Durvalumab (PD-L1) + CROSS boosts MPR to 54% in resectable esophageal adenocarcinoma; tremelimumab (CTLA-4) adds toxicity without benefit

Durvalumab is an anti–PD-L1 antibody, and tremelimumab is an anti–CTLA-4 antibody; in the Phase II RICE trial (NCT04159974), 56 patients with locally advanced, non-metastatic esophageal adenocarcinoma (≥uT3/Nx or uT2/N+) received 2 neoadjuvant doses of durvalumab added to neoadjuvant chemoradiotherapy (CROSS) followed by surgery, then adjuvant durvalumab ± 1 dose of tremelimumab. In the mITT cohort (n=56), grade ≥3 non-hematologic neoadjuvant AEs were 25% (G3 23%, G4 2%), 93% underwent resection with 6% anastomotic leak, and MPR was 54% with pCR 23%—numerically above historical CROSS pCR rates (~23% in the original CROSS trial, albeit mixed histologies) and above the abstract’s propensity-based comparisons with CROSS MPR 37% and perioperative FLOT MPR 25%; 2-year OS/PFS were 72.7%/53.1%. In the randomized adjuvant set (R-mITT; arm A durvalumab n=20 vs arm B durvalumab+tremelimumab n=18), completion of ≥80% planned adjuvant therapy was 80% vs 44%, grade ≥3 toxicity was higher with tremelimumab (G3/G4: 22%/22% vs 15%/0%), and 2-year OS/PFS favored durvalumab alone (95.0%/80.0% vs 82.4%/54.2%); for current perioperative standards, many centers have shifted toward perioperative FLOT (FLOT4 median OS ~50 months; pCR ~16%), while adjuvant nivolumab is established after neoadjuvant CRT + surgery with residual disease (CheckMate 577 median DFS 22.4 vs 11.0 months; HR 0.69).