Cadonilimab in combination with ivonescimab and chemotherapy as first-line (1L) therapy in patients with advanced gastric (G) or gastroesophageal junction adenocarcinoma (GEJA).

e16037 Background: Cadonilimab (AK104), an anti-PD-1/CTLA-4 bispecific antibody, plus chemotherapy (chemo) significantly improved OS versus chemo and had a tolerable safety profile in 1L treatment of advanced G/GEJA patients (pts), including those with low PD-L1 expression. Currently, it has been approved by NMPA. Ivonescimab (AK112), approved in China, is a novel bispecific antibody against PD-1 and VEGF and has shown a clinically significant improvement in efficacy with favorable safety for advanced NSCLC in two phase 3 studies (HARMONi-2 and HARMONi-A). Here, we presented the preliminary safety and efficacy of AK104 combined with AK112 and chemo in previously untreated advanced G/GEJA. Methods: Pts with previously untreated advanced G/GEJA were enrolled. The phase II trial consisted of a dose escalation (part 1) and a dose expansion (part 2). Eligible pts were firstly enrolled into sequential part 1 including 10mg/kg and 15mg/kg AK104 (Q6W, D8) combined with AK112 (20mg/kg, Q3W, D1) and chemo (SOX or XELOX) following the conventional 3+3 design. If the starting dose of 10mg/kg AK104 led to ≥2 dose-limiting toxicities (DLTs), 6mg/kg AK104 would be administered. After the part 1 completed, eligible pts were enrolled into the part 2 and received AK104 (the recommended dose, Q6W, D8) combined with AK112 (20 mg/kg, Q3W, D1) and chemo (SOX or XELOX). The primary outcomes were safety and ORR. Secondary endpoints were DCR, PFS, OS, biomarkers of drug activity and pharmacokinetics. Results: As of 24 January 2025, 21 pts were enrolled with a median age of 59 (range: 39-73). 100.0% were ECOG PS 1, 42.9% were PD-L1 CPS＜5 and 38.1% had liver metastasis. In the part 1, 9 pts (3 at dose level 10mg/kg, 6 at 15mg/kg) were treated and one DLT (grade 3 acneiform rash) was observed at 15 mg/kg dose level. 76.2% (16/21) experienced treatment-related adverse events (TRAEs). The most common TRAEs were weight loss (8/21, 38.1%), neutropenia (5/21, 23.8%), nausea (5/21, 23.8%) and hyperbilirubinemia (5/21, 23.8%). 3 pts had grade 3 TRAEs including neutropenia (1/21, 4.8%), acneiform rash (1/21, 4.8%) and hypokalemia (1/21, 4.8%). Immune-related AEs occurred in 23.8% (5/21) with one reported with grade 3 (acneiform rash). There were no grade 4/5 TRAEs or treatment-related deaths. No new safety signals were identified. At data cut off, 12 pts were evaluable for response. 9 pts were PR and 3 had SD, the ORR and DCR were 75.0% and 100.0%, respectively. Among efficacy evaluable pts who received 15mg/kg AK104, 7 reached PR and 2 were SD. the ORR was 77.8% (7/9) and DCR was 100.0% (9/9). The median PFS and OS were immature. Conclusions: AK104 combined with AK112 and chemo as 1L treatment showed a promising tolerable safety profile and highly encouraging efficacy in pts with advanced G/GEJA. The combinations may further improve benefits for advanced G/GEJA pts in the first-line setting. Clinical trial information: NCT06196697 .