Gastric Cancer EFS Improves With Perioperative Serplulimab

An event-free survival signal in a selected population

ASTRUM-006 was a peer-reviewed phase 3 trial evaluating serplulimab, an anti-PD-1 antibody, alongside perioperative S-1 plus oxaliplatin chemotherapy, commonly called SOX. The relevant population had resectable gastric or gastroesophageal junction adenocarcinoma and PD-L1 expression meeting the trial’s combined positive score threshold.

The trial reported that adding serplulimab improved event-free survival compared with perioperative SOX alone. Event-free survival is particularly relevant here because it captures clinically important outcomes before or after surgery, rather than waiting only for death. Depending on the protocol definition, events may include progression that prevents surgery, local or distant recurrence after resection, or death.

The numerical abstract needed to verify the randomized sample size, hazard ratio, confidence interval, landmark event-free survival rates and median follow-up was not available in the source material supplied for this draft. Those values should therefore be taken directly from the indexed publication before clinical publication or policy use; inserting estimates would create false precision. The same caution applies to pathologic complete response, major pathologic response, surgical completion and overall survival results.

What can be interpreted without extrapolation is the trial’s direction of effect and its biomarker boundary. The evidence concerns serplulimab added to a defined perioperative chemotherapy program in PD-L1-selected disease. It should not be generalized automatically to all resectable gastric cancers, other chemotherapy backbones, or patients who would not have met the protocol’s surgical and fitness requirements.

Who resembles the ASTRUM-006 population

The closest clinical match is an adult with histologically confirmed gastric or gastroesophageal junction adenocarcinoma considered resectable with curative intent, no distant metastatic disease, and adequate performance status and organ function for both gastrectomy and multiagent systemic therapy. A patient treated for metastatic disease, unresectable local disease, squamous histology, or recurrent cancer would represent a different clinical setting.

Resectability alone is not enough. Perioperative trials generally enroll patients considered fit enough to start systemic therapy before surgery and to proceed to a protocol-defined operation if the tumor remains operable. Frailty, major organ dysfunction, uncontrolled autoimmune disease, active infection or prior exposure to immune-checkpoint therapy may affect eligibility and reduce how confidently the findings can be applied. The publication’s full inclusion and exclusion criteria remain the appropriate standard for determining trial similarity.

The chemotherapy backbone also matters. SOX combines the oral fluoropyrimidine S-1 with oxaliplatin and is used more commonly in East Asian gastric cancer practice than in the United States. US clinicians frequently encounter other perioperative regimens, including fluorouracil-, leucovorin-, oxaliplatin- and docetaxel-based treatment. A positive result with SOX does not prove that serplulimab produces the same benefit, toxicity profile or treatment completion rate with another backbone.

Anatomic location requires similar care. The trial included gastric and gastroesophageal junction adenocarcinoma, but that does not make the findings applicable to esophageal squamous cell carcinoma. It also does not resolve whether the magnitude of benefit was consistent across primary sites unless the prespecified subgroup results demonstrate that consistency.

Why the PD-L1 CPS threshold matters

PD-L1 combined positive score is calculated from PD-L1-staining tumor cells, lymphocytes and macrophages relative to the number of viable tumor cells, multiplied by 100. It is not interchangeable with tumor proportion score, which counts only PD-L1-positive tumor cells. The assay, specimen quality, sampling site and timing of biopsy can affect classification.

For treatment selection, “PD-L1 positive” is incomplete unless the cutoff is stated. The ASTRUM-006 result applies most directly to tumors meeting the CPS threshold specified for the trial’s efficacy population. A tumor with detectable staining but a CPS below that boundary should not be assumed to have the same evidence-supported benefit. Likewise, a result reported simply as positive or negative without the score and assay may be insufficient for matching a patient to the study.

CPS is a selection marker, not a guarantee of response. Some patients above a threshold will relapse despite treatment, while some below it may do well. The score should therefore be interpreted as defining the studied population and modifying the estimated probability of benefit—not as a binary measure of individual sensitivity.

Biomarker assessment also needs to fit into broader gastric cancer testing. HER2 status, mismatch-repair or microsatellite-instability status, and other clinically relevant markers can influence systemic treatment choices. ASTRUM-006 should not be read as making PD-L1 the only relevant variable in perioperative decision-making.

Implications for US practice and policy

The trial supports the conclusion that perioperative immune-checkpoint inhibition can improve event-free survival when serplulimab is added to SOX in the population actually studied. Because this was a randomized phase 3 comparison, differences in event-free survival can be attributed more confidently to the assigned strategy than they could in an observational study, subject to adherence, missing data and other trial limitations.

That conclusion is narrower than a class-wide recommendation. The result does not establish interchangeability among PD-1 inhibitors, nor does it demonstrate efficacy equivalent to US-preferred perioperative chemotherapy. The trial itself also does not confer FDA approval. Regulatory review separately considers the complete efficacy dataset, safety, manufacturing, companion-diagnostic questions and relevance to the proposed population.

For health systems, the study highlights the operational importance of reporting an exact CPS, rather than only a qualitative PD-L1 result. Pathology workflows may need to document the assay, specimen and score clearly enough for clinicians to determine whether a patient resembles the trial population. Policymakers and guideline panels will also need mature survival, safety and generalizability data before deciding how the regimen fits among existing perioperative strategies.

Safety requires attention across the entire treatment pathway. Adding checkpoint blockade introduces the possibility of immune-mediated adverse events, while SOX retains the toxicities of fluoropyrimidine and platinum treatment. Perioperative studies must also show that preoperative therapy does not create an unacceptable increase in surgical delay, inability to undergo resection or postoperative complications. Efficacy cannot be considered separately from the proportion of patients who complete surgery and postoperative treatment.

Important uncertainties

Generalizability is the central limitation. Regional differences in screening, tumor biology, surgical technique, lymph-node dissection, supportive care and use of SOX may affect how the results translate to US practice. Trial participants are also usually fitter than many patients encountered in routine care.

Event-free survival is meaningful, but overall survival remains important, particularly when later therapies could alter outcomes. Longer follow-up is needed to determine whether the early advantage persists and translates into fewer deaths. Subgroup estimates may be informative but can be underpowered and should not be interpreted as definitive evidence of benefit or no benefit in small categories.

The full report should also be reviewed for immune-related toxicity, perioperative mortality, treatment discontinuation, surgery rates, protocol deviations, sponsor involvement and author conflicts. These details are necessary for a balanced assessment and cannot be reconstructed safely from a bibliographic citation alone.

Questions clinicians ask

Does this evidence apply to every resectable gastric cancer?

No. It applies most directly to patients resembling the trial population: resectable gastric or gastroesophageal junction adenocarcinoma, fitness for perioperative SOX and surgery, and PD-L1 CPS meeting the prespecified threshold. It should not be extrapolated automatically to metastatic disease, squamous histology, lower CPS categories or different chemotherapy backbones.

Is any detectable PD-L1 staining enough?

Not necessarily. Treatment matching depends on the numerical CPS cutoff used in ASTRUM-006, not merely a laboratory label of “PD-L1 positive.” Clinicians should confirm the reported score, assay and specimen because CPS differs from tumor proportion score and results near a cutoff can be affected by sampling and analytic variability.

Can serplulimab be substituted into another perioperative regimen?

The trial supports serplulimab with the studied SOX program; it does not prove equal benefit with FLOT or another regimen. Substitution would require separate evidence addressing efficacy, overlapping toxicity, surgical completion and postoperative recovery, as well as the relevant US regulatory and guideline status.

Is event-free survival enough to change practice?

A statistically and clinically credible event-free survival benefit can support regulatory or guideline consideration, particularly in curative-intent therapy. Decisions still require the exact effect estimate, follow-up, safety, surgery outcomes and maturity of overall survival data, together with an assessment of whether the population and chemotherapy backbone are applicable locally.

References

1. Serplulimab perioperative immunotherapy improves EFS in PD-L1-positive gastric cancer (ASTRUM-006) — PubMed, 2026 2. Gastric Cancer Treatment (PDQ)—Health Professional Version — National Cancer Institute, n.d. 3. FDA approves pembrolizumab with chemotherapy for HER2-negative gastric or gastroesophageal junction adenocarcinoma — US Food and Drug Administration, 2023