Therapeutic management of resectable mismatch-repair deficient/microsatellite instability-high gastroesophageal...

Mismatch-repair deficient (dMMR)/microsatellite instability-high (MSI-H) gastroesophageal adenocarcinoma (GEA) is associated with favorable prognosis but limited benefit from perioperative chemotherapy. Although immune checkpoint inhibition (ICI)-based approaches have shown promising activity in early-phase studies, the optimal therapeutic combinations and the feasibility of non-operative management (NOM) in resectable disease remain unclear.

This exploratory, retrospective, multicenter, real-world study included 55 patients with resectable dMMR/MSI-H GEA, who were divided into subgroups, characterized and compared according to treatment strategy (chemotherapy, chemoimmunotherapy, immunotherapy and initial resection). Pathological and clinical complete response (pCR, cCR), progression-free (PFS) and overall survival (OS) were evaluated.

Patients received chemotherapy (n=10), chemoimmunotherapy (n=16), immunotherapy (n=13), or initial resection without subsequent systemic treatment (n=16). Forty five (82%) patients had surgical resection. The pCR rate was 22% (n=2/9) in the chemotherapy, 27% (n=4/15) in the chemoimmunotherapy, and 100% (n=5/5) in the immunotherapy (i.e., doublet ICI) subgroup, respectively. The overall cCR rate in NOM patients was 33% (n=3/9), with rates of 37.5% (n=3/8) in the immunotherapy subgroup and 0% (n=0/1) in the chemoimmunotherapy subgroup. Complete response rates (pCR/cCR) were significantly associated with PD-L1 CPS ≥10 (p=0.027).

This real-world study is, to our knowledge, the first to evaluate a broad spectrum of therapeutic combinations in patients with resectable MSI-H/dMMR GEA, showing meaningful antitumor activity, particularly with doublet ICI. Given the challenges of conducting randomized prospective trials in this setting, such real world data may support clinicians in guiding therapeutic decision-making.