$SMMT, $MRK: Ivonescimab’s China-First Test

The central question around ivonescimab has changed. The first question was whether simultaneously targeting PD-1 and VEGF could outperform established checkpoint inhibition. China-only trials supplied a compelling early answer: ivonescimab delivered a large progression-free survival benefit against pembrolizumab in HARMONi-2, while ivonescimab plus chemotherapy later produced a statistically significant overall-survival advantage over tislelizumab plus chemotherapy in HARMONi-6. The question now facing Summit Therapeutics, Merck and the wider pharmaceutical industry is harder: can a China-origin asset preserve its clinical signal when the trial population, standard of care, regulatory expectations and commercial battlefield become global?

That makes the current FDA review a test of a new development model—not a referendum on Chinese science. The model begins with rapid proof generation in China, moves through a high-value Western licensing transaction, and then depends on a multiregional program designed to establish applicability to U.S. patients and medical practice. Summit’s global HARMONi trial is the first major regulatory bridge. But the larger commercial tests are still ahead: HARMONi-3 and HARMONi-7 will ask whether ivonescimab can compete directly with pembrolizumab-based first-line standards in broader Western populations.

“China-first” is a development sequence, not a quality label

The phrase “China-first” is often used too loosely. It can describe at least three very different situations: a China-only pivotal trial submitted directly to the FDA; a China-origin molecule subsequently tested in a genuinely multiregional pivotal trial; or an early-stage Chinese asset acquired before definitive efficacy is known and then rebuilt inside a global program. Those routes do not carry the same regulatory risk.

Ivonescimab belongs in the second category. Akeso discovered and initially developed the bispecific in China. Under the agreement signed in December 2022, Summit committed to a $500 million upfront payment for rights in the United States, Canada, Europe and Japan, plus as much as $4.5 billion in potential milestones and low-double-digit royalties. The upfront consideration was paid in two installments: $300 million in January 2023, partly in Summit shares at Akeso’s election, and $200 million in cash in March 2023. Summit’s license territory has since expanded and now includes North America, South America, Europe, the Middle East, Africa and Japan. The economic bargain was clear: Akeso supplied the molecule, development velocity and early clinical evidence; Summit accepted the cost and execution risk of establishing a global regulatory and commercial pathway.

The final two boxes are decisive. A high-quality result in China can reduce biological risk, but it does not eliminate differences in ancestry, disease biology, prior therapy, subsequent therapy, medical practice, trial conduct or the control arm used in the United States. The FDA’s position is not that foreign evidence is inherently inferior. Its position is that foreign evidence must be relevant to the population and clinical setting in which the drug will be used.

What the China trials have already established

Ivonescimab’s China data are not a single headline. They form a growing body of evidence across multiple settings, with two studies carrying particular strategic importance.

HARMONi-2 remains the source of the most provocative commercial claim because it directly compared ivonescimab monotherapy with pembrolizumab. The prespecified interim PFS effect was large, but higher-grade treatment-related adverse events were also more frequent with ivonescimab—29% versus 16% in the published analysis. HARMONi-6 then raised the level of evidence by demonstrating a statistically significant OS benefit in squamous disease at its prespecified interim OS analysis. In that study, however, serious treatment-related adverse events were reported in 41.4% of the ivonescimab arm versus 34.3% of the control arm, and grade 3 or worse hemorrhage occurred in 2.6% versus 0.8%.

The mechanism therefore carries both the promise and the cost of adding VEGF inhibition: the possibility of deeper antitumor activity, alongside bleeding, hypertension, proteinuria and other toxicity considerations that can shape real-world use. A commercially disruptive profile requires more than superior efficacy. It needs a benefit-risk balance that remains favorable outside highly selected trial populations.

What FDA is really testing

FDA’s September 2024 draft guidance on multiregional oncology development is unusually relevant to ivonescimab, but it is important to describe its status correctly: it is non-binding, marked “not for implementation” and remains a draft. The document recommends that sponsors design evidence capable of supporting FDA’s assessment of applicability to U.S. patients and U.S. medical practice; enroll enough U.S. participants to help interpret benefit and risk; avoid excessive concentration in one region; and use a control arm relevant to U.S. care. It also highlights intrinsic and extrinsic factors that can alter treatment effects, including ancestry, molecular disease patterns, prior and subsequent therapy, access to care and local practice.

The distinction matters because it changes the industry lesson. A negative outcome would not prove that Chinese oncology assets are unlicensable. It would show that the global validation package was not strong enough for a specific indication, endpoint or control arm. Conversely, an approval would not grant every China-only data package an automatic pass. It would validate a more demanding formula: discover and de-risk in China, then generate fit-for-purpose multiregional evidence.

The sintilimab precedent: a warning, not a template

The most cited precedent is sintilimab. In 2022, an FDA advisory committee voted 14–1 that additional clinical study was needed before approval of sintilimab plus chemotherapy for first-line non-squamous NSCLC. ORIENT-11 had been conducted entirely in China. FDA’s concerns included the trial’s applicability to U.S. patients, a control arm that did not reflect U.S. standard care, reliance on a progression endpoint rather than a survival design suitable for the regulatory question, and late engagement with the agency. Lilly subsequently received a complete response letter requesting a multiregional study using an active U.S.-relevant comparator, with an OS-based design discussed.

The lesson was widely simplified into “FDA rejected Chinese data.” That reading is incomplete. FDA rejected the adequacy of that particular evidence package for the proposed U.S. use. The distinction is precisely why HARMONi matters: Summit did run a global randomized Phase III trial, included Western patients and engaged FDA sufficiently for the agency to accept the biologics license application for review.

HARMONi has de-risked geography more than it has de-risked the decision

HARMONi met its co-primary progression-free survival endpoint. At the prespecified analysis of the co-primary OS endpoint in April 2025, however, the hazard ratio was 0.79, with a 95% confidence interval of 0.62–1.01 and p=0.057. Median OS was 16.8 months with ivonescimab plus chemotherapy and 14.0 months with chemotherapy alone. The numerical trend favored ivonescimab, but the analysis did not cross the conventional threshold for statistical significance.

With longer follow-up, Summit reported an updated OS hazard ratio of 0.78 in September 2025, accompanied by a nominal p-value of 0.0332. In July 2026, the company reported a further descriptive improvement to 0.76 in the intention-to-treat population and the same 0.76 hazard ratio in what it defined as the Western subgroup. The Western estimate improved over time—from 0.98 after 9.2 months of follow-up, to 0.84 after 13.7 months, and 0.76 after 23.2 months. The new follow-up accrued primarily in Western patients; the Asian dataset remained locked at its earlier cutoff with 32.7 months of median follow-up.

That consistency is encouraging, but the most recent disclosure remains incomplete. Summit did not provide a confidence interval, median OS or event count for the June 2026 update and said detailed results would be presented at a future medical conference. The Western subgroup consisted of 83 ivonescimab patients and 82 control patients. An identical point estimate in the global and Western populations is directionally reassuring; without the interval around that estimate, it is not possible to judge its precision.

A moving U.S. comparator raises a second issue

HARMONi compared ivonescimab plus platinum-doublet chemotherapy with chemotherapy alone. That was a relevant post-TKI benchmark when the program was designed, but the U.S. treatment landscape evolved. In September 2024, FDA approved amivantamab plus carboplatin and pemetrexed for EGFR exon 19 deletion or L858R-mutated NSCLC after progression on an EGFR TKI, based on MARIPOSA-2. HARMONi completed enrollment in the second half of 2024.

This does not automatically invalidate HARMONi. Regulatory decisions consider when a study was designed, unmet need, the proposed label and the full evidence package. It does mean that commercial relevance cannot be inferred from a placebo-controlled chemotherapy comparison alone. Even with approval, ivonescimab would enter a setting that already has a newly approved targeted-antibody combination. Positioning, tolerability, physician familiarity and comparative evidence would affect uptake.

There are three separate ivonescimab tests

Investors often compress the program into a single binary event. The more useful framework separates regulatory transportability, first-line commercial disruption and mechanistic breadth.

HARMONi-3 is especially important because both the squamous and non-squamous cohorts are powered separately. Summit expects the final PFS event threshold for the squamous cohort in the second half of 2026, together with an early interim OS analysis. Another squamous OS interim is planned for the first half of 2027, when the non-squamous PFS analysis is also expected. These readouts are more consequential to the long-run value of the franchise than the current narrow-label filing alone.

Why $MRK belongs in the headline

Merck generated $31.68 billion in 2025 sales from Keytruda and Keytruda Qlex. Merck lists Keytruda’s principal U.S. compound patent as expiring in 2028 and two additional U.S. active-ingredient patents in 2029, while the separate Keytruda Qlex patent estate currently extends to 2043. Formulation, combination, device and other intellectual-property strategies can influence the pace of erosion. The franchise also spans dozens of approved indications. Ivonescimab is therefore not an immediate binary threat to all of Keytruda.

The strategic risk is narrower and more important: if global HARMONi-3 or HARMONi-7 demonstrates a credible efficacy advantage with an acceptable safety profile, PD-(L)1/VEGF bispecifics could become a new first-line standard in major lung-cancer segments. That would pressure Keytruda not merely through patent expiry, but through therapeutic displacement before or during the loss-of-exclusivity window.

Merck has already responded to the mechanism. In November 2024, it licensed LaNova Medicines’ China-origin PD-1/VEGF bispecific LM-299, now known as MK-2010, for $588 million upfront and up to $2.7 billion in milestones. The transaction is both a hedge and a validation: Merck does not need ivonescimab to fail if it can develop a competitive next-generation asset of its own.

Big Pharma has already voted with capital

Western regulators have not yet delivered a definitive verdict on the category, but pharmaceutical business-development teams have already committed billions of dollars to China-origin PD-(L)1/VEGF programs.

The implication extends beyond one drug class. China’s trial infrastructure, large treatment-naïve populations, increasingly sophisticated biotechnology sector and competitive development economics can make it an efficient place to discover and rapidly test oncology assets. But licensing transactions validate access to innovation, not final regulatory success. The highest-value companies will be those that can preserve speed while redesigning programs early enough around FDA, European and Japanese requirements.

The transmission map: how one FDA decision can change the market

An approval would strengthen the idea that China-origin assets can reach the United States efficiently if a global program is constructed around U.S. applicability. A narrow label, post-marketing commitment or cautious FDA communication would still validate part of the model, but it would reinforce the need for indication-specific evidence. A request for another study would shift negotiating leverage back toward licensors with already-global data and raise the discount applied to China-only efficacy packages.

Equity impact map: the same result means different things to $SMMT and $MRK

What is priced in—and what still needs proof

At the July 24 close of $13.66, Summit’s market value was approximately $10.9 billion using the company’s July 17 basic share count. That valuation cannot be explained by the current post-TKI EGFR-mutant opportunity alone. It embeds substantial expectations for approval, first-line expansion and a broader role for the mechanism.

Yet the shares were also roughly 22% below their first trading close of 2026 and fell 8.81% on July 24, immediately after the company’s quarterly update and additional financing disclosure. The price action suggests that investors continue to demand cleaner proof despite the improved OS point estimate. This is a stock whose valuation can move sharply on the interpretation of an endpoint, not only on whether the endpoint is technically positive.

Scenario matrix: three ways the model can develop

No precise probabilities are assigned here because the full FDA review record, the latest OS dataset and detailed subgroup intervals are not public. The useful exercise is to define the evidence that would distinguish the paths.

The monitoring calendar

Bottom line

Ivonescimab has already done enough to change oncology strategy. It helped move PD-(L)1/VEGF bispecifics from an interesting Chinese development theme into the capital-allocation plans of Merck, Pfizer, BioNTech and Bristol Myers Squibb. It has not yet done enough to prove that the strongest China efficacy signals will reproduce at the same magnitude in Western practice.

The current HARMONi filing is the first regulatory bridge, not the final commercial verdict. Approval would show that a China-born oncology asset can reach FDA through a global evidence package and would strengthen Summit’s position ahead of first-line readouts. But HARMONi-3 and HARMONi-7 will decide whether ivonescimab is mainly a valuable new option in selected settings or a true successor-class challenge to pembrolizumab.

For $SMMT, the appropriate framework is therefore watch the label, then demand global head-to-head proof. For $MRK, the key question is not whether one competitor receives one approval; it is whether PD-(L)1/VEGF becomes a durable new standard before Keytruda’s exclusivity weakens. For the pharmaceutical industry, the lesson is broader: China-first can accelerate discovery and early validation, but global regulatory value is created only when trial design, population and comparator are global from the start.

Related MerlinTrader research

• Summit Therapeutics Stock Hub: pipeline, catalysts and ivonescimab coverage
• ASCO 2026 Investor Calendar

Sources and freshness

This analysis uses public information available through July 27, 2026. Market-price calculations use the July 24, 2026 U.S. close, the latest completed session at publication time. The June 2026 HARMONi update has not yet been disclosed at conference level; its missing confidence intervals, medians and event counts are treated as an explicit uncertainty rather than estimated.

1. Summit: updated global and Western HARMONi OS analysis, July 22, 2026
2. Summit: Q2 2026 results and program timelines
3. Summit: original Akeso license territories and upfront-payment installments
4. Summit: FDA BLA acceptance and PDUFA date
5. ClinicalTrials.gov: HARMONi design and co-primary endpoints
6. The Lancet / PubMed: HARMONi-2 Phase III publication
7. Summit/Akeso: HARMONi-6 overall-survival results
8. FDA: oncology multiregional clinical-development draft guidance
9. FDA: acceptance of foreign clinical studies guidance
10. FDA: sintilimab advisory-committee meeting materials
11. Eli Lilly: sintilimab complete response letter
12. FDA: amivantamab plus chemotherapy approval after EGFR TKI
13. Summit 2026 Q2 Form 10-Q
14. Summit July 2026 at-the-market prospectus
15. Summit investor information: July 24, 2026 closing market data used in this article
16. Merck 2025 Form 10-K
17. Merck–LaNova LM-299 licensing announcement
18. Pfizer: 3SBio SSGJ-707 licensing agreement
19. Bristol Myers Squibb/BioNTech: BNT327 partnership