On 14 July 2026 Celltrion, Inc. (KRX:068270) registered an early termination notice for the European arm of its Phase 3 trial of CT-P51, a biosimilar of Merck's Keytruda, and voluntarily withdrew the trial plan. The EMA had approved that plan on 12 December 2024, less than four months after Celltrion applied. Then, on 24 August 2026, the company filed for Korean marketing approval of the same product.
Kill the European trial, apply for approval six weeks later. That sequence is only coherent under one reading: the evidence a biosimilar sponsor needs to show has just gotten smaller, and Celltrion is moving faster than its filings alone would suggest.
The disclosure is unusually specific for a Korean investment-judgment filing. The trial was a double-blind, randomized, active-controlled Phase 3 comparing CT-P51 against Keytruda in previously untreated metastatic non-squamous non-small cell lung cancer. The stated reason for pulling the EU plan: following recent changes in the global regulatory environment for biosimilar clinical trials, Celltrion revised its clinical and approval strategy to reduce the number of trial subjects. Enrollment for the global Phase 3 has closed, EU countries were removed from the list of participating countries, and with that the European plan no longer served a purpose.
The forward plan is to compare efficacy and safety against Keytruda through the trial as it runs in the non-EU countries.
That is a company shrinking a Phase 3 mid-flight because the rules changed underneath it. Not because the drug failed. Worth separating those, because an early termination headline reads like failure to anyone who doesn't open the filing.
The regulatory shift Celltrion is responding to has been building for about a year. In October 2025 the FDA commissioner announced draft guidance aimed at removing the comparative efficacy study requirement for biosimilars, on the argument that modern analytical characterization already proves similarity better than a clinical trial does. In March 2026 the agency went further with a revised draft Q&A permitting non-US-licensed comparator products in the required studies, which it estimated could cut pharmacokinetic study costs by roughly half, or around $20m per development program.
Understand what that does to the economics. A biosimilar developer's largest costs are the manufacturing scale-up and the comparative clinical trial. The trial is the part that takes years, and in oncology it means enrolling hundreds of cancer patients across dozens of sites to demonstrate something the sponsor already believes is true. Removing or shrinking it compresses both the timeline and the capital at risk. The FDA has said the change could take three to four years off a development program.
For a company like Celltrion, which competes on being early to a molecule rather than on inventing one, three to four years is close to the whole competitive advantage. Every biosimilar developer benefits, so this is not a proprietary edge. But it disproportionately helps the developers with the deepest queue of molecules already in flight, because they get to re-cut many programs at once rather than one.
Celltrion appears to be doing exactly that. In a five-week stretch this summer it amended the US Phase 3 protocol for CT-P51 on 6 July, amended the European Phase 3 for CT-P44 (a Darzalex biosimilar) on 14 July and the Korean one on 3 August, filed CT-P55 (a Cosentyx biosimilar) for European approval on 24 July and US approval on 30 July, and filed three new Phase 3 plans for label expansion of Zymfentra into rheumatoid arthritis across the US, Europe and Korea on 24 July. That is not a sequence of independent decisions. That is a portfolio being re-planned against a new rulebook.
CT-P51 targets the biggest drug in the world. Keytruda's US composition-of-matter patent is expected to expire in 2028, and third-party forecasts put branded sales near $33.7bn in 2028 falling to about $27.4bn in 2029 as biosimilars arrive. Those are outside estimates, not company figures, so treat the precision loosely. The order of magnitude is what counts: no biosimilar opportunity of this size has ever existed.
It will also be the most crowded. Every large biosimilar developer has a pembrolizumab program, and being fourth to market in a category where hospital formularies pick one or two suppliers is worth a fraction of being first. Celltrion's history here is genuinely good. It brought the first Remicade biosimilar to approval in Europe and the US and has repeatedly been early rather than best-funded. Whether that translates when the competition is Amgen and Sandoz on a $30bn molecule is a different question.
Celltrion's own filing carries a line worth quoting to anyone getting excited. It states, as standard risk language, that the statistical probability of a trial drug receiving final approval is around 10%, and that results may fall short of expectations, in which case commercialization plans may change or be abandoned. Biosimilars carry far better odds than that base rate, since the molecule is already proven. The company still puts it in writing.
The pipeline spending happens against a much stronger P&L than Celltrion had two years ago. Revenue reached ₩4,162.5bn in FY2025 against ₩3,557.3bn in FY2024 and ₩2,176.4bn in FY2023. The second quarter of 2026 brought ₩1,393.7bn of revenue and ₩451.8bn of operating income, a 32.4% operating margin, against ₩961.5bn and ₩242.5bn a year earlier. First-half revenue of ₩2,538.7bn is up about 41%.
Note a discrepancy for anyone comparing sources. Press coverage of the second quarter described operating profit of about ₩430bn and revenue of ₩1.3tn, growth of 77.3%. The half-year report filed with DART on 14 August shows ₩451.8bn and ₩1,393.7bn. Celltrion filed a correction to its preliminary results disclosure on 27 July. Where they differ, the half-year report is the number.
Cash generation lags the income statement, which is normal for a company building inventory ahead of launches. Operating cash flow was ₩646.1bn in FY2025 against ₩1,168.5bn of operating income, and inventories sat at ₩2,940.5bn at the end of Q2 2026 against ₩2,712.1bn a year earlier. A biosimilar company stockpiles product before a launch window it cannot control. That is a defensible reason for the gap and also a real risk if a launch slips.
The obvious counter is that easier approval cuts both ways. If it costs less and takes less time to build a biosimilar, more companies build them, and the price erosion on each molecule gets steeper. Celltrion's advantage has partly been that the barrier was high. Lowering it helps its speed and hurts its pricing, and which effect dominates depends on the molecule.
There is also a straightforward regulatory risk in the CT-P51 story. The March 2026 guidance is draft. Draft guidance is not final rule, and the FDA has walked back positions before. Celltrion has now reduced enrollment in a global Phase 3 on the strength of a changing environment. If the final guidance lands more conservatively than the draft, the company could find itself holding a trial too small for the file it wants to submit, and the fix costs a year.
The Korean filing does not de-risk the American one. Korea's MFDS and the FDA are different reviewers with different expectations, and Celltrion's largest market for this product would be the US.
Two dates. First, whether the FDA finalizes the comparative efficacy guidance and in what form. That single document determines how much of the summer's replanning was correct.
Second, when Celltrion files CT-P51 with the FDA. It filed in Korea on 24 August. The gap between the Korean and US submissions is the cleanest read available on how confident the company is that the shrunken trial package satisfies the American reviewer. A US filing within two quarters says the strategy worked. A longer gap says the data set still needs filling in, and the 2028 launch window is tighter than it looks.
kstock reads DART every morning and writes up what moved — the contract, the buyback, the number that does not add up. The daily post and a Saturday roundup, by email.