In a blinded trial, patients do not know what they are taking and neither do their doctors, but a small independent committee sees everything. On August 13, that committee told Taiho and Cullinan to stop pretending.

The trial is REZILIENT3, testing zipalertinib plus platinum chemotherapy against chemotherapy alone as a first-line treatment for non-small cell lung cancer carrying EGFR exon 20 insertion mutations. Investigators randomised 285 previously untreated patients with locally advanced or metastatic disease. At a planned interim analysis, patients on the combination went significantly longer without their disease getting worse, hitting the primary endpoint. The safety profile was described as manageable, and the independent data monitoring committee recommended unblinding the study.

The committee exists because blinding creates an ethical problem the moment a drug starts working. If one arm is clearly doing better, continuing to give the other arm chemotherapy alone becomes hard to defend. If a drug is clearly harming people, worse. So a group with no financial interest in the answer gets access to the unblinded data at prespecified moments, and only they get it. Everyone else, the sponsor included, stays in the dark on purpose.

When that committee recommends unblinding on efficacy grounds, it is saying the question the trial was built to answer has been answered. Keeping people blind after that point stops protecting the science and starts getting in the way. That is a genuinely good outcome. It is also worth noting what has not arrived: Taiho and Cullinan called the improvement clinically meaningful but have released no progression-free survival numbers at all. Until they do, that phrase is just a company’s word for it, and it is fair to wait for the actual curve before deciding how big the win is.

EGFR exon 20 insertions are an uncommon subtype within an already heavily subdivided cancer, and for years one company owned the space. Johnson & Johnson’s Rybrevant was approved for second-line use in 2021 and first-line in 2024. Takeda withdrew its challenger Exkivity in 2023 after a confirmatory trial failed, leaving J&J unopposed until the FDA approved Dizal’s Zegfrovy in 2025. Rybrevant and Lazcluze grew 70% to $546 million in the first half of 2026. Last month AstraZeneca paid $600 million upfront for global rights to Zegfrovy, a drug from a company it had itself spun out in 2017.

Cullinan collects $30 million on a second-line approval and $100 million on a first-line one, and splits US profits with Taiho evenly; the FDA rules on the second-line filing by 27 February. But the number to keep is 70%. A niche defined by a rare mutation grew that fast largely because almost nobody was competing for it. Three companies are about to be, which is how a market that was too small to bother with becomes one worth $600 million upfront.

Sources: Taiho Oncology announcement of REZILIENT3 interim results; ClinicalTrials.gov NCT05973773.