDA, EMA EPAR, company and acquirer disclosures.

[FLAG: drug argument not supplied — human decision required] The competitor-dossier workflow was invoked without a drug argument — the prompt reached this workstream reading literally "the competitor drug." I resolved the target to tovorafenib (OJEMDA, Day One Biopharmaceuticals / Servier; ex-US: Ipsen) because (a) it is the only BRAF-landscape competitor with a post-approval period long enough to generate RWE, (b) the workstream brief specifies "discontinuation and rebound," which is the tovorafenib/pLGG drug-holiday question specifically, and (c) tovorafenib is the anchor competitor in the Fore CI registry (C:\Users\Owner\Documents\Hugh Context Folder\Fore\competitive_intelligence\CI_PROCESS.md, existing dossier data cut 24 Aug 2026). If the intended target was dabrafenib + trametinib or selumetinib, this workstream must be re-run — their RWE bases are much larger and entirely different.

---

1 · Bottom line: the RWE base is thin and should not be over-read

Tovorafenib was approved in the US on 23 April 2024 (accelerated approval) and in the EU on 20 April 2026 (European Commission conditional marketing authorisation; announced by Ipsen 22 April 2026). As of this data cut there is no published prospective registry, no claims/EHR-database utilization or persistence study, and no published disproportionality (FAERS) pharmacovigilance analysis specific to tovorafenib. [FLAG: negative finding — searched PubMed across all 50 tovorafenib-indexed records and web-searched for FAERS studies; absence of a publication is not proof one does not exist in a non-indexed venue or in press.]

What exists post-approval is: - two small real-world case series (n=8 and n=7), both in adults, therefore both entirely off-label; - one retrospective institutional cohort in pediatric LGG (n=236) in which MAPK-inhibitor exposure (including tovorafenib) is examined as a hemorrhage risk factor; - one systematic review pooling hemorrhage signals across MAPK-pathway agents; - spontaneous FAERS reports (n=292 cumulative, verified); - at least one post-approval FDA label revision (08/2025) driven by growth effects; - commercial utilization data (prescription fills, revenue) as an exposure proxy.

Everything on durability, drug holidays, rebound and retreatment remains trial-derived (FIREFLY-1) — it is long-term follow-up of a single-arm phase 2, not real-world evidence, and must be labeled as such in any comparative slide.

---

2 · Real-world exposure denominator (utilization proxy)

Not outcomes data, but it sizes the exposed population underpinning everything below. All five rows below were verified against the Day One releases on 31 Aug 2026.

Metric	Value	Period	Source
OJEMDA US net product revenue	$52.8M	Q4 2025	Day One press release, 24 Feb 2026
OJEMDA US net product revenue	$155.4M (+172% YoY)	FY2025	Day One press release, 24 Feb 2026
OJEMDA prescriptions	1,394	Q4 2025	Day One press release, 24 Feb 2026
OJEMDA prescriptions	4,635 (+181% vs 2024)	FY2025	Day One press release, 24 Feb 2026
2026 US revenue guidance	$225–250M	FY2026	Day One, 11 Jan 2026 (reaffirmed 24 Feb 2026)

Interpretation caveats. "Prescriptions" are fills, not unique patients. Tovorafenib is dosed once weekly and dispensed monthly, so 4,635 FY2025 fills correspond to roughly 390 patient-years of therapy, implying several hundred unique patients treated during 2025 — not 4,635, and not "thousands." An independent cross-check supports the monthly-fill assumption: $155.4M ÷ 4,635 fills ≈ $33.5K net revenue per fill, consistent with a one-month supply rather than a single weekly dose. [FLAG: the fills-to-patients conversion is an analyst calculation, not a disclosed figure — Day One never published a treated-patient count. Do not present a patient number as sourced.] Day One's own characterization ("increasing treatment persistence and expanding prescriber adoption") is a company statement, not a measured persistence metric [FLAG: no published median time-on-therapy, discontinuation rate, or refill-persistence curve from any real-world data source].

Disclosure discontinuity — material for tracking. Servier announced acquisition of Day One on 6 March 2026 and completed the tender offer 23 April 2026 at $21.50/share (~$2.5B equity value) — completion date and price verified against the Servier newsroom release. Day One consequently stopped standalone quarterly reporting; no Q1 or Q2 2026 OJEMDA revenue or prescription figures were located as of this data cut. Ex-US commercialization sits with Ipsen under the exclusive ex-US license announced 25 July 2024: ~$111M upfront comprising ~$71M cash plus a $40M premium equity investment, up to ~$350M in launch and sales milestones, and tiered double-digit royalties starting in the mid-teens. [FLAG: the workstream previously stated "$111M upfront" without the cash/equity split and dated the agreement 23 July 2024; the press release is 25 July 2024. If an execution date distinct from the announcement date matters, confirm from the Day One 8-K/10-Q.] Practical consequence: the utilization proxy that existed through FY2025 is going dark, split across two private/foreign reporters. [FLAG: verify whether Servier or Ipsen disclose OJEMDA-level sales in their annual reports before relying on this metric in future refreshes.]

---

3 · Published real-world series

3a · Adults with BRAF-altered CNS tumors — Mayo Clinic (off-label)

Williams A, Schultz S, Uhm J, Ruff M, Schwartz J, Go R, Gordon R, Keating G, Caron S, Sener U. PATH-94. Use of tovorafenib for BRAF mutant or altered CNS tumors in adults: a single institution case series. Neuro-Oncology 2025;27(Suppl 5):v263. doi:10.1093/neuonc/noaf201.1046 (SNO 2025).

• N = 8 adults treated March 2024 – June 2025, i.e. entirely post-approval and entirely outside the labeled (pediatric) indication. Age 21–66 (median 42); 5M/3F.
• Histologies: pilocytic astrocytoma (n=2), high-grade astrocytoma with piloid features/HGAP (n=2), Erdheim-Chester disease (n=2), pleomorphic xanthoastrocytoma (n=1), unclassified LGG (n=1).
• Alterations: KIAA1549 fusion (n=4), V600E (n=3), GAB2 fusion (n=1).
• Outcomes as reported — read the arithmetic carefully. The abstract reports "radiographic response" in 5 of 8, and that figure comprises 3 patients with stability (one of whom later progressed) plus 2 patients with tumour-size reduction (3 + 2 = 5). Both size reductions were in KIAA1549-fused tumours (a pilocytic astrocytoma and an HGAP). This is therefore a disease-control figure of 5/8, of which only 2/8 shrank — it is NOT 5 responders in addition to 3 stable patients. [CORRECTION: the prior draft listed these additively, implying benefit in 8/8.]
• Toxicity: fatigue (n=2), headache, myalgias, diplopia, cytopenias, transaminitis (n=1 each). 3/8 paused or stopped treatment (intolerable headache; cytopenias; transaminitis).
• [FLAG: response criterion not stated in the abstract] — the abstract does not specify RANO/RAPNO vs. investigator-described "response/stability." Per dossier rules this must not be quoted as an ORR and must not be compared against a RANO/RAPNO-adjudicated figure until the criterion is confirmed. [FLAG: conference abstract, not peer-reviewed full text; n=8, single institution, no central radiology review — hypothesis-generating only.]

3b · Adults with treatment-refractory high- and low-grade gliomas

Ioannou M, Mehta S, Devkota A, et al. Clinical experience with tovorafenib in adults with treatment-refractory high- and low-grade gliomas. Neuro-Oncol Pract 2025;12(6):1051–1057. doi:10.1093/nop/npaf068. PMID 41458920. Peer-reviewed full article (Johns Hopkins; verified December 2025 issue).

• N = 7 adults (5 HGG, 2 LGG) treated through routine clinical care (off-label). All had prior BRAF-targeted therapy; HGG patients had prior radiation and temozolomide.
• 3/7 achieved stable disease or better for ≥4 months.
• Median treatment duration 8 weeks (range 3 weeks – 10 months) — short, and the single most useful real-world durability datapoint currently available in adults.
• 2/7 experienced grade 4 intratumoral haemorrhage.
• 2/7 remained on therapy at 4 and 10 months.
• Authors' conclusion: "limited efficacy in HGG in combination with other standard treatments."
• [FLAG: n=7, single centre, heavily pre-treated, all MAPK-inhibitor-exposed — not generalizable, and not a read on the labeled pediatric relapsed/refractory LGG population. This is a full article, not an abstract; the radiographic response criterion is not given in the abstract and the full text was not obtained for this cut — retrieve it and record the criterion before any comparative use.] [CORRECTION: the prior draft described this source as an abstract while citing it with full volume/page numbers — an internal contradiction.]

Cross-cutting read. Both post-approval adult series are small, off-label, and converge on the same two signals: (i) meaningful activity concentrated in KIAA1549-fusion / pilocytic-type biology, weak activity in adult HGG; (ii) haemorrhage and intolerance driving early discontinuation (3/8 and 2/7 respectively). Combined N across both = 15 adults, of whom 13 had glioma (6 low-grade, 7 high-grade counting the 2 HGAP) and 2 had Erdheim-Chester disease. This is the entire published post-approval adult experience identified in this search.

3c · Pediatric LGG haemorrhage cohort (RWE-grade, but not tovorafenib-specific)

Grin EA, Frome S, Turner J, et al. Incidence and predictors of hemorrhage in pediatric low-grade glioma. J Neurooncol 2026;178(3):96. doi:10.1007/s11060-026-05699-w. PMID 42414678. (Article existence and headline figures verified; full text is paywalled and was not obtained.)

• Retrospective cohort, 236 children with pLGG, 2011–2025 (53.8% boys; median age 7.4 years), 2,234 person-years, median follow-up 8.2 years. Pilocytic astrocytoma 35.6%; genetic syndromes 25.8%; high-risk location (brainstem, optic pathway, hypothalamus) 46.2%.
• Haemorrhage in 12/236 (5.1%); incidence 0.54 per 100 person-years (12 ÷ 2,234 py — arithmetic checks).
• KIAA1549::BRAF fusion carried the strongest association with late intratumoral haemorrhage. [FLAG: the "~6-fold increased risk" effect size and the "median time from diagnosis to haemorrhage 6.4 years" figure could NOT be verified from the abstract or available summaries — obtain full text before quoting either number.]
• MAPK-inhibitor exposure (binimetinib and tovorafenib) reported as associated with haemorrhage on univariate analysis but partially confounded by fusion status. [FLAG: UNVERIFIED — this specific claim, which is the load-bearing sentence for the entire "class effect, not drug effect" argument below, was not confirmable from the abstract. Do not use it in any comparative safety argument until the full text is read.]
• CSF diversion independently increased risk at brainstem/midline locations; no haemorrhages in NF1/genetic-syndrome patients. [FLAG: also unverified from the abstract.]

Why this matters competitively. This is the first real-world attempt to disentangle whether the tovorafenib haemorrhage signal is drug-attributable or tumour-biology-attributable. If the fusion-confounding finding holds, it is genuinely double-edged: it partially defends tovorafenib, but it also implies any type II RAF inhibitor treating a fusion-enriched population inherits an elevated baseline haemorrhage rate. [FLAG: single-institution retrospective; the tovorafenib-exposed subgroup size and event count are not stated in the abstract — obtain full text and extract exposed-N and event-N before using this in a comparative safety argument. Until then this paragraph is a hypothesis, not a finding.]

3d · Systematic review of MAPK-pathway CNS haemorrhage

Damodharan S, Calderon A, Abdelbaki MS. Intratumoral and intracranial hemorrhage associated with MAPK-pathway targeted therapy: a systematic review and mechanistic synthesis. J Neurooncol 2026;178(3):103. doi:10.1007/s11060-026-05714-0. PMID 42455393. (Article existence, authors, journal, volume/article number, and the search window — PubMed/MEDLINE + Embase from inception through May 2026, supplemented by FDA prescribing information — verified. Numerical contents below are paywalled and were not verified.)

• FIREFLY-1 (n=137): any-grade any-site haemorrhage 42%; grade 3–4 haemorrhage 7/137 (5%). Intratumoral haemorrhage 9% of the pooled FDA-label population.
• Comparators: dabrafenib-based regimens in melanoma — ICH 6% (monotherapy); fatal haemorrhage 0.8% (combination). BRAF inhibitor + stereotactic radiosurgery: 3-fold higher odds of ICH vs radiosurgery alone.
• Authors' conclusion: definitions are non-standardized, CNS-specific toxicity capture is inconsistent, and true incidence is undefined; they call for prospective surveillance.

[FLAG: every number in this subsection is unverified — the full text is paywalled. Note that the review's "42% any-grade haemorrhage" does not match the 08/2025 FDA label's "hemorrhagic events occurred in 37% of patients," which I did verify directly. That is most likely a different pooling or data cut, but it could also be a transcription error. Do not quote 42% until the full text is checked against the label's 37%.]

[FLAG: the FIREFLY-1 figures inside this review are trial data, not real-world data — do not present them as RWE. The review's value here is that it establishes haemorrhage as a class-level, not tovorafenib-unique, concern, which is directly relevant to how a competing type II RAF inhibitor will be assessed.]

3e · Practical-management review

Bazer D, Ayanlaja AA, Schreck KC. Practical management of BRAF inhibitors in glioma: toxicity and resistance. CNS Oncol 2026;15(1):2711620. doi:10.1080/20450907.2026.2711620. PMID 42544547. (Authors, journal, volume/issue and DOI verified.) Narrative review; notes durable responses in pediatric LGG versus common resistance and progression in adult LGG and HGG, and stresses proactive toxicity management and dose reduction to avoid treatment interruptions. Expert opinion, not evidence — cite as practice context only.

[FLAG: coverage gap — a SNO/EANO consensus review on clinical management of BRAF-altered glioma in adults and children appears to exist (Neuro-Oncology, doi:10.1093/neuonc/noag135) and was not consulted for this cut. It is guideline-level context, not RWE, but it should be read before the dossier makes any positioning claim about standard of care. Verify the citation before use — it surfaced only in a search-result listing.]

---

4 · Discontinuation, drug holidays and rebound — trial-derived, not real-world

The dossier's "discontinuation and rebound" question is currently answered only by FIREFLY-1 long-term follow-up. Label it accordingly wherever it appears.

Kline C, et al. Clinical stability after tovorafenib treatment in patients with relapsed/refractory pediatric low-grade glioma: updated results from the phase 2 FIREFLY-1 trial. Neuro-Oncology Pediatrics 2026;2(2):wuag022. doi:10.1093/neuped/wuag022. Data cutoff 6 June 2025. Arm 1 n=77 (median follow-up 40.6 months); Arm 2 n=60 (median follow-up 34.0 months). Responses by RAPNO-LGG, independent review committee — criterion confirmed against both the paper and §14 of the 08/2025 US label. All figures below verified against the article.

Endpoint	Value (with denominator)
ORR (RAPNO-LGG, IRC, Arm 1)	53% (40/76 evaluable) — PR 39% (30/76), MR 13% (10/76)
Median DoR	19.4 months (16.6 mo prior-MAPKi; 24.0 mo MAPKi-naïve)
Median PFS (RAPNO)	16.6 months
Median time to next treatment	42.6 months
Entered treatment-free observation	39/77 (51%)
Still off therapy at cutoff	31/39 (79%)
Treatment-free ≥12 months	30/39 (77%)
Median treatment-free interval	Not reached
Rebound (≥25% tumour-size increase within 6 months of last dose vs last on-treatment scan)	12/39 (31%)
— of those, tumour still below baseline	10/12 (83%)
— of those, stabilized without intervention on next scan	9/12 (75%)
Retreated with tovorafenib	8/39 (21%); median 10.5 cycles / 9.0 months; median tumour-size change −38%; all 8 still on retreatment at cutoff

[FLAG: label-vs-update reconciliation — do not quote the two interchangeably.] Both sources report ORR 53% by RAPNO-LGG in 76 evaluable patients, but they are different data cuts and the components differ: the 08/2025 US label gives 95% CI 41–64, PR 29 (38%), MR 11 (14%) and median DoR 18 months (95% CI 12.0–22.8); Kline 2026 (6 Jun 2025 cut) gives PR 30 (39%), MR 10 (13%) and median DoR 19.4 months. Any slide must state which cut it is using.

Growth (the discontinuation driver). Among patients <18 with on- and off-treatment measures (n=81): median height percentile 44th at baseline → 22nd at 12 months → 11th at 24 months on treatment, recovering to the 29th percentile by 12 months off treatment. Among the 74 patients who experienced decreased growth velocity, 67 (91%) had growth recovery and 53 (72%) achieved catch-up growth; median annualized growth velocity rose from 1.40 cm/yr on treatment (IQR 0.86–2.72) to 7.23 cm/yr off (IQR 3.50–10.06). [CORRECTION: the prior draft applied the 91%/72% percentages against the n=81 baseline and flagged an unexplained "n≈74." The denominator is confirmed from the full text as 74 = patients with decreased growth velocity, distinct from the 81 with paired on/off measurements. Flag resolved.]

Safety, pooled arms 1+2 (n=137). 100% had TEAEs; 82% grade ≥3. Common any-grade: hair colour change 77%, CPK elevation 62%, fatigue 61%, anaemia 61%, decreased growth velocity 45% (n=61). Treatment-related SAEs 29/137 (21%) — most often decreased growth velocity (9 patients, 7%) and tumour haemorrhage (5 patients, 4%). Discontinuation for treatment-related AEs 18/137 (13%), predominantly decreased growth velocity (n=6) and intratumoral haemorrhage (n=5, 4%). Intratumoral haemorrhage overall 21/137 (15%), serious in 8 (6%), of which 5 were grade ≥3.

Deaths — state both facts together. Kline reports no treatment-related deaths; five grade 5 TEAEs occurred and were investigator-assessed as unrelated to tovorafenib. However, the 08/2025 US label separately documents a Grade 5 tumour haemorrhage in 1 patient (0.6%) within serious bleeding events. [FLAG: "no treatment-related deaths" must never be presented in a way that implies no fatal haemorrhage occurred. A fatal tumour haemorrhage is on the label; the causality assessment is what is negative, not the event.]

[FLAG: haemorrhage figures differ by data cut and must be labeled.] Intratumoral haemorrhage is 9% per the 08/2025 label and 15% (21/137) per the 6 Jun 2025 Kline cut; any-grade haemorrhagic events are 37% per the label. These are not competing measurements of the same thing — they are sequential cuts of the same trial. Pick one, cite its cut, and do not mix.

Competitive read. The 31% rebound rate is frequently soft-pedalled as "minimal rebound"; the honest statement is: 31% (12/39) of patients who stopped met a ≥25% regrowth threshold within 6 months, and in 75% of those (9/12) the tumour then stabilized without intervention. Both halves belong in any comparison. Equally, only 51% (39/77) of the registrational arm ever reached a drug holiday at all — the majority of the treated population did not. And 13% (18/137) discontinued for toxicity, with growth and haemorrhage the two named causes. [FLAG: all of the above is single-arm phase 2 with no real-world replication — no independent cohort has reproduced the drug-holiday, rebound or retreatment findings outside the trial.]

---

5 · Post-approval regulatory/label activity as a safety signal

• US label revised 08/2025 (Drugs@FDA supplements 217700 s002/s003/s004), with "Recent Major Changes" flagged in §2.1 Dosage and Administration and §5.4 Warnings and Precautions — both dated 08/2025, both concerning effects on growth in pediatric patients (verified verbatim from the label). A dosing/warning change 16 months after launch, in the toxicity domain that also drives discontinuation, is the clearest post-approval safety development to date.
• Current label (rev. 08/2025) figures, pooled safety population N=137 (FIREFLY-1 Arms 1+2), all verified directly: haemorrhagic events 37%, epistaxis 26%, intratumoral haemorrhage 9%, serious bleeding 5% including Grade 5 tumour haemorrhage in 1 patient (0.6%); rash 67% (grade 3 12%); ALT increased 42% (grade 3 4%), AST increased 74% (grade 3 2%); effects on growth in 46% of 133 patients 18 years or younger, 35% grade ≥3; warning that, based on nonclinical data in NF1 models without BRAF alterations, tovorafenib may promote tumour growth in patients with NF1 tumours. [FLAG: the prior draft rendered haemorrhagic events as "37% (52/140)" and growth as "46% (61/133)". The label states 37% of the N=137 safety population and 46% of 133 patients ≤18 — the "/140" denominator could not be reproduced and has been removed. Re-derive any count-form denominator from §6.1 directly.]
• Two NDAs, two presentations. OJEMDA is approved as a tablet (NDA 217700) and as an oral-suspension kit (NDA 218033). This matters for §7: a once-weekly paediatric drug supplied in two formats is a plausible substrate for the dosing-error reports in FAERS.
• Accelerated approval remains unconverted. The label states continued approval "may be contingent upon verification and description of clinical benefit in a confirmatory trial(s)." The confirmatory trial is LOGGIC/FIREFLY-2 (NCT05566795), phase 3 vs standard-of-care chemotherapy, n=418 (actual — enrollment complete), Active, not recruiting, primary completion June 2027 (all verified against the CTG API on 31 Aug 2026). [FLAG: no FDA safety communication, Dear-HCP letter, or MedWatch alert for tovorafenib was located at this data cut, and a targeted search returned none — but the MedWatch archive was not queried directly. Confirm before stating this affirmatively.]
• EU: CHMP positive opinion adopted at the 23–26 Feb 2026 meeting; European Commission conditional marketing authorisation granted 20 April 2026 (Ipsen announced 22 April 2026), EU-27 plus Iceland, Liechtenstein, Norway; monotherapy, ≥6 months of age, BRAF fusion/rearrangement or BRAF V600 mutation, after ≥1 prior systemic therapy. Orphan designation EU/3/21/2434 remains in effect. The conditional MA carries a specific obligation: the company must submit final results of the ongoing randomised study comparing tovorafenib with chemotherapy in paediatric LGG from six months of age (i.e. LOGGIC/FIREFLY-2), with annual reassessment of new safety and efficacy data. The product is under EU additional monitoring, with an agreed risk management plan. EU RWE is effectively zero at this data cut — roughly four months of market presence — but the regulatory machinery that will generate it is now defined. An EU Joint Clinical Assessment has been published. [FLAG: the JCA publication date and content were not verified for this cut; the European Commission published its first JCA on an innovative medicine in June 2026 and Ipsen issued a welcome statement — confirm the date and whether the JCA is tovorafenib-specific before citing.] [CORRECTION: the prior draft recorded "no registry or mandated post-authorisation study identified." That was wrong; see §8.]

---

6 · Off-label and expanded-access use

Off-label use is materially present and measurable, which is unusual this early post-launch and is itself a competitive datapoint: it indicates prescriber willingness to reach for a type II RAF inhibitor outside a narrow pediatric label.

• FAERS: "Off Label Use" is the joint most-frequently reported term — 26 reports of 292 (8.9%), tied with rash (verified). In FAERS, off-label use is coded as a reportable event term, so this is a direct, if crude, indicator. [FLAG: it is a report count, not a use rate — the share of off-label prescribing cannot be inferred from it, because reporting propensity for off-label use is itself non-random.]
• Adults: 15 adults across the two post-approval series (§3a, §3b) — all outside the pediatric label.
• Non-glioma histologies in real-world use: Erdheim-Chester disease (n=2, Mayo series) — a BRAF-driven histiocytosis with no tovorafenib indication in any region.
• Pre-approval expanded access: EAP NCT05760586 ("Expanded Access Program for Tovorafenib (DAY101) in RAF-Altered, Relapsed or Refractory Low-Grade Glioma"), record status now "Approved for Marketing" (verified). A compassionate-use programme has also been advertised in Germany [FLAG: sourced only to a commercial medicine-access broker (everyone.org), not to Day One/Ipsen or a regulator — do not cite externally without primary confirmation].
• Pre-approval single-patient report: a 5-year-old with recurrent spindle cell sarcoma harbouring a novel SNX8-BRAF fusion (JCO Precis Oncol 2023, doi:10.1200/PO.23.00065). This predates approval and is not post-approval RWE — included only because it is frequently miscited as real-world off-label evidence.
• Investigational (not off-label) expansion shaping future real-world use — all records verified against the ClinicalTrials.gov API on 31 Aug 2026: craniopharyngioma NCT05465174 (ph2, n=57 est., recruiting, primary completion Mar 2027); Langerhans cell histiocytosis NCT05828069 (ph2, n=48 est., recruiting, primary completion Sep 2028); vinblastine + tovorafenib in recurrent RAF-altered pLGG NCT06381570 (early ph1, n=57 est., recruiting, primary completion Mar 2027); HGG/DIPG NCT07206849 (ph2, n=79 est., not yet recruiting, start Nov 2026, primary completion Nov 2030); Japanese pediatric bridging study NCT07441707 (ph1, n=6 est., recruiting, started Mar 2026). [CORRECTION: the prior draft omitted primary-completion dates for NCT05828069 and NCT07206849, which are 2028 and 2030 — materially later than the others and relevant to when any of this reaches practice.]
• The melanoma/solid-tumour programmes were terminated: NCT04985604 (ph2, n=23 actual, terminated, primary completion Jul 2024) and NCT07121829 (ph1 ± pimasertib, n=44 actual, terminated, primary completion Dec 2024). The registry whyStopped field for both reads "Sponsor decision." [CORRECTION: flag resolved by querying the field. Do not characterize either as an efficacy failure — the registry gives a business reason, and no efficacy rationale is on the public record.]

---

7 · FAERS-based safety (own extract — read with the stated limits)

Method. openFDA drug/event endpoint, search=patient.drug.openfda.generic_name:"tovorafenib", re-queried 31 Aug 2026; database last_updated 2026-07-30. This is a raw report-count extract. No disproportionality statistic (PRR/ROR/EBGM) was computed and none should be quoted from this table.

Field	Result
Total reports	292
Serious	94 (32.2%)
Non-serious	198 (67.8%)
Reports with a death outcome (seriousnessdeath=1)	12 (4.1%)
Report receipt window	18 Jun 2024 – 30 Jun 2026 (both endpoints verified by ascending/descending sort)
Reports by receipt year	2024 = 69 · 2025 = 179 · 2026 (through 30 Jun) = 44 — sums to exactly 292

[CORRECTION: the prior draft reported the annual split as 2024 ≈76 / 2025 ≈152 / 2026 ≈48 and flagged a "16-report discrepancy." Both the figures and the discrepancy were query artifacts. Re-querying with explicit receivedate:[YYYY0101 TO YYYY1231] range searches returns 69 / 179 / 44, which reconciles exactly to 292. The annual trend is therefore sound and the flag is withdrawn.]

Reporter and geography (previously unextracted, now resolved).

Reporter qualification	n (%)		Country of occurrence	n (%)
Physician	121 (41.4%)		United States	274 (93.8%)
Other health professional	88 (30.1%)		EU (unspecified)	10 (3.4%)
Consumer / non-health professional	76 (26.0%)		Russia	2
Pharmacist	7 (2.4%)		Canada, UK, Israel	1 each
Healthcare-professional-sourced	216 (74.0%)		(3 reports lack a country field)

Most-reported terms (n reports of 292 — all counts re-verified against the API):

#	Term	n	#	Term	n
1	Off label use	26	11	Growth retardation	10
2	Rash	26	12	Intracranial tumour haemorrhage	10
3	Disease progression	24	13	Dermatitis acneiform	9
4	Fatigue	21	14	Nausea	9
5	Accidental underdose	19	15	Anaemia	7
6	Hair colour changes	18	16	Death	7
7	Headache	15	17	Photosensitivity reaction	7
8	Melanocytic naevus	15	18	Weight decreased	7
9	Blood CPK increased	11	19	Blood phosphorus decreased	6
10	Acne	10	20	Asthenia	5

Further terms at n=3–5: constipation, ephelides, erythema, insomnia, product administration error, vomiting (5 each); amenorrhoea, arthralgia, blood bilirubin increased, decreased appetite, epistaxis, haemoglobin decreased, "no adverse event", pruritus, somnolence (4 each); adverse drug reaction, alopecia, anxiety, blister, cellulitis (3 each).

Reading. The profile is concordant with the label and with FIREFLY-1 — hair colour change, rash/acneiform dermatitis/photosensitivity, CPK elevation, fatigue, anaemia — which is reassuring rather than alarming. Three observations deserve attention:

1. Intracranial tumour haemorrhage (n=10) and death (n=7, with 12 reports carrying a death outcome). Consistent with the known haemorrhage signal; FIREFLY-1 reported no treatment-related deaths (though the label records one fatal tumour haemorrhage assessed as unrelated), and FAERS cannot establish causality in a population whose underlying disease can be fatal. Not a new signal, but the one to monitor.
2. Growth retardation (n=10) and amenorrhoea (n=4). Endocrine/growth effects are reaching spontaneous reporting — temporally consistent with the 08/2025 label change. [FLAG: temporal consistency is not evidence of a causal sequence; the label change may also have stimulated reporting of these terms. A pre/post-August-2025 split of these two terms would test that and was not run.]
3. Accidental underdose (n=19) plus product administration error (n=5) = 24 reports, 8.2% of the total. A once-weekly oral supplied as both a tablet (NDA 217700) and an oral-suspension kit (NDA 218033), in a paediatric population, plausibly carries a real-world medication-error burden. In this analyst's view it is the most novel FAERS observation relative to the trial data, and a legitimate design consideration for any competing once-weekly formulation. [FLAG: the "24 / 8.2%" figure is an analyst-constructed composite of two preferred terms, NOT a standardized MedDRA grouping. Before any external use, re-derive it against the Medication Errors SMQ and check for report overlap — a single report can carry both terms, so 19 + 5 may double-count.]

Limits that must travel with these numbers. Spontaneous reports have no exposure denominator; there is no causality assessment; duplicates are not de-duplicated by openFDA; reporting is subject to notoriety, litigation and stimulated-reporting bias. Counts here are not incidences and must never be compared against a trial-derived percentage. [CORRECTION: the prior draft asserted that the 68% non-serious fraction "is a hallmark of programme-driven reporting" from a manufacturer patient-support programme. The reporter data now available do not support that inference — 74% of reports come from healthcare professionals and only 26% from consumers. A high non-serious fraction is still expected for a recently launched product with active manufacturer pharmacovigilance, but the specific consumer/PSP-driven explanation should be dropped unless a manufacturer-vs-direct report split is obtained.] [FLAG: openFDA does not expose a manufacturer-vs-direct (primarysourcecountry/report-source) split in the queried fields; 94% US origin means these counts describe the US experience and say almost nothing about EU practice.]

---

8 · Gaps — where RWE is immature, stated explicitly

Gap	Status at 31 Aug 2026
Real-world durability / time-on-therapy in the labeled pediatric population	None published. All durability is FIREFLY-1. Best real-world duration datapoint is median 8 weeks in 7 refractory adults — a different population.
Real-world persistence, adherence, discontinuation rates	None published. Company language ("increasing treatment persistence") is unquantified.
Real-world response rates by RANO/RAPNO	None. Neither post-approval series states a response criterion.
Rebound after discontinuation outside a trial	None. 31% (12/39) is FIREFLY-1 only, never externally replicated.
Retreatment-after-holiday outside a trial	None. n=8 in FIREFLY-1 only.
Claims/EHR utilization, payer coverage, time-to-access	None published.
Registry / mandated post-authorisation study	CORRECTED — one exists in the EU. The EU conditional MA imposes a specific obligation to submit final results of the ongoing randomised comparative study vs chemotherapy (LOGGIC/FIREFLY-2), with an agreed risk management plan, EU additional monitoring, and annual reassessment. No disease registry with a tovorafenib arm (CBTN, PNOC, SIOPE-LGG) was identified. [FLAG: the EPAR was consulted for the obligation wording but the full Annex II and RMP summary were not read — confirm whether a PASS/PAES distinct from FIREFLY-2 is imposed before describing the obligation set as complete.]
Long-term growth/endocrine outcomes beyond 12 months off therapy	Trial data only, max 24 months on / 12 months off.
Comparative effectiveness vs dabrafenib+trametinib or chemotherapy in practice	None. LOGGIC/FIREFLY-2 (NCT05566795) is randomized vs chemotherapy but is a trial; enrollment complete at n=418; primary completion Jun 2027.
Real-world data in adults with LGG at scale	6 adults with low-grade glioma histology across both series (Mayo 4: 2 pilocytic astrocytoma, 1 PXA, 1 unclassified LGG; Ioannou 2); 13 adults with glioma of any grade; 15 adults total, 2 of whom had Erdheim-Chester disease rather than glioma. [CORRECTION: the prior row read "n=10 LGG/HGG adults with LGG histology," which is internally incoherent and not supported by either series' histology breakdown.]
FAERS disproportionality analysis	Not published; not computed here.
Post-Servier utilization tracking	Disclosure discontinued after Q4 2025.

---

9 · RWE comparison row — plixorafenib column (template, Fore team to complete)

Caveat, per dossier guardrails: tovorafenib figures below are public and cited. The plixorafenib column is left deliberately empty. No plixorafenib RWE figures were supplied to this workstream, and none were researched, inferred, or estimated. Fore must populate this column from cleared internal data and mark each entry "Fore-internal — verify." Note also that plixorafenib is not approved in any region, so several rows below have no plixorafenib analogue by definition — mark those "N/A — pre-approval" rather than leaving them ambiguous.

RWE dimension	Tovorafenib (public, cited, cut 31 Aug 2026)	Plixorafenib — Fore-internal, to be completed
Years since first approval	2.4 yrs US (23 Apr 2024); 0.4 yrs EU (EC decision 20 Apr 2026)	[Fore-internal — verify]
Approval basis / conversion status	US accelerated approval, not yet converted; EU conditional MA with a specific obligation to file final LOGGIC/FIREFLY-2 results; primary completion Jun 2027	[Fore-internal — verify]
Cumulative real-world exposure proxy	4,635 prescription fills FY2025 (fills, not patients — ≈390 patient-years, several hundred unique patients); $155.4M FY2025 US net revenue	[Fore-internal — verify]
Published post-approval real-world series	2 series, n=15 adults total, both entirely off-label	[Fore-internal — verify]
Real-world response criterion used	Not stated in either series [FLAG]	[Fore-internal — verify; must be RANO/RAPNO/iRANO]
Real-world median duration of therapy	8 weeks (n=7 refractory adults); not measured in the labeled population	[Fore-internal — verify]
Drug-holiday / rebound evidence	Trial only (FIREFLY-1, 6 Jun 2025 cut): 51% (39/77) reached a holiday; rebound 31% (12/39); 75% of rebounds (9/12) stabilized untreated	[Fore-internal — verify]
Post-approval label change	Yes — 08/2025, §2.1 dosing + §5.4 warnings, growth effects	[Fore-internal — verify]
FAERS cumulative reports (through 30 Jul 2026 refresh)	292 total; 94 serious (32%); 12 with a death outcome (4%); 74% HCP-reported; 94% US	[Fore-internal — verify]
Distinctive real-world signals	Intracranial tumour haemorrhage (n=10); growth retardation (n=10); dosing/administration error (n=24, 8.2% — analyst composite, verify); off-label use (n=26, 8.9%)	[Fore-internal — verify]
Off-label penetration	Documented: adults, Erdheim-Chester disease, HGG	[Fore-internal — verify]
Registry / mandated post-authorisation study	EU specific obligation (FIREFLY-2 final results) + RMP + additional monitoring; no disease registry arm identified	[N/A — pre-approval, unless a Fore-internal analogue exists]

---

Sources

1. Kline C, et al. Clinical stability after tovorafenib treatment in patients with relapsed/refractory pediatric low-grade glioma: updated results from the phase 2 FIREFLY-1 trial. Neuro-Oncology Pediatrics 2026;2(2):wuag022. doi:10.1093/neuped/wuag022. https://academic.oup.com/neuro-onc-peds/article/2/2/wuag022/8661078 — full text retrieved and figures verified 31 Aug 2026.
2. Williams A, Schultz S, Uhm J, Ruff M, Schwartz J, Go R, Gordon R, Keating G, Caron S, Sener U. PATH-94. Use of tovorafenib for BRAF mutant or altered CNS tumors in adults: a single institution case series. Neuro-Oncology 2025;27(Suppl 5):v263. doi:10.1093/neuonc/noaf201.1046. https://academic.oup.com/neuro-oncology/article/27/Supplement_5/v263/8319150 — abstract retrieved and verified.
3. Ioannou M, Mehta S, Devkota A, et al. Clinical experience with tovorafenib in adults with treatment-refractory high- and low-grade gliomas. Neuro-Oncol Pract 2025;12(6):1051–1057. doi:10.1093/nop/npaf068. PMID 41458920 — citation verified; full text not obtained.
4. Grin EA, Frome S, Turner J, et al. Incidence and predictors of hemorrhage in pediatric low-grade glioma. J Neurooncol 2026;178(3):96. doi:10.1007/s11060-026-05699-w. PMID 42414678 — citation and headline figures verified; full text paywalled, subgroup data unverified.
5. Damodharan S, Calderon A, Abdelbaki MS. Intratumoral and intracranial hemorrhage associated with MAPK-pathway targeted therapy: a systematic review and mechanistic synthesis. J Neurooncol 2026;178(3):103. doi:10.1007/s11060-026-05714-0. PMID 42455393 — citation and search window verified; all numerical contents unverified (paywalled).
6. Bazer D, Ayanlaja AA, Schreck KC. Practical management of BRAF inhibitors in glioma: toxicity and resistance. CNS Oncol 2026;15(1):2711620. doi:10.1080/20450907.2026.2711620. PMID 42544547 — citation verified.
7. OJEMDA (tovorafenib) US Prescribing Information, revised 08/2025. DailyMed SETID ea3a9631-3a66-6a7c-e053-2995a90ae2ad. https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=ea3a9631-3a66-6a7c-e053-2995a90ae2ad ; Drugs@FDA label 217700s002s003s004. https://www.accessdata.fda.gov/drugsatfda_docs/label/2025/217700s002s003s004lbl.pdf — read directly; Recent Major Changes, §14 efficacy and §6.1 safety figures verified verbatim. Oral-suspension kit NDA 218033: https://www.accessdata.fda.gov/drugsatfda_docs/label/2024/218033s000lbl.pdf
8. European Medicines Agency. Ojemda EPAR. https://www.ema.europa.eu/en/medicines/human/EPAR/ojemda — read directly: EC decision 20 April 2026, conditional MA, specific obligation (final results of the ongoing comparative study vs chemotherapy), RMP, additional monitoring, orphan designation EU/3/21/2434. Public assessment report: https://www.ema.europa.eu/en/documents/assessment-report/ojemda-epar-public-assessment-report_en.pdf ; CHMP news item: https://www.ema.europa.eu/en/news/new-medicine-treat-paediatric-low-grade-glioma
9. Singh S, et al. FDA Approval Summary: Tovorafenib for Relapsed or Refractory BRAF-Altered Pediatric Low-Grade Glioma. Clin Cancer Res 2025;31(8):1383. PMID 39808502. https://aacrjournals.org/clincancerres/article/31/8/1383/754518/
10. Kilburn LB, et al. The type II RAF inhibitor tovorafenib in relapsed/refractory pediatric low-grade glioma: the phase 2 FIREFLY-1 trial. Nat Med 2024;30(1):207–217. doi:10.1038/s41591-023-02668-y. PMID 37978284. https://www.nature.com/articles/s41591-023-02668-y — citation, volume/pages and PMID verified. Note an Author Correction exists (Nat Med 2024, doi:10.1038/s41591-024-02910-1) [FLAG: the correction's content was not reviewed — check what it amends before quoting the primary paper]. Note also that the widely quoted 67% ORR from this paper is by RANO-LG, whereas the label and the Kline update use RAPNO-LGG (53%) — never mix the two.
11. openFDA FAERS drug/event API, patient.drug.openfda.generic_name:"tovorafenib", queried 31 Aug 2026; database last_updated 2026-07-30. https://api.fda.gov/drug/event.json — all counts in §7 independently re-run and verified, including the corrected annual split (69/179/44) and the newly extracted reporter-qualification and country distributions.
12. ClinicalTrials.gov API v2, term tovorafenib, retrieved 31 Aug 2026 — NCT05566795, NCT04775485, NCT05760586, NCT05465174, NCT05828069, NCT06381570, NCT07206849, NCT07441707, NCT04985604, NCT07121829. Status, phase, enrollment, primary-completion and whyStopped fields verified record-by-record.
13. Day One Biopharmaceuticals. Fourth Quarter and Full Year 2025 Financial Results, 24 Feb 2026. https://ir.dayonebio.com/news-releases/news-release-details/day-one-reports-fourth-quarter-and-full-year-2025-financial — all revenue and prescription figures verified.
14. Day One Biopharmaceuticals. Preliminary 2025 OJEMDA Net Product Revenue and 2026 Guidance, 11 Jan 2026. https://www.globenewswire.com/news-release/2026/01/11/3216499/0/en/day-one-announces-preliminary-2025-ojemda-net-product-revenue-and-provides-2026-net-product-revenue-guidance.html
15. Ipsen. Ojemda approved in the European Union…, 22 Apr 2026. https://www.ipsen.com/press-release/ojemda-approved-in-the-european-union-as-the-first-targeted-therapy-in-relapsed-or-refractory-pediatric-low-grade-glioma-regardless-of-braf-alteration-3278647/ — press-release date; the EC decision date is 20 Apr 2026 per source 8.
16. Ipsen / Day One. Exclusive ex-U.S. licensing agreement to commercialize tovorafenib, announced 25 Jul 2024 (~$111M upfront = ~$71M cash + $40M premium equity; up to ~$350M milestones; tiered double-digit royalties starting mid-teens). https://www.ipsen.com/press-release/ipsen-and-day-one-enter-into-exclusive-ex-u-s-licensing-agreement-to-commercialize-tovorafenib-for-the-most-common-childhood-brain-tumor-2918482/
17. Servier. Servier completes the acquisition of Day One Biopharmaceuticals. https://servier.com/en/newsroom/servier-completes-the-acquisition-of-day-one-biopharmaceuticals/ — tender offer completed 23 Apr 2026, $21.50/share, ~$2.5B, verified. Announcement 6 Mar 2026: BioPharma Dive, https://www.biopharmadive.com/news/day-one-servier-acquisition-deal-cancer-drug-research-biotech/814065/ ; SEC filings CIK 1845337 (SC TO-T/A).
18. Activity of type II RAF inhibitor tovorafenib in a pediatric patient with recurrent spindle cell sarcoma harboring a novel SNX8-BRAF gene fusion. JCO Precis Oncol 2023. doi:10.1200/PO.23.00065 — pre-approval, included for disambiguation only.
19. (Not consulted — flagged for the next cut.) BRAF-altered glioma in adults and children: a SNO/EANO consensus review on clinical management and future directions. Neuro-Oncology, doi:10.1093/neuonc/noag135. [FLAG: citation surfaced in search results only; verify before use.]

Regulatory & Label Status by Region — Tovorafenib (OJEMDA)

Data cut: 31 August 2026. All figures below are public-domain and sourced to primary regulatory documents (FDA approval letters, FDA-approved labeling, Drugs@FDA/openFDA, the FDA Orange Book bulk data files, the EMA CHMP Assessment Report, the EU HTA Joint Clinical Assessment report, Health Canada records) or dated company press releases. Promotional headlines are separated from approved indication text throughout.

CRITICAL TEMPORAL CAVEAT — read before using any efficacy figure below. The FIREFLY-1 efficacy results that support both the FDA 08/2025 label and the EU conditional MA derive from a single data cutoff of 10 May 2024 (stated explicitly in the CHMP Assessment Report, Table on p.88) [24]. The August 2025 FDA efficacy supplement did not introduce data newer than May 2024. Every ORR/DoR figure in this workstream is therefore approximately 27 months old at the 31 Aug 2026 dossier data cut. Do not describe these numbers as "current" or as "longer follow-up" without naming the 10 May 2024 cut.

Scope note. The workflow that generated this workstream was invoked without a drug argument, so the prompt carried the literal placeholder "the competitor drug." The subject was resolved as tovorafenib (OJEMDA), Day One Biopharmaceuticals (now Servier) from this session's working files and from the CI registry at C:\Users\Owner\Documents\Hugh Context Folder\Fore\competitive_intelligence\CI_PROCESS.md, which lists tovorafenib as the LIVE dossier. [FLAG: drug identity inferred, not passed explicitly — confirm this is the intended target before the dossier is circulated.]

---

1 · Snapshot by region

Region	Status	Type	Key date	Holder
US (FDA)	Approved	Accelerated approval (21 CFR 314.510) — not yet converted	23 Apr 2024	Day One Biopharms (US NDA holder of record) [1][2][7]
EU (EC/EMA)	Approved	Conditional marketing authorisation (one-year, annually renewable)	CHMP opinion 26 Feb 2026; EC decision 20 Apr 2026	Ipsen Pharma [5][6][8]
Canada (Health Canada)	Approved	Notice of Compliance with conditions (NOC/c)	reported in the August 2026 Health Product InfoWatch; PDL addition 27 Aug 2026	Ipsen Biopharmaceuticals Canada Inc. [26]
EU HTA	JCA report published	EU HTA Regulation Joint Clinical Assessment (the first ever completed)	endorsed by Coordination Group 30 Apr 2026; EC procedural review 19 May 2026; published 9 Jun 2026	assessor NCPE (IE), co-assessor IQWiG (DE) [11][12][25]
Japan (PMDA)	No approval located	Phase 1 in Japanese children only	NCT07441707 started 3 Mar 2026	— [19]
UK (MHRA)	No marketing authorisation located	—	—	— [FLAG below]
Other (CH/AU/KR/TW/UAE/RU)	Not verified	—	—	— [FLAG below]

All three approved regions restrict use to BRAF-altered disease after prior systemic therapy, and all three used a contingent instrument (US accelerated approval, EU conditional MA, Canada NOC/c). No region has approved tovorafenib on full/unconditional evidence of clinical benefit.

[FLAG: the exact Health Canada NOC date and submission details were not read from a primary page — canada.ca returned HTTP 403 to automated fetch. The NOC/c is corroborated by three independent canada.ca listings surfaced in search (August 2026 Health Product InfoWatch; the "Qualifying notice for Ojemda", control number 301603; and the Prescription Drug List addition notice dated 2026-08-27). Confirm the exact NOC date, DIN(s) and the Letter of Undertaking terms in the Health Canada Notice of Compliance Database and the NOC/c qualifying-notice page before external use.]

---

2 · United States — FDA

2.1 Approval mechanics — two NDAs, one action date

Tovorafenib was approved as two separate NDAs on the same day, both signed by Martha B. Donoghue, M.D., Acting Associate Director, Pediatric Oncology, OOD/OND/CDER, on 23 April 2024 [1][2]:

NDA	Product	Submission class	Review	Original approval
217700	OJEMDA tablet, 100 mg	Type 1 — New Molecular Entity	Priority	23 Apr 2024 [7]
218033	OJEMDA for oral suspension, 25 mg/mL	Type 3 — New Dosage Form	Priority	23 Apr 2024 [7]

Both were approved under accelerated approval pursuant to §506(c) FDCA and 21 CFR 314.510 [1][2]. Neither application was referred to an advisory committee — FDA stated the application "did not raise significant safety or efficacy issues that were unexpected in the intended population" [1][2].

[FLAG: the submission date "31 August 2023" asserted in the prior draft was not found in the approval letters or the openFDA record retrieved. Verify the NDA receipt date against the Drugs@FDA approval package (or the FDA review memoranda) before asserting it; it is not carried here.]

Rare Pediatric Disease Priority Review Voucher: granted on the tablet NDA only (tracking number PRV NDA 217700) [2]. The PRV request on the suspension NDA 218033 was denied, FDA citing §529(g) FDCA — verbatim: "you have been granted a rare pediatric disease priority review voucher under NDA 217700 … and 'no sponsor of a rare pediatric disease product application may receive more than one priority review voucher issued under any section of [the FDCA] with respect to the drug for which the application is made'" [1]. (Competitive read: a single PRV was monetizable, not two.)

Tovorafenib also carried Breakthrough Therapy and Rare Pediatric Disease designations, and Orphan Drug designation for malignant glioma [10]. [FLAG: exact designation grant dates not verified against FDA's designation databases — verify in the OOPD orphan-designation database and FDA BTD listing before external use.]

2.2 Approved indication — verbatim

"OJEMDA is indicated for the treatment of patients 6 months of age and older with relapsed or refractory pediatric low-grade glioma (LGG) harboring a BRAF fusion or rearrangement, or BRAF V600 mutation.

This indication is approved under accelerated approval based on response rate and duration of response [see Clinical Studies (14)]. Continued approval for this indication may be contingent upon verification and description of clinical benefit in a confirmatory trial(s)." — OJEMDA PI §1, rev. 08/2025 [4]

The accelerated-approval sentence is unchanged in the current 08/2025 label [4], i.e. the accelerated approval had not been converted to traditional approval as of the 31 Aug 2026 data cut.

Boundaries worth noting for competitive positioning: - Age floor 6 months; no upper age limit in the indication text, though the population studied is pediatric LGG (FIREFLY-1 Arm 1 age range 2–21 years; pooled safety range 1–24 years) [4]. - Relapsed or refractory only — not front-line. - BRAF fusion/rearrangement or V600 mutation — broader than V600E-restricted labels, but not BRAF-agnostic. - NF1-associated tumors are a labeled warning (§5.6: "tovorafenib may promote tumor growth in patients with NF1 tumors"), and patients with known or suspected NF1 were excluded from FIREFLY-1 [3][4]. - Contraindications: None [4].

2.3 Approval basis (regulatory endpoints)

Approval rested on FIREFLY-1 (NCT04775485) Arm 1, single-arm, open-label, ORR and DoR by blinded independent central review using RAPNO-LGG criteria — the CNS-appropriate criterion, not RECIST [4][9].

Denominator discipline: the efficacy population is N=76 — "patients who had measurable disease at baseline and who received OJEMDA" [4]. Arm 1 as enrolled/treated was 77 (ITT) [24]; Arm 1 + Arm 2 pooled for safety was 137 [1][4][24]. Always state which of 76 / 77 / 137 a figure refers to.

As originally approved (label rev. 4/2024): ORR 51% (95% CI 40–63), 0 CR / 28 PR (37%) / 11 MR (14%); DoR N=39, median 13.8 months (95% CI 11.3, NE); DoR ≥6 mo 85%; DoR ≥12 mo 23% [3]. A supportive secondary read by RANO-LGG (2011), n=76, was reported in running text (not tabulated): ORR 53% (95% CI 41, 64), 20 PR + 20 MR [3].

Confusion trap: the 4/2024 RANO-LGG ORR (53%, 95% CI 41–64) is numerically identical to the 08/2025 RAPNO-LGG ORR (53%, 95% CI 41–64). These are different criteria, different response compositions, and different label revisions. Never cite "53% (41, 64)" without naming both the criterion and the label revision.

2.4 Dosing (dosage regimen unchanged 2024 → 2025)

380 mg/m² orally once weekly, maximum 600 mg once weekly, with or without food, until disease progression or intolerable toxicity [3][4]. Once-weekly oral dosing is the distinctive regulatory/commercial feature.

Formulation availability by BSA (corrected): tablets are available only for BSA ≥0.90 m² (400 mg / 500 mg / 600 mg bands); the oral suspension spans the full BSA range, 0.30 m² to ≥1.40 m² (125 mg up to 600 mg) — it is not restricted to the low-BSA bands. Below BSA 0.90 m² the suspension is the only option; no dosage is established below BSA 0.30 m² [4]. Missed dose ≤3 days: take it; >3 days: skip [4].

Note for any exposure comparison: FIREFLY-1 actually dosed patients at approximately 420 mg/m² (range 290–476 mg/m², i.e. 0.76–1.25× the approved dose); the 380 mg/m² recommendation was derived post hoc [4]. The efficacy figures were not generated at the labeled dose.

2.5 Accelerated-approval and postmarketing obligations (NDA 218033 letter; parallel obligations on 217700)

Subpart H confirmatory requirement — PMR 4626-1 [1]:

"Conduct a multiregional, randomized clinical trial comparing tovorafenib to physician's choice of chemotherapy in pediatric patients with low grade glioma with RAF fusions or rearrangements, or V600 mutations, intended to verify and describe the clinical benefit of tovorafenib through assessment of overall response rate (ORR) as a primary endpoint and progression-free survival (PFS) and duration of response, as determined by blinded independent central review, as key secondary endpoints."

Committed timetable: interim report 04/2028 · trial completion 10/2031 · final report 04/2032 [1]. Two 180-day AA progress reports per year are required until the final report [1].

The trial serving this PMR is LOGGIC/FIREFLY-2 (NCT05566795) — phase 3, tovorafenib vs. SoC chemotherapy, primary endpoint ORR vs SoC chemotherapy, secondary PFS/EFS/TTNT/OS. Registry status at the data cut: ACTIVE_NOT_RECRUITING, N=418 actual, start 27 Feb 2023, primary completion 06/2027, study completion 06/2031, last update posted 14 May 2026 [13]. Enrollment completion was announced 8 May 2026 [14].

Material regulatory observation — line-of-therapy mismatch. FIREFLY-2 enrolls patients who "meet indication for first-line systemic therapy," excluding "prior or ongoing nonsurgical anticancer therapy for this indication" [13]. The approved US indication is relapsed or refractory [4]. The confirmatory trial therefore verifies clinical benefit in a treatment-naïve population, one line earlier than the population the accelerated approval covers. PMR 4626-1's own text does not restrict the population to relapsed/refractory [1], so this is compliant as written — but it means a positive FIREFLY-2 does not directly re-demonstrate benefit in the labeled R/R setting, and it is the plausible vehicle for an eventual front-line label expansion. [FLAG: this is an inference about FDA's conversion logic, not an FDA statement — do not present as FDA's position. Verify against the 180-day AA progress reports or any published FDA correspondence before external use.]

Other §505(o)(3) postmarketing requirements [1] — note these are required, i.e. FDA made statutory findings that a serious risk warranted them:

PMR	Subject	Committed milestones
4626-2	Mouse carcinogenicity study	completion 03/2025; final report 08/2025
4626-3	Rat carcinogenicity study (SPA required)	final protocol 09/2024; completion 12/2027; final report 05/2028
4626-4	Long-term growth & development — height, weight, height velocity, height SDS, adrenarche, menarche, Tanner stage; monitoring to discontinuation or a minimum of 5 years, whichever first	interim 04/2028; completion 10/2031; report 04/2032
4626-5	Gonadal toxicity in adolescent/young adult patients — pubertal + hormonal assessment, min. 5 years, incl. time to gonadal-function recovery off-treatment	draft protocol 12/2024; final protocol 04/2025; interim 04/2028; completion 10/2031; report 04/2032
4626-6	DDI PK trial with a strong CYP2C8 inhibitor	draft protocol 12/2024; final protocol 05/2025; completion 11/2025; report 04/2026
4626-7	DDI PK trial — tovorafenib on CYP3A4/2C8/1A2/2B6/2C9/2C19 substrates	completion 06/2026; report 12/2026
4626-8	DDI PK trial — tovorafenib on BCRP substrates	draft protocol 12/2024; final protocol 06/2025; completion 06/2026; report 12/2026

FDA's statutory finding, verbatim [1]: "only a clinical trial (rather than a nonclinical or observational study) will be sufficient to assess a known serious risk of long-term adverse effects on the growth and development of pediatric patients, and to assess a signal of a serious risk of gonadal toxicity in adolescent and young adult patients…" Note the asymmetry in FDA's own language: growth is a known serious risk; gonadal toxicity is a signal of a serious risk (PMR 4626-5's own text says "the potential serious risk of gonadal toxicity").

Postmarketing commitments [1]: 4626-9 — complete FIREFLY-1 including follow-up for the 137 Arm 1+2 patients, characterising ORR/DoR per RAPNO with IRC assessment for Arm 1 (interim 03/2025; completion 05/2027; report 12/2027); 4626-10 — CYP2C8 inducer DDI PK trial; 4626-11 — analytical/clinical validation to support an in-vitro companion diagnostic; 4626-12 — dissolution profile data (CMC), interim report 10/23/2024.

[FLAG: completion status of PMRs 4626-2 through 4626-8 and PMCs 4626-9/-10/-11/-12 is not verified. Milestones for 4626-2, -6, -9, -10, -11 and -12 have passed as of the data cut. Query the FDA Postmarketing Requirements and Commitments database (accessdata.fda.gov/scripts/cder/pmc) for current status before any board or investor use — commitment slippage is a live competitive datapoint.]

2.6 Post-approval label history — Drugs@FDA record

NDA	Supplement	Class	Approval date
217700 & 218033	SUPPL 1	Labeling	26 Jun 2024
217700 & 218033	SUPPL 2	Efficacy	27 Aug 2025
217700 & 218033	SUPPL 3	Labeling	27 Aug 2025
217700 (only)	SUPPL 4	Labeling	27 Aug 2025

Source: Drugs@FDA via the openFDA drug/drugsfda endpoint, retrieved 31 Aug 2026 [7] — verified against the retrieved JSON. All supplements carried STANDARD review priority; only the two ORIG-1 submissions were PRIORITY. The resulting label is revised 08/2025 and flags Recent Major Changes: Dosage and Administration (2.1), 08/2025; Warnings and Precautions (5.4), 08/2025 [4].

(a) §2.1 Patient Selection — companion diagnostic now exists. The 4/2024 label stated: "An FDA approved test for the detection of BRAF fusion or rearrangement, or BRAF V600 mutation in relapsed or refractory pediatric LGG is not currently available" [3]. The 08/2025 label replaces this with a pointer to fda.gov/companiondiagnostics [4]. FoundationOne CDx was approved as the companion diagnostic for OJEMDA on 17 January 2025 [15][16] — closing PMC 4626-11 in substance and removing a real-world diagnostic-access friction point.

(b) §5.4 Effect on Growth — the labeled growth estimate rose sharply. This is the single most consequential label change to date:

	Label rev. 4/2024 [3]	Label rev. 08/2025 [4]
Treatment-emergent growth AEs	15% of patients ≤18 y (denominator not stated)	46% of 133 patients ≤18 y
Severity band	Grade 3 events in 5%	Grade 3 or higher in 35%
Permanent discontinuation for growth	2% (n=2)	3%
Dose interruption / reduction	not stated	5% / 2.3%
Median Δ height percentile	not reported	−14 (z −0.6) at 12 mo (N=107); −20 (z −0.9) at 18 mo (N=95)
Growth velocity on treatment	"recovered after interruption"	median annualized 0.86–1.8 cm/yr over 2 yr (81 evaluable); 4.2 cm/yr in 17 patients ≥90 days off-treatment

The 08/2025 text also softened the recovery claim from "Growth velocity recovered" to "Growth velocity improved after interruption" [3][4].

[FLAG: the 15%→46% delta is NOT a strictly matched comparison, on two counts. (i) The 4/2024 sentence gives no explicit denominator. (ii) The severity bands differ — 4/2024 reports "Grade 3 events in 5%", 08/2025 reports "35% were Grade 3 or higher". Both figures derive from FIREFLY-1 Arms 1+2 (N=137; 133 patients ≤18 y in the 2025 text), and the change plausibly reflects more systematic ascertainment rather than a new toxicity — but the label does not state the mechanism, and note that the underlying data cut did NOT advance (10 May 2024 for efficacy; the safety cut is not stated in the label). Characterise this as "the labeled estimate rose materially at the 08/2025 revision," NOT as "toxicity worsened over time." Obtain the SUPPL 2/3 review documents before any causal characterisation.]

Warnings that did and did not change (verified by programmatic diff of the two label bodies): §5.1 Hemorrhage, §5.2 Skin Toxicity/Photosensitivity, §5.3 Hepatotoxicity and §5.6 NF1 Associated Tumors are textually identical between 4/2024 and 08/2025 (pooled safety population, see denominator note below) [3][4]. §5.5 Embryo-Fetal Toxicity also changed, though not flagged in Recent Major Changes: the word "nonhormonal" was deleted from the male-partner advice (4/2024 "Advise male patients … to use effective nonhormonal contraception"; 08/2025 "…to use effective contraception"), alongside two typographical fixes. The female advice still requires nonhormonal contraception. So §5.4 growth is the only materially revised warning — but it is not the only revised warning.

Pooled safety denominator (corrected). §6.1 defines the safety population supporting §§5.1–5.3 as 172 patients: 140 patients "with relapsed or refractory pediatric LGG or advanced solid tumors harboring a RAF alteration" dosed by BSA, plus 32 adult patients with advanced solid tumors at a flat 600 mg [4]. The 140 is not "140 pediatric LGG patients." Separately, the pediatric LGG safety set is FIREFLY-1 Arms 1+2, N=137 [4]. Among the 172, 86% were exposed ≥6 months and 49% ≥1 year.

(c) §14 Clinical Studies — efficacy updated at the same N and the same data cut. The Efficacy (SUPPL 2) supplement refreshed FIREFLY-1 results without changing the denominator or the indication:

FIREFLY-1 Arm 1, RAPNO-LGG by independent review, N=76	Label 4/2024 [3]	Label 08/2025 [4]
ORR (95% CI)	51% (40, 63)	53% (41, 64)
CR / PR / MR	0 (0) / 28 (37%) / 11 (14%)	CR row not tabulated (EPAR confirms CR=0) / 29 (38%) / 11 (14%)
DoR denominator	N=39	N=40
Median DoR (95% CI)	13.8 mo (11.3, NE)	18 mo (12.0, 22.8)
DoR ≥6 mo	85%	row dropped
DoR ≥12 mo	23%	65%
DoR ≥18 mo	not reported	50%
Median time to response	5.3 mo (range 1.6–11.2)	5.4 mo (range 1.6–17.5)

08/2025 exploratory subgroups: ORR 53% in BRAF fusion/rearrangement (n=64); 50% in BRAF V600E (n=12); 51% with prior MAPK-targeted therapy (n=45); 55% without (n=31). By RANO-LGG (2011), n=76: ORR 54% (95% CI 42, 65), 23 PR + 18 MR [4]. Patients had received a median of 3 prior systemic regimens (range 1–9); 74% had a KIAA1549:BRAF fusion, 16% a V600E mutation, 11% "other" BRAF alteration [4].

Competitive read: FDA accepted a materially improved DoR (median 13.8 → 18 months; ≥12-month DoR 23% → 65%) without converting the accelerated approval. Durability alone did not move FDA off Subpart H — the randomized comparison is the gate. [FLAG: this is an inference about FDA's reasoning, not an FDA statement.]

[FLAG — MR and criterion dependence, the single biggest cross-drug comparison hazard in this dossier. MR (minor response, 25–49% reduction) is a RAPNO-LGG/RANO-LGG category with no RECIST equivalent; 11 of 40 responders (27.5%) in the 08/2025 label are MRs, not PRs. Worse, the same 76 patients yield materially different ORRs depending on criterion: the CHMP Assessment Report reports RAPNO 52.6%, RANO-LGG 53.9%, and RANO-HGG 71.0% — an ~18-point swing from criterion choice alone [24]. The EU Joint Clinical Assessment, when it needed a comparable measure, used a CR+PR composite excluding MR [25]. Any cross-drug ORR comparison must state the criterion AND whether MR is included, or the comparison is invalid.]**

2.7 US regulatory exclusivity and patents — VERIFIED

The Orange Book HTML product page returns HTTP 404, but the Orange Book bulk data files were retrieved successfully and are authoritative [23]. Both NDAs carry identical entries:

NDA	Type	Code	Expiry
217700 & 218033	Exclusivity	NCE (new chemical entity, 5 yr)	23 Apr 2029
217700 & 218033	Exclusivity	ODE-478 (orphan drug, 7 yr)	23 Apr 2031

NDA	Patent	Expiry	Drug substance	Drug product	Submitted
217700 & 218033	8,293,752	4 Aug 2031	Y	Y	8 May 2024
217700 & 218033	10,426,782	23 Jun 2035	—	Y	8 May 2024

Applicant of record in the Orange Book products file: DAY ONE BIOPHARMS / DAY ONE BIOPHARMACEUTICALS INC, both products RLD and RS, RX [23].

[FLAG: no pediatric-exclusivity (PED) extension appears in the retrieved exclusivity file, and no patent-use codes are listed. Orange Book data are a monthly snapshot — re-pull before any IP, LOE or generic-entry statement, and confirm with the commercial/IP workstream, which owns this axis. Note also that listed patent expiry is not the same as loss of exclusivity: ODE-478 to Apr 2031 and the 8,293,752 patent to Aug 2031 are the near-term gates; 10,426,782 (drug product only) runs to Jun 2035.]

---

3 · European Union — EMA / European Commission

3.1 Approval mechanics

• CHMP positive opinion: 26 February 2026, recommending a conditional marketing authorisation [5][6][24].
• European Commission decision: 20 April 2026, granting a one-year conditional marketing authorisation (annually renewable) [5]; announced by Ipsen 22 April 2026 [8].
• Marketing authorisation holder: Ipsen Pharma, 70 rue Balard, 75015 Paris [5]. Ipsen licensed ex-US rights from Day One in 2024 [8][17].
• Orphan designation: EU/3/21/2434, designated 20 May 2021 [5].
• ATC code: L01EC04; under additional monitoring (black inverted triangle) [5].
• Territory: all 27 EU Member States plus Iceland, Liechtenstein and Norway [8].
• Procedure number EMEA/H/C/006140/0000; CHMP Assessment Report EMA/67438/2026, adopted 26 February 2026 [24].

3.2 EU indication — verbatim

"Ojemda is indicated as monotherapy for the treatment of patients 6 months of age and older with paediatric low-grade glioma (LGG) harbouring a BRAF fusion or rearrangement, or BRAF V600 mutation, who have progressed after one or more prior systemic therapies" [5][8]

3.3 US vs EU wording — the differences that matter

Axis	FDA [4]	EU [5][8]
Regulatory instrument	Accelerated approval (Subpart H)	Conditional MA (one-year, annually renewable)
Prior-therapy qualifier	"relapsed or refractory"	"who have progressed after one or more prior systemic therapies"
Monotherapy stated	not stated in §1	"as monotherapy" explicit
Age	≥6 months	≥6 months (same)
Biomarker	BRAF fusion/rearrangement or V600	same

The EU wording is operationally more specific: it names an explicit prior-line requirement (≥1 prior systemic therapy) rather than the clinical descriptor "relapsed or refractory," and it locks monotherapy use. In practice both restrict to second-line-or-later BRAF-altered pLGG. Canada's NOC/c indication follows the EU construction ("who have received one or more prior systemic therapies") [26].

3.4 Conditional MA specific obligations — VERBATIM, from Annex II

The CHMP Assessment Report §1.9.7 sets out the Article 14-a specific obligations. There are two, both due 30 April 2032 [24]:

"In order to confirm the efficacy and safety of tovorafenib in the treatment of patients 6 months of age and older with paediatric low-grade glioma (LGG) harbouring a BRAF fusion or rearrangement, or BRAF V600 mutation, the MAH shall conduct and submit the final report of the phase III randomised, parallel-group, two-arm study (FIREFLY-2) to evaluate the efficacy and safety of tovorafenib monotherapy versus standard of care (SoC) chemotherapy in patients with paediatric low-grade glioma harbouring an activating rapidly accelerated fibrosarcoma gene (RAF) alteration requiring first-line systemic therapy." — due 30 April 2032

"The MAH shall generate additional PK data in paediatrics patients below 2 years of age and submit an updated population PK model incorporating these data, including an assessment of systemic exposure and, if necessary, revised dosing recommendations for this subgroup of patients." — due 30 April 2032

Two competitive observations. First, the EU obligation shares the same front-line-population asymmetry as the FDA PMR: the confirming study is explicitly in patients "requiring first-line systemic therapy," while the authorised indication is post-progression. Second, the under-2-years PK obligation is a labeled soft spot — the EU authorised down to 6 months of age on acknowledged incomplete PK in the youngest patients. [FLAG: these obligations are quoted from the CHMP Assessment Report §1.9.7, which reproduces the Annex II content. If Annex II verbatim text is needed for a regulatory filing, retrieve the EPAR Product Information Annex II PDF itself.]

3.5 EU evidence base — the FDA and EMA figures are the SAME estimate

The prior draft asserted a "76 (FDA) vs 77 (EMA)" denominator difference. This is incorrect and has been removed. The CHMP Assessment Report states explicitly [24]:

"As of the data cutoff date (10 May 2024), 77 patients were enrolled and treated in Arm 1 of study FIREFLY-1 (ITT population). However, the efficacy analysis is based on 76 patients with measurable disease at baseline per RAPNO-LGG criteria."

The EPAR's primary RAPNO table (RAPNO Evaluable Analysis Set, Arm 1, N=76) reports: CR 0; PR 29 (38.2%); MR 11 (14.5%); SD 22 (28.9%); PD 13 (17.1%); NE 1 (1.3%); 40 confirmed responders; ORR 52.6% (95% CI 40.8, 64.2); median DoR 18.0 months (12.0, 22.8); clinical benefit rate 57.9% [24].

These are the identical figures in the FDA 08/2025 label (40 responders = 29 PR + 11 MR; median DoR 18 mo, 95% CI 12.0–22.8). 40/76 = 52.6%; FDA rounds to 53% and rounds the CI to (41, 64). The difference is rounding convention only — not a different denominator, not an independent estimate, and not a different data cut.

[FLAG: the EMA public news item [6] describes "an uncontrolled, open-label, phase 2 study with 77 patients," of whom "forty patients (52.6%) achieved a response." That sentence is internally inconsistent — 40/77 is 51.9%; the 52.6% figure is 40/76. This is EMA's own imprecision in a lay-facing news item. Cite the CHMP Assessment Report [24], not the news item, for any denominator. Never present 52.6% and 53% as two corroborating estimates.]

3.6 EU HTA — first-ever Joint Clinical Assessment (report now read)

Tovorafenib is the first medicine to complete a Joint Clinical Assessment under the EU HTA Regulation (EU) 2021/2282. Procedural chronology [25]: endorsed by the Member State Coordination Group on 30 April 2026 (Art. 12(2)); European Commission procedural review 19 May 2026 (Art. 28(d)); published 9 June 2026 [11]. Assessor: NCPE, Ireland (Emer Fogarty). Co-assessor: IQWiG, Germany (Beate Wieseler) — the prior draft's attribution is confirmed [25]. Ipsen publicly welcomed the publication [18].

Substance — this is the most competitively significant finding in the workstream. The JCA scope defined 8 PICO questions across 3 populations. Comparator results were submitted for only two (PICO 5 and PICO 7); PICO 7 was excluded by the assessors. Data were included in the JCA report for exactly one of eight PICO questions [25]:

Population	PICO	Comparator	Results submitted	Included in JCA
1 — full claimed indication	1–4	individualised treatment; carboplatin+vincristine; vinblastine; etc.	no ("No comparator data available")	no
2 — BRAF V600E, >1 yr	5	dabrafenib + trametinib	yes	yes
2 — BRAF V600E, >1 yr	6	individualised treatment	no	no
3 — BRAF fusion/rearr. or V600 non-E	7	trametinib	yes	no
3	8	individualised treatment	no	no

The assessors also refused to present the RANO-LGG ORR analysis at all, stating verbatim: "Due to the clear bias arising from the methodological flaws and inappropriate analysis methods used by the HTD, the results of the analysis of ORR based on RANO-LGG criteria are not presented in the assessment" [25] (HTD = health technology developer, i.e. the sponsor). The report records several further points where "the assessors disagree with this conclusion" of the sponsor.

The single included comparison (PICO 5, unanchored matching-adjusted indirect comparison of FIREFLY-1 vs. Bouffet 2023) showed no statistically significant difference on any outcome, on a vanishingly small effective sample size [25]:

• Effective sample size (ESS) after MAIC weighting: 5.81–6.64 in the base case (from an already-small V600E subgroup).
• 6-month PFS (RANO-LGG, IRC): RR 1.09 (0.92, 1.30) base case.
• 12-month PFS (RANO-LGG, IRC): RR 0.92 (0.61, 1.38) base case.
• Objective response, CR+PR by RANO-HGG, INV-assessed: OR 0.56 (0.08, 3.84), p=0.551.
• Objective response, CR+PR by RANO-HGG, IRC-assessed: OR 7.26 (0.98, 53.68), p=0.052.
• OS, generic HRQoL, disease-specific HRQoL, and disease symptoms: "Relative effectiveness results not available."

Note that the ORR point estimate flips direction (OR 0.56 vs OR 7.26) purely on whether the assessment is investigator- or IRC-read — a direct illustration of the criterion/reader sensitivity flagged in §2.6.

Read-across for Fore: the first EU JCA established that (i) a single-arm registrational package plus an unanchored MAIC yields no usable comparative evidence for 7 of 8 PICOs; (ii) assessors will exclude analyses outright for methodological inadequacy rather than present them with caveats; (iii) OS and HRQoL gaps are recorded explicitly as absent. Any Fore EU filing built on a single-arm CNS dataset should assume this treatment.

[FLAG: the JCA content above was extracted from the 154-page report PDF programmatically. The quantitative table values (ESS, RRs, ORs) were read from a text extraction of a complex multi-column table and should be re-checked against the published PDF by a human before any external or investor use. The report's own framing — a JCA presents relative effectiveness and does NOT issue a value judgment or a reimbursement recommendation; national HTA bodies do that — must be preserved: do not characterise the JCA as a "negative opinion." Source PDF: https://health.ec.europa.eu/publications/joint-clinical-assessment-report-tovorafenib-ojemda_en]

---

4 · Japan — PMDA

No PMDA approval, and no PMDA filing, was located as of the 31 Aug 2026 data cut. The only Japan-specific regulatory activity identified is an early-phase trial: NCT07441707, "A Study to Assess a Medicine Called Tovorafenib in Japanese Children and Young Adults With Brain Tumours" — Phase 1, RECRUITING, estimated enrolment 6, start 3 March 2026, primary completion 31 July 2030 [19].

A Japanese phase 1 bridging study starting in 2026 with a 2030 primary completion implies Japan approval is years away and is consistent with PMDA requiring Japanese PK/safety data before accepting the global package. [FLAG: that inference about PMDA's requirements is not sourced to PMDA.]

[FLAG: "no approval located" is a negative search result, not proof of absence. Verify directly against the PMDA approved-products database (pmda.go.jp) and Ipsen's Japan/Asia regulatory disclosures before stating in any external document that tovorafenib is unapproved in Japan.]

---

5 · Other regions

• Canada (Health Canada): APPROVED — see §1. Notice of Compliance with conditions (NOC/c), 100 mg tablets and 300 mg/bottle oral suspension (25 mg/mL reconstituted); indication as monotherapy in patients ≥6 months with relapsed or refractory pLGG harbouring a BRAF fusion or rearrangement, or BRAF V600 mutation, who have received one or more prior systemic therapies. Authorised under a Letter of Undertaking signed by Ipsen Biopharmaceuticals Canada Inc., including a commitment to provide confirmatory data from the phase 3 randomised trial of tovorafenib vs. SoC chemotherapy (i.e. FIREFLY-2). Reported in the August 2026 Health Product InfoWatch; qualifying notice control number 301603; added to the Prescription Drug List 27 Aug 2026 [26]. [FLAG: exact NOC date and DINs unverified — canada.ca refused automated fetch (HTTP 403). Confirm in the Health Canada NOC Database before external use.]
• United Kingdom (MHRA): no marketing authorisation located, and a targeted search at the data cut returned nothing. A secondary source stated that as of June 2024 no MAA had been submitted to MHRA, and that the International Recognition Procedure (IRP), effective 1 January 2024, would allow MHRA to rely on the EMA or FDA decision [20]. Given the EC decision of 20 Apr 2026, an IRP route is available. [FLAG: MHRA status is unverified against gov.uk's "Marketing authorisations granted" listings or the MHRA Products database — check both before asserting anything about UK availability.]
• UAE, Switzerland, Russia, Taiwan, South Korea, Australia: a single patient-facing commercial blog claimed approval in the UAE and orphan-drug designation in Russia, Switzerland, Taiwan, Japan, South Korea and Australia [20][21]. [FLAG: this source is a medicines-sourcing company blog, not a regulator or the sponsor. NONE of these claims are verified against Swissmedic, TGA, MFDS, TFDA, or UAE MOHAP primary records. Do NOT reproduce them in the dossier without primary confirmation — treat as a research lead only.] Partial counter-evidence: Ipsen's H1 2026 results describe Ojemda's first sales as "mainly related to specialized access schemes in Rest of World and launch in Germany" [27] — i.e. named-patient/early-access routes rather than marketing authorisations, which is what one would expect if the blog's "approval" claims are overstated.

---

6 · Corporate overlay affecting who holds the regulatory assets

• Ipsen holds ex-US rights, licensed from Day One in 2024 [8][17], is the EU marketing authorisation holder [5], and — via Ipsen Biopharmaceuticals Canada Inc. — the Canadian NOC/c holder [26].
• Servier completed its acquisition of Day One Biopharmaceuticals on 23 April 2026 — $21.50/share, ~$2.5 bn total equity value — thereby acquiring the US NDAs and the US commercial franchise [22]. Day One shares ceased trading on Nasdaq after the close on 22 April 2026 [22].
• Commercial scale so far: Ipsen reported Ojemda H1 2026 sales of €14.5 m (vs €0.9 m H1 2025), described as first sales from specialized access schemes in Rest of World plus the German launch [27]. This is a launch-stage asset outside the US.

The result: US regulatory obligations (the AA PMRs, the 180-day reports, the conversion decision) now sit with Servier; EU conditional-MA obligations and the Canadian Letter of Undertaking sit with Ipsen; all depend on the same trial, LOGGIC/FIREFLY-2, whose lead sponsor of record remains Day One Biopharmaceuticals [13]. Servier co-announced FIREFLY-2 enrollment completion on 8 May 2026 [14].

[FLAG: whether the NDA holder of record has been formally transferred to a Servier entity at FDA is NOT verified — both the openFDA Drugs@FDA record and the Orange Book products file still list "DAY ONE BIOPHARMS / DAY ONE BIOPHARMACEUTICALS INC" as applicant as of the 31 Aug 2026 data cut [7][23]. Describe Servier as the owner of Day One, not as the NDA holder, until confirmed.]

---

7 · Approved indication vs. promotional headline

Headline (Ipsen, 22 Apr 2026): "Ojemda® approved in the European Union as the first targeted therapy in relapsed or refractory pediatric low-grade glioma regardless of BRAF alteration" [8].

Approved EU indication: monotherapy in pLGG "harbouring a BRAF fusion or rearrangement, or BRAF V600 mutation, who have progressed after one or more prior systemic therapies" [5][8].

Two readings: 1. Charitable / likely intended: "regardless of which BRAF alteration" — i.e. active across fusion/rearrangement and V600, in contrast to V600E-restricted regimens. This is defensible against the label. 2. Literal: "regardless of BRAF alteration" reads as BRAF-alteration-agnostic, which the indication does not support — a confirmed BRAF fusion, rearrangement, or V600 mutation is a mandatory eligibility criterion in all three approved regions, and in the US a companion diagnostic now exists to establish it [4][15].

For Fore's purposes: the approved territory is BRAF-altered, prior-treated pLGG in the US, EU and Canada. Any competitive analysis that treats tovorafenib as label-approved in BRAF-wild-type disease, in front-line, or outside pLGG is reading the press release, not the label. The Ipsen headline also uses "relapsed or refractory" — FDA's phrase — where the EU indication actually reads "progressed after one or more prior systemic therapies" [5].

---

8 · Regulatory-competitive read for Fore

(Assessments in this section are analytical inferences drawn from the cited primary documents, not statements by any regulator.)

1. Tovorafenib is not fully approved in any region. US = accelerated approval, unconverted; EU = one-year conditional MA; Canada = NOC/c. All three hang on LOGGIC/FIREFLY-2 [1][5][13][26]. The addition of Canada strengthens rather than weakens this point: three regulators independently declined to grant unconditional approval on the same single-arm package.
2. The verification timeline is long. FDA's committed interim report is 04/2028 and final report 04/2032; the EU specific obligations are due 30 April 2032; registry primary completion is 06/2027 and study completion 06/2031 [1][13][24]. Conversion is unlikely before 2028 on the sponsor's own committed schedule.
3. The efficacy dataset is frozen at a 10 May 2024 data cut [24] and is ~27 months old at this dossier's data cut. Neither the FDA 08/2025 supplement nor the EU approval introduced newer efficacy data. The next genuine data event is the FIREFLY-1 PMC 4626-9 final report (committed 12/2027) or FIREFLY-2.
4. The confirmatory trial is in a different line of therapy than the approved indication (front-line vs relapsed/refractory) [4][13][24] — a genuine regulatory asymmetry present in both the FDA PMR and the EU specific obligation, and simultaneously the vehicle for a front-line label expansion that would materially enlarge the competitive footprint.
5. Pediatric growth and gonadal toxicity are FDA-designated serious risks carrying required ≥5-year monitoring trials (PMRs 4626-4, 4626-5) [1], and the labeled growth estimate rose from 15% to 46% of 133 patients ≤18 y (Grade 3 5% → Grade ≥3 35%) at the 08/2025 revision [3][4]. In a chronically-dosed pediatric population this is the most differentiating labeled liability — but see the matched-comparison FLAG in §2.6(b).
6. Carcinogenicity is an open required question — mouse and rat studies were both mandated, the rat study under Special Protocol Assessment with a final report not due until 05/2028 [1].
7. The label carries a DDI burden: avoid moderate/strong CYP2C8 inhibitors and inducers, avoid certain CYP3A substrates, avoid hormonal contraceptives (tovorafenib can render them ineffective; females must use nonhormonal contraception) [4] — a real-world constraint in the adolescent/young-adult segment, compounded by the labeled embryo-fetal and gonadal risks.
8. Diagnostic access is established for tovorafenib (FoundationOne CDx, 17 Jan 2025) [15][16] — a competitor entering later faces an established tissue-NGS testing pathway that identifies the same biomarker population. [FLAG: "established" ≠ "solved"; real-world pediatric tissue-NGS access, turnaround and reimbursement were not assessed in this workstream.]
9. EU market access precedent is set, and it is a cautionary one. Tovorafenib is the first product through the EU HTA Regulation's JCA [11][12][25]. The assessment included comparative data for 1 of 8 PICO questions, excluded the sponsor's RANO-LGG ORR analysis for methodological inadequacy, produced no OS or HRQoL comparative results, and returned an unanchored MAIC with an effective sample size under 7 and no statistically significant differences. This is directly precedent-setting for any Fore EU filing resting on single-arm CNS data.
10. Response-criterion choice is worth ~18 ORR points on the identical patients (RAPNO 52.6% / RANO-LGG 53.9% / RANO-HGG 71.0%) [24]. This is the most exploitable — and most dangerous — comparability axis in the whole competitive picture.

---

9 · Plixorafenib comparison column — regulatory axes (TEMPLATE — Fore team to complete)

Per the guardrails, no plixorafenib data is researched, inferred, or supplied here. No comparison figures were provided to this workstream. The rows below are the regulatory axes on which the head-to-head should be built; the plixorafenib column must be completed by the Fore team from cleared internal data and each entry labeled "Fore-internal — verify."

Regulatory axis	Tovorafenib (public, cited)	Plixorafenib (Fore-internal — verify)
US regulatory status / instrument	Accelerated approval, 23 Apr 2024, unconverted [1][2][7]	[to be completed by Fore]
US indication verbatim	R/R pediatric LGG, BRAF fusion/rearrangement or V600, ≥6 months [4]	[to be completed by Fore]
Tumor types / histologies covered	pLGG only	[to be completed by Fore]
Line of therapy	≥2L (relapsed or refractory) [4]	[to be completed by Fore]
Age range	≥6 months (studied 2–21 y in Arm 1) [4]	[to be completed by Fore]
Biomarker requirement + CDx	BRAF fusion/rearrangement or V600; FoundationOne CDx approved 17 Jan 2025 [15]	[to be completed by Fore]
Dosing schedule / formulations	380 mg/m² once weekly, max 600 mg; 100 mg tablet (BSA ≥0.90 m²) + 25 mg/mL suspension (BSA 0.30 m² and up) [4]	[to be completed by Fore]
Response criterion accepted by FDA	RAPNO-LGG by BICR (RANO-LGG 2011 supportive) [4][9]	[to be completed by Fore — must be RANO/RAPNO/iRANO, not RECIST]
Approval-basis N + ORR + DoR + data cut	N=76 (of 77 treated); ORR 53% (95% CI 41–64), 29 PR + 11 MR (MR = 27.5% of responders); median DoR 18 mo (12.0, 22.8); data cut 10 May 2024 [4][24]	[to be completed by Fore — state N, criterion, MR handling, data cut]
Confirmatory obligation + timeline	PMR 4626-1 / FIREFLY-2 NCT05566795; FDA interim 04/2028, final 04/2032; EU specific obligation due 30 Apr 2032 [1][13][24]	[to be completed by Fore]
Boxed warning / key labeled warnings	No boxed warning; hemorrhage, skin/photosensitivity, hepatotoxicity, growth, embryo-fetal, NF1 tumor growth; Contraindications: None [4]	[to be completed by Fore]
Pediatric long-term safety PMRs	Growth (4626-4) and gonadal toxicity (4626-5), ≥5 yr monitoring [1]	[to be completed by Fore]
US exclusivity / patents	NCE to 23 Apr 2029; ODE-478 to 23 Apr 2031; patents 8,293,752 (4 Aug 2031) and 10,426,782 (23 Jun 2035) [23]	[to be completed by Fore]
EU status / instrument	Conditional MA (one-year, renewable), EC decision 20 Apr 2026; MAH Ipsen Pharma [5][8]	[to be completed by Fore]
EU indication verbatim	monotherapy, pLGG, BRAF fusion/rearrangement or V600, progressed after ≥1 prior systemic therapy [5]	[to be completed by Fore]
EU HTA / JCA outcome	First-ever JCA; comparative data included for 1 of 8 PICOs; MAIC ESS <7; no OS/HRQoL results [25]	[to be completed by Fore]
Canada	NOC/c, Ipsen Biopharmaceuticals Canada Inc.; Letter of Undertaking → FIREFLY-2 [26]	[to be completed by Fore]
Japan / other regions	No PMDA approval located; Japanese Ph1 NCT07441707 ongoing [19]	[to be completed by Fore]
US designations	Breakthrough, Rare Pediatric Disease, Orphan (malignant glioma); PRV granted on NDA 217700, denied on 218033 [2][10]	[to be completed by Fore]

---

10 · Open flags carried from this workstream

1. Drug identity was inferred, not passed — confirm tovorafenib is the intended target before circulation.
2. NDA submission date (31 Aug 2023) not corroborated in the retrieved sources; not asserted here.
3. PMR/PMC completion status unverified for 4626-2 through 4626-12; several milestones have passed. Query FDA's PMR/PMC database — slippage is a live datapoint.
4. Growth-AE 15% → 46%: not a matched comparison (no 2024 denominator; Grade 3 vs Grade ≥3 bands). Characterise as "the labeled estimate rose at the 08/2025 revision," not "toxicity worsened."
5. The §5.4 growth revision is not the only warning change — §5.5 dropped "nonhormonal" from the male-partner contraception advice without a Recent Major Changes entry.
6. JCA quantitative values were machine-extracted from a complex PDF table — human re-check required before external use. Preserve the framing that a JCA issues no value judgment.
7. Health Canada NOC exact date, DINs and Letter of Undertaking terms unverified (canada.ca 403 to automated fetch).
8. MHRA, Japan/PMDA, Switzerland, Australia, Korea, Taiwan, UAE, Russia unverified against primary regulators; the UAE-approval and multi-country orphan-designation claims rest on a single commercial blog and must not be reproduced. Ipsen's own H1 2026 language ("specialized access schemes") suggests these are access routes, not approvals.
9. NDA-holder-of-record transfer to Servier not confirmed — openFDA and the Orange Book both still list Day One at the data cut.
10. Orange Book is a monthly snapshot — re-pull before any IP/LOE statement; no pediatric exclusivity is currently listed.
11. MR inclusion and criterion choice: 11/40 responders are minor responses; ORR moves ~18 points across RAPNO / RANO-LGG / RANO-HGG on the same 76 patients. Any cross-drug ORR comparison must state both or it is invalid.
12. The FIREFLY-2 line-of-therapy observation, and all §8 "competitive read" items, are inferences — not FDA, EMA, HTACG or Health Canada statements. Do not attribute.

---

References

1. FDA. NDA 218033 Accelerated Approval Letter, Ojemda (tovorafenib) for Oral Suspension. Signed Martha B. Donoghue, M.D., 23 April 2024. Reference ID 5368915. Local copy: C:\Users\Owner\AppData\Local\Temp\claude\C--Users-Owner-Documents-Hugh-Context-Folder-Fore\06918b60-6b30-474c-84c4-6907c80d5eef\scratchpad\tovo_ltr.pdf (extracted text: ltr.txt).
2. FDA. NDA 217700 Accelerated Approval Letter, Ojemda (tovorafenib) Tablet. Signed Martha B. Donoghue, M.D., 23 April 2024. Reference ID 5368906. Local copy: …\scratchpad\tovo_tab_ltr.pdf (extracted text: tabltr.txt).
3. FDA. OJEMDA (tovorafenib) — Prescribing Information, rev. 4/2024. Reference ID 5368915. https://www.accessdata.fda.gov/drugsatfda_docs/label/2024/217700s000lbl.pdf — local copy …\scratchpad\tovo_lbl2024.pdf (text: tovo_lbl2024.txt).
4. FDA. OJEMDA (tovorafenib) — Prescribing Information, rev. 08/2025 (217700 s002/s003/s004). Reference ID 5649878. https://www.accessdata.fda.gov/drugsatfda_docs/label/2025/217700s002s003s004lbl.pdf — local copy …\scratchpad\lbl2025.pdf (text: lbl2025.txt).
5. EMA. Ojemda (tovorafenib) — EPAR medicine overview. https://www.ema.europa.eu/en/medicines/human/EPAR/ojemda — retrieved 31 Aug 2026.
6. EMA. "New medicine to treat paediatric low-grade glioma," news item, 27 February 2026. https://www.ema.europa.eu/en/news/new-medicine-treat-paediatric-low-grade-glioma — lay-facing; contains an internally inconsistent 40/77 vs 52.6% statement, see §3.5.
7. FDA Drugs@FDA, via openFDA drug/drugsfda endpoint, retrieved 31 August 2026 (local: …\scratchpad\da.json). Sponsor of record for both NDAs: DAY ONE BIOPHARMS. Submission histories as tabulated in §2.6.
8. Ipsen. "Ojemda® approved in the European Union…regardless of BRAF alteration," press release, 22 April 2026. https://www.ipsen.com/press-release/ojemda-approved-in-the-european-union-as-the-first-targeted-therapy-in-relapsed-or-refractory-pediatric-low-grade-glioma-regardless-of-braf-alteration-3278647/
9. FDA. "FDA grants accelerated approval to tovorafenib…," 23 April 2024. https://www.fda.gov/drugs/resources-information-approved-drugs/fda-grants-accelerated-approval-tovorafenib-patients-relapsed-or-refractory-braf-altered-pediatric
10. Day One Biopharmaceuticals. "Day One's OJEMDA™ (tovorafenib) Receives US FDA Accelerated Approval…," 23 April 2024. https://ir.dayonebio.com/news-releases/news-release-details/day-ones-ojemdatm-tovorafenib-receives-us-fda-accelerated
11. European Commission, DG SANTE. "Joint clinical assessment report published on tovorafenib (Ojemda)," 9 June 2026. https://health.ec.europa.eu/latest-updates/joint-clinical-assessment-report-published-tovorafenib-ojemda-2026-06-09_en · landing page: https://health.ec.europa.eu/publications/joint-clinical-assessment-report-tovorafenib-ojemda_en
12. National Centre for Pharmacoeconomics (Ireland). "Major milestone under the EU HTA Regulation: First Joint clinical assessment of benefit complete." https://www.ncpe.ie/major-milestone-under-the-eu-hta-regulation-first-joint-clinical-assessment-of-benefit-complete/
13. ClinicalTrials.gov. NCT05566795 (LOGGIC/FIREFLY-2). Phase 3; ACTIVE_NOT_RECRUITING; N=418 actual; start 2023-02-27; primary completion 2027-06; completion 2031-06; lead sponsor Day One Biopharmaceuticals; primary outcome ORR vs SoC chemotherapy. Retrieved via CTG API v2, 31 Aug 2026 (last update posted 2026-05-14); local: …\scratchpad\ff2.json.
14. Day One Biopharmaceuticals / Servier. "Day One and Servier Complete Enrollment in Pivotal Phase 3 FIREFLY-2 Trial…," 8 May 2026. https://www.prnewswire.com/news-releases/day-one-and-servier-complete-enrollment-in-pivotal-phase-3-firefly-2-trial-of-tovorafenib-as-a-front-line-treatment-for-pediatric-low-grade-glioma-plgg-302766277.html
15. Foundation Medicine. "U.S. FDA Approves FoundationOne®CDx as a Companion Diagnostic for OJEMDA™ (tovorafenib)…," 17 January 2025. https://www.foundationmedicine.com/press-release/fda-approval-foundationone-cdx-ojemda
16. Targeted Oncology / OncLive / CancerNetwork coverage of the 17 Jan 2025 FoundationOne CDx approval. https://www.targetedonc.com/view/fda-approves-foundationone-cdx-as-companion-diagnostic-for-tovorafenib-in-pediatric-low-grade-glioma
17. Ipsen. "Ipsen and Day One enter into exclusive ex-U.S. licensing agreement…," 2024. https://www.ipsen.com/press-release/ipsen-and-day-one-enter-into-exclusive-ex-u-s-licensing-agreement-to-commercialize-tovorafenib-for-the-most-common-childhood-brain-tumor-2918482/
18. Ipsen. "Ipsen welcomes the European Union's publication of the Joint Clinical Assessment for Ojemda® (tovorafenib)." https://www.ipsen.com/update/ipsen-welcomes-the-european-unions-publication-of-the-joint-clinical-assessment-for-ojemda-tovorafenib-3309159/
19. ClinicalTrials.gov. NCT07441707, tovorafenib in Japanese children and young adults with brain tumours. Phase 1; RECRUITING; N=6 estimated; start 2026-03-03; primary completion 2030-07-31. Retrieved 31 Aug 2026 (local: …\scratchpad\c.json).
20. Everyone.org. "Tovorafenib's approval in Europe and around the world." https://everyone.org/blog/tovorafenib-approval-in-europe — secondary, commercial, non-regulatory source; cited only as the origin of the unverified MHRA-status claim.
21. Everyone.org, ibid. — origin of the unverified UAE-approval and Russia/Switzerland/Taiwan/Japan/South Korea/Australia orphan-designation claims. Not confirmed against any primary regulator.
22. Servier. "Servier completes the acquisition of Day One Biopharmaceuticals," 23 April 2026. https://servier.com/en/newsroom/servier-completes-the-acquisition-of-day-one-biopharmaceuticals/ ($21.50/share; ~$2.5 bn equity value; Nasdaq trading ceased after close 22 Apr 2026). Related SEC filings: Day One Biopharmaceuticals SC TO-T / SC 14D9, CIK 0001845337, FY2026.
23. FDA Orange Book bulk data files (products.txt, exclusivity.txt, patent.txt), retrieved 31 August 2026; local: …\scratchpad\ob\. NDA 217700/218033: NCE exp. 23 Apr 2029; ODE-478 exp. 23 Apr 2031; patents 8,293,752 (exp. 4 Aug 2031, drug substance + drug product) and 10,426,782 (exp. 23 Jun 2035, drug product), both submitted 8 May 2024. (Supersedes the prior draft's "not verified / HTTP 404" note — the 404 affected only the HTML product page.)
24. EMA / CHMP. Ojemda — CHMP Assessment Report, EMA/67438/2026, procedure EMEA/H/C/006140/0000, adopted 26 February 2026. https://www.ema.europa.eu/en/documents/assessment-report/ojemda-epar-public-assessment-report_en.pdf — local copy …\scratchpad\ojemda_epar.pdf (text: ojemda_epar.txt). Source for: FIREFLY-1 data cutoff 10 May 2024; the 77-enrolled / 76-evaluable denominator explanation; the RAPNO primary efficacy table (ORR 52.6%, median DoR 18.0 mo); the RAPNO 52.6% / RANO-LGG 53.9% / RANO-HGG 71.0% criterion comparison; and the §1.9.7 Article 14-a specific obligations, both due 30 April 2032.
25. Member State Coordination Group on Health Technology Assessment. Joint Clinical Assessment Report of Tovorafenib, Version 1.0, European Union, Brussels, 2026. Endorsed by the Coordination Group 30 April 2026 (Art. 12(2), Reg. (EU) 2021/2282); EC procedural review 19 May 2026 (Art. 28(d)); published 9 June 2026. Assessor: NCPE Ireland (Emer Fogarty); Co-assessor: IQWiG Germany (Beate Wieseler). https://health.ec.europa.eu/publications/joint-clinical-assessment-report-tovorafenib-ojemda_en (report PDF retrieved and extracted 31 Aug 2026).
26. Health Canada. Health Product InfoWatch, August 2026 (Ojemda NOC/c notice) https://www.canada.ca/en/health-canada/services/drugs-health-products/medeffect-canada/health-product-infowatch/august-2026.html · Qualifying notice for Ojemda, control number 301603 https://www.canada.ca/en/health-canada/services/drugs-health-products/drug-products/notice-compliance/conditions/qualifying-notice-ojemda-301603.html · Prescription Drug List addition notice, 2026-08-27. Accessed via search result summaries only — canada.ca returned HTTP 403 to direct automated fetch; exact NOC date unverified.
27. Ipsen. H1 2026 results (Ojemda net sales €14.5 m H1 2026 vs €0.9 m H1 2025; "first sales, mainly related to specialized access schemes in Rest of World and launch in Germany"). Local copy: …\scratchpad\ipsen_h1.pdf (text: ipsen_h1.txt).

Commercial, Competitive Position & IP — Tovorafenib (OJEMDA®)

Data cut: 31 August 2026. All figures below are public-domain (SEC filings, FDA Orange Book August 2026 edition, Federal Register, EMA EPAR, company press releases, state drug-price-transparency filings, peer-reviewed and conference abstracts). Derived arithmetic is labeled as such. Analyst judgment is labeled [CI judgment] and is not a sourced finding. No plixorafenib data appears in this workstream; the comparison column in §9 is an unfilled template for the Fore team.

Single most important change since the last CI cycle: tovorafenib is no longer controlled by a small-cap US biotech. Servier acquired Day One Biopharmaceuticals outright on 23 April 2026 for $21.50/share cash (~$2.5B). The asset now sits inside a foundation-owned French pharma alongside VORANIGO (vorasidenib), creating a two-product glioma franchise with a shared neuro-oncology call point — and, as a side effect, US sales of OJEMDA are no longer separately disclosed to the public.

---

1. Ownership, developer/marketer chain, and deal economics

Stage	Party	Terms / date	Source
Molecule origin	Millennium Pharmaceuticals (Takeda) + Sunesis Pharmaceuticals — as TAK-580 / MLN2480	Patent priority 29 Jun 2007	US 8,293,752 front page
Asset-out	Millennium/Takeda → Day One (subsidiary)	16 Dec 2019; $1.0M cash + 9,857,143 Series A preferred shares (FV $9.9M) — i.e. ~$10.9M total consideration	Day One FY2025 10-K, Note 7
Underlying license	Viracta Therapeutics (assigned by Takeda); royalty stream sold to XOMA (US) LLC 3 Dec 2024	Tiered mid-single-digit % royalties on net sales; up to $34.0M remaining development/regulatory milestones; $2.0M milestone paid Dec 2025; $9.0M approval milestone paid May 2024	Day One FY2025 10-K, Note 7
Ex-US rights	Ipsen (exclusive, all territories outside US)	25 Jul 2024. ~$111M upfront ($71M cash + $40M equity at premium); up to ~$375M launch/sales milestones (as re-stated at FX on 31 Dec 2025; ~$350M at signing); tiered double-digit royalties starting mid-teens %; royalty term ≥10 yrs from first commercial sale per country; Ipsen may terminate for convenience after 23 Jul 2026 on 6 months' notice	Ipsen/Day One PR 25 Jul 2024; 10-K
Day One inbound M&A	Mersana Therapeutics acquired by Day One	Tender offer expired 5 Jan 2026; statutory merger completed 6 Jan 2026. $25.00/share cash + one non-tradable CVR worth up to $30.25/share; ~$129M equity value at close, up to ~$285M including CVRs. Brought in Emi-Le (B7-H4-directed ADC, Phase 1 in adenoid cystic carcinoma)	Day One/Mersana PR 13 Nov 2025 and 6 Jan 2026
US + global control	Servier Pharmaceuticals LLC (Servier S.A.S. guarantor)	Merger agreement 6 Mar 2026; tender offer opened 26 Mar 2026; 88,180,910 shares (85.34%) tendered; closed 23 Apr 2026 via DGCL §251(h). $21.50/share, ~$2.5B equity value; ~68% premium to 5 Mar 2026 close, ~86% to 1-month VWAP. No CVR	Day One 8-K filed 23 Apr 2026; Servier PR 6 Mar 2026

[FLAG: the ~$2.5B equity value is a fully diluted figure (options, RSUs, ESPP) from the Servier release. It does not reconcile against the tender arithmetic — 88,180,910 shares at 85.34% implies ~103.3M shares outstanding, and 103.3M × $21.50 ≈ $2.22B. Do not present the two numbers as if they should tie.]

[FLAG: the ~68% / ~86% premium figures were not independently verified in this pass. Verify against the Schedule TO or the 14D-9 before external use.]

[FLAG — do not read as a contradiction: §1 records the royalty stream sold to XOMA in Dec 2024 and §6 lists XOMA as a current assignee of US 8,293,752, yet §7 lists Viracta as PTE co-applicant. The reconciliation is chronological — the PTE applications were filed in June 2024, six months before the XOMA transaction, so Viracta is the applicant of record. Confirm with counsel that the post-transaction assignee change does not affect the pending extension.]

Servier's ~$2.5B therefore purchased more than tovorafenib. It also acquired Emi-Le and the Mersana ADC platform, brought in three months earlier. Any attempt to impute a standalone valuation to tovorafenib from the headline deal price must first net out the Mersana consideration.

Ancillary monetization: Day One sold its Rare Pediatric Disease Priority Review Voucher for $108.0M on 29 May 2024 (undisclosed buyer); $8.1M of proceeds went to Viracta to extinguish PRV obligations, after a $5.0M March 2024 payment to reduce those obligations. [FLAG: PRV sale price and the Viracta payment amounts trace to the 10-K only and were not independently corroborated in this pass.]

CI read [CI judgment]: the total inbound cost of the molecule to Day One was ~$11M plus milestones; the PRV alone returned ~10× that. The economics of this asset were driven as much by the pediatric-rare-disease regulatory toolkit as by the drug.

Disclosure consequence (important for future tracking): Day One filed Form 25-NSE (23 Apr 2026) and Form 15-12G (4 May 2026), ending Exchange Act reporting. Q4 2025 is the last separately reported US OJEMDA revenue. Servier is governed by the Fondation Internationale de Recherche Servier, is not listed, and reports only annual group/therapy-area results (FY ends 30 September; FY2024/25 group revenue €6.9B, oncology €2.21B = ~32% of group (2.21 ÷ 6.9 = 32.0%; the previously stated 32.2% was a rounding error)). Going forward, the only product-level public number is Ipsen's quarterly ex-US line.

---

2. Pricing

Tovorafenib is priced flat per package regardless of dose or patient body-surface area — the price does not rise as a child grows.

Effective	Tablet package (100 mg; 16-, 20- or 24-count) = 28-day supply	Oral suspension (single-use bottle = 7-day supply)	Implied annual WAC (×13 packages/yr — my calculation)
7 May 2024 (launch)	$33,916	$8,479	~$440,900
1 Jan 2025	$35,272.64 (+4.0%)	$8,818.16	~$458,500
1 Jan 2026	$38,702.90 (+9.7% YoY; +14.1% cumulative since launch)	$9,675.73	~$503,100

Correction: the 2026 oral suspension WAC was previously recorded as "not separately captured at this cut." It is disclosed in the same Colorado prescriber notice already cited: $9,675.73. Internal consistency check: 4 bottles × $9,675.73 = $38,702.92, matching the tablet package price to two cents — the flat-per-28-days structure holds across both presentations in 2026.

NDCs: 82950-0001-16 / -20 / -24 (tablets); 82950-0012-01 (suspension). Sources: North Dakota Insurance Dept. drug-transparency filing (Day One, 3 May 2024); Day One Connecticut prescriber notice (data as of 01/01/2025, TOVO-US-0241); Day One Colorado prescriber notice (data as of 01/01/2026); Managed Healthcare Executive, "FDA Approves Ojemda for Children with Brain Tumors" (2024), reporting launch WAC of $33,916/28 days and annual cost "tops $440,000."

Manufacturer's stated rationale, verbatim from the ND filing: "Price balances a variety of factors, including the product's clinical value, the size of the patient population and the significant costs required to develop a drug."

Payer mix: Day One's chief commercial officer guided at launch to ~60% commercial / ~40% Medicaid for eligible patients (Managed Healthcare Executive, 2024). [FLAG: this is a launch-era management expectation reported secondhand, not an audited realized mix; no later company disclosure of realized payer mix was found.]

Implied gross-to-net (my calculation — DO NOT PUBLISH AS A FINDING): FY2025 net revenue $155.4M ÷ 4,635 prescriptions = $33,528 net per prescription against a 2025 WAC of $35,272.64 → ~95% net-to-gross, i.e. only ~5% GTN erosion.

[FLAG — CONTRADICTION, unresolved: a ~5% GTN erosion is arithmetically incompatible with the ~40% Medicaid payer mix stated immediately above. Medicaid statutory rebates alone are a minimum 23.1% of AMP for brand drugs, plus the CPI penalty, which at a 40% Medicaid mix would imply blended erosion of at least ~9–10%, before commercial rebates, 340B, copay assistance and distribution fees. At least one input is wrong: the launch-era payer-mix expectation, the assumption that one disclosed "prescription" equals one 28-day package, or the use of tablet WAC as the gross basis. A human with payer-analytics judgment must resolve this before any GTN figure enters the dossier.]

[FLAG: separately, the Q4-2025-only ratio ($52.8M ÷ 1,394 = $37,876/Rx) exceeds 2025 WAC, which is arithmetically impossible for pure net revenue per 28-day package — it likely reflects channel inventory build, reserve true-ups, or multi-package fills. Do not cite the quarterly ratio under any circumstances.]

---

3. US launch uptake and net sales trajectory

US net product revenue (Day One was the US marketer through 22 Apr 2026):

Period	Net product revenue	Note
Q2 2024	$8.2M	partial quarter; approval 23 Apr 2024
Q3 2024	$20.1M	first full quarter; +145% QoQ
Q4 2024	$28.9M (derived: $57.2M − $8.2M − $20.1M)
FY 2024	$57.2M
Q1 2025	$30.5M (derived: $102.6M 9M − $38.5M − $33.6M)
Q2 2025	$33.6M	+310% YoY
Q3 2025	$38.5M	+15% QoQ
Q4 2025	$52.8M	+37% QoQ
FY 2025	$155.4M	+172% YoY
FY 2026 guidance	$225–250M (US only)	~53% growth at midpoint; reaffirmed 24 Feb 2026, ~10 days before the Servier deal was announced

FY2025 total and the +172%/+53% growth rates independently verified against the 11 Jan 2026 preliminary-revenue release and the 24 Feb 2026 results release. All quarterly subtractions and growth rates re-derived and confirmed; the four 2025 quarters sum exactly to $155.4M.

Prescription volumes (the cleanest available demand proxy): 4,635 in FY2025, +181% vs 2024 (→ ~1,650 in 2024, derived); 1,394 in Q4 2025.

Patients on therapy (my calculation, flag as an estimate): at 28-day fills, Q4 2025's 1,394 prescriptions annualize to ~5,576 Rx/yr ÷ 13 = ~429 patients on therapy exiting 2025. Against Day One's own stated addressable pool of 2,000–3,000 relapsed/progressive/refractory pLGG patients per year at steady state, that is roughly 14–21% penetration after ~20 months on market. 2026 guidance at midpoint implies ~545 average patients on therapy. [FLAG: derived from disclosed Rx counts and a 28-day-fill assumption; the company never published a patients-on-therapy figure, and the same "one prescription = one 28-day package" assumption that is in doubt in §2 is load-bearing here. Treat as an order-of-magnitude estimate only.]

Cost structure at scale: FY2025 cost of product revenue $14.7M = 9.5% of net revenue (inventory + intangible amortization + third-party royalties). FY2025 R&D $148.1M, SG&A $120.6M, net loss $107.3M; cash/equivalents/ST investments $441.1M at 31 Dec 2025. [FLAG: the FY2025 R&D/SG&A/net-loss figures postdate the Mersana close (6 Jan 2026) only by days, so FY2025 opex does not yet carry the ADC program. FY2026 economics, had they been disclosed, would have.]

Ex-US trajectory (Ipsen — the only ongoing public product-level disclosure):

	H1 2026	H1 2025	Q2 2026	Q2 2025
Ojemda total	€14.5m	€0.9m	€9.2m	€0.5m
— Europe	€4.3m	€0.7m	€4.0m	€0.4m
— Rest of World	€10.2m	€0.2m	€5.2m	€0.2m
— North America	—	—	—	—

All four Ojemda totals independently verified against Ipsen's H1 2026 release (30 Jul 2026).

Ipsen's own commentary: "Ojemda: first sales, mainly related to specialized access schemes in Rest of World and launch in Germany." Ipsen's H1 2026 disclosure further indicates that as of July 2026 only German authorities had initiated an HTA, and only Germany had awarded reimbursement status — which Germany grants automatically to orphan-designated products. Ex-US is therefore still a named-patient/early-access business with a single meaningful EU launch market, and that market's reimbursement was administratively granted rather than negotiated on value. Ipsen upgraded full-year group guidance on the back of Iqirvo and Bylvay, not Ojemda.

Ex-US regulatory/access milestones: CHMP positive opinion 27 Feb 2026; European Commission conditional marketing authorisation, EMA EPAR authorisation date 20 April 2026 (Ipsen's release says 22 April), across EU-27 plus Iceland, Liechtenstein, Norway. MAH Ipsen Pharma; ATC L01EC04. EU Joint Clinical Assessment published 9 Jun 2026, which explicitly acknowledged "the absence of a clearly established clinical comparator for much of the pLGG patient population" — helpful framing for Ipsen in national P&R negotiations, but it also means no comparative-effectiveness anchor, which several EU HTA bodies penalize on price. [FLAG: EPAR 20 April vs Ipsen PR 22 April — likely decision date vs. notification date; immaterial but use "April 2026" in external documents.]

Authorised EU indication, verbatim from the EMA EPAR: "Monotherapy for the treatment of patients 6 months of age and older with paediatric low-grade glioma (LGG) harbouring a BRAF fusion or rearrangement, or BRAF V600 mutation, who have progressed after one or more prior systemic therapies."

[FLAG: Ipsen's EU headline reads "first targeted therapy in relapsed or refractory pediatric low-grade glioma regardless of BRAF alteration." Set against the authorised wording quoted above — which requires a BRAF fusion/rearrangement or BRAF V600 mutation — the headline is promotional shorthand for "regardless of which BRAF alteration," not a broader label. Never repeat the headline phrasing; quote the indication.]

---

4. Commercial model and access infrastructure

• Sales force: 18 representatives covering ~200 accounts that treat >90% of US pLGG patients (Day One FY2025 10-K). This is one of the most concentrated call points in oncology. [CI judgment] a footprint this small is why Servier can absorb the product with minimal incremental infrastructure. [FLAG: the 18-rep and ~200-account figures trace to a single unverified 10-K read; confirm directly in Item 1.]
• No owned manufacturing. Single-source CMOs: STA Pharmaceutical Hong Kong (API), Quotient Sciences–Philadelphia (drug product), Experic (packaging), Sharp Packaging Services. Single-source dependency is a disclosed supply risk. [FLAG: CMO list unverified outside the 10-K.]
• Companion diagnostic: FoundationOne®CDx approved as CDx on 17 January 2025 — nine months after drug approval. The first three quarters of launch ran with no FDA-approved CDx. [CI judgment] this is a plausible contributor to the modest 2024 ramp, but it is a hypothesis, not a demonstrated cause.
• Distribution via specialty distributors/specialty pharmacy; customer concentration is extreme — two customers = 97.6% of FY2025 revenue. [FLAG: the FY2024 10-K/A reports two customers at 94.3% of FY2024 net product revenue. The FY2025 figure of 97.6% was not independently verified in this pass — confirm before use.]

---

5. Market positioning within the class

What tovorafenib uniquely owns: it is the only FDA-approved agent for BRAF fusion/rearrangement-driven pLGG, and fusions are the clear majority of RAF-altered pLGG (predominantly KIAA1549::BRAF). It is also the only pLGG systemic therapy with once-weekly dosing, food-independent, available as tablet and oral suspension (an underrated advantage in a 6-month-and-older population).

[FLAG — denominator, unresolved: the previous draft asserted "fusions ~85–90% of RAF-altered pLGG" alongside "BRAF V600E only ~10–20%," which is unsourced and internally inconsistent (the two ranges sum to 95–110%). Published pLGG series report KIAA1549::BRAF fusion in roughly 30–40% of all pLGG (higher, ~65–75%, within pilocytic/pilomyxoid astrocytoma specifically) and BRAF V600E in roughly 7–15% of pLGG, which puts fusions at approximately 70–80% and V600E at 15–25% within the BRAF-altered subset. Do not use any of these ranges externally until the Fore team fixes one authoritative denominator with a named cohort and N — ideally the FIREFLY-1 molecular breakdown or a single named series. This figure will get quoted back in a board deck.]

Where it does not compete:

Agent	Sponsor	Status in pLGG	Overlap with tovorafenib
Dabrafenib + trametinib (Tafinlar®+Mekinist®)	Novartis	Full approval Mar 2023, pediatric ≥1 yr, BRAF V600E LGG requiring initial systemic therapy	Direct but narrow: BRAF V600E is the minority of RAF-altered pLGG (see denominator flag above). Critically, it sits first-line while OJEMDA's label is relapsed/refractory (2L+) — an asymmetry that caps OJEMDA until FIREFLY-2 reads out
Selumetinib (Koselugo®)	AstraZeneca	Approved for NF1-PN only; used off-label in pLGG	MEK inhibitors are active in both V600E and fusion-driven pLGG; none is FDA-approved as pLGG monotherapy
Mirdametinib (Gomekli®)	SpringWorks (a Merck KGaA company)	Approved for NF1-PN; SJ901 (NCT04923126) Phase 1/2 in pediatric/young-adult LGG ongoing — see corrected efficacy data below	[CI judgment] the most credible medium-term threat to OJEMDA's franchise: an approved, brain-penetrant, already-marketed pediatric MEK inhibitor with a live LGG registration path and a large-pharma owner
First-generation type I½ BRAF V600 inhibitors (vemurafenib/Zelboraf, encorafenib/Braftovi)	Genentech, Pfizer	Not approved in pLGG	Mechanistically disadvantaged in fusion-driven disease (paradoxical MAPK activation). Corrected from "type I RAF inhibitors" — these agents are ATP-competitive type I½ binders, and the type I½ vs type II distinction is the entire basis of the paradoxical-activation argument
Next-gen RAF pipeline: JZP815 (Jazz/Redx), naporafenib (Erasca/Novartis), NST-628 (Nested), PF-07799933/ARRY-440 (Pfizer), BDTX-4933 (Black Diamond)	—	Phase 1–3, adult solid tumors	No current pLGG overlap; relevant to the broader BRAF-altered CNS space

CORRECTED — mirdametinib SJ901 efficacy. The prior draft stated "63% ORR in patients with measurable tumors, median time to response 5.4 months." The 63% figure could not be traced to any retrievable source and appears to have been paired with a time-to-response value drawn from a different report. Replaced with the sourced primary data:

• N = 35 patients, enrolled June 2021 – December 2024; ages 2.4–21.9 yrs (median 8.4); MEKi-naive, recurrent/progressive pLGG with biopsy-proven MAPK pathway activation, BRAF V600 excluded.
• Response assessed by RAPNO-LGG criteria. 26 of 35 (74%) achieved an objective response (minor response or better): 2 CR, 3 major response, 10 PR, 11 minor response. Median time to response 5.4 months (range 1.7–24.1) among responders. Data cutoff May 2025. (Neuro-Oncology 2025;27(Suppl 5):v147, abstract CTP-08.)
• A later reported cut (23 September 2025) gives 87% minor-response-or-better, 55% PR-or-better, 26% major-response-or-better, 6% CR, with median time to PR-or-better 10.7 months. [FLAG: the N behind the September 2025 cut was not established in this pass — obtain the presentation and record its denominator before citing any of these percentages.]

[FLAG — CRITICAL COMPARABILITY WARNING: SJ901's headline "ORR" means minor response or better under RAPNO-LGG, a lower threshold than the RANO-LGG-assessed ORR that supported tovorafenib's registration. These numbers are not comparable and must never appear side by side in a response-rate table without an explicit criteria footnote. If a cross-agent comparison is required, the only defensible pairing is PR-or-better against PR-or-better, with both criteria named. Applies with equal force to the plixorafenib column in §9 — CNS response is RANO/RAPNO/iRANO, never RECIST.]

[CI judgment] Why SJ901 matters strategically: the trial excludes BRAF V600, which means it enrolls squarely into the fusion/MAPK-altered population that tovorafenib currently owns alone. If it reaches registration, the displacement is direct rather than adjacent.

Day One's FY2025 10-K competition section names Fore Biotherapeutics and plixorafenib among next-generation BRAF inhibitors it tracks. (Recording the fact of the listing only — no plixorafenib data is reproduced here, per the embargo guardrail. The Fore team should read the paragraph directly at Day One FY2025 10-K, Item 1, "Competition," to see exactly how the competitor characterizes the program.)

The strategic boundary that matters most for Fore: tovorafenib's expansion beyond pediatric LGG has effectively stalled. - The Phase 2a FIRELIGHT-1 monotherapy substudy in patients ≥12 yrs with MAPK-aberrant solid tumors had enrollment concluded because, despite responses, "a limited duration of response in this relatively rare patient population was observed" (Day One Q3 2023 results, 2 Nov 2023). - The Phase 1b/2 tovorafenib + pimasertib substudy (DAY101-102b) was terminated in December 2024 (sponsor decision). - The FY2025 10-K describes no registrational program in adults or in BRAF-altered CNS tumors outside pLGG. The entire forward clinical bet on tovorafenib is FIREFLY-2. (Note: Day One's adult-oncology ambition as of 2026 ran through the acquired Mersana ADC, Emi-Le, not through tovorafenib.)

Servier franchise effect. Servier already markets VORANIGO® (vorasidenib), the IDH1/2 inhibitor approved by FDA in August 2024 for grade 2 IDH-mutant astrocytoma/oligodendroglioma in patients ≥12 years, acquired via the ~$2B Agios oncology purchase and viewed as a €1B+ peak-sales asset central to Servier's stated €3B oncology-by-2030 ambition. Tovorafenib is Servier's second glioma-relevant acquisition in five years. [CI judgment] the practical consequence is that the competitor product is now carried by a company with an established, funded, molecularly-stratified neuro-oncology field organization and payer relationships — a stronger commercial adversary than a single-marketed-product biotech, even though the drug's label has not changed. (Corrected from "single-product biotech": at acquisition Day One held tovorafenib plus the Mersana ADC pipeline.)

Accelerated-approval overhang. OJEMDA's US approval is accelerated, based on response rate and duration of response. FIREFLY-2/LOGGIC (NCT05566795) is the confirmatory trial: ~400 patients across ~140 sites in the US, Canada, Europe, Australia, South America, the Middle East and Asia, ages 6 months to 25 years, newly diagnosed LGG with an activating RAF alteration, randomized to tovorafenib once weekly versus one of four standard-of-care chemotherapy regimens, run in collaboration with the SIOPe Brain Tumour Group LOGGIC Consortium. Enrollment completed 8 May 2026 after just over three years; topline anticipated mid-2027.

Endpoint history: the primary efficacy endpoint was originally specified on RANO-LGG (BMC Cancer, January 2024) and is stated as RAPNO-LGG in the 8 May 2026 Day One/Servier release. [FLAG: the change itself is corroborated in direction by those two primary sources, but the prior draft's attribution — "updated in June 2024 following FDA feedback during NDA review" — is not corroborated. Verify the date and the FDA-feedback causation against the FY2025 10-K clinical section or the NCT05566795 record-history tab before repeating it. State the fact of the change; do not state the cause.]

Two commercial implications: (i) continued approval is contingent on FIREFLY-2, and (ii) a positive readout would move tovorafenib from 2L into 1L and displace carboplatin/vincristine-type chemotherapy — the event that would most change the competitive landscape, landing mid-2027.

---

6. Intellectual property — Orange Book position

Authoritative source: FDA Orange Book data files, patent.txt / exclusivity.txt / products.txt, file date 14 August 2026 (August 2026 edition).

Products: NDA 217700 (tovorafenib tablet, 100 mg) and NDA 218033 (tovorafenib for oral suspension, 25 mg/mL), both approved 23 Apr 2024, both RLD and RS.

Patent	Title	OB expiry	Drug substance	Drug product	Use code	Submitted
US 8,293,752	Compounds useful as Raf kinase inhibitors	4 Aug 2031	Y	Y	(none)	8 May 2024
US 10,426,782	Pharmaceutical formulations of a pan-RAF kinase inhibitor	23 Jun 2035	—	Y	(none)	8 May 2024

Both patents are listed against both NDAs. (Independently corroborated at the summary level: third-party Orange Book trackers confirm exactly two listed US patents, none expired, Paragraph IV challenges open from 23 April 2028, and last outstanding exclusivity expiring 2031.)

Two structural observations: 1. Only two patents are Orange Book–listed, and neither carries a patent use code — i.e. there is no listed method-of-use patent. The entire OB barrier is one composition-of-matter patent plus one formulation patent. [CI judgment] this is a thinner OB wall than is typical for a ~$500K/yr oncology product. 2. US 8,293,752 is a genus claim, not a species claim. Priority 29 Jun 2007; filed 30 Jun 2008; granted 23 Oct 2012; original assignees Millennium Pharmaceuticals + Sunesis Pharmaceuticals; current assignees XOMA (US) LLC + Day One Biopharmaceuticals (co-owned). The Google Patents record shows the independent claims covering a genus of Raf-kinase-inhibitor compounds with variable substitution, and tovorafenib is not named in the specification by any of its later names. [FLAG: genus-claim breadth is a validity and infringement-scope question not litigated here; a proper freedom-to-operate/validity read requires patent counsel, not a CI dossier. Nothing in this section is an FTO or validity opinion.]

Term-adjustment note explaining the US/ex-US gap: a 20-year term from the 30 Jun 2008 filing would expire ~30 Jun 2028. The Orange Book expiry of 4 Aug 2031 reflects ~1,131 days of US Patent Term Adjustment. Foreign counterparts have no PTA and the 10-K states they expire in 2028 — a three-year ex-US composition-of-matter gap versus the US.

---

7. Patent term extension (Hatch-Waxman §156) — status and arithmetic

Primary source: 90 FR 60722–60724, "Determination of Regulatory Review Period for Purposes of Patent Extension; OJEMDA," Docket No. FDA-2024-E-3865, published 29 December 2025. All facts in the bulleted list below were independently verified against the Federal Register notice in this pass.

• PTE applicants: Day One Biopharmaceuticals, Inc. and Viracta Therapeutics, Inc. (confirming co-ownership of US 8,293,752 at the time of filing — see the chronology flag in §1).
• Patent: US 8,293,752.
• IND effective 14 Oct 2010; NDA 217700 submitted 31 Aug 2023; approved 23 Apr 2024. ✅ verified
• FDA-determined regulatory review period: 4,942 days — 4,705 days testing phase + 237 days approval phase. ✅ verified
• Redetermination comment deadline 27 Feb 2026; due-diligence petition deadline 29 Jun 2026. ✅ verified
• Applicants seek the full 5 years of extension.
• Two PTE applications were filed with USPTO in June 2024 (one for NDA 217700 tablet, one for NDA 218033 suspension); the FY2025 10-K states they remained under USPTO review as of the filing date, and that approval would extend the US patent to August 2036.

Will the full 5 years be granted? Almost certainly yes — here is the arithmetic (my calculation, independently re-derived): - §156 credit = ½ testing phase + full approval phase, counting only time after patent issuance (23 Oct 2012). Pre-issuance testing time is ~735–740 days (14 Oct 2010 → 23 Oct 2012), leaving ~3,965–3,970 creditable testing days. ½(3,967) + 237 ≈ 2,220 days ≈ 6.08 years, which exceeds the 5-year statutory cap. - The 14-years-from-approval cap binds at 23 Apr 2038; 4 Aug 2031 + 5 years = 4 Aug 2036, comfortably inside it. - Conclusion: expect US 8,293,752 to be extended to ~4 August 2036, at which point the composition-of-matter patent — not the formulation patent — becomes the last US barrier.

[FLAG: as of this data cut the USPTO has not been confirmed to have issued the extension certificate; the August 2026 Orange Book still shows the unextended 4 Aug 2031 date. Treat Aug 2036 as highly likely but not of record, and label it as such in every downstream use. The due-diligence petition window under 21 CFR 60.30 closed 29 June 2026 and no petition was found.]

[FLAG — statutory precision, corrected: the prior draft said "only one patent may be extended per approved product." The actual rules are 35 USC 156(c)(4) (no more than one patent may be extended for the same regulatory review period) and 156(a) (a patent may be extended only once). NDA 217700 and NDA 218033 were approved the same day against the same patent, which is not the ordinary fact pattern. Whether that permits or forecloses parallel extension is a question for patent counsel, not CI. The practical expectation — one extension, to ~4 Aug 2036 — is unaffected, but the reasoning should not be stated as settled.]

---

8. Regulatory exclusivity and LOE waterfall

United States (Orange Book exclusivity file, Aug 2026 edition; both NDAs):

Code	Meaning	Expiry
NCE	New Chemical Entity, 5 yr	23 Apr 2029 (NCE-1 Paragraph IV gate: 23 Apr 2028)
ODE-478	Orphan Drug Exclusivity, 7 yr	23 Apr 2031

Designations underpinning these: FDA orphan drug designation for malignant glioma, September 2020; FDA rare pediatric disease designation, 2021; breakthrough therapy designation based on PNOC014 Part A results. [FLAG: the 10-K dates the rare pediatric disease designation July 2021 in one section and May 2021 in another; use "2021" externally or verify against the FDA OOPD designation database.]

NEW — ODE scope was narrowed by statute in February 2026, and this is favorable to Fore. The Consolidated Appropriations Act, 2026 (P.L. 119-75), §6605 (the RARE Act, within the Mikaela Naylon Give Kids a Chance Act) amended the Orphan Drug Act to replace "same disease or condition" with "same approved use or indication within such rare disease or condition." This abrogates the Eleventh Circuit's 2021 Catalyst Pharmaceuticals v. Becerra decision and codifies FDA's longstanding narrower reading. Critically, it applies retroactively — "regardless of the date on which the drug was so designated, and regardless of the date on which the drug was approved."

Why it matters here: tovorafenib's ODE-478 rests on a September 2020 orphan designation for malignant glioma — a far broader condition than its approved indication. Under the pre-February-2026 Catalyst reading, a sponsor could have argued ODE reached across malignant glioma generally. Post-amendment, ODE-478 is statutorily limited to the approved use — relapsed/refractory BRAF-altered pLGG in patients ≥6 months. [FLAG: have IP/regulatory counsel confirm how FDA is applying the amendment in practice to already-listed ODE codes. The Orange Book exclusivity file displays the code and expiry but not the scope, so the narrowing is not visible in the data file itself.]

US LOE waterfall:

Date	Event
23 Apr 2028	NCE-1: earliest ANDA submission with Paragraph IV certification
23 Apr 2029	NCE expires
23 Apr 2031	ODE-478 expires — last regulatory exclusivity (now scope-limited to the approved indication per CAA 2026 §6605)
4 Aug 2031	US 8,293,752 expires absent PTE
23 Jun 2035	US 10,426,782 (formulation) expires
~4 Aug 2036	US 8,293,752 expires with the expected 5-year PTE — the operative US LOE (expectation, not of record)
2038	Synthesis-method patent family expires (not OB-listed)
2040 / 2041	Additional formulation and method-of-treating-pLGG families if granted (pending: US, EP, CN, JP, HK, SG, MX, KR, CA, AU, BR, IL, NZ, ZA)

Third-party estimates converge on this: Pharsight/GreyB models an estimated generic entry of 23 June 2035 (i.e., keyed to the formulation patent, not crediting the PTE). Our read is that the PTE, once granted, pushes the operative date to August 2036.

European Union / ex-US: - Conditional marketing authorisation, April 2026, MAH Ipsen Pharma, ATC L01EC04. ✅ verified against the EPAR. Conditional MA carries annual renewal and specific obligations — FIREFLY-2 is almost certainly the specific obligation. [FLAG: the EPAR's list of specific obligations was not retrieved; pull it if the Fore team needs the exact EU conversion conditions.] - EU orphan designation granted 20 May 2021 ✅ verified against the EPAR. - Foreign composition-of-matter counterparts expire 2028 (no PTA outside the US) — granted in DE, FR, GB, BE, CH, DK, ES, IE, IT, NL, AU, BR, CA, CN, IN, JP, KR, MX, RU, SG, ZA, TW, HK. - Formulation family expires 2035 (granted DE, FR, GB, BE, BR, CH, ES, IN, IE, IT, LU, MC, JP, CN). - Ex-US protection therefore rests on a different stack than the US. [FLAG — analysis, not verified filings: with the base compound patent expiring 2028 and first EU MA on 20 April 2026, an SPC would run ~5 years from patent expiry (the 5-year cap binds long before the 15-years-from-first-MA ceiling), extending EU compound protection to roughly mid-2033, plus a possible 6-month paediatric extension on PIP compliance. Separately, EU orphan market exclusivity normally runs 10 years from MA (→ ~April 2036), extendable to 12 with a completed PIP. Neither a filed SPC nor EMA's formal maintenance of orphan status at authorisation was confirmed.]

⏰ ACTIONABLE — near-term, checkable: an SPC must be applied for within six months of the grant of the marketing authorisation. With MA granted 20 April 2026, the SPC filing window closes around 20 October 2026 — inside the current planning horizon, not a distant analytical caveat. Assign someone to check the national SPC registers (minimum: DE, FR, GB, IT, ES) after that date. If no SPC is filed, ex-US compound protection ends in 2028 with no extension, and the whole ex-US barrier collapses to orphan market exclusivity plus the 2035 formulation family.

Practical LOE conclusion: tovorafenib is not a near-term genericization story. The binding US constraint is ODE to April 2031 (now indication-scoped) and the extended compound patent to ~August 2036; ex-US the compound patent is weaker (2028) but orphan market exclusivity and formulation patents carry it to a broadly similar horizon, conditional on the SPC filing above. [CI judgment] the commercial risk to this franchise between now and 2031 is competitive and clinical, not patent-cliff.

---

9. Competitive scorecard — comparison template

Tovorafenib column is sourced and current to 31 Aug 2026, with flagged items marked. The plixorafenib column is deliberately blank: those figures are Fore-internal/embargoed and must be entered and verified by the Fore team. Do not populate from any external source.

Dimension	Tovorafenib (OJEMDA®) — verified public	Plixorafenib — Fore-internal, to be completed and verified by Fore
Marketer, US	Servier Pharmaceuticals (since 23 Apr 2026)	☐
Marketer, ex-US	Ipsen (license, since 25 Jul 2024)	☐
Approved indication	r/r pLGG ≥6 months with BRAF fusion/rearrangement or BRAF V600 (US accelerated approval; EU conditional MA, indication quoted verbatim in §3)	☐
Approval basis / endpoint	Response rate + DoR (FIREFLY-1, RANO-LGG); confirmatory FIREFLY-2 primary endpoint now stated as RAPNO-LGG	☐ (CNS response must be RANO/RAPNO/iRANO, never RECIST — and state the criteria AND the response threshold, per the §5 comparability flag)
Adult / broader BRAF-CNS indication	None. Adult solid-tumor monotherapy substudy stopped for limited DoR (2023); MEK combo terminated Dec 2024	☐
Line of therapy	2L+ today; 1L contingent on FIREFLY-2 topline mid-2027 (enrollment complete 8 May 2026)	☐
Dosing	Once weekly, ± food; tablet and oral suspension	☐
Companion diagnostic	FoundationOne®CDx, approved 17 Jan 2025	☐
US list price	~$503K/yr WAC (2026: $38,702.90 tablets / $9,675.73 suspension per bottle); flat per package regardless of BSA	☐
Implied GTN	~95% net-to-gross (FY2025, derived) — [FLAG: contradicts the ~40% Medicaid mix; see §2. Do not publish.]	☐
US net sales	$57.2M (2024) → $155.4M (2025) → $225–250M guided (2026); no US disclosure after Q4 2025	☐
Ex-US net sales	€14.5m H1 2026 (Ipsen), mostly early access + Germany (only market with HTA started / reimbursement awarded as of Jul 2026)	☐
Patients on therapy	~429 exiting 2025 (derived, flagged); ~14–21% of Day One's stated 2,000–3,000 addressable	☐
Commercial footprint	18 US reps / ~200 accounts / >90% of US pLGG patients (10-K, unverified)	☐
Royalty burden	Mid-single-digit % to XOMA; COGS 9.5% of net revenue	☐
US regulatory exclusivity	NCE Apr 2029; ODE Apr 2031, now indication-scoped per CAA 2026 §6605	☐
US composition patent	US 8,293,752, Aug 2031 → ~Aug 2036 with expected 5-yr PTE (not yet of record)	☐
Other OB patents	US 10,426,782 (formulation) Jun 2035; no listed method-of-use patent	☐
Ex-US compound patent	2028 (no PTA); SPC filing window closes ~20 Oct 2026 — check the registers	☐
Owner's strategic posture	Inside Servier's glioma franchise with VORANIGO; €3B oncology-by-2030 target; acquisition also brought in Mersana/Emi-Le	☐

---

10. Watch list — what would move this position

1. FIREFLY-2 topline, mid-2027. The single highest-leverage event. Success converts accelerated → full approval and moves tovorafenib to first line. Failure exposes the product to expedited withdrawal. [FLAG: the prior draft quantified the first-line expansion against a "~1,100/yr newly-diagnosed BRAF-altered incidence" figure that appears in no citation and cannot be reconciled with Day One's disclosed 2,000–3,000 r/r pool. Source it or drop it — the "roughly triples the addressable pool" claim rests entirely on it and must not be presented to leadership until the denominator is real.]
2. USPTO grant of the 5-year PTE on US 8,293,752. Confirms Aug 2036 vs. Jun 2035 as the operative US LOE. Check the Orange Book patent file and the USPTO "Patents extended under 35 U.S.C. 156" list quarterly.
3. Mirdametinib (Gomekli, Merck KGaA/SpringWorks) in LGG. An approved pediatric brain-penetrant MEK inhibitor with an active LGG registration path (SJ901/NCT04923126) and deep-pocketed ownership. Because SJ901 excludes BRAF V600, it enrolls directly into the fusion population tovorafenib currently owns alone. Track the next data cut, its N, and whether SpringWorks/Merck KGaA opens a registrational LGG cohort.
4. ⏰ EU SPC filing, deadline ~20 October 2026. Six-month statutory window from the 20 April 2026 MA. A filed SPC carries ex-US compound protection to ~mid-2033; no filing leaves it at 2028. This is the nearest-term checkable IP event on the list.
5. Servier's development choices for tovorafenib beyond pLGG. Servier has the neuro-oncology infrastructure and balance sheet to restart adult or broader BRAF-altered CNS development that Day One abandoned. Watch ClinicalTrials.gov for new Servier-sponsored tovorafenib studies and Servier's annual results (FY ends 30 September — the next disclosure falls shortly after this data cut) for pipeline statements. Note Servier also inherited Emi-Le and the Mersana ADC platform, which competes for the same oncology development budget.
6. Ipsen's termination-for-convenience right, exercisable from 23 July 2026 on six months' notice. Ipsen's H1 2026 commentary treats Ojemda as a nascent launch, not a priority growth driver — a handback or renegotiation is not the base case but is now contractually available.
7. EU national P&R outcomes following the 9 Jun 2026 JCA. As of July 2026 only Germany had started an HTA and awarded reimbursement (automatic for orphan products). The JCA's acknowledgement that there is no established comparator cuts both ways and will produce divergent country-level prices as other markets begin assessment.
8. US drug-pricing policy. The 10-K flags an HHS proposal to test payment models specifically for accelerated-approval drugs, aimed at reducing Medicare spend on products without confirmed clinical benefit — directly applicable to OJEMDA until FIREFLY-2 reads out. [FLAG: the current status of this proposal as of the 31 Aug 2026 data cut was not re-verified in this pass; it is repeated from the 10-K risk factors. Given that the adjacent PRV item in the prior draft turned out to be six months stale for exactly this reason, re-verify before use.]
9. ~~RPD PRV program sunset~~ — RESOLVED, item corrected. The prior draft carried the FY2025 10-K's statement that it was "currently uncertain whether the program will be renewed." That was superseded before this data cut. The Rare Pediatric Disease PRV program was reauthorized on 3 February 2026 by the Consolidated Appropriations Act, 2026 (P.L. 119-75), §6604 (Mikaela Naylon Give Kids a Chance Act), extending FDA authority to award RPD PRVs through 30 September 2029; FDA granted its first voucher under the restored program in March 2026. §6604 also directs GAO to report to Congress on program effectiveness within five years. The PRV incentive is available for pediatric CNS filings in this space, and the 2029 sunset is the date to plan against.

---

Residual gaps

• No public figure for realized payer mix, median duration of therapy, or discontinuation rate — only prescription counts. The unresolved GTN/payer-mix contradiction in §2 flows directly from this.
• No US OJEMDA revenue disclosure after Q4 2025; 2026 US performance vs. the $225–250M guidance is now unobservable.
• Servier has not published a peak-sales expectation for OJEMDA, nor an Ojemda-specific segment figure.
• PTE certificate issuance unconfirmed; EU SPC filings and EMA orphan-exclusivity maintenance unconfirmed (SPC window closes ~20 Oct 2026).
• No NCCN or COG guideline citation retrieved at this cut; guideline positioning in 2L pLGG should be verified directly before any external claim about "standard of care" status (the phrase appears only in Day One's own management commentary).
• The N behind the September 2025 SJ901 data cut (87% / 55%) was not established; only the May 2025 cut (N=35) is fully sourced.
• Several §4 operational figures (18 reps, ~200 accounts, 97.6% two-customer concentration, CMO list) rest on a single unverified 10-K read.
• One authoritative denominator for BRAF fusion vs V600E share of BRAF-altered pLGG has not been fixed; two inconsistent unsourced ranges were removed in this pass and not replaced with a citable figure.

Head-to-Head Comparison Framework vs Plixorafenib

Data cut for this workstream: 31 August 2026. All competitor figures below are public and cited to a primary source (FDA label, peer-reviewed publication, EMA EPAR, or dated company press release). The plixorafenib column contains no data — only labeled template cells for the Fore team.

Adversarial verification status (added 31 Aug 2026): every figure in §3–§9 was re-checked against the cited primary source during fact-check. The FDA label figures (§14, §5.1–5.4, §6.1, §7, §12.1–12.3), the Nature Medicine FIREFLY-1 figures, the SNO 2025 press-release figures, the EMA regulatory dates and indication text, the Servier/Day One transaction terms, the FY2025 revenue figure, the FoundationOne CDx date, the adult phase 1 design, and the FIREFLY-2 enrollment milestone all reconciled exactly. Eight items were corrected or re-framed; they are listed in §12. Source recency at this data cut: US label revised 8/2025 (~12 months old; confirmed current on DailyMed); Nature Medicine primary publication data cut 5 Jun 2023 (~3.2 years old); most recent efficacy cut 6 Jun 2025 (SNO 2025, press release only).

---

0. Two framing flags — read before using this table

[FLAG: competitor drug not named at invocation] This pipeline run supplied the literal string "the competitor drug" rather than a drug name, so no target was specified to any workstream. I resolved the target to tovorafenib (OJEMDA®; formerly DAY101 / TAK-580 / MLN2480 / BIIB024) because it is the only agent listed as LIVE in Fore's own competitive-intelligence registry (C:\Users\Owner\Documents\Hugh Context Folder\Fore\competitive_intelligence\CI_PROCESS.md, which records it as "LIVE (draft, 41 flags)"; dabrafenib+trametinib and selumetinib are "Planned", next-gen type II / pan-RAF is "Watch") and it is the closest public comparator to plixorafenib in BRAF-altered CNS tumors. The Fore team must confirm this was the intended target before the dossier is circulated. If the intended target was dabrafenib+trametinib, selumetinib, belvarafenib, claturafenib, NST-628, or BDTX-4933, the row structure in §3–§9 transfers unchanged; only the competitor column must be repopulated. Note also that the other six workstreams in this run received the same unnamed placeholder and may have resolved it differently — reconcile before synthesis.

[FLAG: plixorafenib column is empty by design] No comparison slide or figure set was supplied. Per the pipeline guardrail, plixorafenib data is Fore-internal/embargoed and was not researched, inferred, or estimated. Every plixorafenib cell below is a placeholder marked ▢ Fore-internal — to complete. Any number that appears in that column later must be tagged "Fore-internal — verify" and traced to a cleared internal source before external use. This guardrail also forbids inferring plixorafenib's dosing schedule, population, or binding mode in the prose sections — §10 has been corrected where it previously did so.

---

1. Rules of engagement for completing the plixorafenib column

Apply these to every cell, or the table will produce a false comparison:

#	Rule	Why
R1	Never enter a bare percentage. Format every efficacy/safety cell as n/N (x%), [criterion], [reader: BICR or investigator], [data cut date].	The competitor column carries three different ORRs for the same trial (§2, hazard H2), and its labeled ORR itself moved between label revisions (H9).
R2	CNS response must be RANO / RAPNO / iRANO — never RECIST. If the plixorafenib CNS dataset was read by RECIST, say so explicitly and do not place it on the same row as a RAPNO figure.	Fore's own SNO 2026 CNS work was SME-corrected on exactly this point (the study used RECIST, with some CNS lesions read by RANO).
R3	State the response-evaluable denominator, not the enrolled N.	FIREFLY-1 has five denominators in play, all verified against source: 77 enrolled (Arm 1) [2], 69 RANO-HGG–evaluable [2], 76 in the label efficacy population (patients with measurable disease at baseline per RAPNO-LGG) [1, §14], 137 in the pLGG safety table [1, §6.1], 172 in the pooled safety population [1, §6.1].
R4	Match populations before comparing. Age band, line of therapy, prior MAPK-inhibitor exposure, BRAF alteration class, and NF1 status must all be stated in the same cell block (§4).	See hazard H4.
R5	Do not compute a cross-trial delta. This is a side-by-side descriptive framework, not an indirect treatment comparison. Any ITC/MAIC requires a separate, statistically specified analysis.	Both datasets are single-arm; no randomized bridge exists.
R6	Leave a cell blank rather than approximate it. A ▢ not established cell is a finding; a guessed cell is a liability.	Fore's audit-shortcuts standard.
R7	Never carry a response depth claim (CR rate, CR+PR rate) across criteria. Cite the criterion in the same breath as the number.	New — see H10. The CR count for this trial is 0 under RAPNO-LGG and 12 under RANO-HGG. Same patients.

---

2. Comparability hazards specific to this pairing

These are the traps a reader of the finished table will fall into. Keep this section adjacent to the table.

H1 — RAPNO-LGG counts minor response as a response. The FDA-label ORR for tovorafenib is 53% (95% CI 41–64), 40/76, by RAPNO-LGG per blinded independent central review, and that figure decomposes as partial response 29/76 (38%) + minor response 11/76 (14%); complete responses: 0 (0%) [1, §14] (component percentages are individually rounded and sum to 52%, not 53%; the 40/76 count is exact). Minor response (≥25% to <50% reduction) is a RAPNO-LGG–specific category with no RANO-HGG, RANO 2010, or RECIST equivalent. If the plixorafenib ORR counts only CR+PR, the correct comparator cell is 38% (29/76) by RAPNO-LGG, not 53%. Both must appear in the table, on separate rows — and see H10 before making any claim about complete responses. [FLAG: the "≥25% to <50%" MR definition should be verified against the RAPNO-LGG methodology publication (Fangusaro et al., Lancet Oncol 2020) rather than the label, which does not define the threshold.]

H2 — Three ORRs exist for one trial. All from FIREFLY-1, same patients, different criteria/cuts: RANO-HGG 46/69 (67%, 95% CI 54–78) [2]; RANO-LGG 2011 41/76 (54%, 95% CI 42–65) [1, §14]; RAPNO-LGG 40/76 (53%, 95% CI 41–64) [1, §14]. The 67% figure is the one that reaches lay press and company decks; it is not the labeled efficacy claim and must not be the comparator row. It also rests on the oldest data cut in the set (5 Jun 2023).

H3 — Metric swap: 16.6 and 19.4 months each appear as both DoR and PFS. Lock these pairings (all three verified verbatim against source): - Nat Med primary analysis (cut 5 Jun 2023), RANO-HGG: median DoR 16.6 mo (95% CI 11.6–NR); median PFS 19.4 mo (95% CI 16.9–NR) [2]. - SNO 2025 three-year update (cut 6 Jun 2025), RAPNO-LGG: median DoR 19.4 mo (95% CI 13.8–27.2); median PFS 16.6 mo (95% CI 10.9–22.0) [3]. - FDA label 8/2025, RAPNO-LGG: median DoR 18 mo (95% CI 12.0–22.8), n=40 responders [1, §14].

H4 — The populations may not overlap at all. FIREFLY-1's efficacy population is pediatric (median age 8.5 y, range 2–21), NF1-excluded, heavily pretreated (median 3 prior systemic regimens; 59% prior MAPK-pathway inhibitor), and fusion-dominant (74% KIAA1549::BRAF; only 16% BRAF V600E) [1, §14]. If plixorafenib's CNS dataset is adult/adolescent, V600-enriched, or NF1-inclusive, several rows below are structurally non-comparable and should be greyed out rather than filled.

H5 — Dose actually administered ≠ approved dose. FIREFLY-1 patients received doses spanning 290–476 mg/m² weekly (0.76–1.25× the approved recommended dosage); the approved dose is 380 mg/m² weekly [1, §12.2, §2.3]. Safety and efficacy in the trial therefore reflect a spread of exposures around the label dose. [FLAG: an earlier draft of this row stated a median of "approximately 420 mg/m² weekly." That median does not appear in the label text retrieved (§12.2 gives only the 290–476 mg/m² exposure–response range). Do not cite a median administered dose until it is located in a primary source; the range is the defensible figure.]

H6 — The reported incidence of growth-velocity toxicity rose sharply between label revisions. Growth-velocity toxicity was reported as 15% of patients 18 years of age or younger (Grade 3 in 5%; permanent discontinuation 2%, n=2) in the original 4/2024 label [4, §5.4] and as 46% of 133 patients ≤18 y, 35% Grade ≥3 (permanent discontinuation 3%) in the current 8/2025 label [1, §5.4]. Always cite the 8/2025 revision. Framing discipline: this is a change in reported incidence under longer follow-up and (apparently) fuller ascertainment — it is not evidence that the drug became more toxic, and it is not a like-for-like comparison, because the 4/2024 label states no denominator for its 15% while the 8/2025 label specifies n=133. The 4/2024 label also asserted "Growth velocity recovered after interruption," whereas 8/2025 documents only partial recovery. Present it as "the label's growth-toxicity characterisation was materially revised," not as "the safety profile worsened."

H7 — Growth toxicity is age-conditional. Reduced growth velocity is only a differentiator where the age ranges overlap. In an adult population it is not an applicable row — mark it n/a (population), not 0%.

H8 — Single-arm, accelerated/conditional approvals on both sides of the Atlantic. The tovorafenib evidence base carries no randomized comparator; the confirmatory trial has not read out (§8). Any "standard of care" framing is commercial positioning, not a randomized finding.

H9 — The labeled efficacy claim itself changed between label revisions. (new) The 4/2024 label reported ORR 51% (95% CI 40–63); CR 0, PR 28 (37%), MR 11 (14%); median DoR 13.8 mo (11.3–NE) in 39 responders; median TTR 5.3 mo [4, §14]. The 8/2025 label reports ORR 53% (41–64); CR 0, PR 29 (38%), MR 11 (14%); median DoR 18 mo (12.0–22.8) in 40 responders; median TTR 5.4 mo [1, §14]. Baseline demographics were also restated (e.g. race "not reported" fell from 26% to 18%). Any competitor efficacy figure sourced to a 2024-vintage secondary article is stale — cite the 8/2025 label.

H10 — The "zero complete responses" claim is criterion-specific and must never be stated bare. (new — this was an error in the prior draft) Complete responses in FIREFLY-1 Arm 1 are 0/76 (0%) by RAPNO-LGG [1, §14] but 12/69 (17%) by RANO-HGG, alongside 34 PR (49%), in the same trial and the same patients [2]. A bare "tovorafenib produced no complete responses" is false as written and is precisely the criterion-mixing error rule R2 exists to prevent; it would not survive competitor challenge. Any depth-of-response comparison must state the criterion on both sides of the table.

---

3. Comparison framework — Table A: Asset & class

Attribute	Tovorafenib (OJEMDA) — public, cited	Plixorafenib — Fore-internal
Class	"Type II RAF kinase inhibitor of mutant BRAF V600E, wild-type BRAF, and wild-type CRAF kinases" — quoted verbatim from label [1, §12.1]	▢ Fore-internal — to complete (state binding mode: type I / type I½ / paradox-breaker / dimer-disrupting)
Prior identifiers	DAY101, TAK-580, MLN2480, BIIB024 [5]	▢ Fore-internal — to complete
Originator / current owner	Day One Biopharmaceuticals (US); Servier completed acquisition 23 Apr 2026, $21.50/share, ~$2.5B equity value, a 68% premium to the 5 Mar 2026 close [6]; ex-US rights licensed to Ipsen [6]	▢ Fore-internal — to complete
BRAF alteration classes addressed (claimed)	BRAF fusion/rearrangement and BRAF V600 mutation (per approved indication) [1, §1]	▢ Fore-internal — to complete (specify class I / II / III coverage and evidence level)
Paradoxical MAPK activation	Type II mechanism described in the review literature as avoiding the paradoxical activation seen with type I BRAF inhibitors [5] — [FLAG: this is a mechanistic claim from review/company sources, not an FDA-label statement; label §12.1 characterises the class but makes no paradox claim. Do not present as label-supported.]	▢ Fore-internal — to complete
CNS penetration	Described as "CNS-penetrant" by the developer, by the review literature [5], and in the peer-reviewed adult phase 1 publication [7]; [FLAG: no quantitative human CSF or brain:plasma ratio appears in the US label §12.3 [1]. If a numeric value is needed for the comparison, it must be sourced from a preclinical/clinical pharmacology publication and labeled as such — the descriptor alone is not a comparable figure.]	▢ Fore-internal — to complete (give brain:plasma or CSF:plasma with species/method)

---

4. Table B: Baseline characteristics of the population supporting the efficacy claim

Tovorafenib column = FIREFLY-1 (NCT04775485) efficacy population, N=76, from FDA label 8/2025 §14 [1]. Every row in this table was re-verified verbatim against the current label during fact-check.

Baseline characteristic	Tovorafenib — FIREFLY-1 (N=76)	Plixorafenib — Fore-internal
Trial / registration	FIREFLY-1 (PNOC026), NCT04775485; multicenter, open-label, single-arm [1, §14]	▢ to complete (trial name + NCT)
Tumor type	Relapsed or refractory pediatric low-grade glioma	▢ to complete
Median age (range)	8.5 y (2–21)	▢ to complete
Sex	53% male	▢ to complete
Race/ethnicity	61% White, 7% Asian, 2.6% Black or African American, 3.9% multiple races, 8% other race, 18% not reported; 3.9% Hispanic/Latino. (Restated from the 4/2024 label, which reported 53% White / 26% not reported for the same 76 patients — see H9.)	▢ to complete
Performance status	Karnofsky/Lansky 80–100 in 93%	▢ to complete
Prior systemic regimens	Median 3 (range 1–9); all had ≥1 prior systemic therapy with documented radiographic progression	▢ to complete
Prior MAPK-pathway inhibitor	45/76 (59%)	▢ to complete
BRAF alteration mix	KIAA1549::BRAF fusion 56/76 (74%); BRAF V600E 12/76 (16%); other BRAF alteration incl. duplication/rearrangement 8/76 (11%). Note for §5: the label's exploratory "BRAF fusion or rearrangement" subgroup is n=64 — i.e. the 56 fusions plus the 8 "other" duplications/rearrangements (64 + 12 V600E = 76). 56 and 64 are not in conflict.	▢ to complete
Tumor location	Optic pathway 51%; deep midline structures 12%; brain stem 8%; cerebellum 7%; cerebral hemisphere 5%	▢ to complete
Key exclusions	Verbatim: "Patients with tumors harboring additional activating molecular alteration(s) (e.g., IDH1/2 mutations, FGFR mutations, etc.) or patients with known or suspected diagnosis of neurofibromatosis type 1 (NF1) were excluded." [1, §14]	▢ to complete
Measurability requirement	≥1 measurable lesion at baseline per RAPNO-LGG criteria [1, §14 table footnote]	▢ to complete
Imaging cadence	"Tumor assessments were performed every 12 weeks." [1, §14]	▢ to complete (cadence materially affects TTR/PFS)

---

5. Table C: Efficacy — the row set that must be criterion-matched

Efficacy row	Tovorafenib (cite + criterion + denominator)	Plixorafenib — Fore-internal
Labeled ORR (regulatory claim)	53% (95% CI 41–64), 40/76, RAPNO-LGG, BICR, label rev. 8/2025 [1, §14]	▢ to complete — n/N (%), criterion, reader, cut
CR+PR-only ORR, RAPNO-LGG (H1)	38% (29/76); CR = 0/76, PR = 29/76 [1, §14]	▢ to complete
CR+PR-only ORR, RANO-HGG (H10)	67% (46/69); CR = 12/69 (17%), PR = 34/69 (49%) [2] — the same patients, a different criterion, and a non-zero CR count	▢ to complete
Minor response (RAPNO-only category)	11/76 (14%) [1, §14]	▢ to complete or n/a (criterion)
ORR, RANO-LGG 2011, BICR	54% (95% CI 42–65), 41/76 (PR 23, MR 18) [1, §14]	▢ to complete
ORR, RANO-HGG, independent review (primary publication)	67% (95% CI 54–78), 46/69, data cut 5 Jun 2023 [2] — see H2, H10	▢ to complete
Median DoR	Label 8/2025 RAPNO-LGG: 18 mo (95% CI 12.0–22.8), n=40 [1]; 3-y cut RAPNO-LGG: 19.4 mo (13.8–27.2) [3]; Nat Med RANO-HGG: 16.6 mo (11.6–NR) [2]; (superseded 4/2024 label: 13.8 mo, n=39 [4])	▢ to complete
DoR ≥12 mo / ≥18 mo	65% / 50% (of 40 responders), RAPNO-LGG [1, §14]	▢ to complete
Median time to response	5.4 mo (range 1.6–17.5), RAPNO-LGG [1, §14; concordant in 3]; 3.0 mo (range 2.6–16.6) by RANO-HGG [2]	▢ to complete
Median PFS	RAPNO-LGG, 3-y cut: 16.6 mo (95% CI 10.9–22.0) [3]; RANO-HGG, primary: 19.4 mo (16.9–NR) [2] — see H3	▢ to complete
Clinical benefit rate	93%, RANO-HGG [2]. [FLAG: the paper states 93%; the "64/69" numerator in the prior draft was back-calculated (0.93 × 69 = 64.2), not quoted. Do not present a derived numerator as a source figure — quote 93% or retrieve the exact n from the paper's table.]	▢ to complete (define CBR components)
ORR — BRAF fusion/rearrangement	53% (n=64), exploratory, RAPNO-LGG [1, §14]; (4/2024 label: 52%, same n=64 [4]). See §4 note on why n=64 ≠ 56	▢ to complete
ORR — BRAF V600E	50% (n=12), exploratory [1, §14] — note the denominator: 12 patients. A 50% point estimate on n=12 carries a very wide interval and should never be compared to a large-N figure without one.	▢ to complete
ORR — prior MAPK-targeted therapy	51% (n=45) vs 55% (n=31) without [1, §14]; (4/2024 label: 49% vs 55% [4])	▢ to complete
Median follow-up / study duration	Median study duration 40.6 mo; data cutoff 6 Jun 2025 [3]	▢ to complete
Off-treatment durability	Median time to next treatment 42.6 mo (95% CI 36.7–NE); of 39 entering treatment-free observation, 30/39 (77%) treatment-free ≥12 mo; retreatment n=8, median retreatment duration 9 mo, median maximum tumor reduction −38.3% [3] — all verified verbatim	▢ to complete
Overall survival	Not reported in the FDA label, the Nature Medicine primary publication, or the SNO 2025 release [1, 2, 3] — [FLAG: verify against the SNO 2025 presentation slides/abstract; a relapsed pLGG population may not reach an OS estimate. Record as "not reported," never as "no OS benefit."]	▢ to complete
Adult efficacy	No adult efficacy claim. Adult phase 1 (n=149; Q2D n=110, QW n=39) established RP2D of 600 mg once weekly or 200 mg every other day; that study was dose-finding with melanoma expansion. Grade ≥3 AEs in expansion: 58/80 (73%) Q2D vs 9/19 (47%) QW [7]	▢ to complete

---

6. Table D: Safety & tolerability

Tovorafenib safety population per US label 8/2025 §6.1: pooled 172 patients — 140 BSA-dosed with r/r pediatric LGG or advanced solid tumors harbouring a RAF alteration, plus 32 adults at a flat 600 mg dose; the pLGG-specific adverse-reaction table is N=137 [1, §6.1]. Every percentage below was re-verified against the label during fact-check.

Safety row	Tovorafenib	Plixorafenib — Fore-internal
Boxed warning	None [1]	▢ to complete
Contraindications	"None." [1, §4]	▢ to complete
Serious adverse reactions	45% (N=137); viral infection 9%, pneumonia 4%, sepsis 4% [1, §6.1]	▢ to complete
Fatal adverse reaction	Tumor hemorrhage in 1 patient, reported as 1% of the N=137 pLGG population [1, §6.1] and as Grade 5 in 0.6% (n=1) of the N=172 pooled population [1, §5.1]. These are the same single event expressed against two denominators — do not read as two deaths.	▢ to complete
Permanent discontinuation for AE	7% (all-cause AE, N=137); recurring causes tumor hemorrhage and reduced growth velocity [1, §6.1]. Separately, Nat Med reports 9/137 (7%) discontinuing for treatment-related AEs [2]. These measure different things and coincidentally round to the same figure — do not describe them as "concordant" or cite them as mutual confirmation.	▢ to complete
Dose interruption	57% (rash, pyrexia, vomiting, hemorrhage each ≥5%) [1, §6.1]	▢ to complete
Dose reduction	24% (rash, fatigue each ≥2%) [1, §6.1]	▢ to complete
Grade ≥3 treatment-related AEs	42% (N=137) [2]	▢ to complete
Rash	77% all grades / 12% Grade 3–4 (N=137) [1, §6.1]; warnings section, pooled N=172: rash 67%, Grade 3 12%, interruption 15%, reduction 7%, discontinuation 1% (n=2) [1, §5.2]. The 77% and 67% are the same toxicity on different denominators — never place both in a comparison row.	▢ to complete
Hair color changes	76% / 0% Grade 3–4 [1, §6.1]	▢ to complete
Fatigue	55% / 4% [1, §6.1]	▢ to complete
Viral infection	55% / 7% [1, §6.1]	▢ to complete
Vomiting / nausea / constipation	50% / 4%; 33% / 0%; 33% / 0% [1, §6.1]	▢ to complete
Headache	45% / 1% [1, §6.1]	▢ to complete
Hemorrhage	42% / 5% (N=137) [1, §6.1]; pooled N=172: any hemorrhagic event 37%, epistaxis 26%, intratumoral hemorrhage 9%, serious bleeding 5%, permanent discontinuation 2%, Grade 5 tumor hemorrhage 0.6% (n=1) [1, §5.1]	▢ to complete
Pyrexia	39% / 4% [1, §6.1]	▢ to complete
Dermatitis acneiform / dry skin / pruritus	31% / 1%; 36% / 0%; 26% / 1% [1, §6.1]	▢ to complete
Photosensitivity	12% pooled, Grade 3 in 0.6% (n=1) [1, §5.2]	▢ to complete
Hepatotoxicity	AST increased 74% (Grade 3 2%); ALT increased 42% (Grade 3 4%); bilirubin increased 23% (Grade 3 0.6%, n=1); median time to onset 14 days (range 3–280) [1, §5.3]	▢ to complete
Reduced growth velocity	46% of 133 patients ≤18 y; 35% Grade ≥3; interruption 5%, reduction 2.3%, permanent discontinuation 3%. Median change in height percentile −14 (z −0.6) at 12 mo (N=107) and −20 (z −0.9) at 18 mo (N=95). Partial recovery off treatment: annualized growth velocity 0.86–1.8 cm/y on treatment over a 2-year period vs 4.2 cm/y median at ≥90 days off treatment (n=17) [1, §5.4]. [FLAG: the "n=81" on-treatment denominator in the prior draft was not found in the retrieved label text — the label states the 0.86–1.8 cm/y range without that n. Verify or drop the n.] See H6, H7	▢ to complete — or n/a (population) if adult
Grade 3–4 lab abnormalities (≥2%)	Decreased phosphate, decreased hemoglobin, increased CPK, increased ALT, decreased albumin, decreased lymphocytes, decreased leukocytes, increased AST, decreased potassium, decreased sodium [1, §6.1]	▢ to complete
Class/organ-specific warnings	Six warnings, per label §5: Hemorrhage; Skin Toxicity Including Photosensitivity; Hepatotoxicity; Effect on Growth; Embryo-Fetal Toxicity; NF1-Associated Tumors ("tovorafenib may promote tumor growth in patients with NF1 tumors," based on nonclinical NF1 models without BRAF alterations) [1, §5]	▢ to complete
Cardiac	"A mean increase in the QT interval >20 milliseconds was not observed" at the recommended dosage [1, §12.2]	▢ to complete
Secondary malignancy signal	Not listed in label warnings [1] — [FLAG: type I BRAF inhibitors carry cutaneous SCC/keratoacanthoma warnings; absence here must be stated as "not in label", never as "proven absent." Follow-up in this population is also short relative to secondary-malignancy latency.]	▢ to complete

---

7. Table E: PK, PD & formulation

Parameter	Tovorafenib [1, §2.3, §12.2, §12.3]	Plixorafenib — Fore-internal
Dose & schedule	380 mg/m² orally once weekly, maximum 600 mg, until progression or intolerable toxicity; not established below BSA 0.30 m²	▢ to complete (state mg, frequency, continuous vs intermittent)
Formulations	100 mg film-coated tablets; powder for oral suspension 25 mg/mL. [FLAG: the "bottle delivers 300 mg / 12 mL" detail was not confirmed in the sections retrieved — verify against label §16 before use.]	▢ to complete
Pediatric-usable formulation	Yes — oral suspension, BSA-banded dosing tables	▢ to complete
Food effect	No clinically significant change in Cmax/AUC with a high-fat meal; may be taken with or without food. [FLAG: the "Tmax delayed 3 h → 6.5 h" specific was not confirmed in the sections retrieved — verify against §12.3.]	▢ to complete
Tmax	3 h (1.5–4)	▢ to complete
Steady-state Cmax / AUC	6.9 µg/mL (23%) / 508 µg·h/mL (31%); steady state at 12 days; dose-proportional; no clinically significant accumulation	▢ to complete
Terminal half-life	~56 h (33%) — enables weekly dosing	▢ to complete
Apparent clearance / Vd	0.7 L/h/m² (31%) / 60 L/m² (23%)	▢ to complete
Protein binding	97.5% bound to human plasma proteins	▢ to complete
Metabolism	Primarily aldehyde oxidase and CYP2C8; CYP3A/2C9/2C19 minor	▢ to complete
Excretion	65% recovered in feces (8.6% unchanged); 27% in urine (0.2% unchanged)	▢ to complete
Key DDIs	Avoid strong/moderate CYP2C8 inhibitors and inducers; avoid coadministration of hormonal contraceptives; avoid certain CYP3A substrates where minimal concentration changes may lead to serious therapeutic failures; inhibits BCRP at clinically relevant concentrations; inhibits CYP2C8/2C9/2C19/3A and induces CYP3A/2C8/1A2/2B6/2C9/2C19 in vitro [1, §7]	▢ to complete
Organ impairment	No dose adjustment for mild hepatic impairment; no clinically significant differences with mild–moderate renal impairment (eGFR ≥30)	▢ to complete
Exposure–response	ORR exposure–response not clinically significant over 290–476 mg/m² (0.76–1.25× approved dose); higher exposure associated with increased rash, ALT/AST elevation, CPK elevation, and reduced height-for-age z-score [1, §12.2]	▢ to complete
Required pre-treatment testing	Confirm BRAF fusion/rearrangement or V600 mutation; evaluate ALT, AST and bilirubin before initiating [1, §2.1, §2.2]	▢ to complete

---

8. Table F: Regulatory status by region

Item	Tovorafenib	Plixorafenib — Fore-internal
FDA status	Accelerated approval, 23 April 2024 [8], based on response rate and duration of response [1, §1]	▢ to complete (IND / designations / filing plan)
FDA indication wording	"…patients 6 months of age and older with relapsed or refractory pediatric low-grade glioma (LGG) harboring a BRAF fusion or rearrangement, or BRAF V600 mutation" [1, §1]	▢ to complete
Confirmatory obligation	Continued approval contingent on verification of clinical benefit in a confirmatory trial [1, §1]. Confirmatory trial = FIREFLY-2 / LOGGIC (NCT05566795), phase 3 randomized, front-line, ages 6 months–25 years, tovorafenib once weekly vs four standard-of-care chemotherapy regimens, primary endpoint ORR by RAPNO-LGG, ~400 patients across ~140 sites [9, 10]. Enrollment completed 8 May 2026; primary analysis expected ~12 months after last patient enrolled, with preliminary insights in 2027 and topline expected by mid-2027 [10]. [FLAG: as of the 31 Aug 2026 data cut, searching found no evidence of conversion to full approval — verify directly against Drugs@FDA before any external claim. The 2027 readout timeline makes conversion before then improbable but unconfirmed.]	▢ to complete
Companion diagnostic	FoundationOne CDx approved 17 January 2025 as the first and only CDx for OJEMDA [11]. (The original 4/2024 label stated verbatim: "An FDA approved test for the detection of BRAF fusion or rearrangement, or BRAF V600 mutation in relapsed or refractory pediatric LGG is not currently available." [4, §2.1])	▢ to complete
EMA status	Conditional marketing authorisation; European Commission decision 20 April 2026; CHMP positive opinion 26 February 2026; MAH Ipsen Pharma; EMA product number EMEA/H/C/006140 [12]. [FLAG: the Ipsen press release announcing the approval is dated 22 Apr 2026 and some secondary coverage gives the CHMP opinion as 27 Feb 2026 — those are announcement dates. The EMA product page dates (20 Apr / 26 Feb) are authoritative and are the ones used here.] [FLAG: the "MAA submitted 26 Feb 2025" date in the prior draft is not shown on the EMA product page and is exactly one year before the CHMP opinion date — a pattern consistent with a transcription error. Verify against the EPAR procedural steps or drop the row.]	▢ to complete
EU indication wording	Verbatim: "Monotherapy for treatment of patients 6 months of age and older with paediatric low-grade glioma harbouring a BRAF fusion or rearrangement, or BRAF V600 mutation, who have progressed after one or more prior systemic therapies" [12]. [FLAG: the Ipsen press release headline describes the approval as covering pLGG "regardless of BRAF alteration" [13]. The charitable reading is "regardless of which BRAF alteration (fusion or V600)," but as written it contradicts the authorised indication — which requires a BRAF alteration — and must not be reproduced in any Fore document.]	▢ to complete
EU specific obligation	"Submit final results from an ongoing clinical study in patients with paediatric low-grade glioma from six months of age comparing the effectiveness and safety of Ojemda with chemotherapy"; annual reassessment of benefit–risk [12]	▢ to complete
EU orphan designation	EU/3/21/2434, designated 20 May 2021 [12]	▢ to complete
EU HTA	Ipsen states it is the first medicine to complete the EU Joint Clinical Assessment process [13] — [FLAG: company-sourced "first" claim; not independently verified. Attribute to Ipsen or drop the superlative.]	▢ to complete
PMDA / other regions	Not established in the sources reviewed — [FLAG: Japan, UK MHRA, Health Canada, TGA, China NMPA status not verified; do not assert absence of approval, only absence of evidence found at this data cut.]	▢ to complete
US orphan / pediatric exclusivity	[FLAG: not verified in this workstream. Orphan-drug exclusivity, rare pediatric disease priority review voucher status, and Orange Book patent listings should be pulled from Drugs@FDA / the Orange Book by the competitive_ip workstream and cross-checked here.]	▢ to complete

---

9. Table G: Commercial position (context rows — detail belongs to the IP/commercial workstream)

Item	Tovorafenib	Plixorafenib — Fore-internal
US marketer	Day One Biopharmaceuticals; acquired by Servier, tender offer completed 23 April 2026, $21.50/share, ~$2.5B equity value, 68% premium to the 5 Mar 2026 close [6]	▢ to complete
Ex-US rights	Licensed to Ipsen [6]; Ipsen is the EU MAH [12]	▢ to complete
Reported revenue	Preliminary FY2025 net product revenue $155.4M, +172% year-over-year (announced 11 Jan 2026), with 2026 US guidance of $225–250M [14]. Full-year results confirmed the $155.4M figure in Feb 2026 [14]	▢ to complete
List price	Launch WAC reported at ~$33,916 per 28-day supply, with annual cost reported to exceed $440,000 [15]. Day One's Colorado prescriber-transparency filing effective 1 Jan 2026 lists WAC figures ranging from $9,675.73 to $38,702.90 across tablet configurations/pack sizes [16]. [FLAG (two issues): (a) the prior draft quoted only the $38,702.90 top-of-range figure — the filing is a range and quoting only its maximum is selectively unfavourable; (b) citation [16] returned HTTP 404 when fetched during verification, so the figures are corroborated only via a search-index snapshot. Re-source the filing before any external use, and do not annualize any package price without checking the carton-to-dose mapping in label §16 across BSA bands.]	▢ to complete
Positioning claim	Marketed as the first FDA-approved therapy specifically for BRAF-altered relapsed/refractory pLGG [8]	▢ to complete

---

10. Differentiation axes the completed table should resolve

When the plixorafenib column is filled, these are the questions the framework exists to answer. Each maps to rows above. Each is written as a question about the competitor's public position — none asserts or infers a plixorafenib property.

1. Population adjacency or overlap? Pediatric r/r pLGG (tovorafenib's labeled space) vs whatever age band and tumor set plixorafenib addresses. If the overlap is small, the competitive threat is a label-expansion threat (via FIREFLY-2 front-line and any adult program), not a head-to-head threat today.
2. Alteration-class coverage. Tovorafenib's public evidence is fusion-dominant (74% KIAA1549::BRAF) with only 12 V600E patients; its label covers fusion/rearrangement and V600. Whether plixorafenib addresses classes tovorafenib does not is the sharpest potential differentiator — and it is answerable only from cleared Fore data.
3. Response depth — criterion-conditional, handle with care. Under the labeled RAPNO-LGG analysis there were 0 complete responses in 76 patients, with 14 of the 53 percentage points contributed by minor response (H1). Under RANO-HGG in the same trial there were 12 complete responses in 69 patients (17%) [2] (H10). Depth of response may be where a differentiated claim could live, but any such claim must name the criterion on both sides; a bare "no complete responses" claim is false and will not survive challenge. (Corrected — the prior draft stated the zero-CR figure without its criterion.)
4. Durability vs treatment-free interval. Tovorafenib's strongest public durability data are the off-treatment figures (TTNT 42.6 mo; 30/39 (77%) treatment-free ≥12 mo) [3]. Any plixorafenib durability comparison should be framed on the same axis — and note that these come from a press release, not a peer-reviewed source (flag 11).
5. Chronic-therapy tolerability in growing children. 46% growth-velocity events with 35% Grade ≥3 and 3% discontinuing for it [1, §5.4] is the most material public tolerability difference to examine — but only where populations overlap (H7), and it must be presented with the label's own recovery data (annualized velocity 0.86–1.8 cm/y on treatment vs 4.2 cm/y off) rather than as an unqualified liability. (Reworded from "most exploitable public tolerability liability" — objective framing.)
6. Dosing convenience. Once-weekly dosing on a ~56 h half-life with a pediatric oral suspension is a genuine competitive asset for tovorafenib. The question for the Fore team is what plixorafenib's schedule is and what it implies — this workstream does not know it and does not assume it. (Corrected — the prior draft asserted "a BID continuous regimen would need a countervailing efficacy or breadth argument," which infers a plixorafenib dosing regimen and breaches the SCOPE guardrail in §0.)
7. DDI burden. Avoid CYP2C8 inhibitors and inducers, and avoid coadministration of hormonal contraceptives — a real constraint in adolescent and young-adult patients, and one that interacts with the embryo-fetal toxicity warning (non-hormonal contraception required).
8. Regulatory maturity. Tovorafenib holds accelerated (US) and conditional (EU) authorisations with an unread-out confirmatory trial. A plixorafenib program should be positioned against tovorafenib's 2027 FIREFLY-2 readout (topline expected by mid-2027), not against today's label alone.

---

11. Open flags carried by this workstream

1. [FLAG: target drug never named] — pipeline invoked with the literal placeholder "the competitor drug"; resolved to tovorafenib from Fore's own CI registry (verified: CI_PROCESS.md lists it as the only LIVE dossier target). Confirm before circulation; reconcile with the other six workstreams.
2. [FLAG: plixorafenib column empty] — no figures supplied; every cell is a placeholder. Nothing in this workstream should be read as a plixorafenib claim. One inferred plixorafenib property (a BID continuous regimen, §10 axis 6) was found and removed during fact-check.
3. [FLAG: accelerated-approval conversion status] — no evidence of conversion to full approval found at the 31 Aug 2026 cut; verify against Drugs@FDA.
4. [FLAG: EU indication headline discrepancy] — "regardless of BRAF alteration" press phrasing contradicts the authorised EMA indication text; likely means "regardless of which BRAF alteration," but must not be reproduced.
5. [FLAG: no quantitative CNS-penetration figure in the label] — "CNS-penetrant" is a developer/review/phase-1-publication characterisation, not a label statement and not a comparable number.
6. [FLAG: WAC figures — dead link and selective quotation] — citation [16] 404s on fetch; the filing is a range ($9,675.73–$38,702.90), not a single figure. Re-source; do not annualize without the label §16 carton-to-dose mapping.
7. [FLAG: PMDA and other non-US/EU regions unverified] — absence of evidence, not evidence of absence.
8. [FLAG: OS not reported] — no overall-survival estimate in the label, the Nature Medicine paper, or the SNO 2025 release; verify against the SNO 2025 presentation. Record as "not reported," never "no OS benefit."
9. [FLAG: US orphan exclusivity / Orange Book patents not verified here] — hand off to the competitive_ip workstream and cross-check.
10. [FLAG: label-revision drift affects efficacy as well as safety] — always cite the 8/2025 US label. Growth toxicity moved 15%/Grade 3 5% → 46% of 133/Grade ≥3 35%; the labeled ORR moved 51% → 53%, median DoR 13.8 mo → 18 mo, and baseline race ascertainment was restated (H6, H9). Any secondary source written in 2024 is stale.
11. [FLAG: SNO 2025 figures are from a company press release] — the 6 Jun 2025 cut (ORR 40/76, DoR 19.4 mo, PFS 16.6 mo, TTNT 42.6 mo, 30/39 treatment-free ≥12 mo, retreatment n=8) was verified verbatim against the release but the release is the only source; upgrade to the presented abstract/slide or a peer-reviewed publication before external use.
12. [FLAG (new): criterion-conditional CR count] — 0 CR by RAPNO-LGG (n=76) vs 12 CR (17%) by RANO-HGG (n=69), same trial. Never state either without its criterion (H10, R7).
13. [FLAG (new): median administered dose unverified] — the "~420 mg/m² weekly" median was removed; only the 290–476 mg/m² range is label-supported (H5).
14. [FLAG (new): EU MAA submission date unverified] — "26 Feb 2025" is not on the EMA product page and duplicates the CHMP opinion date one year prior.
15. [FLAG (new): three derived-or-unconfirmed detail figures] — CBR numerator "64/69" (back-calculated from 93%), growth-velocity on-treatment "n=81", and the suspension "300 mg / 12 mL" and food-effect "Tmax 3 h → 6.5 h" specifics were not confirmed in retrieved primary text. Verify or drop.

---

12. Fact-check change log (added 31 Aug 2026)

#	Change	Severity
C1	§10 axis 3 and H-series: "zero complete responses" made criterion-conditional; new H10 and R7 added; new §5 row showing RANO-HGG CR 12/69 (17%), PR 34/69	High — the prior wording was factually false as a bare claim
C2	§10 axis 6: removed the inferred plixorafenib "BID continuous regimen"; reframed as an open question	High — SCOPE guardrail breach
C3	H6 retitled and reframed: "reported incidence rose" rather than "safety profile worsened"; noted the 4/2024 label states no denominator for its 15%	Medium — objectivity
C4	§9 list price: corrected from a single $38,702.90 figure to the $9,675.73–$38,702.90 range; flagged citation [16] as returning HTTP 404 on fetch	Medium — citation integrity
C5	§6 fatal AE row: noted the 1% (N=137) and 0.6% (N=172) figures are the same single event on two denominators	Medium — double-counting risk
C6	§6 discontinuation row: removed "concordant" — label 7% is all-cause, Nat Med 9/137 is treatment-related	Medium — conflation
C7	H5: removed the unsourced "~420 mg/m²" median administered dose	Medium — unsupported figure
C8	New H9 documenting efficacy drift between label revisions (ORR 51%→53%, DoR 13.8→18 mo, demographics restated); §5 annotated with superseded 4/2024 values	Medium — temporal
C9	§8: EMA dates confirmed authoritative (EC 20 Apr 2026, CHMP 26 Feb 2026) against press-release dates (22 Apr, 27 Feb); "MAA submitted 26 Feb 2025" flagged as unverified	Low — temporal precision
C10	§4: added the note reconciling 56 fusions vs the n=64 "fusion or rearrangement" exploratory subgroup	Low — apparent contradiction resolved
C11	§5 CBR row: flagged "64/69" as back-calculated rather than quoted; §6 growth row: flagged "n=81"; §7: flagged suspension volume and food-effect Tmax specifics	Low — derived figures presented as sourced
C12	§8 EU HTA: "first medicine to complete the JCA process" attributed to Ipsen rather than stated as fact	Low — company claim
C13	Enrichment from verification: FIREFLY-2 N≈400 across ~140 sites, four SoC chemo comparators, topline by mid-2027; adult phase 1 Grade ≥3 AE split; Servier 68% premium; 2026 revenue guidance $225–250M; label §5 six-warning list and NF1 nonclinical basis quoted	—
C14	Added verification-status and source-recency header, R3 expanded to five verified denominators, R7 added	—

CITATIONS: ["[1] OJEMDA (tovorafenib) tablets / for oral suspension, US Prescribing Information, revised 8/2025 — DailyMed setid ea3a9631-3a66-6a7c-e053-2995a90ae2ad. https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=ea3a9631-3a66-6a7c-e053-2995a90ae2ad — VERIFIED 31 Aug 2026: sections 1, 2.1-2.3, 4, 5.1-5.6, 6.1, 7, 12.1-12.3 and 14 retrieved and quoted directly; confirmed as the current revision.","[2] Kilburn LB, Khuong-Quang D-A, Hansford JR, et al. The type II RAF inhibitor tovorafenib in relapsed/refractory pediatric low-grade glioma: the phase 2 FIREFLY-1 trial. Nature Medicine 2024;30(1):207-217 (published online 17 Nov 2023). https://pmc.ncbi.nlm.nih.gov/articles/PMC10803270/ — VERIFIED 31 Aug 2026: N=77 enrolled, 69 RANO-HGG-evaluable, ORR 67% (54-78) with 12 CR (17%) and 34 PR (49%), CBR 93%, median DoR 16.6 mo (11.6-NR), median PFS 19.4 mo (16.9-NR), median TTR 3.0 mo (2.6-16.6), Grade >=3 TRAE 42%, 9 (7%) discontinued for TRAEs. Data cutoff 5 Jun 2023.","[3] Day One Biopharmaceuticals. 'Day One Announces Three Year Follow-Up Data From OJEMDA (tovorafenib) Phase 2 FIREFLY-1 Trial at the 2025 Society for Neuro-Oncology (SNO) Annual Meeting.' Press release, 24 Nov 2025. Primary IR posting: https://ir.dayonebio.com/news-releases/news-release-details/day-one-announces-three-year-follow-data-ojemdatm-tovorafenib (mirrors: globenewswire.com/news-release/2025/11/24/3193443, biospace.com) — VERIFIED 31 Aug 2026: ORR 53% (40/76) RAPNO-LGG BICR, median DoR 19.4 mo (13.8-27.2), median PFS 16.6 mo (10.9-22.0), median TTNT 42.6 mo (36.7-NE), 39 entered treatment-free observation with 30/39 (77%) off therapy >=12 mo, retreatment n=8 median 9 mo with -38.3% median maximum tumor reduction, median study duration 40.6 mo, data cutoff 6 Jun 2025, no OS reported. NOTE: press release only; not peer-reviewed.","[4] OJEMDA (tovorafenib) US Prescribing Information, original approval label, revised 4/2024, FDA Reference ID 5368915. https://www.fda.gov/media/180679/download — VERIFIED 31 Aug 2026 (PDF text extracted): section 5.4 'treatment-emergent adverse effects on growth occurred in 15% of patients 18 years of age or younger, including Grade 3 events in 5% of patients... permanently discontinued... in 2% of patients (n=2)'; section 2.1 'An FDA approved test... is not currently available'; section 14 Table 10 ORR 51% (40, 63), CR 0, PR 28 (37%), MR 11 (14%), DoR 13.8 mo (11.3, NE) in N=39, TTR 5.3 mo. Cited only for superseded figures. (The accessdata.fda.gov/drugsatfda_docs/label/2024/218033s000lbl.pdf mirror returned HTTP 404 on 31 Aug 2026.)","[5] Day One Biopharmaceuticals product page, tovorafenib (DAY101) — mechanism, prior identifiers (TAK-580 / MLN2480 / BIIB024), CNS penetration and type II RAF characterisation. https://www.dayonebio.com/tovorafenib-day101/ ; corroborated by the review: 'Type II RAF inhibitor tovorafenib for the treatment of pediatric low-grade glioma.' Expert Review of Clinical Pharmacology 2024;17(11). https://pubmed.ncbi.nlm.nih.gov/39412085/ — company/review source; NOT label-supported for the paradox-avoidance or CNS-penetration claims.","[6] Servier. 'Servier completes the acquisition of Day One Biopharmaceuticals.' Press release, 23 Apr 2026. https://servier.com/en/newsroom/servier-completes-the-acquisition-of-day-one-biopharmaceuticals/ — VERIFIED 31 Aug 2026: $21.50/share, ~$2.5B equity value, 68% premium to the 5 Mar 2026 close, tender offer expired 22 Apr 2026 and closed 23 Apr 2026; Day One markets OJEMDA in the US and has licensed ex-US rights to Ipsen. See also SEC SC TO-T filings, CIK 1845337.","[7] Phase 1 study of the pan-RAF inhibitor tovorafenib in patients with advanced solid tumors followed by dose expansion in patients with metastatic melanoma. Cancer Chemotherapy and Pharmacology, 2023. https://link.springer.com/article/10.1007/s00280-023-04544-5 ; PMC10261210 — VERIFIED 31 Aug 2026: n=149 (Q2D n=110, QW n=39); RP2D 200 mg every other day or 600 mg weekly; Grade >=3 AEs in expansion 58/80 (73%) Q2D vs 9/19 (47%) QW; describes tovorafenib as CNS-penetrant.","[8] Day One Biopharmaceuticals. 'Day One's OJEMDA (tovorafenib) Receives US FDA Accelerated Approval for Relapsed or Refractory BRAF-altered Pediatric Low-Grade Glioma (pLGG).' Press release, 23 Apr 2024. https://ir.dayonebio.com/news-releases/news-release-details/day-ones-ojemdatm-tovorafenib-receives-us-fda-accelerated — corroborated by FDA: https://www.fda.gov/drugs/resources-information-approved-drugs/fda-grants-accelerated-approval-tovorafenib-patients-relapsed-or-refractory-braf-altered-pediatric (VERIFIED: accelerated approval granted 23 April 2024).","[9] van Tilburg CM, Kilburn LB, Perreault S, et al. LOGGIC/FIREFLY-2: a phase 3, randomized trial of tovorafenib vs. chemotherapy in pediatric and young adult patients with newly diagnosed low-grade glioma harboring an activating RAF alteration. BMC Cancer 2024;24:147. NCT05566795. https://bmccancer.biomedcentral.com/articles/10.1186/s12885-024-11820-x ; PMC10826080","[10] Day One / Servier. 'Day One and Servier Complete Enrollment in Pivotal Phase 3 FIREFLY-2 Trial of Tovorafenib as a Front-Line Treatment for Pediatric Low-Grade Glioma (pLGG).' Press release, 8 May 2026. https://www.prnewswire.com/news-releases/day-one-and-servier-complete-enrollment-in-pivotal-phase-3-firefly-2-trial-of-tovorafenib-as-a-front-line-treatment-for-pediatric-low-grade-glioma-plgg-302766277.html — VERIFIED 31 Aug 2026: ~400 participants across ~140 sites, ages 6 months to 25 years, tovorafenib once weekly vs four SoC chemotherapy regimens, primary endpoint ORR (with DoR) by RAPNO-LGG, 'primary analysis is expected to occur approximately 12 months after the last patient is enrolled; preliminary insights are expected to be available in 2027', topline expected by mid-2027.","[11] Foundation Medicine. 'U.S. FDA Approves FoundationOne CDx as a Companion Diagnostic for OJEMDA (tovorafenib).' Press release, 17 Jan 2025. https://www.foundationmedicine.com/press-release/fda-approval-foundationone-cdx-ojemda — VERIFIED 31 Aug 2026: approval date 17 January 2025; first and only CDx for OJEMDA.","[12] European Medicines Agency. Ojemda (tovorafenib) EPAR / medicine overview. https://www.ema.europa.eu/en/medicines/human/EPAR/ojemda — VERIFIED 31 Aug 2026: conditional marketing authorisation, marketing authorisation issued / EC decision 20 April 2026; CHMP opinion 26 February 2026; MAH Ipsen Pharma, 70 rue Balard, 75015 Paris; EMEA/H/C/006140; orphan designation EU/3/21/2434 of 20 May 2021; authorised indication and specific obligation quoted verbatim in section 8. MAA submission date NOT shown on this page.","[13] Ipsen. 'Ojemda approved in the European Union as the first targeted therapy in relapsed or refractory pediatric low-grade glioma regardless of BRAF alteration.' Press release, 22 Apr 2026. https://www.ipsen.com/press-release/ojemda-approved-in-the-european-union-as-the-first-targeted-therapy-in-relapsed-or-refractory-pediatric-low-grade-glioma-regardless-of-braf-alteration-3278647/ ; and 'Ipsen welcomes the European Union's publication of the Joint Clinical Assessment for Ojemda (tovorafenib).' https://www.ipsen.com/update/ipsen-welcomes-the-european-unions-publication-of-the-joint-clinical-assessment-for-ojemda-tovorafenib-3309159/ — headline phrasing flagged as inconsistent with the EMA authorised indication text; JCA 'first' claim is company-sourced and unverified.","[14] Day One Biopharmaceuticals. 'Day One Announces Preliminary 2025 OJEMDA Net Product Revenue And Provides 2026 Net Product Revenue Guidance.' Press release, 11 Jan 2026. https://ir.dayonebio.com/news-releases/news-release-details/day-one-announces-preliminary-2025-ojemdatm-net-product-revenue — VERIFIED 31 Aug 2026: $155.4M preliminary FY2025 net product revenue, +172% YoY, 2026 US guidance $225-250M; confirmed in the Q4/FY2025 results release (Feb 2026).","[15] Managed Healthcare Executive. 'FDA Approves Ojemda for Children with Brain Tumors.' https://www.managedhealthcareexecutive.com/view/fda-approves-ojemda-for-children-with-brain-tumors — VERIFIED 31 Aug 2026: launch WAC ~$33,916 per 28-day supply; annual cost reported to top $440,000; specialty distribution via Biologics by McKesson and Onco360. Secondary trade-press source.","[16] Day One Biopharmaceuticals. 'Information for Colorado Prescribers of Prescription Drugs' — OJEMDA WAC effective 1 Jan 2026. https://www.dayonebio.com/wp-content/uploads/Ojemda-Colorado_01.01.26.pdf — RETURNED HTTP 404 ON DIRECT FETCH 31 Aug 2026. Content corroborated only via search-index snapshot, which gives a WAC range of $9,675.73 to $38,702.90 across tablet configurations. MUST BE RE-SOURCED before external use (try the Colorado DOI drug-transparency filings or the equivalent Connecticut/North Dakota filings, e.g. https://www.dayonebio.com/wp-content/uploads/Ojemda-Connecticut_VF.pdf).","[17] Kline C, et al. Type II RAF inhibitor tovorafenib in relapsed/refractory pediatric low-grade glioma (pLGG): reversible decreases in growth velocity in the phase 2 FIREFLY-1 trial. J Clin Oncol 2024;42(16_suppl):10036 / Neuro-Oncology LGG-40. https://ascopubs.org/doi/10.1200/JCO.2024.42.16_suppl.10036 ; PMC11183845 — earlier-cut growth-velocity analysis; SUPERSEDED by the 8/2025 label figures. Do not cite for current incidence.","[18] Fore internal process document (not a data source): C:\Users\Owner\Documents\Hugh Context Folder\Fore\competitive_intelligence\CI_PROCESS.md — VERIFIED 31 Aug 2026: registry table lists 'Tovorafenib (Ojemda) | Type II pan-RAF | LIVE (draft, 41 flags)' as the only LIVE entry; dabrafenib+trametinib and selumetinib are 'Planned'. Used only to resolve the missing drug-name argument, and to confirm the plixorafenib-data guardrail quoted in section 0."]