MATTERHORN's final results: durvalumab plus FLOT extends survival in gastric cancer

The final results of MATTERHORN show that adding the immunotherapy durvalumab to perioperative FLOT chemotherapy lengthened overall survival in patients with resectable gastric and gastro-oesophageal junction adenocarcinoma, with a 22% reduction in the risk of death (hazard ratio 0.78; 95% CI 0.63 to 0.96; p=0.021) [s1]. That result crosses the statistical line the trial's earlier interim analysis had not reached, and the investigators now describe the regimen as a new standard treatment option for this cancer [s1][s2].

The trial matters because the current standard, perioperative FLOT — fluorouracil, leucovorin, oxaliplatin and docetaxel given before and after surgery — cures some patients but leaves recurrence rates stubbornly high [s2]. MATTERHORN tested whether layering a PD-L1-blocking antibody onto that backbone could push more patients into durable remission. Its first report, in 2025, answered the near-term question; this one answers the one that counts most.

What the trial did

MATTERHORN was a global, randomised, double-blind, placebo-controlled phase 3 trial run at 147 centres in 20 countries [s1]. It enrolled 1,258 adults with resectable, previously untreated gastric or gastro-oesophageal junction adenocarcinoma and randomly assigned 948 of them 1:1 — 474 to durvalumab plus FLOT and 474 to placebo plus FLOT [s1]. Durvalumab was given at 1,500 mg every four weeks: two neoadjuvant and two adjuvant cycles alongside FLOT, then ten further cycles of the antibody alone [s2]. The population's median age was 62 years, and 682 participants (72%) were male [s1]. The primary endpoint had been event-free survival; overall survival was the key secondary endpoint and the one reported as the main result here [s1].

What it found

At the earlier interim analysis, durvalumab had already improved event-free survival: the two-year estimate was 67.4% versus 58.5% with placebo (hazard ratio 0.71; 95% CI 0.58 to 0.86; P<0.001), and pathological complete response — no residual cancer found at surgery — nearly tripled, from 7.2% to 19.2% (relative risk 2.69; 95% CI 1.86 to 3.90) [s2]. But at that point overall survival was not yet convincing: two-year survival was 75.7% versus 70.4%, a difference that did not clear the trial's stringent significance threshold [s2].

The final analysis closes that gap. Overall survival was significantly improved with durvalumab (HR 0.78; 95% CI 0.63 to 0.96; p=0.021) [s1]. Safety was consistent with the earlier report — grade 3 or 4 adverse events had occurred in 71.6% of the durvalumab group and 71.2% of the placebo group, essentially identical — and treatment-related deaths remained rare, in 6 of 475 participants (1%) on durvalumab and 2 of 469 (under 1%) on placebo [s1][s2].

How to read it

The honest framing is that this is a modest absolute gain on a hard endpoint, not a transformation. A hazard ratio of 0.78 means the survival curves separate meaningfully, but the trial's own two-year survival figures — roughly a five-point gap at the interim — set expectations for the size of the benefit an individual patient might see [s1][s2]. What makes it persuasive is the coherence of the package: a higher pathological complete response rate, better event-free survival, and now longer overall survival all point the same way [s1][s2].

Two caveats belong in view. The trial was funded by AstraZeneca, which makes durvalumab; the endpoints were centrally defined and the design double-blind, which are meaningful safeguards, but the funding is relevant context [s1]. And the adjuvant tail of treatment is long — ten additional cycles of antibody after surgery — so the toxicity and cost of the regimen extend well beyond the chemotherapy phase, even though serious adverse events were no more common overall [s1][s2].

Why it matters

Perioperative chemotherapy for gastric cancer had not seen a practice-changing addition in years, and immunotherapy's record in this disease has been mixed — strong in some advanced settings, disappointing in others. MATTERHORN is the first phase 3 trial to show that adding a checkpoint inhibitor to FLOT improves survival in the curative, resectable setting, which is where the stakes are highest because the goal is cure rather than control [s1].

It joins a run of trials reshaping gastric and junction cancer from several directions at once — including intraperitoneal chemotherapy for peritoneal spread in DRAGON-01 and the first-line approval of zanidatamab for HER2-positive disease. Each targets a different slice of a cancer that remains a leading cause of cancer death worldwide.

What to watch

The open questions are which patients drive the benefit — whether it concentrates in those with high PD-L1 expression or extends across the board — and how the perioperative regimen compares with strategies that reserve immunotherapy for after surgery. Longer follow-up will also show whether the survival curves keep diverging or converge over time [s1].

This article describes research and regulatory developments and is not medical advice. Treatment decisions are for patients and their treating clinicians.

Sources

• Perioperative durvalumab plus FLOT for resectable gastric and gastro-oesophageal junction adenocarcinoma (MATTERHORN): final results — The Lancet, online 9 September 2026
• Perioperative Durvalumab in Gastric and Gastroesophageal Junction Cancer — New England Journal of Medicine, online 1 June 2025