CAR-like T-cells in combination with a PD-1 inhibitor and S-1 plus oxaliplatin (SOX) was effective in patients with treatment-naive metastatic gastric or gastroesophageal junction (GEJ) adenocarcinoma, according to findings from a phase 2 study (ChiCTR2200061306) presented during 2026 Gastrointestinal Cancers Symposium.
The primary end point of objective response rate (ORR) was 76.0% (95% CI, 61.8%-86.9%) in the CAR-like T-cell group vs 50.0% (95% CI, 35.5%-64.5%) in the control group (P = .007). The disease control rate was 96.0% (95% CI, 86.3%-99.5%) vs 76.0% (95% CI, 61.8%-86.9%) between both arms respectively (P = .008).
The best response rates between the CAR-like T-cell group vs the control group was a partial response in 76% vs 50%, stable disease in 20% vs 26%, progressive disease in 4% vs 20%, and were not evaluable in 0% vs 4%. The median duration of response (DOR) was 10.3 months (95% CI, 6.6-14.0) vs 6.0 months (95% CI, 0.0-18.4; P = .414).
The median progression-free survival (PFS) was 10.3 months (95% CI, 9.3-11.3) in the CAR-like T-cell group vs 7.4 months (95% CI, 6.5-8.3) in the control group (HR, 0.64; 95% CI, 0.40-1.02; P = .0505). The subgroup analysis showed that the CAR-like T-cells was more effective in those who had 2 or more metastatic sites (HR, 0.385), peritoneum metastases (HR, 0.289), and a PD-L1 combined positive score (CPS) of 5 or more (HR, 0.221).
Any-grade adverse effects of any grade occurred in both groups The most common included thrombocytopenia, lymphopenia, leukopenia, elevated aspartate aminotransferase/alanine aminotransferase, anemia, neutropenia, and hypoproteinemia. Treatment-related adverse effects of grade 3 or higher, as well as serious adverse effects, were comparable between both arms.
Regarding exploratory outcomes, patients in the CAR-like T-cell group were stratified to either group A, having a PFS of at least 1 year, or group C, having a PFS of less than 6 months. In the CAR-like T-cell group, there was a more favorable efficacy in patients with tumors that had high levels of PD-L1, integrin αvβ3, and neuropilin 1.
“CAR-like T cells are cytokine activated T lymphocytes engineered with the tumor-targeting and penetrating peptide iRGD, which is anchored to the T-cell surface via lipid insertion, eliminating the need for genetic manipulation,” Qin Liu, MD, PhD, Department of Oncology, Nanjing Drum Tower Hospital & Group’s Suqian Hospital, in Nanjing China, said during the presentation. “In preclinical studies, we demonstrated that these cells showed synergistic antitumor effects with PD-1 blockade.”
Key Takeaways From the Phase 2 Study
- The combination of CAR-like T-cells, PD-1 inhibitor, and SOX demonstrated clinical activity in patients with previously untreated metastatic gastric or GEJ adenocarcinoma.
- The ORR was 76.0% (95% CI, 61.8%-86.9%) in the CAR-like T-cell group vs 50.0% (95% CI, 35.5%-64.5%) in the control group (P = .007).
- The most common any-grade AEs included thrombocytopenia, lymphopenia, leukopenia, elevated aspartate aminotransferase/alanine aminotransferase, anemia, neutropenia, and hypoproteinemia.
What was the design of the phase 2 study?
Patients were randomly assigned 1:1 to either the CAR-like T-cell plus 1 Gy of radiation therapy plus PD-1 monoclonal antibody plus SOX for 6 cycles or less followed by PD-1 monoclonal antibody plus S-1 (n = 50) or PD-1 monoclonal antibody plus SOX for 6 cycles or less followed by PD-1 monoclonal antibody plus S-1 (n = 50). The primary end point was ORR, with secondary end points of PFS, DOR, and overall survival. Key eligibility criteria included having previously untreated, unresected, or metastatic gastric/GEJ adenocarcinoma; no known HER2-positive status; and an ECOG performance status of 0 to 1.
The data cutoff was November 3, 2025. The median follow-up was 15.8 months.
What were the baseline characteristics?
Overall, 75% of patients were male, 76% were younger than 70 years old, 63% had an ECOG performance status of 1, and 93% had gastric cancer. Additionally, 58% of patients had 2 or more metastatic sites, 43.5% had peritoneum metastasis, 39% had a PD-L1 CPS of less than 1, and 99% had microsatellite stable/mismatch repair–proficient disease.
Reference
Liu Q, Zhang L, Shao J, et al. SOX (tegafur plus oxaliplatin) plus anti-PD-1 antibody with or without CAR-like T cell immunotherapy in the treatment of patients with previously untreated metastatic gastric cancer/gastroesophageal junction adenocarcinoma: a multicenter, open-label, randomized, controlled, phase II trial. J Clin Oncol. 2026;44(suppl 2):291. doi:10.1200/JCO.2026.44.2_suppl.291
