Cadonilimab Plus Chemo Retains Clinically Meaningful Survival Advantage in Frontline Gastric/GEJ Adenocarcinoma

The combination of cadonilimab (AK104) and XELOX upheld clinically meaningful long-term survival benefit vs placebo plus XELOX alone as first-line therapy for patients with advanced gastric or gastroesophageal junction (GEJ) adenocarcinoma regardless of PD-L1 expression, according to final results from the phase 3 COMPASSION-15 trial (NCT05008783) that were presented during the 2025 ESMO Congress.¹

At a median follow-up of 33.9 months, the median overall survival (OS) was 13.9 months (95% CI, 12.0-16.8) in the experimental arm (n = 305) vs 11.1 months (95% CI, 10.2-12.2) in the control arm (n = 305) of the intention-to-treat (ITT) population (HR, 0.61; 95% CI, 0.51-0.73; P < .001). The median progression-free survival (PFS) was 7.0 months (95% CI, 6.7-8.4) in the experimental arm vs 5.3 months (95% CI, 4.5-5.6) in the control arm (HR, 0.51; 95% CI, 0.42-0.62; P < .001). The median duration of response (DOR) was 8.9 months (95% CI, 7.0-11.1) in the experimental arm (n = 199) vs 4.5 months (95% CI, 4.2-5.5) in the control arm (n = 149; HR, 0.45; 95% CI, 0.35-0.59; P < .001).

“The efficacy and safety results in the final analysis demonstrated the long-term OS benefits of cadonilimab plus chemotherapy and will further strengthen its leading position as a standard of care in first-line treatment for [patients with] gastric/GEJ adenocarcinoma,” Lin Shen, MD, lead study author and vice president of Clinical Oncology at Beijing Cancer Hospital and Peking University and deputy director at Beijing Institute for Cancer Research at the United Family New Hope Oncology Center in China, said in a presentation delivered during the meeting.

How Is Cadonilimab Used in Gastric/GEJ Adenocarcinoma Treatment?

PD-(L)1 inhibitors have shown improved survival when administered in combination with chemotherapy for patients with untreated gastric/GEJ adenocarcinoma, but the therapeutic benefit is not as enriched in patients with low PD-L1 expression.

Cadonilimab is a humanized bispecific antibody-drug candidate targeting PD-1 and CTLA-4. In September 2024, the National Medical Products Administration approved the agent for use in combination with fluoropyrimidine and platinum-based chemotherapy for the first-line treatment of patients with locally advanced, unresectable or metastatic gastric/GEJ adenocarcinoma.² The approval was based on interim findings from the COMPASSION-15 trial, which, at a median follow-up of 18.7 months, showed that the regimen produced a median OS of 14.1 months in the ITT population and 15.3 months in patients with a PD-L1 combined positive score (CPS) of 5 or greater, reducing the risk of death by 44% (HR, 0.66; 95% CI, 0.54-0.81) and 42% (HR, 0.58; 95% CI, 0.41-0.82) vs chemotherapy alone, respectively.³

What Was the Trial Design and Focus of This Analysis?

Findings from the prespecified final analyses were presented at a data cutoff of December 6, 2024.¹

To be eligible, patients between the ages of 18 and 75 had to have previously untreated, locally advanced, unresectable or metastatic, HER2-negative gastric/GEJ adenocarcinoma with an ECOG performance status of 0 or 1, measurable tumor lesion per RECIST 1.1 criteria, and life expectancy of at least 3 months.

Following a 1:1 randomization, patients received either 10 mg/kg of cadonilimab every 3 weeks plus XELOX for a maximum of 6 cycles, followed by cadonilimab maintenance, or placebo plus XELOX followed by placebo maintenance in the same doses and schedules as in the experimental arm. XELOX consisted of 1000 mg/m² of capecitabine administered orally, twice daily on days 1 through 14 every 3 weeks, plus 130 mg/m² of intravenous oxaliplatin every 3 weeks.

The primary end point was OS in the ITT population. Secondary end points included OS in the CPS 5 or greater population, PFS, objective response rate (ORR), investigator-assessed DOR in the ITT and CPS 5 or greater populations, safety, pharmacokinetics, anti-drug antibody assessment, and health-related quality of life.

The final analyses were planned once 39 months had passed and 443 OS events had occurred. The 1-sided P value boundary was .021, and the HR boundary was .825.

Were the Patient Demographics Reflective of Who I Might See in My Clinic?

Baseline characteristics were well balanced between the arms. In the experimental arm (n = 305), the median age was 63.7 years (range, 29-75,) and most patients were under the age of 65 (55.7%) and male (78.4%). ECOG performance status was predominantly 1 (77.0%), and a minority of patients had recurrent disease (21.3%). The primary tumor location was the stomach in 79.3% of cases and the GEJ in 20.7%. Most patients had metastatic disease (96.1%) with sites in the liver (47.2%) and lung (16.7%). PD-L1 CPS expression was either less than 1% (23.6%), between 1% and 4% (27.9%), between 5% and 9% (14.4%), or 10% or greater (23.6%); 32 (10.5%) samples were missing.

What Were the Landmark OS and PFS Rates?

In the ITT population the 12- and 24-month OS rates in the cadonilimab arm were 55.6% and 36.0%, respectively, vs 44.9% and 14.8% in the placebo arm. In the PD-L1 CPS 5 or greater population, the 12- and 24-month OS rates in the cadonilimab arm were 57.8% and 43.7%, respectively, vs 45.0% and 13.5% in the placebo arm. In the PD-L1 CPS 5 or less population, the 12- and 24-month OS rates in the cadonilimab arm were 53.8% and 31.4%, respectively, vs 45.8% and 15.6% in the placebo arm.

In the ITT population, the 12- and 24-month PFS rates in the cadonilimab arm were 35.4% and 22.3%, respectively, vs 11.2% and 4.2% in the placebo arm. In the PD-L1 CPS 5 or greater population, the 12- and 24-month PFS rates in the cadonilimab arm were 39.9% and 28.5%, respectively, vs 10.1% and 2.7% in the placebo arm. In the PD-L1 CPS 5 or less population, the 12- and 24-month PFS rates in the cadonilimab arm were 32.2% and 16.2%, respectively, vs 12.7% and 6.1% in the placebo arm.

How Did the Regimen Perform According to PD-L1 CPS Cutoffs?

“OS benefits were consistently observed at all prespecified CPS cutoffs,” Shen said. In the PD-L1 CPS 5 or greater population, the median OS was 16.8 months (95% CI, 11.3-25.1) with cadonilimab (n = 116) vs 10.8 months (95% CI, 9.0-12.9) with placebo (n = 140; HR, 0.49; 95% CI, 0.36-0.65; P < .001). In the PD-L1 CPS 5 or less population, the median OS was 13.2 months (95% CI, 11.2-15.6) with cadonilimab (n = 157) vs 11.3 months (95% CI, 10.1-13.0) with placebo (n = 147; HR, 0.76; 95% CI, 0.59-0.99; P = .0019).

“The results for ORR, DOR, and PFS were consistent between the interim and final analyses,” Shen stated.

In the PD-L1 CPS 5 or greater population, the median DOR was 8.5 months (95% CI, 5.7-21.4) in the cadonilimab arm (n = 79) vs 4.9 months (95% CI, 4.2-5.7) in the placebo arm (n = 74; HR, 0.43; 95% CI, 0.28-0.65; P < .001). In the PD-L1 CPS 5 or less population, the median DOR was 9.4 months (95% CI, 5.6-11.0) in the cadonilimab arm (n = 98) vs 4.3 months (95% CI, 3.2-5.7) in the placebo arm (n = 64; HR, 0.52; 95% CI, 0.36-0.77; P < .001).

In the PD-L1 CPS 5 or greater population, the median PFS was 7.0 months (95% CI, 5.6-10.5) in the cadonilimab arm (n = 116) vs 5.5 months (95% CI, 4.5-5.8) in the placebo arm (n = 140; HR, 0.46; 95% CI, 0.34-0.63; P < .001). In the PD-L1 CPS 5 or less population, the median PFS was 6.9 months (95% CI, 5.7-8.5) in the cadonilimab arm (n = 157) vs 4.6 months (95% CI, 4.3-5.6) in the placebo arm (n = 167; HR, 0.61; 95% CI, 0.46-0.80; P < .001).

Were Any New Safety Signals Seen With Longer Follow-Up?

The most common grade 3 or greater treatment-related adverse effects (TRAEs) that occurred in at least 5% of patients in the experimental and control arms, respectively, were decreased platelet count, decreased neutrophil count, anemia, decreased white blood cell count, and hypokalemia.

TRAEs occurred in 99.0% (grade ≥3, 66.9%) of patients in the experimental arm (n = 305) vs 97.4% (grade ≥3, 53.6%) of those in the control arm (n = 304). Serious TRAEs and TRAEs leading to discontinuation occurred in 40.0% and 23.9% of those in the experimental arm, respectively, vs 26.0% and 6.6% of those in the control arm. There were 6 (2.0%) and 8 (2.6%) treatment-related deaths in the experimental and control arms, respectively; all 6 were due to cadonilimab in the experimental arm and 6 were due to placebo in the control arm.

“No new long-term adverse effects were identified; the safety profile remained essentially consistent with that observed at the interim analysis,” Shen concluded.

Disclosures: Shen disclosed consulting with AstraZeneca, Boehringer Ingelheim, MSD, Servier, and Transcenta Holding Limited; research funding from BeiGene, Innovent, NovaRock Biotherapeutics Limited, Roche, and Rongchang Pharmaceuticals.

References

1. Shen L, Zhang Y, Li Z, et al. Cadonilimab (cado) plus chemotherapy (chemo) versus chemotherapy as first-line (1L) treatment for advanced gastric or gastroesophageal junction (G/GEJ) adenocarcinoma: final results of the phase III COMPASSION-15 trial. Presented at: 2025 ESMO Congress; October 17-25, 2025; Berlin, Germany. Abstract 2098MO.
2. Akeso's cadonilimab receives second indication approval from NMPA for first-line treatment of gastric/GEJ cancer in all-comers population. News release. Akeso. September 30, 2024. Accessed October 21, 2025. https://www.akesobio.com/en/media/akeso-news/akesos-cadonilimab-receives-second-indication-approval-from-nmpa-for-first-line-treatment-of-gastricgej-cancer/
3. First-line cadonilimab plus chemotherapy in HER2-negative advanced gastric or gastroesophageal junction adenocarcinoma: a randomized, double-blind, phase 3 trial. Nat Med. 2025;31(4):1163-1170. doi:10.1038/s41591-024-03450-4