INFINITY: A multicentre, single-arm, multi-cohort, phase II trial of tremelimumab and durvalumab as neoadjuvant treatment of patients with microsatellite instability-high (MSI) resectable gastric or gastroesophageal junction adenocarcinoma (GAC/GEJAC).

358 Background: In resectable GAC/GEJAC, MSI status is associated with better survival and potential lack of benefit from chemotherapy. Given the high responsiveness of MSI tumors to immunotherapy, neoadjuvant or definitive dual CTLA-4/PD(L)-1 inhibition may allow omission of chemotherapy or surgery. Methods: INFINITY is a multicentre, single-arm, multi-cohort phase II trial (NCT04817826) investigating the activity and safety of tremelimumab+durvalumab as neoadjuvant (Cohort 1) or definitive (Cohort 2) treatment for MSI, mismatch repair deficient (dMMR) and EBV-negative resectable GAC/GEJAC. In Cohort 1, patients (pts) received a 12-week treatment with single high dose tremelimumab 300 mg and durvalumab 1500 mg q4 weeks (T300/D) for 3 cycles followed by surgery. The primary endpoint was pCR rate (ypT0N0) with negative ctDNA after T300/D. Secondary endpoints: disease-free survival, overall survival, quality of life. Exploratory: correlation of pCR with clinical variables, PDL-1 CPS assessed by IHC 22C3, tumor mutational burden (TMB) by Foundation One, liquid biopsies and other biomarkers. Cohort 2 investigates non operative management after same treatment regimen. Results: Overall, 18 pts with MSI/dMMR resectable cT2-4 any N GAC/GEJAC were recruited in Cohort 1. One withdrew consent and 2 achieved a complete clinical-pathological response at radiology and endoscopy (ongoing) and refused surgery. Among 15 evaluable patients, 1 had disease progression and 14 underwent resection. pCR rate was 60% (9/15) and major-complete pathological response (<10% viable cells) was 80%. All pts with pCR had negative ctDNA status pre-surgery. pCR rate was 1/6 (17%) in T4 vs 8/9 (89%) in T2-3 tumors (p=0.011), whereas no correlation was found with baseline N status. PDL-1 CPS was not associated with outcomes and TMB had a non-significant trend of correlation with pCR (median TMB 26 in non-pCR vs 40 in pCR group, p=0.2). Grade≥3 immune-related AEs occurred in 3 pts of safety population (n=18): colitis, pneumonitis, liver toxicity, all resolved with high dose steroids and did not impair surgery. Two pts died after surgery for other reasons than disease or AEs, whereas no disease relapses were observed in remaining pts. QoL and additional translational analyses on RNA-seq, digital spatial profiling and ctDNA monitoring will be presented. Conclusions: Pre-operative T300/D for 3 months was safe and provided promising proof of efficacy in MSI, dMMR GAC/GEJAC pts. These results open the way to investigate NOM in pts with clinical, pathological and molecular (ctDNA minimal residual disease) complete response after T300/D. Enrollment in Cohort 2 has started after IDMC evaluation and protocol. amendment to include only pts with cT2-3 tumors confirmed at staging laparoscopy. Clinical trial information: NCT04817826 .