ASCO-GI – Jazz sees a new Ziihera Herizon

Jazz, BeOne and Zymeworks hope that their HER2 x HER2 bispecific MAb Ziihera could become the new drug of choice in first-line HER2-positive gastric cancer, after reporting positive results from the phase 3 Herizon-GEA-01 trial.

Still, investors seemed less convinced by the data, released unexpectedly on Tuesday ahead of full presentation at the ASCO-GI meeting on Thursday, presumably owing to an embargo breach. For one, the results fell short of those from a phase 2 trial. And there are also questions about the role of Tevimbra, which was combined with Ziihera in one of the Herizon-GEA-01 arms; BeOne’s PD-1 blocker doesn’t appear to add much in terms of efficacy, but was linked with an uptick in adverse events.

In addition, while the groups claimed a consistent benefit regardless of PD-L1 expression, there are hints that the result might be being driven by PD-L1-negatives.

Jazz’s stock sank 4% on Tuesday, while Zymeworks, which originated Ziihera, fell 5%. BeOne, which as well as owning Tevimbra has rights to Ziihera in Asia, was flat.

Triplet vs doublet

Herizon-GEA-01 tested Ziihera plus chemo, with or without Tevimbra, versus Herceptin plus chemo, in patients with HER2-positive front-line gastric and oesophageal cancers. The co-primary endpoints were progression-free and overall survival.

On PFS, results were almost identical with the doublet and triplet, with medians of 12.4 months, representing around a 35% reduction in the risk of progression or death versus control.

On OS the triplet looked to have a slight numerical edge, with a median of 26.4 months and a hazard ratio of 0.72 versus control, compared with 24.4 months and 0.80 respectively with the doublet. Indeed, at this first interim OS analysis, the doublet didn’t hit statistical significance – as disclosed when the partners toplined the data in November.

Herizon-GEA-01 in first-line HER2+ gastroesophageal adenocarcinoma

Notes: *co-primary endpoints; **didn’t hit statistical significance at first interim OS analysis; data cutoff Oct 2025. Source: ASCO-GI & company releases.

The companies promised another interim OS analysis of Ziihera plus chemo in mid-2026, so the doublet could still have a chance to show an OS benefit.

The investigators concluded that Ziihera could replace Herceptin in the front line, but results look less impressive than a phase 2 trial that showed mPFS of 12.5 months and mOS of 36.5 months among 46 patients with HER2-expressing gastroesophageal adenocarcinoma receiving Ziihera plus chemo.

Ahead of the Herizon-GEA-01 readout, Evercore ISI’s Jonathan Miller said that “mPFS should be into the teens, and mOS could be north of 30 months”. Still, some slippage between phase 2 and phase 3 is to be expected.

Adverse events

That said, perhaps most striking, in the context of similar efficacy results, is an increase in toxicity with the triplet versus the doublet.

72% of patients receiving Ziihera/Tevimbra/chemo had grade 3 or higher adverse events, and 12% discontinued owing to adverse events, versus 59% and 9% respectively with the doublet.

The most common grade 3 or higher treatment-related adverse event was diarrhoea (25% with the triplet, and 20% with the doublet).

PD-L1 expression

Tevimbra plus chemo is already approved in the US for first-line gastric cancer, but only for HER2-negative patients with PD-L1 expression of 1% or higher. A Keytruda/Herceptin/chemo combo is indicated for first-line HER2-positive disease, but again only in PD-L1 ≥1% expressers.

In Herizon-GEA-01, the partners claimed a consistent benefit regardless of PD-L1 status, which could help Ziihera, with or without Tevimbra, reach a market that PD-(L)1 inhibitors cannot.

The groups didn’t disclose full details of this subgroup analysis, but BeOne gave results with the triplet, which suggested that the benefit is being driven by PD-L1-negatives. In this subgroup, the hazard ratios were 0.47 for PFS and 0.49 for OS, while in PD-L1-positives the hazard ratios were 0.65 for PFS and 0.82 for OS. Around a third of patients in the trial had PD-L1 expression <1%.

It therefore seems possible that any Ziihera approval could be limited to PD-L1-negative patients. Jazz is planning a “rapid” FDA submission, and BeOne is set to file a supplemental BLA in the US for Tevimbra, as well as filing Tevimbra and Ziihera in China.

One potential upcoming threat could be AstraZeneca and Daiichi’s HER2 ADC Enhertu, whose Destiny-Gastric05 trial has an exploratory PD-L1-negative cohort; however, results look some way off.