1512MO Perioperative camrelizumab (C) combined with rivoceranib (R) and chemotherapy (chemo) versus chemo for locally advanced resectable gastric or gastroesophageal junction (G/GEJ) adenocarcinoma: The first interim analysis of a randomized, phase III trial (DRAGON IV)

The efficacy of standard perioperative chemotherapy for patients (pts) with localized resectable G/GEJ cancer still needs to be further improved. This ongoing multicentred, randomized, open-label, phase III study compared efficacy and safety of perioperative C plus R and chemo (SOXRC) versus chemo alone (SOX) in localized resectable G/GEJ adenocarcinoma. Pts with T3-4aN+M0 were randomized 1:1 to SOXRC (C 200 mg IV D1 + R 250 mg PO QD D1-21 + oxaliplatin 130 mg/m2 IV D1 + S-1 PO BID D1-14, Q3W) or SOX for 3 cycles pre/post D2 surgery, then R plus C (SOXRC) or S-1 (SOX) up to 17 cycles per investigator's choice. Stratification was by tumour location (GEJ vs G) and bulky N (yes vs no). Blinded independent review committee (BIRC) assessed pathological complete response (pCR, ypT0) and investigators assessed EFS were primary endpoints. The secondary endpoints were major pathological response (MPR), total pCR (ypT0N0), R0 resection, DFS and OS. The planned sample size was 512 pts. The main analysis of pCR would be conducted after the first 360 randomized pts had the opportunity for D2 surgery. Of the first 360 randomized pts (SOXRC n=180; SOX n=180), 71.4% had gastric cancer and 28.6% had GEJ cancer; 66.7% had T4 and 100% had N+ with balanced baseline between arms. 179 pts received neoadjuvant therapy, 164 completed all neoadjuvant therapy and 155 underwent surgery (SOXRC); 177, 169 and 156 did the same (SOX). In the ITT population, BIRC-assessed pCR was 18.3% (95% CI 13.0-24.8) for SOXRC and 5.0% (95% CI 2.3-9.3) for SOX, with a statistically significant improvement of 13.7% (95% CI 7.2-20.1, p<0.0001); MPR was 51.1% vs 37.8%; tpCR was 16.7% vs 4.4%. In the surgery set, R0 resection was 98.7% (SOXRC) vs 94.2% (SOX). Surgical complications occurred at rates of 27.7% (SOXRC) and 30.1% (SOX). Preoperative grade ≥3 treatment-emergent adverse events occurred at rates of 36.3% (SOXRC) and 16.3% (SOX). The trial showed perioperative C combined with R and chemo significantly improved pCR compared to chemo alone with a tolerable safety profile for localized resectable G/GEJ adenocarcinoma.