A surprise interim trial miss in ivonescimab’s closely watched Harmoni-3 trial has sent Summit Therapeutics’ shares tumbling, casting a shadow over the company’s momentum as the star PD-1xVEGF bispecific heads into the ASCO annual meeting with a prestigious spot on the plenary session.
The squamous cohort of Harmoni-3 has apparently missed statistical significance on progression-free survival (PFS) in a recently added interim analysis. Summit said Thursday an independent data monitoring committee recommended the study continue as planned and remain double-blinded.
The trial is evaluating ivonescimab combined with chemotherapy against Merck & Co.’s Keytruda plus chemo for first-line treatment of metastatic non-small cell lung cancer.
Summit’s shares plunged more than 25% on the news in early trading Friday morning. Analysts were broadly surprised and scrambled to figure out what could have happened in the black box.
The market’s negative reaction is reasonable, Leerink Partners analysts wrote in a May 1 note to clients. The global Harmoni-3 trial is so important that Summit has tweaked its design at least twice: first to separate squamous and nonsquamous patients into two cohorts, and second, announced in February, to add a new interim analysis for PFS in the squamous group in the second quarter.
At the time, Summit framed the addition as an opportunity to speak with health authorities about a potential earlier submission, while the trial continues toward its final PFS and interim overall survival readouts, which the company still expects in the second half of 2026.
Missing statistical significance at the interim analysis isn’t the end of the world.
“To achieve statistical significance, there was a meaningfully higher bar than the upcoming planned final PFS analysis based on the minimal alpha spent on the interim analysis,” Summit said in its April 30 press release.
But it’s still a major disappointment for investors. Harmoni-2 and Harmoni-6, two Chinese first-line NSCLC trials run by ivonescimab’s original developer, Akeso, both met their PFS endpoints at interim analyses. The first study showed ivonescimab monotherapy beat Keytruda by 49% on PFS in PD-L1-positive NSCLC. The second linked ivonescimab plus chemotherapy to a 40% PFS advantage over BeOne Medicines’ Tevimbra plus chemotherapy in squamous NSCLC.
Analysts tried to crunch the numbers, but without knowing the exact statistical plan, any analysis so far is guesswork.
“With likely minimal (~0.001) alpha spent on the interim, the bar was indeed high and most of the alpha is preserved for the final, we believe supporting eventual success,” Citi analysts wrote in a Thursday note.
Citi analysts suspected the total number of PFS events accrued was lower than expected, causing the miss on statistical significance. But they couldn’t rule out that the magnitude of PFS improvement was simply weak.
In Leerink’s base-case scenario, missing statistical significance suggests the PFS effect size is about 10 percentage points lower than that seen in Harmoni-6. But the team cautioned that “considerable uncertainty remains,” citing multiple unknowns, most crucially Summit’s statistical design.
Harmoni-3’s latest PFS miss comes as an overall survival update from Harmoni-6 is set for a plenary session at the American Society of Clinical Oncology annual meeting, which kicks off later this month in Chicago.
The result’s plenary placement excited investors. In an April 21 note, Leerink analysts said ivonescimab combined with chemotherapy likely drove a strong overall survival benefit, with at least a 28% improvement over Tevimbra plus chemotherapy.
Analysts at both Leerink and Evercore ISI suggested that the stakes are now even higher for the Harmoni-6 readout.
“[A] degradation in PFS benefit from the Chinese to global trial would make the combination a less compelling option,” Leerink wrote in a Friday note.
“With investors expecting a degradation in benefit from the Chinese to global studies, the bar for excitement is higher,” the team added for the comparison between Harmoni-3 and -6.
Similarly, an Evercore team wrote in a Thursday note that, “If the deterioration isn’t as dramatic as previously thought, that could be the saving grace until we get H-3 final PFS & interim OS in 2H26.”
