An experimental lung cancer medicine developed by Akeso and Summit Therapeutics reduced the risk of death by 34% in a late-stage study of patients with squamous non-small-cell lung cancer, according to data released ahead of the American Society of Clinical Oncology annual meeting.
In the Phase 3 Harmoni-6 trial, conducted in China, ivonescimab combined with chemotherapy produced median overall survival of 27.9 months, compared with 23.7 months for patients receiving a standard PD-1 immunotherapy, tislelizumab, plus chemotherapy. The four-month survival difference was statistically significant. A global Phase 3 study is still underway, and researchers said those results will be important in assessing whether the findings apply outside China.
Ivonescimab is a bispecific antibody designed to target both PD-1, a protein involved in immune evasion by cancer cells, and VEGF, a protein linked to blood-vessel growth that can support tumors. The drug has attracted attention because it could compete with established checkpoint inhibitors such as Merck’s Keytruda, though some clinicians and analysts caution that earlier oncology approaches with promising data have not always produced broad clinical success.
Safety remains an area of review. Bleeding events occurred in nearly one-quarter of patients receiving ivonescimab, about twice the rate in the control group, though severe bleeding was reported in less than 3% of patients.
Experts described the trial as encouraging for patients in China while noting uncertainties about the clinical significance of a four-month median survival gain and the unusually strong performance of the control group. Summit’s share price has risen sharply since licensing the drug outside China, but it has recently declined amid questions about global applicability.
The results come as drugmakers pursue PD-1/VEGF combinations and other new lung cancer treatments. Upcoming global data from Summit may clarify ivonescimab’s role in future treatment decisions. Regulators have not yet reviewed the complete dataset.