Zipalertinib

Zipalertinib is an oral EGFR inhibitor built to hit a rare but historically tough mutation - EGFR exon 20 insertions - that drives roughly 2% of lung cancers, an estimated 3,000-4,000 newly diagnosed U.S. patients per year. Taiho Oncology and Cullinan Therapeutics are testing it in two Phase 3 trials: REZILIENT3 in first-line metastatic disease and NCT07128199 in the adjuvant setting after surgery. The FDA accepted a New Drug Application for zipalertinib in pretreated exon 20 NSCLC with a PDUFA target action date of February 27, 2027 [9]. If the data hold, zipalertinib would compete with J&J's amivantamab and Dizal's already-approved sunvozertinib as an oral, exon-20-selective option.

Novel compound, never approved anywhere. Phase 2 results in pretreated patients (Piotrowska et al., JCO 2025) reported an objective response rate (ORR - the fraction of patients with measurable tumor shrinkage) near 40% in EGFR exon 20 insertion NSCLC patients who had progressed on platinum chemo, including those previously treated with amivantamab [1]. FDA granted Breakthrough Therapy Designation to CLN-081/zipalertinib in 2022 for the same setting [2]. Taiho and Cullinan completed a rolling NDA submission in early 2026; FDA accepted the NDA with a PDUFA action date of February 27, 2027 [9]. Two Phase 3 trials are now active and not recruiting: REZILIENT3 (NCT05973773) tests first-line zipalertinib plus carboplatin/pemetrexed versus chemo alone in 285 patients [3], and NCT07128199 tests adjuvant zipalertinib versus placebo after complete resection in 360 patients with stage IB-IIIA disease. The adjuvant trial's listed eligibility refers to 'uncommon EGFR mutations,' which conventionally includes exon 20 insertions plus G719X, S768I, and L861Q - if the trial covers the broader class, the addressable population is meaningfully larger than ex20ins alone; the writeup defers to the trial record on the final eligibility list [4]. The first-line REZILIENT3 readout is the next major trial catalyst, on top of the 2027 PDUFA. Cullinan retains 50% U.S. economic rights via its 2021 co-development deal with Taiho; ex-U.S. rights belong to Taiho. An adjuvant disease-free survival (DFS - time from surgery until cancer returns or patient dies) readout sits years out given the follow-up required in resected disease.

EGFR is a switch on cell surfaces that tells cells to grow when triggered by signaling proteins outside the cell. In some lung cancers, the EGFR gene is mutated so the switch stays stuck on, and cells multiply without stopping. The two common EGFR mutations - exon 19 deletions and L858R - respond well to tyrosine kinase inhibitors (TKIs - small molecules that block the kinase enzyme inside the receptor) such as osimertinib. Exon 20 insertions are the awkward sibling: a few extra amino acids wedge into a loop region just behind the drug-binding pocket (the αC-β4 loop in the kinase domain's C-terminal lobe), distorting the pocket geometry so classic EGFR TKIs slide off [5]. Zipalertinib was designed to fit that altered pocket while leaving normal wild-type EGFR largely alone. That selectivity matters commercially. Drugs that hit wild-type EGFR cause severe rash and diarrhea, which is how mobocertinib - Takeda's first-generation exon 20 inhibitor - ended up too toxic at therapeutic doses and was pulled after a failed confirmatory trial [6]. Preclinical work showed zipalertinib has roughly 10-fold selectivity for mutant over wild-type EGFR, and clinical data so far back that profile with lower rates of grade 3 rash and diarrhea than mobocertinib reported [5][1]. The mechanism is well-validated by human genetics, by other approved EGFR TKIs in different mutation contexts, and by amivantamab's approval in this exact patient population.

NCT07128199 is the adjuvant Phase 3, sponsored by Taiho. 360 patients with completely resected stage IB-IIIA NSCLC harboring uncommon EGFR mutations are randomized to zipalertinib or placebo as adjuvant therapy. Primary endpoint is investigator-assessed disease-free survival (DFS) [4]. The design is conventional and well-powered for a biomarker-selected adjuvant trial; the bigger concern is that DFS in early-stage NSCLC takes years to mature, so this readout sits far out. The more immediately relevant trial is REZILIENT3 (NCT05973773), a Phase 3 first-line study of 285 patients randomized to zipalertinib plus carboplatin/pemetrexed versus chemo alone in advanced EGFR exon 20 insertion NSCLC [3]. Progression-free survival (PFS - time from randomization until tumor growth or death) is the efficacy readout that matters commercially. Overall survival is a secondary endpoint; no OS signal has yet been reported. REZILIENT3 has finished enrollment. The design mirrors what worked for amivantamab in PAPILLON, which showed amivantamab plus chemo beat chemo alone in this same population [7]. Enrollment was always the hardest part because exon 20 insertions are roughly 10% of EGFR-mutant NSCLC and about 2% of all NSCLC (~3,000-4,000 newly diagnosed U.S. patients per year). The Phase 1/2 REZILIENT1 (NCT04036682) continues with 284 patients and is the data foundation for the accelerated approval filing in pretreated patients.

Our model gives this drug a 17% chance of eventually being approved. That starts from the historical approval rate for Phase 3 drugs in this area - about 48% - then adjusts up or down based on ten specific facts about the trial and sponsor. The biggest factors pulling the number down are the sponsor's weak approval track record, limited earlier-phase results, and a randomized trial design; the one factor pushing it up is an above-average number of secondary endpoints. The remaining facts are close to typical for this stage and leave the estimate near where it started.

Efficacy risk: REZILIENT3's chemo combination needs to clear a higher bar than the single-agent Phase 2 because PAPILLON already established amivantamab plus chemo as a front-line standard [7], and sunvozertinib monotherapy posted a first-line PFS hazard ratio of 0.65 versus chemo in WU-KONG28 [11]. A modest PFS improvement won't displace amivantamab; zipalertinib needs oral convenience plus competitive efficacy. Safety risk: on-target wild-type EGFR toxicity is the class killer. Mobocertinib's October 2023 withdrawal after EXCLAIM-2 failed was driven partly by GI and skin toxicity at therapeutic doses [6]. Zipalertinib's selectivity profile looks better in early data, but Phase 3 surfaces tolerability problems smaller studies miss, especially in combination with carboplatin/pemetrexed. Execution risk: a rare patient population, but REZILIENT3 has completed enrollment and the adjuvant trial is active-not-recruiting, so most of this risk is behind them. Commercial risk: even with a positive readout, payers will compare zipalertinib head-to-head against amivantamab plus chemo, amivantamab plus lazertinib [10], and now sunvozertinib. Sunvozertinib (Dizal's ZEGFROVY) is the most direct analog - oral, selective exon 20 TKI, already FDA-approved (accelerated, July 2023) in pretreated exon 20 NSCLC, with positive Phase 3 first-line data presented at ASCO 2026 [11]. Zipalertinib's path to differentiation rests on a chemo-combo regimen (rather than monotherapy) and tolerability claims that have to hold up at scale. The oral formulation is a real advantage over IV amivantamab, but J&J's subcutaneous amivantamab (approved December 2024) narrows the convenience gap [12]. Resistance is the other open question: data show EGFR T790M and C797S can emerge on zipalertinib, limiting durability in patients who respond [8].

Worth watching, not a lottery ticket. The signal to watch is REZILIENT3's PFS readout, followed by the February 27, 2027 PDUFA action on the pretreated indication [9]. Hazard ratio (HR) is the standard trial metric: a value below 1.0 means fewer progression events in the zipalertinib arm, and an HR of 0.6 would mean a 40% reduction in progression risk versus chemo. If zipalertinib plus chemo posts an HR meaningfully better than 0.6, this becomes a real oral alternative to amivantamab and Cullinan re-rates hard. A 0.7-0.8 HR still gets approval but commercial uptake gets capped - particularly now that sunvozertinib monotherapy showed an HR of 0.65 in first-line WU-KONG28 [11], setting a benchmark a chemo-combo arm has to clear without the chemo toxicity penalty. Anything above 0.8 is a problem given the class's safety baggage. For Cullinan specifically (NASDAQ: CGEM), zipalertinib is the asset that justifies the valuation - 50% U.S. economics with Taiho running the trial is a deal structure that pays if the readout is positive. Taiho is a private subsidiary of Otsuka Holdings, so the cleaner equity exposure is Cullinan. The mobocertinib withdrawal narrative will be a headwind on launch even if the data is clean; physicians remember the toxicity story and it affects formulary placement. The adjuvant trial (NCT07128199) is the long-term option value but won't read out for years. Front-line is the catalyst that matters in 2026, and the PDUFA in February 2027 is the near-term gate.