Akeso and Summit Therapeutics’ ivonescimab reduced the risk of death by 27% compared with Keytruda in patients with treatment-naïve, PD-L1-positive advanced non-small cell lung cancer in the phase 3 Harmoni-2 trial conducted in China.
The result, to be presented at the 2026 World Conference on Lung Cancer in Seoul, marks another positive readout from ivonescimab, whose Keytruda head-to-head win two years ago on the trial’s primary endpoint of progression-free survival sparked a gold rush for similar PD-(L)1xVEGF bispecific antibodies.
“In Keytruda’s monotherapy setting, having such strong data help[s] physicians understand that this is an approach capable of beating current PD-1 standards of care in this indication,” Akeso CEO Michelle Xia, Ph.D., told Fierce.
As a trailblazer, ivonescimab now offers comfort to all PD-1xVEGF developers after demonstrating a head-to-head overall survival (OS) advantage over Keytruda, Xia said.
“We’ve become something of a bellwether for the entire PD-1xVEGF space,” Xia said. “By proving that this works, we’re delivering a boon to everyone who has bought into PD-1xVEGF, whether that’s Summit, Pfizer, AbbVie, Merck & Co., or BMS.”
A closer look at the readout: Patients in the ivonescimab arm lived a median of 30.8 months, versus 22.6 months for Keytruda.
The 27% OS improvement compares favorably to the 22.3% reduction that ivonescimab delivered at an unplanned interim analysis requested by Chinese authorities as part of Akeso’s new drug application. At the time, investors were disappointed by the magnitude of the benefit. Back then, the 398-patient trial had accrued 157 deaths, with a high statistical significance bar of 0.0001, according to Xia. Now, with a total of 234 deaths and an Aug. 20 data cutoff, the OS endpoint has reached statistical significance with a p-value of 0.009.
After Akeso’s headline announcement earlier this month, analysts at Citi and Leerink Partners modeled an OS benefit above 30% based on incomplete knowledge of Akeso’s statistical analysis plan and various assumptions of accrued death events, while those at Jefferies and Evercore ISI were more conservative in their expectations.
For reference, in the Keynote-042 global phase 3 trial of more than 1,200 patients, Keytruda monotherapy only improved OS by 19% versus chemo in first-line PD-L1-positive NSCLC.
Strong PD-L1-high data
Since Harmoni-2’s initial readout in 2024, critics have argued the trial is less relevant in clinical practice, especially for patients with low PD-L1 expression, for whom Keytruda plus chemo has become the standard of care.
But Xia argued that the trial was designed to prove ivonescimab is better than Keytruda, and that, with a green light from local authorities, it was “a smart way” to pursue a broad label for the drug in China.
In an important message from Harmoni-2, ivonescimab delivered a particularly strong OS benefit in PD-L1-high patients with a tumor proportion score of at least 50%, a setting where Keytruda monotherapy remains the standard of care globally and where chemotherapy offers limited additional benefit when added to the PD-1 inhibitor.
Within the PD-L1-high subgroup, ivonescimab improved OS by 42% versus Keytruda, whereas the PD-L1-low group recorded a 15% advantage.
The PD-L1-high setting “offers a relatively clean test” and proof-of-concept that the PD-1xVEGF bispecific construct “will confer incremental clinical benefit over PD-1 alone for patients where VEGF-A antagonism has weak clinical benefit,” Leerink analysts wrote in an Aug. 28 note, while expressing skepticism about a strong OS outcome for ivonescimab in this patient subgroup.
The strong PD-L1-high results bode well for Summit’s global Harmoni-7 trial in the same setting.
To Xia, ivonescimab’s 35% OS improvement in patients with squamous NSCLC marks another exciting data point. The result again supports the idea that the PD-1xVEGF bispecific mechanism differs from inhibition of PD-1 and VEGF via separate medicines. Bevacizumab, the best-established VEGF inhibitor in NSCLC, is generally avoided in patients with squamous histology because of the risk of serious bleeding.
Separately, in the Chinese Harmoni-6 trial, ivonescimab plus chemotherapy previously delivered a significant 34% reduction in the risk of death versus BeOne Medicines’ PD-1 inhibitor Tevimbra and chemo in first-line squamous NSCLC.
When the Harmoni-6 results were presented during the plenary session at ASCO 2026, the invited discussant, Julie Brahmer, M.D., from Johns Hopkins, criticized the trial for excluding patients whose tumors significantly invade, surround or encase major blood vessels, all of which increase the risk of bleeding. Those patients were included in Harmoni-2, which enrolled 45% of participants with squamous histology, a proportion that Xia said mirrors real-world patient characteristics.
While PD-1xVEGF drug developers have their eyes on Keytruda, the first-line NSCLC landscape is shifting. Also at ASCO 2026, Merck’s partner Kelun-Biotech reported that a combination of Keytruda and its TROP2-directed antibody-drug conjugate, sac-TMT, reduced the risk of disease progression or death by 65% versus Keytruda alone in a phase 3 trial of treatment-naïve, PD-L1-positive NSCLC in China. Overall survival data were immature, although the combination showed a favorable trend.
Now, all eyes are on Summit’s Harmoni-3 trial, which pairs ivonescimab with chemo in first-line NSCLC, with its final PFS readout from the squamous cohort expected this year. The study recently missed the statistical significance threshold at an interim analysis, triggering a selloff in Summit shares.
Ivonescimab’s goal is to replace PD-1 immunotherapy, while ADCs are becoming better chemotherapy, and the two are not necessarily at odds, Xia said.
If both PD-1xVEGF-chemo and Keytruda-ADC work, “the next frontier could well be ivo plus an ADC,” Xia said. “Beyond that, if ADCs prove safe enough, there’s every possibility of exploring ivo plus ADC plus chemo. But let’s conquer one fortress at a time.”
Just a few days before Harmoni-2’s OS readout, Merck posted the first global clinical trial for its PD-1xVEGF candidate, MK-2010. The phase 2 study will combine the ivonescimab competitor with sac-TMT in advanced solid tumors.