Background
Phase 3, randomized, double-blind, multicenter trial; N=588 (China/Thailand; 57 China hospitals in the analysis); PD-L1-positive (CPS ≥5), locally advanced, resectable gastric/gastro-oesophageal junction adenocarcinoma; perioperative serplulimab + SOX vs perioperative chemotherapy.
Results
Interventions and follow up: Arm A: Neoadjuvant serplulimab 4.5 mg/kg IV + SOX (oxaliplatin 130 mg/m² day 1 + S-1 40–60 mg PO BID days 1–14) ×3 cycles → surgery → adjuvant serplulimab ×up to 17 cycles (n=292)
Arm B: Neoadjuvant placebo + SOX ×3 → surgery → adjuvant SOX ×5 cycles (n=296)
Primary endpoint: Investigator-assessed event-free survival (EFS); tested first in CPS ≥10, then ITT (CPS ≥5)
mFollow up: 35.9 mo (ITT); 42.7 mo (CPS ≥10)
Results: EFS (CPS ≥10): NR vs 42.0 mo; HR 0.65 (95% CI 0.47–0.90); P=.0082
EFS (ITT, CPS ≥5): NR vs 35.9 mo; HR 0.73 (95% CI 0.56–0.94); P=.015
pCR: 21.6% (serplulimab arm)
OS: NR (immature)
Arm B: Neoadjuvant placebo + SOX ×3 → surgery → adjuvant SOX ×5 cycles (n=296)
Primary endpoint: Investigator-assessed event-free survival (EFS); tested first in CPS ≥10, then ITT (CPS ≥5)
mFollow up: 35.9 mo (ITT); 42.7 mo (CPS ≥10)
Results: EFS (CPS ≥10): NR vs 42.0 mo; HR 0.65 (95% CI 0.47–0.90); P=.0082
EFS (ITT, CPS ≥5): NR vs 35.9 mo; HR 0.73 (95% CI 0.56–0.94); P=.015
pCR: 21.6% (serplulimab arm)
OS: NR (immature)
Adverse events
Grade ≥3 treatment-related (any): 47% vs 59%
Discontinuation (treatment-related): 7% vs 11%
Discontinuation (treatment-related): 7% vs 11%
Conclusions
Perioperative serplulimab + SOX with a chemotherapy-sparing serplulimab-monotherapy adjuvant phase significantly improved EFS vs perioperative chemotherapy in PD-L1-positive resectable gastric/GEJ adenocarcinoma, with a better safety profile. OS data are awaited.
Key Limitations
Almost entirely China-enrolled (generalizability); OS immature; asymmetric adjuvant design (serplulimab monotherapy vs continued SOX) confounds the perioperative comparison; SOX (not FLOT) chemotherapy backbone; restricted to PD-L1 CPS ≥5.
Clinical Context
Serplulimab (HANSIZHUANG) has an NDA accepted with priority review by China's NMPA for neo/adjuvant gastric cancer; it is not FDA/EMA-approved in this setting. Perioperative durvalumab + FLOT (MATTERHORN) became the first FDA-approved (2025) and EMA-approved (2026) perioperative immunotherapy for resectable gastric/GEJ cancer. ASTRUM-006 adds CPS-selected, chemo-sparing-adjuvant evidence to the emerging perioperative immunotherapy paradigm (cf. the mixed KEYNOTE-585 results).