JPMorgan has released a research report following Akeso's (09926) updated overall survival (OS) data from its HARMONi-2 trial, noting that the results are notably robust. The hazard ratio (HR) stands at 0.73 with a confidence interval of 0.57 to 0.95 and a p-value of 0.009. The median OS for the ivonescimab group reached 30.8 months, surpassing the pembrolizumab group's median OS of 22.6 months by more than eight months.

The investment bank has maintained its "Overweight" rating on Akeso with a target price of HK$125, based on a valuation as of December 2027. This valuation employs a discounted cash flow methodology, assuming a terminal growth rate of 3% and a weighted average cost of capital of 9.6%.

Where to begin with the analysis? JPMorgan points out that the median OS for pembrolizumab in the HARMONi-2 trial is approximately two months longer than that observed in the KEYNOTE-042 China study. Given the emergence of numerous novel non-small cell lung cancer (NSCLC) therapies, including antibody-drug conjugates (ADCs), since KEYNOTE-042 commenced in 2014, this outcome is not unexpected. The safety profiles of the two groups are broadly comparable, with serious treatment-related adverse event (TRAE) rates of 29.9% for the ivonescimab group versus 21.6% for the pembrolizumab group, and TRAE rates leading to permanent discontinuation at 4.1% and 5%, respectively.

The significance of these findings extends beyond the immediate results. JPMorgan believes that the updated OS data from HARMONi-2 will have positive implications for two global Phase III trials: the squamous cell cohort of HARMONi-3 and the HARMONi-7 trial. In the HARMONi-2 study, the OS HR for squamous cell patients was 0.65, notably lower than the OS HR of 0.79 observed in non-squamous cell patients. Furthermore, for patients with PD-L1 TPS equal to or greater than 50%, the OS HR was 0.58, and this patient population is precisely the target for HARMONi-7.

The report also clarifies several aspects of the HARMONi-2 statistical plan. The current WCLC data represents the results of the second planned interim analysis. The earlier OS data (OS HR=0.777) was from the first interim analysis, which consumed minimal alpha. The second interim analysis was pre-specified with 232 events and a one-sided alpha of 0.0139. As of the data cutoff date of August 20, 2026, with a median follow-up of approximately 36 months, 234 OS events had occurred, resulting in a one-sided alpha of 0.0141. At the time of the first OS interim analysis, hypothetically in February 2025, the median follow-up was only about 17 months, significantly lower than the median OS of both groups, rendering the first interim analysis OS HR of 0.777 relatively immature.