Lirafugratinib: FDA approval and the ReFocus evidence

September 23, 2026 / What was approved

FDA approved Lyrfigtu (lirafugratinib) for adults with previously treated, unresectable, locally advanced or metastatic cholangiocarcinoma with an FGFR2 fusion or other rearrangement. This completed event enters History directly. The September 22 announcement concerned only conditional acceptance of the proposed brand name; it was not drug approval. The earlier September 25 target therefore remains a historical expectation, not the date of the actual action. [1] [2] [5] [6]

Elevar is the licensee responsible for further development and commercialization; Nasdaq-listed Relay is the licensor. The RLAY association identifies that contractual relationship, not a second approval or a claim that Relay directly commercializes the medicine. FDA’s notice does not establish a tissue-agnostic indication or efficacy after a prior FGFR inhibitor. Those populations require their own evidence.

Biology / FGFR2 selection

FGFR2 is a receptor tyrosine kinase. A fusion or rearrangement can create persistent growth signaling, giving a selected tumor a targetable dependency. Lirafugratinib was developed as an irreversible, relatively selective FGFR2 inhibitor. The distinction between a fusion, a mutation and amplification matters: they are different molecular findings, and the approved population cannot be expanded by treating every FGFR2 alteration as equivalent. [2] [4] [8]

Analytical interpretation: molecular matching explains why this study selected patients, but does not guarantee response. Prior therapy, coexisting alterations and acquired resistance may influence the result. The pivotal study’s good-performance-status population also limits extrapolation to people with substantial functional impairment. Molecular eligibility and fitness for treatment are separate questions.

Molecule / Dose and PK–PD limits

The approved regimen is 70 mg orally once daily until progression or unacceptable toxicity. ReFocus established a continuous once-daily regimen through dose escalation before expansion. Laboratory selectivity is a pharmacological property; it does not mean the drug has no toxicity in normal tissues. The clinical frequency of oral, skin, nail and eye events must be evaluated alongside the target-selectivity rationale. [1] [4] [8]

This report does not have a verified final-label exposure–response model, receptor-occupancy threshold, food-effect estimate or full interaction table. It therefore does not infer dose changes from biochemical potency. Frequent interruptions and reductions in the clinical evidence are especially relevant: the prescribed starting dose and the dose actually sustained over time are not interchangeable measures of exposure.

ReFocus / Trial and analysis populations

ReFocus (NCT04526106) is a multicenter, open-label Phase 1/2 program without a randomized comparator. Its pivotal cholangiocarcinoma cohort had prior systemic therapy and no prior FGFR inhibitor. The original poster specifies adults with measurable disease and ECOG performance status 0–1. The historical pivotal dataset contains 116 treated participants, while the independent-review primary efficacy set contains 114; two were unavailable for that analysis. This is not evidence that exactly two patients withdrew. [1] [3] [4] [7]

Confirmed response by independent RECIST v1.1 review was the primary endpoint; duration of response was a key secondary measure. The original planned pivotal sample, complete missing-scan rules, treatment-policy estimand and full statistical analysis plan were not recovered. Independent review reduces assessment discretion but cannot supply the counterfactual outcome that randomization would provide. No background anticancer combination or control arm is assumed.

Prior clinical evidence / Response and durability

The September 27, 2024 analysis reported three complete and fifty partial responses among 114 independently reviewed patients. FDA later summarized response using 116 patients and a rounded 46% estimate. The two versions are retained separately; this report does not assert that the difference represents improved or worsened efficacy. A final review document is needed to reconcile denominator and confidence-interval conventions. [1] [3] [4]

Median response duration was 11.8 months, with a 95% interval of 7.5–13.0 months. Response duration describes responders, not all treated participants. Historical median progression-free survival was 11.3 months (9.2–14.8). Neither statistic demonstrates superiority over another FGFR inhibitor, chemotherapy or supportive care, because the study did not randomize patients between those alternatives.

The ASCO abstract reports median overall survival of 22.8 months with a 17.3–27.2 interval; the later conference poster gives 18.1–27.2 for the same median. The reason for that interval discrepancy remains unresolved. Both source versions are identified here rather than silently replacing one. Follow-up, censoring and subsequent treatment must be understood before interpreting survival.

Safety / Treatment burden and denominators

The original poster separates the pivotal safety cohort of 116 from the broader 385-person solid-tumor safety population. The latter contains the former and is not an independent replication. In the pivotal cohort, every participant had a treatment-related adverse event and 67 had a grade 3 or higher treatment-related event. Related deaths were zero; this does not establish zero all-cause deaths or zero serious adverse events. [1] [3] [4]

High rates of interruption and reduction make tolerability an important part of the efficacy interpretation. Sustained treatment can require substantial dose management even when permanent discontinuation is less common. FDA’s approval summary highlights ocular toxicity, hyperphosphatemia with soft-tissue mineralization, and embryo-fetal toxicity. The final prescribing information and regulatory safety review were not recovered for independent extraction; the historical poster is not substituted for them.

Capital / Licensing and research execution

Relay’s June 30, 2026 license note reports $18.7 million of consideration to date: $5.0 million at execution, $3.7 million for transferred materials and $10.0 million of milestones. Potential remaining regulatory and commercial milestones of up to $485 million are contingent, not cash in hand or an estimate of research funding. The agreement assigns further development to Elevar, so Relay’s corporate cash should not be presented as a dedicated budget for this program. [5] [9]

Execution now requires publication of the final label and review, reliable dose-management guidance and completion of the remaining clinical evidence. A separate non-cholangiocarcinoma trial, NCT07359820, has been described, but its planned enrollment and future results cannot enlarge the approved label. This dossier has not independently established Elevar’s unrestricted cash or a fully diluted capitalization; no financing sufficiency conclusion is drawn.

Sources / References

1. FDA · Lyrfigtu approval · September 23, 2026; retrieved September 24
2. Elevar · Approval announcement · September 23, 2026
3. Hollebecque et al. · ASCO GI 2026 abstract 476 · analysis cutoff September 27, 2024
4. Hollebecque et al. · ReFocus original poster · CCF May 1–3, 2026; September 27, 2024 analysis
5. Relay · SEC June 30, 2026 financial statements · Elevar license note
6. Elevar · Conditional brand-name acceptance, not approval · September 22, 2026
7. ClinicalTrials.gov · ReFocus NCT04526106
8. Elevar · Non-confidential scientific presentation · February 2026
9. ASCO 2026 · Non-CCA study design, TPS4251; separate NCT07359820