lirafugratinib

Lirafugratinib (formerly RLY-4008) is a covalent, highly selective FGFR2 inhibitor that Relay Therapeutics out-licensed to Elevar Therapeutics in December 2024 for global development, with Relay eligible for up to $500 million in milestones plus double-digit royalties [7]. The most important credentialing datapoint is from Relay's ReFocus Phase 1/2 (NCT04526106): at the 70 mg recommended Phase 2 dose in FGFR-inhibitor naive FGFR2-fusion cholangiocarcinoma (CCA), confirmed ORR was 82.4%, and in the pivotal chemo-exposed cohort IRC-assessed ORR was 46.5% with a median duration of response of 11.8 months [9]. Elevar filed a CCA NDA that FDA accepted with priority review, PDUFA target action September 27, 2026 [11]. In parallel, Elevar is running a Phase 2 basket study (NCT07359820) in patients with non-CCA solid tumors carrying FGFR2 fusions or rearrangements, targeting 30 patients with objective response rate (ORR) as the primary endpoint [3]. The commercial thesis rests on two ideas: selectivity for FGFR2 over other FGFR family members should reduce the hyperphosphatemia and diarrhea that dog pan-FGFR drugs, and FGFR2 fusions occur across many rare tumor types beyond the crowded CCA market where pemigatinib and futibatinib already compete [5][6]. That thesis is now Elevar's to prove, and Relay retains upside through royalties without carrying the development burn.

Novel small molecule, awaiting first approval. The CCA registrational program is at NDA stage: Elevar's NDA was accepted with priority review and a PDUFA target action date of September 27, 2026, for second-line treatment of pretreated CCA with FGFR2 fusions or rearrangements [11]. This is the near-term binary event for the entire out-license thesis. In parallel, Elevar is running the Phase 2 non-CCA basket study NCT07359820, sponsored by Elevar Therapeutics and currently recruiting toward an enrollment target of 30 patients [3]. Public FDA designations for the non-CCA basket trial have not been identified in current filings; FDA priority review attaches to the specific CCA indication and cannot be assumed to extend to non-CCA solid tumors. A definitive readout timeline for the non-CCA population has not been publicly disclosed. Given the small single-arm design, the basket trial will function as a signal-finding study rather than a filing-ready dataset; a strong ORR would enable expansion cohorts or a registrational basket. Post-deal context: Relay Therapeutics announced significant restructuring and headcount reductions in 2025 as it refocused around its remaining pipeline [10], which affects the assumption that Relay 8-K flow will serve as a clean, timely disclosure vehicle for Elevar's operational milestones. Investors modeling milestone triggers should also track Elevar's own communications where available.

FGFR2 (fibroblast growth factor receptor 2) sits on the surface of cells and receives growth signals from fibroblast growth factors. Under normal conditions, when a growth factor binds, the receptor briefly fires and then shuts off, telling the cell it is time to grow or migrate. In some cancers, particularly a subset of intrahepatic cholangiocarcinoma (bile duct cancer), the FGFR2 gene fuses with a partner gene (BICC1 is the most common) and produces a receptor stuck in the 'on' position, constantly signaling for the cell to divide. Blocking FGFR2 kinase activity in these tumors starves the cancer of its main growth signal. The mechanism is well validated: pemigatinib (Pemazyre, Incyte, 2020) and futibatinib (Lytgobi, Taiho, 2022) both won FDA accelerated approval in FGFR2-fusion cholangiocarcinoma on this exact logic [5][6]. The differentiator for lirafugratinib is not its binding mode per se (like futibatinib, lirafugratinib is a covalent, irreversible inhibitor, binding Cys491 in the ATP pocket), but rather how it achieves selectivity: molecular dynamics work identified differential motion of the P-loop between FGFR1 and FGFR2, and the drug was designed so that covalent bond formation is efficient in FGFR2 but disfavored in FGFR1 [12]. This gives roughly 250-fold selectivity over FGFR1 and about 5,000-fold over FGFR4 [12]. That matters because pan-FGFR drugs cause hyperphosphatemia through FGFR1 blockade (FGFR1 regulates phosphate handling via the FGF23 axis), and diarrhea and hepatotoxicity through FGFR4. A cleaner tolerability profile allows sustained dosing at levels that hit the FGFR2 fusion protein hard, which is the setup Elevar is now testing outside CCA.

NCT07359820 is a single-arm Phase 2 study enrolling 30 patients with non-cholangiocarcinoma solid tumors that carry an FGFR2 fusion or rearrangement [3]. Primary endpoint is objective response rate (ORR) assessed by independent review committee per RECIST v1.1, the standard imaging-based response criteria in solid tumors. Sponsor is Elevar Therapeutics; the trial is currently recruiting. The design rests on the ReFocus Phase 1/2 (NCT04526106) result under Relay: in FGFR-inhibitor naive FGFR2-fusion CCA patients treated at the 70 mg RP2D, confirmed ORR was 82.4%, and in the pivotal chemo-exposed CCA cohort IRC-assessed ORR was 46.5% with median DOR of 11.8 months [9]. Those numbers establish a clean mechanism-hits-target proof point in one tumor type and are the foundation for the tumor-agnostic extension. Design rationale: FGFR2 fusions occur outside CCA at very low prevalence across many tumor types (breast, gastric, pancreatic, gallbladder, and others), which makes a tumor-agnostic basket the only practical way to study efficacy in this population. The risk with n=30 across mixed tumor histologies is statistical: a single ORR number pooled across biology-distinct tumors can mask heterogeneity where FGFR2-driven addiction varies. Registered basket trials for tissue-agnostic labels (larotrectinib in NTRK fusions, entrectinib) succeeded when ORR was in the 60% to 80% range across histologies; a mixed 20% to 40% signal would leave the label pathway uncertain and probably require an expansion cohort. Watch for how quickly Elevar can enroll 30 patients: FGFR2 fusion in non-CCA tumors is rare enough that recruitment pace itself is a readout on program traction. Absence of a comparator arm and reliance on single-arm ORR are typical for rare biomarker-selected oncology, but they also mean the trial reads only as well as its point estimate.

Our model estimates a 8% chance this drug is eventually approved. It starts from the historical base rate for Phase 2 drugs in this area (about 13%), then adjusts using ten facts about the trial and sponsor. What moves the number most: it is helped by more secondary endpoints than usual and a non-randomized design; it is held back by the sponsor's thin or weak approval record and weak or limited earlier-phase results. The other facts land near average for this stage, so they leave the estimate roughly where the base rate put it.

Efficacy risk is real. FGFR2 fusion outside cholangiocarcinoma has been reported in scattered case series across breast, gastric, and pancreatic tumors, but the fraction of these tumors where the fusion is a true oncogenic driver rather than a passenger event is not well characterized. A mixed-histology basket could dilute the signal if only some tumor types are truly FGFR2-addicted. Safety risk is on-target: even selective FGFR2 inhibitors produce some hyperphosphatemia, retinal pigment epithelial detachment (a reversible vision-blurring side effect documented across FGFR drugs, caused by fluid accumulating under the retina), and nail changes. The 2026 Ann Oncol and Clin Cancer Res papers document acquired resistance mutations including gatekeeper mutations in the FGFR2 kinase domain (mutations at the drug-binding site of the kinase that physically block the drug from docking, a recurring resistance mechanism across targeted oncology drugs), meaning even responders will progress and duration of response is bounded [1][2]. Execution risk sits with Elevar: it is a smaller, less established oncology developer than Taiho or Incyte, and the transition of the program from Relay in late 2024 created execution risk on trial start-up and site activation [7]. The infigratinib (Truseltiq) precedent is directly instructive here: it received FDA accelerated approval for FGFR2-fusion CCA in 2021 but was voluntarily withdrawn by the sponsor in 2023 after failure to enroll the confirmatory trial, and formal withdrawal was completed in 2024 [13]. Elevar inherits the confirmatory-trial obligation and must execute where a prior sponsor did not. Commercial risk is the tightest constraint. Even with approval in a rare, well-selected population, addressable patients per year in the US for non-CCA FGFR2 fusion tumors number in the low thousands. Payer reimbursement will hinge on comparative context against pemigatinib and futibatinib in overlapping settings, and pricing power in a competitive FGFR2-selective class is not obvious.

The most important near-term event is the CCA NDA PDUFA action date of September 27, 2026 [11]. Priority review with the ReFocus efficacy profile (82.4% confirmed ORR at RP2D in FGFR-inhibitor naive patients, 46.5% IRC-assessed in the pivotal cohort with median DOR 11.8 months [9]) makes accelerated approval the base-case expectation, but a CRL or label carve-out is the risk. The out-license economics (up to $500 million in milestones plus double-digit royalties [4][7]) mean approval triggers a meaningful non-dilutive payout to Relay. Beyond CCA, the signal to look for is single-agent ORR from NCT07359820 in FGFR-inhibitor naive patients across the enrolled non-CCA histologies [3]. An ORR above 40% pooled, with responses spread across at least three tumor types, would be a genuine positive and support Elevar's tissue-agnostic path; an ORR in the 15% to 25% range across mixed histology would be commercially disappointing and force a retreat to CCA and specific-histology expansion cohorts. Caveat on the Relay disclosure channel: Relay announced material restructuring and headcount reductions in 2025 as it refocused its remaining pipeline [10], which raises questions about whether Relay's own 8-K cadence is still a clean proxy for program progress. Cross-check Elevar communications where available. Check back after the September 27, 2026 PDUFA decision and again 12 to 18 months into NCT07359820 enrollment for an initial ORR readout. Elevar's cadence of trial start-up and enrollment through mid-2027 will tell the story on execution.