ESMO Congress 2026

Dostarlimab is a PD-1-targeting monoclonal antibody, one of the ten approved anti-PD-1 antibodies worldwide^[2], and it received marketing approval in the United States or European Union in 2021^[6]. The pivotal published phase 3 evidence for this asset comes from a randomized, double-blind, placebo-controlled trial in primary advanced or recurrent endometrial cancer: among 494 randomized patients, 23.9% (118) had dMMR/MSI-H tumors, with progression-free survival among the primary endpoints in this population^[3]; earlier, in a phase 1 single-arm study, single-agent dostarlimab (500 mg every 3 weeks for 4 doses, then 1000 mg every 6 weeks) was evaluated for clinical activity and safety in recurrent or advanced dMMR endometrial cancer^[5]. The precedent most directly relevant to this readout is the phase 2 study in dMMR locally advanced rectal cancer: single-agent dostarlimab every 3 weeks for 6 months, with patients achieving clinical complete response (cCR) proceeding without chemoradiotherapy and surgery, and sustained cCR at 12 months after completion among the primary endpoints^[1].

What is being presented at ESMO26 is the first interim analysis of the GERCOR-GONO DEWI phase II study, presented by Thierry André: patients with localized dMMR/MSI-H gastroesophageal adenocarcinoma (GEA) received dostarlimab, and those achieving cCR entered watch-and-wait, with the primary endpoint being 1-year cCR without surgery. Per the OncoAlert abstract, among 24 evaluable patients the 1-year cCR was 62.5% (15/24); at a median follow-up of 16.20 months, no relapse or death occurred among those with 1-year cCR; grade 3-4 dostarlimab-related AEs were 21% (5/24); 9/24 underwent surgery, all with R0 resection. It should be noted that the numbers across the on-site posts are inconsistent: other posts report on 43 patients, with 1-year cCR with organ preservation of 67.4% (29/43), no relapse or death among these 29 W&W patients, surgery in 14/43 (32.5%), and follow-up of 14.9 months; yet another post cites 25 pts enrolled, 33% cCR at 1 year, 1 pt with PD, and that the study continues. The exact figures await confirmation from the formal presentation.

Placing this readout against existing knowledge, the step forward lies in extending the path of deferring surgery after immunotherapy-induced cCR from dMMR localized rectal cancer^[1] to another dMMR/MSI-H solid tumor site - gastroesophageal adenocarcinoma - as one on-site comment put it, 'after rectal, this is the next frontier.' The underlying rationale builds on the sensitivity of dMMR/MSI-H tumors to PD-1 blockade, supported in both endometrial cancer and rectal cancer^[1][3][5]. The on-site discussion suggested its significance goes beyond efficacy numbers, toward a potential shift in treatment philosophy: from 'treat, then operate' to 'treat, assess deeply, and preserve the organ when oncologically safe'; the challenges are how confidently true complete responders can be identified and how durable the response remains over time.

The limitations are equally clear: small sample size (24 evaluable patients; 43 by the full-cohort reading), short follow-up (median 14.9-16.2 months), a single-arm phase II interim analysis, and the inconsistencies across posts themselves suggest the data are still evolving; roughly one-third of patients still underwent surgery (9/24 or 14/43), surgery omission remains an investigational strategy, and longer follow-up is needed to verify the durability of responses.