Highlights
Microsatellite instability-high/deficient mismatch repair gastroesophageal adenocarcinoma accounted for a small but clinically important subset of cases in this single-center cohort, underscoring the importance of routine MSI testing.
Immune checkpoint inhibitor-based therapy produced high response rates in both metastatic and locoregional disease, with many patients achieving objective response or complete clinical/pathologic remission.
Among locoregional responders, some patients avoided immediate surgery through active surveillance, and short- to intermediate-term outcomes after complete response were encouraging.
Lynch syndrome was more common among responders, while HER2 status, PD-L1 combined positive score, and CLDN18.2 expression did not clearly distinguish responders from nonresponders in this dataset.
Clinical background and unmet need
Gastroesophageal adenocarcinoma (GEA) remains a lethal malignancy despite incremental progress with perioperative chemotherapy, surgery, radiation, and biomarker-directed therapy. Although most patients present with either locally advanced or metastatic disease, treatment is increasingly guided by molecular subtype. Among these subtypes, deficient mismatch repair (dMMR) or microsatellite instability-high (MSI-H) tumors are especially relevant because they generate a high number of neoantigens and are therefore more likely to be recognized by the immune system. This biological feature creates a strong rationale for immune checkpoint inhibitor (ICI) therapy.
Even so, the practical implications of MSI-H/dMMR status in GEA are not fully settled. Prior data have suggested that MSI-H tumors may derive limited benefit from conventional cytotoxic chemotherapy in some settings while showing substantial sensitivity to PD-1 blockade, but the evidence base has often been fragmented, derived from mixed upper gastrointestinal cohorts, or limited by small sample sizes. In locally advanced disease, an important open question is whether some patients with deep immunotherapy responses might avoid or delay surgery without compromising oncologic outcomes. The present study addresses these gaps by describing treatment pathways and outcomes after ICI therapy in a contemporary cohort from a tertiary cancer center.
Study design and patient population
This retrospective cohort study was conducted at a single tertiary cancer center in Texas between June 2020 and August 2025. Of 3316 patients diagnosed with GEA, 1638 who underwent MSI testing were included in the analysis. The investigators focused on patients with dMMR tumors and characterized their clinical course according to disease stage and treatment received.
Among the 1638 tested patients, 83 individuals (5.1%) had dMMR tumors. The mean age was 62.5 years, and most patients were male (73.5%). The dMMR group was divided into metastatic disease (34 patients) and locoregional disease (49 patients). ICI-based treatment consisted of nivolumab plus ipilimumab or pembrolizumab, reflecting contemporary checkpoint blockade strategies. The main outcomes included overall survival, best objective response, clinical complete response (cCR), pathologic complete response (pCR), and clinicopathological factors associated with ICI response.
Because this was an observational study rather than a randomized trial, treatment selection was determined by real-world clinical decision-making. That design makes the results highly relevant to practice, but it also introduces the usual limitations of retrospective research, including selection bias, incomplete data capture, and potential confounding by indication.
Key findings
The prevalence of dMMR in this tested cohort was 5.1%, which is consistent with dMMR/MSI-H being an uncommon but meaningful biologic subset of GEA. Survival outcomes favored patients with dMMR tumors, particularly those who underwent curative-intent surgery. The reported 3-year overall survival rate was 84.7% for patients with dMMR who had curative-intent surgery and 68.5% for those without curative-intent surgery. In comparison, among patients with proficient mismatch repair, the corresponding 3-year overall survival rates were 72.3% with curative-intent surgery and 32.3% without curative-intent surgery. While cross-group comparisons should be interpreted cautiously in a retrospective cohort, these figures reinforce the clinical importance of both molecular subtype and multimodal local therapy.
Among the 83 dMMR patients, 34 had metastatic disease and 49 had locoregional disease. In the metastatic subgroup, 30 patients received ICI-based therapy, and 26 were evaluable for response. Best objective response was achieved in 16 of 26 evaluable patients (61.5%), indicating substantial activity in advanced disease. This response rate is notable in a biomarker-selected population and supports the use of checkpoint blockade as a major systemic option for dMMR metastatic GEA.
In the locoregional subgroup, 25 patients received ICI-based therapy, and 23 were evaluable for pCR/cCR. Twelve of 23 patients (52.2%) achieved pCR or cCR, meeting the study’s definition of ICI responsiveness. This finding is especially clinically meaningful because it suggests that a large fraction of dMMR locoregional tumors may undergo deep regression after immunotherapy alone or as part of a multimodality plan.
Among the 25 locoregional patients treated with ICI, 10 proceeded to surgery. All 10 achieved R0 resection, meaning no residual microscopic tumor remained at the surgical margins. Three of these 10 surgical patients (30%) had a pathologic complete response. The remaining patients who did not undergo immediate surgery were managed according to response: 9 achieved cCR and were transitioned to active surveillance. This is the most practice-changing aspect of the study, because it suggests that nonoperative management may be feasible in carefully selected patients with durable immunotherapy-induced complete response.
Time to response was also clinically informative. For patients who achieved cCR or pCR, response occurred after 2 to 4 cycles of nivolumab plus ipilimumab or after 6 to 12 cycles of pembrolizumab. This difference likely reflects regimen choice, dosing schedules, and treatment sequencing rather than a direct head-to-head efficacy comparison, which the study was not designed to assess. Importantly, among patients who achieved cCR or pCR, none experienced recurrence or death during a median follow-up of 26.0 months after response, with an interquartile range of 16.3 to 40.0 months. Although follow-up is still limited for a cancer in which late recurrence can occur, the durability of these complete responses is encouraging.
The investigators also explored biomarkers beyond MSI status. Responders and nonresponders did not differ significantly in ERBB2 (HER2) status, PD-L1 combined positive score, or CLDN18.2 expression. This suggests that, within the dMMR population, these commonly discussed biomarkers may not add much discriminatory value for predicting ICI response. In contrast, Lynch syndrome was more common among responders, raising the possibility that germline mismatch repair deficiency and its associated immunobiology may correlate with heightened sensitivity to checkpoint blockade. This observation is hypothesis-generating rather than definitive, but it may have implications for future biomarker refinement.
Clinical interpretation
For clinicians, the most important message is that MSI-H/dMMR biology in GEA identifies a subgroup with distinctly favorable responsiveness to immunotherapy. In metastatic disease, the response rate observed here supports checkpoint blockade as a central treatment strategy. In locoregional disease, the data go further by suggesting that some patients may achieve complete clinical or pathologic remission robust enough to justify surveillance instead of immediate surgery.
That possibility is clinically attractive but should be approached cautiously. Surgery remains the standard curative modality for resectable GEA, and omitting it outside carefully selected circumstances could expose patients to undertreatment if residual microscopic disease is missed. The present study provides real-world support for nonoperative management, but it does not establish it as a universal standard. The safest interpretation is that cCR after ICI in dMMR GEA may be a rational criterion for discussion in a multidisciplinary setting that includes surgical oncology, medical oncology, gastroenterology, radiology, and pathology.
The survival differences reported by surgical status and mismatch repair status also highlight an ongoing challenge in gastrointestinal oncology: how best to integrate systemic response with local control. In dMMR disease, especially when response is deep and durable, the traditional assumption that all curable locoregional tumors must proceed directly to surgery may need to be revisited. However, prospective validation is necessary before a response-adapted nonoperative strategy can be widely implemented.
Strengths and limitations
This study has several strengths. It reflects contemporary practice at a high-volume tertiary center, includes both metastatic and locoregional settings, and provides granular outcome data including cCR, pCR, surgical status, and short-term durability after response. The inclusion of biomarker correlates beyond MSI status also adds translational value.
Several limitations should temper interpretation. First, the study is retrospective and single-center, which limits causal inference and generalizability. Second, the number of dMMR patients was modest, especially after stratification by disease stage and treatment pathway. Third, the cohort included different ICI regimens, so regimen-specific efficacy cannot be compared directly. Fourth, response assessment in nonoperative management relies on clinical and radiographic evaluation, which may miss microscopic residual disease. Finally, the follow-up period, while encouraging, remains insufficient to fully define long-term cure rates after immunotherapy-induced cCR in locoregional GEA.
How this fits with the broader evidence base
The findings are directionally consistent with the broader checkpoint inhibitor literature in MSI-H/dMMR solid tumors, where durable benefit is common. Across gastrointestinal cancers, MSI-H status has become one of the clearest predictive biomarkers for PD-1 blockade. In GEA specifically, the study adds valuable real-world evidence that extends beyond trial populations and supports the concept that a subset of patients may achieve deep, durable remission with immunotherapy.
Current guidelines generally endorse MSI/MMR testing in advanced and resectable GEA because the result has both prognostic and therapeutic implications. This study strengthens the argument that testing is not merely academic; it can materially change the treatment course, including the possibility of organ preservation in selected locoregional cases. At the same time, the data do not replace existing standards of care. Rather, they suggest a path toward more individualized treatment in which response-adapted strategies may eventually complement or, in highly selected patients, substitute for surgery.
Conclusion
In this single-center retrospective cohort of dMMR GEA, ICI-based therapy produced high response rates in both metastatic and locoregional disease. The durability of complete responses, especially in patients managed without immediate surgery, suggests that nonoperative strategies may be feasible for selected individuals with careful surveillance. The study also underscores the need for better predictive biomarkers and prospective trials to determine which patients can safely omit surgery and how best to sequence systemic and local therapy.
Funding and clinicaltrials.gov
The abstract does not specify funding sources or a clinicaltrials.gov registration number. These details should be confirmed in the full article.
References
1. Okui J, Prakash LR, Lyu HG, et al. Immunotherapy Response in Microsatellite Instability-High Gastroesophageal Adenocarcinoma. JAMA Surgery. 2026;161(6):600-607. PMID: 41984470.
2. National Comprehensive Cancer Network. NCCN Clinical Practice Guidelines in Oncology: Gastric Cancer. Current version.
3. Le DT, Durham JN, Smith KN, et al. Mismatch repair deficiency predicts response of solid tumors to PD-1 blockade. Science. 2017;357(6349):409-413.
4. Andre T, Shiu KK, Kim TW, et al. Pembrolizumab in microsatellite-instability-high advanced colorectal cancer. N Engl J Med. 2020;383:2207-2218.
5. The Cancer Genome Atlas Research Network. Comprehensive molecular characterization of gastric adenocarcinoma. Nature. 2014;513:202-209.