Placement on the plenary session of the upcoming ASCO 2026 annual meeting has significantly elevated enthusiasm around Akeso’s first-line non-small cell lung cancer data for its Summit Therapeutics-partnered ivonescimab.
As the first China-developed asset to score this primetime slot in the oncology event’s more than six decades of history, ivonescimab will showcase its overall survival capability from the phase 3 Harmoni-6 trial. The study previously showed that the first-in-class PD-1xVEGF inhibitor, alongside chemo, beat BeOne Medicines’ PD-1 inhibitor Tevimbra and chemo on progression-free survival (PFS) in Chinese patients with newly diagnosed squamous NSCLC.
But with honor comes scrutiny. The premium status means the Harmoni-6 results will be put under the microscope and carefully critiqued, Leerink Partners analyst Daina Graybosch, Ph.D., said in an interview with Fierce previewing the readout.
“It’s totally going to get minced,” Graybosch said. “That’s the risk of getting in the plenary. If you had the same data in the regular oral [session], I’m sure they’ll get less of a critical discussant.”
Some expectations are so high before the plenary presentation on May 31 that one physician KOL bet Graybosch a cup of coffee that ivonescimab will report an overall survival improvement of above 50%.
Fifty percent would be a lofty outcome “that’s going to exceed everybody’s expectations,” Graybosch said. Even the KOL admitted that it’s merely a guess—which they aren’t seriously confident.
For now, even information on whether Harmoni-6 achieved statistical significance on overall survival remains confidential. If it did, it would be the first regimen to demonstrate a survival benefit over a PD-1/chemotherapy combination in first-line NSCLC in a phase 3 trial, meeting oncology’s gold standard for clinical efficacy.
There are mainly two camps of investors, Graybosch observed.
When Harmoni-6’s results were first published in The Lancet, its statistical plan, which was apparently not properly redacted, showed that the stopping boundary for an interim analysis of OS was a hazard ratio (HR) of 0.722. That means that if ivonescaimb showed an OS improvement of 27.8% along with other criteria, the trial could claim success at the endpoint.
One group of investors equates the plenary inclusion to statistical significance and a better outcome than the efficacy threshold. Therefore, they believe ivonescimab will show an HR between 0.65 and 0.7.
The other half of investors still think the data comes from an unplanned, ad hoc analysis that doesn’t bear statistical significance, just like what Akeso did with Harmoni-2 at the request of Chinese regulators. Referring to past studies in the immuno-oncology space, these investors are more pessimistic about the absolute effect size, suspecting that ASCO chose the study for plenary because it comes from China—which is a rising source of biotech innovation—and that it’s debatable.
Following that logic, Graybosch’s team performed a simulation based on a few recent immunotherapy readouts in first-line NSCLC, with several not in the exact same squamous indication as Harmoni-6. Based on pooled PFS and overall response rate data, the estimated OS HR for Harmoni-6 would land at around 0.83 and 0.87, according to Leerink’s calculation.
Nevertheless, Graybosch, in an April 21 note, sided with the first group of investors, arguing that a plenary means ivonescimab plus chemo “likely drove a strong benefit in OS,” with a statistically significant HR better than the 0.722 bar. But attention should be paid beyond the size of the HR, she argued. And even the bullish KOL cautioned against focusing on topline OS without context.
“I used to think that anything under 0.7 would be a really big win, but the devil’s in the details,” she said.
The Leerink analyst said she would keep a close eye on whether patients got proper subsequent therapies. That would be important for translating the China-only Harmoni-6 result to a global population, as effective second-line treatment typically would blunt the OS effect of first-line therapy.
“Our team notes that poor access to subsequent therapy is more likely to impact earlier, less mature OS analysis—like this look at Harmoni-6 where we expect a median follow-up of ~19-20 months,” Graybosch’s team wrote in a May 27 note.
In her interview with Fierce, Graybosch initially suggested that HR matters more than the absolute median OS length because HR helps estimate for a global population. But she immediately changed her mind because a poorly performing control arm could artificially jack up the treatment effect of ivonescimab.
In the phase 3 Rationale-307 trial conducted in China in first-line squamous NSCLC, Tevimbra and chemo led to a median OS of 22.8 months. In the global Keynote-407 trial, which was done years earlier than the BeOne study, Keytruda plus chemo showed a median OS of 15.8 months at an interim analysis, or 17.1 months at final analysis.
The bullish KOL also flagged that Harmoni-6 largely enrolled Chinese male smokers, which could limit its results’ translatability to a more heterogeneous global population.
The stakes are now even higher for Harmoni-6 after Akeso’s partner Summit recently effectively telegraphed that the global Harmoni-3 trial surprisingly missed PFS during an interim analysis of the squamous cohort, as the trial continues toward its final PFS analysis. The update caught industry watchers of the PD-(L)1xVEGF space off guard because Harmoni-6 had met PFS at interim.
Assuming that the Harmoni-3 miss suggests a 10-percentage-point degradation of ivonescimab’s effect size from China to a global population, Graybsoch said the OS improvement bar for Harmoni-6 is “reasonable” at 30% “if it’s a really clean study,” in which patients got proper second-line treatment. A 20% improvement in the global population could still be statistically significant and meaningful, she said.
“But let’s say [HR is] 0.65, but you have a lot of funny business that the discussant points out that makes this even less likely to translate globally, I think that’s where it’s hard to know what’s going to happen [to Harmoni-3].”
In addition to other PD-(L)1xVEGF bispecifics, ivonescimab is now also facing competition from Merck & Co. and Kelun-Biotech’s sac-TMT. The TROP2 antibody-drug conjugate, combined with Keytruda, improved PFS by 65% versus Keytruda alone in first-line, PD-L1-positive NSCLC in the OptiTROP-Lung05 China study, according to results to be presented at ASCO 2026.
Because Keytruda monotherapy is used in the U.S. mainly in the PD-L1-high population, that’s the only group that matters for a direct ivonescimab/sac-TMT comparison, Graybosch said. For now, only PFS data are available from the two regimens, which Graybosch said wouldn’t be enough for a proper comparison because OS and toxicity are also important to evaluate their potential.
Unlike Summit’s bold plan with Harmoni-3 to test ivonescimab and chemo against Keytruda and chemo in an all-inclusive first-line NSCLC setting, Merck’s strategy with sac-TMT is more targeted. The phase 3 TroFuse-023 trial is combining Keytruda and sac-TMT in a first-line maintenance setting in squamous NSCLC. The TroFuse-007 trial is focused on the combo in PD-L1-high NSCLC.
Merck has divulged very little on its plan for its own PD-1xVEGF drug, MK-2010. Graybosch suggested that Merck could do a study pairing MK-2010 with sac-TMT and platinum chemo against the standard Keytruda-chemo regimen in first-line nonsquamous NSCLC, or run a TROP2 biomarker-enriched study without the help of the extra chemo.
“We think that sac-TMT has an ability to be best-in-class, but it’s also important to put it in the right tumor types and to have the right strategy in those tumor types,” Dean Li, M.D., Ph.D., president of Merck Research Laboratories, said during an investor call in February. “There will be tumor types where we do not believe that a biomarker will be needed, but we also believe that there are places where that biomarker will be needed, especially if you look at how good the comparator you have to go against.”
Meanwhile, Akeso and Summit investors find themselves caught in a paradox. They hope Harmoni-6 delivers a bang at the ASCO plenary session—partly to validate the upcoming Harmoni-3 global readout. But as Graybosch noted, if the OS bar in this China-only cohort is set too high, it’d only be downside risk from here.