Durvalumab Plus FLOT Improves Resectable Gastric Cancer

A clinically meaningful event-free survival gain

The phase 3 MATTERHORN trial provides randomized evidence that adding the PD-L1 inhibitor durvalumab to perioperative FLOT can reduce the risk of progression, recurrence, or death for patients with resectable gastric or gastroesophageal junction adenocarcinoma. The event-free survival hazard ratio was 0.71, with a 95% confidence interval of 0.58 to 0.86 and a p-value below 0.001.

At 24 months, estimated event-free survival was 67.4% with durvalumab plus FLOT and 58.5% with FLOT alone. The absolute difference of 8.9 percentage points helps translate the relative effect into a clinically recognizable outcome, although longer follow-up is needed to determine how much of that separation persists and whether it produces a durable overall survival advantage.

The pathologic findings point in the same direction. Pathologic complete response—no residual viable tumor identified in the resected primary tumor and sampled lymph nodes—occurred in 19.2% of participants assigned durvalumab and 7.2% assigned placebo. That represents a 12-percentage-point absolute increase and supports the biological activity of introducing immunotherapy before resection rather than reserving it exclusively for the postoperative period.

The randomized, placebo-controlled design supports a causal interpretation for the treatment strategy in patients resembling those enrolled. It does not establish that every component is necessary, that the benefit is uniform across molecular subtypes, or that pathologic complete response itself guarantees cure.

How MATTERHORN tested the perioperative strategy

MATTERHORN enrolled 948 adults with resectable gastric or gastroesophageal junction adenocarcinoma. Participants were randomly assigned to receive durvalumab or placebo with perioperative FLOT, the combination of fluorouracil, leucovorin, oxaliplatin, and docetaxel. Treatment began before surgery, resumed after recovery, and continued with durvalumab or placebo maintenance as specified in the protocol.

This design matters when interpreting the results. The trial evaluated an entire perioperative pathway, not a single isolated intervention. It therefore cannot determine whether the event-free survival benefit arose mainly from preoperative immune priming, postoperative treatment, maintenance therapy, or the combination of these phases.

The primary endpoint was event-free survival, which incorporated progression that prevented surgery, local or distant recurrence, and death. Pathologic complete response was an important secondary measure. At a median follow-up of approximately 31.5 months, the event-free survival analysis was statistically positive, while longer observation remained important for characterizing overall survival and late toxicity.

FLOT was an appropriate active comparator. The earlier FLOT4 trial established perioperative FLOT as a major treatment option for fit patients with locally advanced, resectable gastric or gastroesophageal junction adenocarcinoma by demonstrating better survival than older epirubicin- and platinum-containing regimens. MATTERHORN therefore asked whether immune checkpoint inhibition could improve an established modern chemotherapy backbone rather than comparing against an obsolete or minimally active regimen.

Safety was broadly consistent with the intensity expected from perioperative FLOT. Grade 3 or 4 adverse events occurred in 71.6% of the durvalumab group and 71.2% of the control group. Similar aggregate rates do not make the regimens interchangeable: durvalumab introduces immune-mediated risks involving organs such as the thyroid, lungs, liver, bowel, skin, and adrenal glands. These toxicities can complicate surgical timing or postoperative recovery even when their overall incidence is comparatively low.

Where the findings fit in practice

The key practical change is that immunotherapy can now be considered in treatment planning before surgery, rather than deferred until pathology is available. That makes early multidisciplinary coordination more consequential. Medical oncology, surgical oncology, gastroenterology, pathology, radiology, and nutrition teams need a shared plan for staging, treatment fitness, restaging, operative timing, and postoperative resumption.

The evidence applies most directly to patients who are sufficiently fit for FLOT and are treated with curative intent. It should not be automatically extrapolated to frail adults, patients unable to receive docetaxel or oxaliplatin, tumors considered unresectable, or patients managed with a different preoperative regimen. FLOT itself is demanding, and trial eligibility generally selects for better performance status and organ function than are encountered across the full clinical population.

MATTERHORN also differs from evidence supporting adjuvant nivolumab after neoadjuvant chemoradiation and surgery for residual esophageal or gastroesophageal junction cancer. Those strategies begin from different preoperative regimens, eligibility criteria, tumor distributions, and postoperative findings. They should not be treated as interchangeable checkpoint-inhibitor templates.

Biomarker interpretation remains unsettled. The trial was not limited to patients selected by PD-L1 expression, but an overall all-comer benefit does not prove equal benefit at every combined positive score. Mismatch-repair deficiency, microsatellite instability, HER2 status, Epstein-Barr virus association, and other molecular features may identify groups with different baseline prognosis or sensitivity to immunotherapy. Subgroup analyses can generate useful hypotheses but are rarely powered to define exclusions on their own.

In the United States, clinical adoption must also account for the current FDA indication, guideline placement, payer coverage, infusion capacity, and the ability to recognize immune-related adverse events promptly. A positive phase 3 trial supports benefit; it does not independently settle regulatory or access questions.

Important uncertainties and limitations

Follow-up remains relatively short for a curative-intent gastric cancer trial. Event-free survival is clinically important, but mature overall survival will better establish whether earlier disease control translates into longer life. Pathologic complete response is supportive and mechanistically informative, yet it is not a substitute for survival outcomes.

The study cannot isolate the contribution of preoperative durvalumab from postoperative or maintenance treatment. It also offers limited certainty for small molecular subgroups, older or frailer patients, and people treated outside experienced multidisciplinary centers. Toxicity percentages may not fully capture the operational consequences of endocrine monitoring, corticosteroid treatment, surgical postponement, or prolonged immune-mediated events.

AstraZeneca, the manufacturer of durvalumab, sponsored MATTERHORN, and reported author relationships should be considered when appraising the study. Randomization, placebo control, a clinically meaningful primary endpoint, and concordant pathologic findings strengthen confidence, but independent follow-up and mature survival data remain valuable.

Questions clinicians ask

Does PD-L1 expression determine who should receive the regimen?

MATTERHORN was not restricted to a PD-L1-selected population, so the overall result supports a biomarker-unselected treatment effect. It does not show that benefit is identical across PD-L1 strata, and subgroup estimates should be interpreted cautiously unless supported by adequate sample size, interaction testing, and consistent outcomes.

Can pathologic complete response guide postoperative durvalumab?

Not from this trial alone. MATTERHORN tested a planned perioperative strategy that included treatment after surgery; it was not designed to randomize patients according to pathologic response or determine whether those achieving complete response can safely omit subsequent durvalumab.

What does the study say about patients with MSI-high tumors?

Mismatch-repair-deficient or microsatellite instability-high gastric cancers may be especially immunotherapy-sensitive, but they represent a relatively small molecular subgroup. MATTERHORN’s overall findings do not resolve the optimal chemotherapy intensity, treatment duration, or surgery strategy for that population, so dedicated analyses and other trials remain relevant.

Should durvalumab plus FLOT replace every perioperative regimen?

No. The evidence supports the combination for appropriately selected patients resembling the trial population, particularly those fit for FLOT. Decisions still depend on tumor location and stage, surgical candidacy, molecular findings, comorbidities, treatment access, regulatory status, and whether an alternative evidence-based pathway such as chemoradiation is being used.

References

1. Perioperative durvalumab plus FLOT chemotherapy in resectable gastric cancer (MATTERHORN phase 3) — PubMed, 2026 2. Durvalumab in Combination With FLOT Chemotherapy in Participants With Resectable Gastric and Gastroesophageal Junction Cancer (MATTERHORN) — ClinicalTrials.gov, 2020 3. Perioperative chemotherapy with fluorouracil plus leucovorin, oxaliplatin, and docetaxel versus fluorouracil or capecitabine plus cisplatin and epirubicin for locally advanced, resectable gastric or gastro-oesophageal junction adenocarcinoma (FLOT4): a randomised, phase 2/3 trial — The Lancet, 2019