Summit looks to repeat ivo's Chinese success

Less than a year after Summit's anti-PD-1 x VEGF bispecific ivonescimab sent shockwaves through the biopharma industry the heat is on to prove that this was no fluke. The Harmoni study is this project's first phase 3 to have been conducted globally, and it's due to yield data in mid-2025.

The readout is being billed by some analysts as this year's biggest biotech catalyst. It will show whether ivonescimab's Chinese data can be replicated in a true global trial; the drug last May secured its first Chinese approval, as part of a chemo combo in second-line EGFR-mutant lung cancer on the back of Akeso's Harmoni-A trial, and Harmoni, run by Summit, has an analogous design.

True, the second-line NSCLC setting is a relative outlier, and ivonescimab's big success so far has come in the first-line, where the drug beat Keytruda in the separate Harmoni-2 trial. That too was conducted in China by Akeso, and has so far delivered only PFS data, which were scrutinised at length at last year's World Lung conference.

High profile

Ivonescimab has enjoyed a high profile since Summit paid Akeso $500m to license it in 2022, triggering numerous copycat deals. While Harmoni-2 should reveal a first cut of overall survival data late this year or in early 2026, for now the primary focus remains on Harmoni.

Harmoni tests ivonescimab plus chemo in patients who harbour an EGFR mutation but have already failed a third-generation EGFR inhibitor like Tagrisso. That's a very similar design to the Chinese Harmoni-A, except the latter included patients who had progressed after earlier-generation EGFR inhibitors too, not just Tagrisso.

Notably, while Harmoni-A succeeded on its primary endpoint of PFS versus chemo alone, its OS benefit was only numerical and not significant, amounting to a 2.6-month median benefit, and a 0.80 hazard ratio at 52% data maturity. Unlike Harmoni-A, Harmoni tests PFS and OS as primary endpoints, so one question is whether, and how comprehensively, the latter is met.

It's also worth considering how the absolute median OS numbers in Harmoni-A fit into the real world. The result – 17.1 months versus 14.5 months – is very close to that seen in Johnson & Johnson's Mariposa-2 study. In the latter Rybrevant plus chemo scored median OS of 17.7 months versus 15.3 months for chemo alone, and a 0.73 hazard ratio, according to an update at ESMO.

Cross-trial comparisons

Source: OncologyPipeline.

Though this benefit seems marginal, and the p value of 0.039 didn't clear the bar for significance at this interim analysis, Evercore ISI analysts see Mariposa-2 as the bar to success in second-line EGFRm NSCLC. 

Another relevant comparator is Innovent's Orient-31 trial of Tyvyt combined with an Avastin biosimilar plus chemo. This Chinese trial yielded median OS numbers that were rather higher than in either Mariposa-2 or Harmoni-A, but chemo control did far better too, and the result was nowhere near significant.

However, it's by no means clear that regulators will want to see a significant OS benefit before approving a drug in this setting, so perhaps a strong result on PFS in Harmoni, plus numerical superiority on OS, will suffice. Importantly, the FDA last year approved Rybrevant plus chemo in this setting based on Mariposa-2's PFS benefit alone.

That said, numerous companies have already found that Chinese data don't always presage a success in the west. There are plenty of other pivotal readouts ahead for ivonescimab, but investor faith will be badly shaken if Harmoni results don't at least approximate to those in Harmoni-A.