Highlights
Microsatellite instability-high/deficient mismatch repair (MSI-H/dMMR) tumors accounted for 5.1% of 1,638 tested gastroesophageal adenocarcinomas in this single-center cohort.
Immune checkpoint inhibitor-based therapy produced high activity in dMMR disease, with best objective response in 61.5% of evaluable metastatic cases and pCR/cCR in 52.2% of evaluable locoregional cases.
In locoregional disease, some patients achieved durable complete responses without surgery, supporting the possibility of nonoperative management in carefully selected responders.
Lynch syndrome was more frequent among responders, whereas HER2, PD-L1 combined positive score, and CLDN18.2 expression did not clearly distinguish responders from nonresponders.
Clinical Background and Disease Burden
Gastroesophageal adenocarcinoma (GEA), which includes cancers of the stomach, gastroesophageal junction, and distal esophagus with adenocarcinoma histology, remains a major cause of cancer-related morbidity and mortality. Even with advances in perioperative chemotherapy, surgery, and targeted therapy, outcomes vary widely by molecular subtype. One of the most clinically important biomarkers is mismatch repair deficiency (dMMR), often reflected as microsatellite instability-high (MSI-H) status. These tumors generate a high neoantigen burden and are therefore more likely to respond to immune checkpoint inhibition (ICI).
Although MSI-H/dMMR is established as a predictive biomarker across several solid tumors, the evidence base in GEA has been less comprehensive than in colorectal cancer. This gap matters because treatment decisions differ substantially between metastatic and locoregional disease. In metastatic disease, immunotherapy may prolong survival and delay chemotherapy exposure. In locoregional disease, a deep response could potentially allow organ-preserving strategies, including surveillance rather than immediate surgery. However, the durability of such responses and the characteristics of patients most likely to benefit have not been fully defined.
Study Design and Methods
This retrospective cohort study was conducted at a single tertiary cancer center in Texas between June 2020 and August 2025. Among 3,316 patients diagnosed with GEA, 1,638 underwent MSI testing and were included in the analysis. The study focused on patients with dMMR tumors and described treatment pathways and outcomes after ICI-based therapy.
The principal ICI regimens were nivolumab plus ipilimumab or pembrolizumab. Outcomes included the incidence of dMMR, overall survival, best objective response, clinical or pathological complete response (cCR/pCR), and clinicopathological factors associated with response. The study also examined whether patients underwent curative-intent surgery, remained on active surveillance, or experienced recurrence or death after complete response.
Because this was an observational study, treatment selection was not randomized. Therefore, the findings are best interpreted as real-world effectiveness data rather than proof of comparative efficacy between immunotherapy regimens or between surgery and nonoperative management.
Key Findings
Of the 1,638 patients tested, 83 had dMMR tumors, corresponding to 5.1% of the cohort. The mean age was 62.5 years, and 73.5% were male. Among the 83 dMMR cases, 34 had metastatic disease and 49 had locoregional disease. This distribution is clinically important because the expected goals of therapy differ markedly between these two settings.
Overall survival was substantially better in dMMR disease than in proficient mismatch repair (pMMR) disease. In patients with dMMR tumors, the 3-year overall survival rate was 84.7% with curative-intent surgery and 68.5% without curative-intent surgery. By comparison, among patients with pMMR tumors, the 3-year overall survival rate was 72.3% with curative-intent surgery and 32.3% without curative-intent surgery. While the abstract does not provide hazard ratios or confidence intervals, the survival separation strongly suggests that dMMR biology identifies a subgroup with favorable treatment sensitivity and longer-term disease control, especially when local therapy is integrated appropriately.
Response to ICI-based therapy was striking. In the metastatic dMMR group, 30 patients received ICI-based therapy, and 26 were evaluable for best objective response; 16 of these 26 patients, or 61.5%, achieved a best objective response. This is a robust signal of activity for advanced disease. In the locoregional dMMR group, 25 patients received ICI-based therapy, and 23 were evaluable; 12 of 23 patients, or 52.2%, achieved pCR/cCR, meeting the study’s definition of ICI responsiveness. These response rates are notable because deep responses are uncommon in unselected GEA and support the biologic enrichment offered by MSI testing.
The locoregional cohort is especially practice-relevant. Of the 25 locoregional patients treated with ICI, 10 proceeded to surgery. All 10 underwent R0 resection, meaning no microscopic residual tumor remained at the surgical margins. Three of these 10 patients, or 30%, had a pCR. Among the remaining patients, 9 achieved cCR and were transitioned to active surveillance instead of immediate surgery. This finding is particularly important because it suggests that some dMMR locoregional GEA patients may avoid the morbidity of esophagectomy or gastrectomy if a durable complete response is confirmed clinically and radiographically.
Durability also matters. Patients who achieved cCR or pCR did so after a relatively short exposure to ICI therapy: 2 to 4 cycles of nivolumab plus ipilimumab or 6 to 12 cycles of pembrolizumab. With a median follow-up after cCR or pCR of 26.0 months, none experienced recurrence or death. Although follow-up remains modest and the sample is small, this is a clinically meaningful durability signal and argues against dismissing these complete responses as transient radiologic phenomena.
The study also explored potential biomarkers of response beyond MSI status. Individuals who were ICI responsive and nonresponsive did not differ significantly in ERBB2 (HER2), PD-L1 combined positive score, or CLDN18.2 expression. In contrast, Lynch syndrome was more common among responders. This is biologically plausible, as germline mismatch repair defects can produce strongly immunogenic tumors. Still, the result should be viewed as hypothesis-generating rather than definitive, because the number of patients is limited and biomarker ascertainment may not have been uniform.
Clinical Interpretation and Expert Commentary
This study reinforces a key practical point: MSI-H/dMMR is one of the most actionable biomarkers in GEA. The response rates observed here compare favorably with historical outcomes for chemotherapy alone, especially in locoregional disease where the possibility of cure raises the stakes of response selection. The data also support routine MSI/dMMR testing early in the diagnostic pathway so that treatment can be tailored before patients receive ineffective therapy or undergo potentially avoidable surgery.
For metastatic disease, the findings support continued use of ICI-based therapy as an important treatment backbone in dMMR GEA. For locoregional disease, the study goes one step further by suggesting that complete responders may be candidates for nonoperative management under close surveillance. This is not yet a universal standard, but it is an increasingly relevant concept in GI oncology, particularly in the context of organ preservation and quality of life.
At the same time, caution is warranted. The study is retrospective, single-center, and relatively small, especially after stratifying by disease stage and response status. Without randomized comparison, it is not possible to determine whether nivolumab plus ipilimumab is superior to pembrolizumab or whether surgery can be safely omitted outside of a highly selected framework. In addition, pCR and cCR are not interchangeable outcomes: pathological confirmation is more definitive, whereas clinical complete response depends on imaging, endoscopy, biopsy sampling, and judgment about residual disease.
Generalizability may also be limited by the tertiary referral setting, where access to expert multidisciplinary care, intensive endoscopic assessment, and prolonged surveillance may exceed what is available in routine practice. The study also does not fully resolve how to manage discordance between radiographic response and residual microscopic disease, or how best to define the surveillance schedule for patients managed nonoperatively.
Nonetheless, the work is highly clinically relevant because it moves beyond biomarker prevalence and into real-world treatment pathways. It also highlights an important translational challenge: not all MSI-H/dMMR tumors behave identically. Better predictors are still needed to identify which patients will achieve durable complete response, which will require surgery despite initial shrinkage, and which will benefit from combination approaches.
Conclusion
In this single-center cohort, dMMR/MSI-H gastroesophageal adenocarcinoma showed high sensitivity to immune checkpoint inhibition in both metastatic and locoregional settings. The durability of cCR and pCR, with no recurrences or deaths during follow-up, raises the possibility that selected patients with locoregional disease may avoid surgery. These findings strengthen the case for universal MSI testing in GEA and for prospective studies to define immunotherapy-driven, response-adapted treatment strategies.
Funding and ClinicalTrials.gov
The abstract did not report funding support or a ClinicalTrials.gov registration number. These details were not provided in the supplied source.
References
1. Okui J, Prakash LR, Lyu HG, et al. Immunotherapy Response in Microsatellite Instability-High Gastroesophageal Adenocarcinoma. JAMA Surg. 2026;161(6):600-607.
2. Le DT, Uram JN, Wang H, et al. PD-1 Blockade in Tumors with Mismatch-Repair Deficiency. N Engl J Med. 2015;372:2509-2520.
3. Marabelle A, Le DT, Ascierto PA, et al. Efficacy of Pembrolizumab in Patients With Noncolorectal MSI-H/dMMR Cancer: Results From the KEYNOTE-158 Study. J Clin Oncol. 2020;38:1-10.
4. Janjigian YY, Shitara K, Moehler M, et al. First-line nivolumab plus chemotherapy versus chemotherapy alone in advanced gastric, gastro-oesophageal junction, and oesophageal adenocarcinoma: CheckMate 649. Lancet. 2021;398:27-40.
5. National Comprehensive Cancer Network. NCCN Clinical Practice Guidelines in Oncology: Gastric Cancer and Esophageal and Esophagogastric Junction Cancers. Current version.