Fresh off a phase 3 win in biliary tract cancer, Akeso CEO Michelle Xia, Ph.D., is touting the PD-1xVEGF bispecific antibody ivonescimab’s ability beyond non-small cell lung cancer, even as she continues to defend the drug’s data against skepticism.
In a China phase 3 trial, ivonescimab plus chemotherapy prolonged patients’ lives compared with the standard combination of AstraZeneca’s Imfinzi and chemo for the first-line treatment of advanced biliary tract cancer, Akeso announced Tuesday.
The company described the overall survival (OS) win on the Harmoni-GI1 trial’s primary endpoint as “clinically meaningful and statistically significant.” The study also met its key secondary endpoints of progression-free survival and objective response rate, according to Akeso.
The positive readout—marking ivonescimab’s first successful late-stage trial outside lung cancer—validates the drug’s novel bispecific construct and underpins its broad potential across tumor types, Xia told Fierce.
At the same time, the chief executive pushed back against new debates around dataset fluctuations in NSCLC, arguing that some investor speculations overlook ivonescimab’s efficacy advantages.
Positive biliary tract cancer readout
Although detailed data remain under wraps, Citi analysts in a Tuesday note said Harmoni-GI1’s outcome “further broadens the scope of evidence supporting ivonescimab’s unprecedented track record in likely defining a new global standard in cancer immunotherapy.”
While a Jefferies analyst called the win “unexpected” because biliary tract cancer is difficult to treat, Xia said she had always felt confident about the readout because its phase 2 data were already so encouraging.
Previously, phase 2 data from just 22 patients linked ivonescimab and chemo to a median OS of 16.8 months. Back in 2022, the FDA approved Imfinzi-chemo as the first immunotherapy-based treatment for biliary tract cancer after the regimen extended median OS to 12.9 months versus 11.3 months for chemo alone in the phase 3 Topaz-1 trial.
Roche tried adding the VEGF inhibitor bevacizumab (Avastin) to its PD-L1 inhibitor Tecentriq and chemo but observed no OS benefit.
Now, Harmoni-GI1 “once again proved that ivonescimab’s mechanism of action is different from a simple combination of two drugs,” Xia said.
As its bispecific construct enhances both PD-1 and VEGF inhibitions, “our position is that ivonescimab holds broad-spectrum potential” across tumor types, Xia said, pointing to the company’s ongoing Harmoni-GI2 trial in first-line pancreatic cancer and Harmoni-BC1 in first-line triple-negative breast cancer.
Xia declined to offer any additional description of the data as the detailed results are being prepared for presentation at the European Society for Medical Oncology annual meeting this October. But she pointed to how Imfinzi became the standard of care in biliary tract cancer by demonstrating just a 20% reduction in patients’ risk of death.
Outside of China, the ECOG-ACRIN Cancer Research Group recently registered a new phase 2/3 trial slated to begin early next year testing ivonescimab plus chemo in biliary tract cancer likely among a global population. While Akeso’s U.S. partner, Summit Therapeutics, is not the sponsor, the trial is designed to be registrational, Xia said.
A significant unmet need remains in biliary tract cancer. Akeso finished enrolling Harmoni-GI1’s nearly 700 participants in just 11 months, as the cancer is tracking at 100,000 new incidents annually in China alone and growing, Xia noted.
From a global perspective, the indication’s global market is valued at between $1 billion to $4 billion, which indicates room for expansion, she said, citing market research.
Addressing concerns about deteriorating NSCLC data
Despite the nice surprise from biliary tract cancer, “it doesn’t really matter to the [Summit] thesis,” Jefferies’ analyst Raisal Khurshid, wrote in an Aug. 25 note. Public attention around ivonescimab—and its Western developer, Summit—remains laser-focused on the upcoming Harmoni-3 readout in first-line NSCLC, which is expected this year.
Investors were already unnerved after the study had missed statistical significance on progression-free survival during a recently added interim analysis for the squamous cohort. Then, when Chinese authorities approved ivonescimab plus chemo for first-line squamous NSCLC, the label revealed an updated PFS analysis of the China-only Harmoni-6 trial, which showed that the PFS advantage had deteriorated from 40% during a February 2025 data cut-off to 28% during a September 2025 cut-off.
The data raised concerns that, even if Harmoni-3 meets statistically significant PFS at its final analysis, the magnitude of benefit would not be convincing enough to get the FDA on board with an accelerated approval before a mature OS readout.
In an Aug. 14 note, Leerink Partners analysts suggested that “it is probable” that the OS outcome from Harmoni-6 will also degrade with longer follow-up, and that the global Harmoni-3’s OS improvement will fall below 20%. Based on a February 2026 cut-off, the China-only Harmoni-6 study reflected an OS improvement of 34%.
To be clear, as Harmoni-6 hit PFS and OS endpoints at interim analysis, those results became the final analysis, and all follow-ups are therefore exploratory and will not change the study’s statistical significance status. Meeting OS, the gold standard in oncology drug development, is sufficient to support ivonescimab plus chemo as a new standard of care no matter how PFS may change, Xia argued.
It’s not uncommon for immunotherapies to have smaller effect sizes in longer-term analysis following a primary analysis, Xia noted. For example, in the phase 3 Keynote-407 trial in first-line squamous NSCLC, Keytruda plus chemo’s PFS benefit over chemo alone declined slightly from 44% at an initial analysis to 43% during the protocol-specified final analysis with additional follow-up; its OS benefit also dropped from 36% to 29%. Nevertheless, these data do not challenge Keytruda’s position, Xia noted.
However, Summit’s investors are concerned about the phenomenon’s read-through to Harmoni-3, especially as they already expect some efficacy decreases in translation of Chinese lung cancer data to a global population.
In response, Xia highlighted a flattening of the OS curve in Harmoni-6 as the trial progresses, which she said suggests a longer-term treatment effect that for now doesn’t seem likely to experience a sudden step-down in the future.
The Akeso CEO also brought up the Harmoni-A trial for previously treated EGFR-mutated NSCLC. In that China study, ivonescimab’s OS improvement over chemotherapy first shrank from 28% to 20% before meeting statistical significance by posting a 26% win at the final analysis. To Xia, current investor debates around Harmoni-3 are pure guess work, which depends on what data one chooses to look at.
“Scientific questions ultimately demand scientific proof,” Xia said. “You can’t just make simple linear extrapolations to determine scientific outcomes.”
Should Summit find a Big Pharma partner?
Summit added Harmoni-3’s interim PFS analysis for the squamous cohort at the beginning of 2026 so it could start a conversation with regulators early. Xia suggested that the timing of the analysis was not optimal.
The Chinese Harmoni-6 trial enrolled faster than the global Harmoni-3, so the Summit analysis was a bit premature, Xia said. The Akeso team remains ever more confident in Harmoni-3, she added.
This marks the second time that Summit had to deal with the aftermath of an unideal trial outcome because of a design or logistical fumble. Previously, the Harmoni trial in EGFR-mutated NSCLC—the first global phase 3 data for ivonescimab—failed to meet OS at its final analysis. To convince the FDA, Summit is resorting to post-hoc OS analyses because the Western portion of the study recruited patients slower than expected, leading to initially immature OS data in that key subgroup in its bid for FDA approval.
The Harmoni-3 case once again raises questions around Summit’s capability and whether the small biotech can properly manage a large asset as ivonescimab, the first-in-class PD-1xVEGF bispecific, where Big Pharma is swarming.
“I think you’re right. From my observation, Summit does need more resources and more investments,” Xia said.
Akeso has been advancing its “IO 2.0” strategy revolved around novel combinations, which Xia said represents the future of immuno-oncology. Summit has signed some partners for ivonescimab, including with Revolution Medicines on the latter’s RAS(ON) inhibitor daraxonrasib, which on Wednesday won a quick FDA approval in pancreatic cancer after making waves in the oncology world. But the Florida biotech’s internal pipeline and clinical program for novel combinations remain thin compared with its Big Pharma competitors.
Industry watchers are anxiously waiting for the Harmoni-3 readout as the PD-1xVEGF space heats up quickly. On Tuesday, ClinicalTrials.gov published Merck & Co.’s first global clinical trial for its PD-1xVEGF candidate, MK-2010. The phase 2 study aims to combine the drug, licensed from the SBP Group (previously Sino Biopharmaceutical), with the New Jersey pharma’s TROP2 antibody-drug conjugate, sac-TMT, which is partnered with China’s Kelun-Biotech, in various solid tumors.
Earlier this month, AbbVie disclosed a global phase 3 study for its PD-1xVEGF candidate, ABBV-1480, which it licensed from China’s RemeGen. Targeting Harmoni-3, the planned indication is first-line squamous NSCLC.