Tovorafenib

Tovorafenib is approved in the United States (accelerated approval, 23 April 2024), the European Union (conditional marketing authorisation, April 2026), and Japan (16 September 2026). The US and EU indications are limited to relapsed or refractory (EU: progressed after one or more prior systemic therapies) pediatric low-grade glioma with a BRAF fusion or rearrangement or BRAF V600 mutation, based on the single-arm phase 2 FIREFLY-1 trial. The FDA listed the accelerated approval as ongoing (clinical benefit not yet verified) as of 16 September 2026. No approved indication, supplemental new drug application, or EU variation for first-line use was identified. The first-line trial LOGGIC/FIREFLY-2 (tovorafenib vs standard-of-care chemotherapy) completed enrollment of approximately 400 participants in May 2026; the sponsor expects topline data by mid-2027. FIREFLY-2 is also the confirmatory trial for the US accelerated approval (interim report due April 2028, final report April 2032) and the specific obligation for the EU conditional authorisation (final report due 30 April 2032). The EMA overview page gives 20 April 2026 as the authorisation date, while the Ipsen announcement is dated 22 April 2026. Day One Biopharmaceuticals became a wholly owned subsidiary of Servier after the tender offer closed on 22 April 2026.

The US and EU labels restrict tovorafenib to relapsed or refractory disease (US) or disease that progressed after one or more prior systemic therapies (EU). Neither label includes first-line (treatment-naive) systemic therapy of pLGG, so first-line use in the US and EU is outside the approved indication until FIREFLY-2 results are reviewed. The Japanese indication statement issued with the September 2026 approval (low-grade glioma with a BRAF gene mutation or fusion gene) does not name a line of therapy, although the supporting data came from FIREFLY-1 in relapsed or progressive disease.

Tovorafenib is an oral, CNS-penetrant type II RAF kinase inhibitor of BRAF V600E, wild-type BRAF, and wild-type CRAF, with in vitro IC50 values of 7.1, 10.1, and 0.7 nM, respectively. Type I BRAF inhibitors such as dabrafenib act on monomeric BRAF V600E, do not block signaling from the dimer-dependent KIAA1549:BRAF fusion, and can cause paradoxical MAPK pathway activation in BRAF fusion-driven or BRAF wild-type tumors. Tovorafenib binds RAF monomers and dimers; preclinical studies in pediatric low-grade astrocytoma models showed equipotent activity against BRAF V600E and KIAA1549:BRAF with brain penetrance, and the EMA assessment found no paradoxical activation in BRAF fusion models. Tovorafenib is relatively sparing of ARAF, and in NF1 loss-of-function models without a BRAF alteration it increased ERK phosphorylation at low concentrations, which underlies the NF1-associated tumor warning. The mechanistic data apply to both the approved relapsed or refractory indication and the investigational first-line setting (FIREFLY-2).

Based on population PK modelling, tovorafenib at the labeled dose has a steady-state Cmax of 6.9 µg/mL and AUC of 508 µg·h/mL, reaches steady state in 12 days, and has a terminal half-life of approximately 56 hours with no clinically significant accumulation on once-weekly dosing. Median Tmax is 3 hours for tablets and oral suspension; a high-fat meal delays Tmax to 6.5 hours without clinically significant changes in Cmax or AUC. Plasma protein binding is 97.5%. Aldehyde oxidase and CYP2C8 are the primary metabolic enzymes (CYP2C8 fraction metabolized estimated at 49.23% in the PBPK model), and 65% to 66.1% of a radiolabeled dose is recovered in feces and 27% to 28.7% in urine, with less than 0.2% excreted unchanged in urine. In adults (phase 1, 600 mg once weekly), the mean terminal half-life was approximately 70 hours. No formal PK studies were conducted in the pediatric glioma population; exposure was estimated from sparse sampling, and the EMA considered predictions for participants younger than 2 years unreliable (three observations), requiring additional PK data by 30 April 2032. The popPK model estimated 21.5% higher apparent clearance in males than in females.

All labeled safety data derive from relapsed or refractory disease (FIREFLY-1, 137 participants with pLGG; pooled population of 172 participants including adults with solid tumors at a 600 mg flat dose); the first-line FIREFLY-2 trial is active, not recruiting, with primary completion estimated for June 2027, and no first-line safety results were identified. The US label has no contraindications and no boxed warning; the EU SmPC lists hypersensitivity as the only contraindication. Warnings and precautions cover hemorrhage (37% of the pooled population, including intratumoral hemorrhage in 9% and one fatal tumor hemorrhage), skin toxicity (rash 67%, grade 3 rash 12%; photosensitivity 12%), hepatotoxicity (increased AST 74% and ALT 42%), effect on growth (46% of 133 participants aged 18 years or younger, 35% grade 3 or higher), embryo-fetal toxicity, and possible tumor growth in NF1. In FIREFLY-1, serious adverse reactions occurred in 45% of participants; adverse reactions led to dose interruption in 57%, dose reduction in 24%, and permanent discontinuation in 7%. Growth velocity recovered after treatment stopped in 34 of 38 evaluable participants (89%) in a pooled analysis of three studies. The EU risk management plan lists growth retardation and intratumoral hemorrhage as important identified risks, decreased fertility as an important potential risk, and long-term safety as missing information.

The approved US and EU dose is 380 mg/m2 orally once weekly (maximum 600 mg), with the 100 mg tablet used for BSA of 0.90 m2 or more and the 25 mg/mL oral suspension for BSA of 0.30 to 0.89 m2; no dose is established for BSA below 0.3 m2. Treatment continues until disease progression or intolerable toxicity (EU: also loss of clinical benefit). The approved dose is lower than the 420 mg/m2 dose studied in PNOC014, FIREFLY-1 (actual range 290 to 476 mg/m2), and the first-line FIREFLY-2 trial; FDA selected 380 mg/m2 because exposure-response analyses showed no clinically significant relationship with response across the studied range, while higher exposure was associated with rash, liver enzyme elevation, and creatine phosphokinase elevation. The approved indications concern relapsed or refractory disease; no first-line dosing recommendation exists outside FIREFLY-2.

Labeled supportive measures comprise ultraviolet protection (sunscreen, sunglasses, protective clothing; the EU SmPC specifies SPF 50 or higher), dermatologic consultation and supportive care for skin reactions, liver function monitoring before treatment, at 1 month, and every 3 months, routine growth monitoring, and effective nonhormonal contraception. In FIREFLY-1 (relapsed or refractory pLGG), the protocol recommended a gentle skin care routine with thick unscented moisturizers and SPF 30 or higher sunscreen for all participants, and 102 of 128 participants with rash (80%) received standard rash treatments, primarily topical steroids or antibiotics, oral antihistamines, and emollients. In a pooled growth analysis of 38 participants, 4 (11%) received a gonadotropin-releasing hormone analogue and 2 (5%) received growth hormone concurrently with tovorafenib. No published supportive care guideline specific to first-line tovorafenib use was identified.

No randomized comparison of tovorafenib with any other therapy has been reported as of September 2026. The phase 3 LOGGIC/FIREFLY-2 trial, which compares tovorafenib with investigator's choice of standard chemotherapy in first-line RAF-altered pLGG, completed enrollment of approximately 400 participants in May 2026, with preliminary results expected in 2027. The only formal indirect comparison identified is the EU Joint Clinical Assessment (JCA) of tovorafenib in relapsed or refractory pLGG, endorsed on 30 April 2026. Of 8 PICO questions, the developer submitted comparative data for 2 (PICO 5 and PICO 7), and the assessors included 1: an unanchored matching-adjusted indirect comparison (MAIC) of FIREFLY-1 against a phase I/II dabrafenib plus trametinib study in the BRAF V600E subpopulation. The effective sample size was 5.81 to 14.26. For independent review committee (IRC)-assessed progression-free survival (PFS) by RANO-LGG, the hazard ratio was 4.88 (95% CI 2.14 to 11.14; nominal p = 0.011), a direction unfavorable to tovorafenib; objective response comparisons were inconclusive; and the odds ratio for grade 3 or higher adverse events was 12.65 (95% CI 2.30 to 69.58). The assessors stated that these estimates "should not necessarily be interpreted as causal effects of treatment." The comparison with trametinib (TRAM-01) was excluded because the comparator was reported only in conference abstracts.

Independent reviews published in 2025 and 2026 compare agents on mechanism and indication rather than head-to-head efficacy. Type I BRAF inhibitors (dabrafenib, vemurafenib) are restricted to BRAF V600-mutant tumors because of paradoxical MAPK activation in BRAF fusion-driven tumors, whereas tovorafenib is active in both fusion and V600 tumors. Reviewers describe once-weekly dosing without food restrictions and an available oral suspension for tovorafenib, compared with once- or twice-daily dosing for MEK and BRAF inhibitors. Reviewers also describe differences in adverse event profiles: reduced growth velocity has been reported only with tovorafenib, and intratumoral hemorrhage was reported in 9% of the pooled pediatric tovorafenib safety population, while pyrexia is more frequent with dabrafenib plus trametinib and cardiac and ocular monitoring is emphasized for MEK inhibitors. A 2026 systematic review found rebound regrowth after MAPK inhibitor discontinuation in at least 23 of 131 evaluable participants (17.6%) and 1 rebound event among 24 evaluable FIREFLY-1 participants who entered a drug holiday, while noting short post-discontinuation follow-up.

Pediatric low-grade glioma (pLGG) is a heterogeneous group of WHO CNS grade 1 and 2 glial, glioneuronal, and neuronal tumors that account for 30% to 40% of pediatric CNS tumors. The 2021 fifth edition of the WHO classification (WHO CNS5) distributes these tumors across three families: pediatric-type diffuse low-grade gliomas (4 types, including diffuse low-grade glioma, MAPK pathway-altered), circumscribed astrocytic gliomas (including pilocytic astrocytoma, the most common type), and glioneuronal and neuronal tumors. Diagnosis is integrated and layered, combining histology with molecular findings.

The disease is driven predominantly by alterations that activate the RAS/MAPK pathway. In a population-based cohort from Toronto with molecular profiling of 477 tumors, a driver alteration was identified in 84%; KIAA1549-BRAF fusion was the most frequent alteration (35%), followed by BRAF p.V600E (17%), with germline NF1, FGFR1/2 (6.1%), and rare RAF1, NTRK2, ALK, and ROS1 alterations accounting for most of the remainder. Whole-genome sequencing of 96 pilocytic astrocytomas found MAPK pathway alterations in 100%. Germline NF1 is the main hereditary predisposition, with 15% to 20% of children with NF1 developing an optic pathway glioma.

No identified publication reports the incidence of RAF-altered pLGG requiring first-line systemic therapy directly. The available components are as follows.

• Overall incidence, United States: the CBTRUS childhood and adolescent report (2017 to 2021 diagnoses) reports an annual incidence of 0.98 per 100,000 for pilocytic astrocytoma in ages 0 to 19 years and projects 730 new pilocytic astrocytoma cases in 2025. The LOGGIC/FIREFLY-2 design publication cites 1,200 to 1,500 new pLGG cases per year in the United States. Day One Biopharmaceuticals cites an annual incidence of 1.3 to 2.1 per 100,000.
• Overall incidence, Europe: in Great Britain, the incidence of pilocytic astrocytoma at ages 0 to 14 years was 8.6 per million in 2006 to 2010.
• RAF-altered fraction: KIAA1549-BRAF fusion was found in 35% and BRAF p.V600E in 17% of profiled tumors in a population-based Toronto cohort. Day One Biopharmaceuticals estimates that approximately 1,100 patients under 25 years of age are newly diagnosed with BRAF-altered pLGG in the United States each year.
• Fraction requiring nonsurgical therapy: in the UK CCLG CNS9702 population-based cohort, 165 of 639 children (25.8%) started nonsurgical treatment at diagnosis (113 chemotherapy, 52 radiotherapy). In the German HIT-LGG-1996 cohort of 1,031 children, 184 proceeded to nonsurgical treatment at diagnosis and 35.2% ultimately received nonsurgical treatment. In SIOP-LGG 2004, 1,057 of 3,417 registered, previously untreated patients (30.9%) received chemotherapy. Reviews state that up to half of patients require adjuvant therapy.

No identified publication reports the prevalence of RAF-altered pLGG requiring first-line systemic therapy. CBTRUS projects that 42,864 children and adolescents aged 0 to 19 years in the United States were living with a primary brain or other CNS tumor in 2025, including 21,591 with glioma and 9,193 with pilocytic astrocytoma; these figures exclude survivors who have reached adulthood. Day One Biopharmaceuticals reports a registry-based estimate (SEER and CBTRUS) of 26,000 patients under 25 years of age with BRAF-altered pLGG in the United States as of January 1, 2017, and an addressable pool of 2,000 to 3,000 patients per year with recurrent, progressive, or refractory pLGG at steady state.

Long-term overall survival in pLGG is high, and tumor-related mortality is uncommon. In 4,040 children in SEER (1973 to 2008), 20-year overall cancer-specific survival was 87% and the 20-year cumulative incidence of death due to glioma was 12%, falling to 7% after age 22 years. In a population-based Ontario cohort of 1,202 children, 93 deaths (7.7%) occurred over a median of 12.73 years, with 20-year overall survival of 90.1%. Upfront radiotherapy was associated with higher late mortality in both cohorts (hazard ratios 3.9 and 3.3).

Progression is frequent among children who need systemic therapy. In COG A9952 (274 randomized children younger than 10 years), 5-year event-free survival was 39% with carboplatin and vincristine and 52% with TPCV, and was 19% in children younger than 1 year. In SIOP-LGG 2004 (497 randomized non-NF1 patients), 5-year progression-free survival was 46% with vincristine and carboplatin and 45% with added etoposide. In HIT-LGG-1996, 10-year progression-free survival after chemotherapy was 44%. Approximately 40% to 50% of children require subsequent chemotherapy lines.

Molecular subtype modifies prognosis. In the Toronto cohort, 5-year progression-free survival was 69% for KIAA1549-BRAF and 52% for BRAF p.V600E tumors. BRAF V600E tumors had 10-year progression-free survival of 27% versus 60.2% for wild-type tumors, and 5-year progression-free survival after first-line chemotherapy of 30.4%. BRAF p.V600E with CDKN2A deletion was associated with 10-year progression-free survival of 0% and 10-year overall survival of 41%. Low-risk tumors (fusions or germline NF1) often stop growing, with few progressions after 10 years. Malignant transformation was observed in fewer than 1% of the HIT-LGG-1996 cohort.

pLGG is described as predominantly a single-pathway disease in which somatic or germline alterations converge on constitutive activation of the RAS/RAF/MEK/ERK (MAPK) pathway. KIAA1549::BRAF fusions arise from a tandem duplication at 7q34 that removes the N-terminal regulatory domain of BRAF, producing RAS-independent, dimeric RAF signaling. BRAF p.V600E acts as a phosphomimetic that produces constitutive monomeric kinase activity. Germline NF1 loss removes a negative regulator of RAS. Tumors without an identified driver show similar levels of MAPK pathway up-regulation. Sustained MAPK activation also induces oncogene-induced senescence mediated by p16(INK4a) and the senescence-associated secretory phenotype, which was found in 46 of 52 pilocytic astrocytomas (88.5%) and is proposed to explain slow growth and rare progression. CDKN2A deletion, which co-occurs with BRAF p.V600E, is associated with escape from senescence. Type I BRAF inhibitors are limited to BRAF V600E tumors because they can paradoxically activate MAPK signaling in RAF fusion-driven tumors.

Diagnosis combines MRI with histopathology and molecular testing in an integrated, layered report, as endorsed by WHO CNS5. Molecular assays used for pLGG include immunohistochemistry for BRAF p.V600E, fluorescence in situ hybridization, RNA-based multiplex fusion panels (NanoString), next-generation sequencing panels, and DNA methylation profiling. Consensus reviews state that testing must detect structural variants and fusions, copy-number alterations, and single nucleotide variants. In a German population-based cohort of more than 1,200 children, multi-omic integration refined the diagnosis through a DNA methylation class in 50% and detected diagnostic or therapeutically relevant alterations in 47%. In the United States, FoundationOne CDx was approved as a companion diagnostic for OJEMDA in January 2025. Biopsy may be omitted for visual pathway gliomas with characteristic MRI findings, particularly in children with NF1; 115 of 1,031 children (11.2%) in HIT-LGG-1996 and 68 of 497 patients (13.7%) in the SIOP-LGG 2004 randomized trial had no histological verification.

Among 240 five-year survivors in a St. Jude cohort of 361 children, the 15-year cumulative incidence was 18% for monocular blindness, 22% for hearing loss, 29% for growth hormone deficiency, 33% for thyroid hormone deficiency, 26% for ACTH deficiency, 53% for overweight or obesity, and 38% for at least one seizure; 34% of 182 survivors tested had an IQ below 85, compared with 16% in the normative sample. Among children with progressive hypothalamic/chiasmatic tumors, the cumulative incidence of monocular blindness was 57% at 15 years. In a Dutch nationwide cohort of 117 children treated for optic pathway glioma, 18.8% had binocular severe visual impairment or blindness after a median of 8.3 years. In the Childhood Cancer Survivor Study, survivors of astrocytoma and glial tumors showed radiotherapy dose-dependent impairment in attention, processing speed, and memory, and cranial radiotherapy of 50 Gy or more was associated with a 25-year cumulative incidence of CNS subsequent neoplasms of 7.1%, compared with 1.0% without radiotherapy. Chemotherapy is associated with myelosuppression, peripheral neuropathy, allergic reactions, and other toxicities; grade 3 to 4 peripheral nervous system toxicity occurred in 19% of children receiving carboplatin and vincristine in COG A9952.

No study was identified that reports health-related quality of life (HRQoL) specifically in RAF-altered pLGG at the start of first-line systemic therapy, and no published patient-reported outcome (PRO) comparison of first-line chemotherapy versus a MAPK pathway inhibitor was located. Available evidence comes from survivorship cohorts, optic pathway glioma (OPG) cohorts, single-arm chemotherapy trials, and HTA testimony.

In a Mayo Clinic cohort of 121 survivors (median follow-up 21.9 years), mean EORTC QLQ-C30 global QoL was 78, compared with 76.4 in a healthy adult reference population; post-operative radiation, deep tumor location, and recurrence were associated with poorer HRQoL, and participants with recurrence reported more financial problems. In a German survey of 49 survivors, survivors rated their QoL higher than peers, but their parents did not. Among 36 children aged 10 years or younger with OPG, parent-proxy vision-specific QoL was lower with greater vision loss. In a Canadian phase 2 trial of weekly vinblastine in 54 chemotherapy-naive children, QoL was reported as not affected during 70 weeks of intravenous treatment. The ongoing phase 3 LOGGIC/FIREFLY-2 trial is collecting PedsQL-Core, PedsQL-Cancer, and PROMIS measures in both arms. In the NICE appraisal of dabrafenib plus trametinib (TA977), patient experts described regular hospital travel, costs, and time commitments with current treatments, and no child-specific utility values were identified; the committee judged that adult utility decrements likely underestimate the HRQoL loss in children.

Caregiver burden is documented for pediatric brain tumors generally: 85% of 40 caregivers of children newly hospitalized with a brain tumor rated distress 5 or greater on a 0 to 10 scale, and in a five-site matched study of 301 caregivers of brain tumor survivors 1 to 5 years after treatment, mothers reported higher caregiver burden, including financial impact, than mothers of comparison youth.

No study reporting total direct medical costs of pLGG care in the United States was identified. A Day One-sponsored U.S. study of linked claims and electronic health records (Optum Market Clarity) followed 154 patients aged 18 years or younger with pLGG for 36 months: most patients (95% to 98%) had office or outpatient visits in each 6-month interval, 18% to 28% visited an emergency room, and 9% to 27% had an inpatient stay; mean inpatient stay duration increased from 5.1 days at baseline to 8.1 days during follow-up. Over the study period, 74% of patients filled prescriptions for anti-infectives, 56% for antiemetics, and 65% received pLGG-directed treatment (most often brain surgery); chemotherapy was recorded in 14%. Tumor histology and molecular status could not be characterized.

For long-term survivorship, a Taiwanese national claims study of 33,105 five-year survivors of childhood cancer or benign brain tumor (median follow-up 7 years) reported median annual medical expenses of US $614.08 for brain cancer survivors (n = 2,241) and US $336.89 for benign brain tumor survivors (n = 7,825), versus US $203.90 for 64,754 matched individuals without cancer. No publication was identified that quantifies infusion visit, central venous access, or hospitalization costs of first-line carboplatin and vincristine or vinblastine in pLGG.

No study was identified that measures family financial burden specifically in pLGG. Evidence from pediatric brain tumors and pediatric cancer overall indicates the following. At a U.S. cancer center, 16% of 202 patients younger than 21 years who underwent neurosurgery for a cancer-related diagnosis had financial toxicity before surgery, and the 2-year cumulative incidence after surgery was 19%; among 86 patients who survived at least 2 years, 34% experienced financial toxicity, and most affected patients had commercial insurance. In a U.S. single-site survey of 254 caregivers of children with cancer, approximately one third reported that a parent quit or changed work because of the child's cancer, and this was associated with a 13.4-point higher financial burden score (0 to 100 scale) at 1 to 5 years after diagnosis. A systematic review of 35 studies reported frequent parental job loss or job quitting, particularly among mothers, with effects persisting into early survivorship. In a Canadian cohort of 28 families, median first-year indirect costs were $9,668 and median income loss was 55 days. In the Childhood Cancer Survivor Study (3,023 adult survivors; 16% with CNS tumors), survivors with 4 impaired neurocognitive domains had odds ratios of 3.12 for debt collection and 3.77 for bankruptcy compared with survivors with no impaired domains.

Tovorafenib use in relapsed or refractory pLGG requires a BRAF fusion or rearrangement or BRAF V600 mutation, and FoundationOne CDx (a tissue-based next-generation sequencing test of 324 genes) was approved by the FDA as its companion diagnostic in January 2025. The Medicare advanced diagnostic laboratory test payment for FoundationOne CDx (code 0037U) was $3,500.00 during its initial period (July 1, 2018 to March 31, 2019); a current 2026 payment rate was not retrieved. No U.S. cost-effectiveness analysis of molecular testing for tovorafenib eligibility was identified.

Reported per-sample costs in other settings were US$1,000 for an outsourced RNA pan-cancer NGS panel (Jordan, 32 pediatric CNS tumors) and, for 82 pediatric LGGs, an average of $405 with a tiered approach (immunohistochemistry, targeted fusion panel or FISH, then targeted RNA sequencing) versus $745 with upfront targeted RNA sequencing. A Canadian microsimulation cost-utility analysis of BRAF fusion testing at diagnosis (100,000 simulated patients; 2018 Canadian dollars; targeted therapies not modeled) found testing dominant, with a gain of 0.38 quality-adjusted life-years (QALYs) and a $1,384 reduction in lifetime costs, driven by avoided radiation-related adverse events.