Abstract
Background
Immune checkpoint inhibitors (ICIs) provide durable benefit in high microsatellite instability (MSI-H)/mismatch repair deficiency (dMMR) gastrointestinal (GI) tumors, but real-world evidence across colorectal cancer (CRC) and non-CRC primaries remains limited.
Materials and methods
We conducted a multicenter retrospective study including 122 patients with advanced MSI-H/dMMR GI tumors treated with ICIs in eight Spanish university hospitals. The primary endpoint was overall survival (OS). Secondary endpoints included progression-free survival (PFS), response, and safety.
Results
Median age at ICI initiation was 70.4 years; 79.5% had an Eastern Cooperative Oncology Group (ECOG) performance status (PS) of 0-1. CRC accounted for 54.9% of cases, gastroesophageal adenocarcinoma (GEA) for 34.4%, and other GI tumors for 10.7%. At a median follow-up of 33 months, median PFS was 46.0 months [95% confidence interval (CI) 30.6-61.4 months] and median OS was 53.2 months (95% CI 42.6-63.7 months), with 36-month rates of 57.1% for PFS and 62.8% for OS. The objective response rate (ORR) was 77.7% and the disease control rate (DCR) was 91.1%, with comparable efficacy across CRC, GEA, and other GI tumor subgroups. In multivariable analysis, ECOG PS was the only independent prognostic factor for both OS and PFS. Treatment was generally well tolerated; 47.5% of patients experienced adverse events, grade ≥3 in 10.7%, with no treatment-related deaths.
Conclusions
In this multicenter European real-world cohort, ICIs demonstrated clinically meaningful and durable benefit in advanced MSI-H/dMMR GI tumors, confirming pivotal trial results in routine practice. ECOG PS emerged as the main independent prognostic factor, underscoring its central role in patient selection.
Key words: microsatellite instability, mismatch repair deficiency, gastrointestinal tumors, immune checkpoint inhibitors, real-world evidence, prognosis
Highlights
-
•
This was a European real-world cohort of MSI-H/dMMR GI tumors treated with ICIs.
-
•
The ORR was 77.7% and DCR was 91.1%, with prolonged PFS and OS.
-
•
ECOG PS was the only independent prognostic factor.
-
•
Similar outcomes were seen across colorectal, gastroesophageal, and other GI tumors.
-
•
A manageable safety profile with infrequent grade ≥3 adverse events was observed.
Introduction
Gastrointestinal (GI) tumors with mismatch repair deficiency (dMMR) or high microsatellite instability (MSI-H) represent a distinct molecular subtype characterized by hypermutation, high neoantigen load, and prominent immune infiltration, which confer exceptional sensitivity to immune checkpoint inhibitors (ICIs).1,2 Initial clinical evidence came from the phase II KEYNOTE-016 trial, where programmed cell death protein 1 (PD-1) blockade achieved durable responses across dMMR tumors regardless of origin.3,4 Subsequent studies confirmed these findings across different GI malignancies, particularly in metastatic colorectal cancer (CRC), including KEYNOTE-1645 and CheckMate-142,6, 7, 8, 9 while the basket trial KEYNOTE-158 extended pembrolizumab efficacy to non-CRC tumors including gastroesophageal and other GI primaries.10, 11, 12 The phase III KEYNOTE-177 trial subsequently established pembrolizumab as the standard of care in the first-line setting for dMMR/MSI-H metastatic CRC, demonstrating significant improvements in survival compared with chemotherapy.13, 14, 15 More recently, CheckMate-8HW showed that dual checkpoint blockade with nivolumab plus ipilimumab significantly improved both progression-free survival (PFS) and overall survival (OS) compared with chemotherapy and also outperformed nivolumab monotherapy in the first-line setting.16,17
On the basis of these data, international guidelines, including those of the European Society for Medical Oncology and the National Comprehensive Cancer Network, recommend routine MSI/MMR testing in all GI malignancies and position ICIs as the preferred treatment option for advanced MSI-H/dMMR tumors.18,19 Despite this robust evidence, real-world data specifically addressing the heterogeneity of MSI-H/dMMR GI tumors remain limited. Most pivotal trials have focused on CRC, with non-CRC GI tumors underrepresented, and strict eligibility criteria may not fully reflect routine clinical practice.20 Furthermore, although clinical factors such as Eastern Cooperative Oncology Group (ECOG) performance status (PS), primary tumor resection, or metastatic burden have been suggested as potential prognostic variables in pan-cancer cohorts, their prognostic impact within MSI-H/dMMR GI tumors treated with ICIs in real-world settings remains incompletely characterized.21,22
We therefore conducted the GAIN study, a retrospective multicenter analysis including patients with CRC and non-CRC GI tumors, to evaluate the real-world outcomes of ICI therapy in advanced MSI-H/dMMR disease. The objectives of this study were to assess PFS and OS, treatment safety, and to explore baseline clinical factors associated with improved outcomes in this biologically distinct population.
Materials and methods
Study design and population
The GAIN study was a retrospective, multicenter analysis conducted in eight Spanish university hospitals, including seven centers belonging to the Galician Research Group on Gastrointestinal Tumors (GITuD) and Hospital Universitario Marqués de Valdecilla in Santander. Eligible patients were adults with histologically confirmed advanced or metastatic dMMR/MSI-H GI tumors who had received at least one cycle of ICI therapy in routine clinical practice between November 2015 and August 2023. dMMR or MSI-H status was determined according to local institutional standards, using immunohistochemistry and/or PCR-based or next-generation sequencing assays. Patients with insufficient data for efficacy or safety analyses were excluded. A total of 122 patients were included in the final analysis (Supplementary Figure S1, available at https://doi.org/10.1016/j.esmorw.2025.100676).
Ethics
The study was conducted in accordance with the Declaration of Helsinki and Good Clinical Practice guidelines. The protocol was approved by the Comité de Ética de la Investigación con Medicamentos de Galicia (CEIm-G; code 2020/458) and by the institutional review boards of all participating centers. Written informed consent was obtained from all living patients; for deceased patients, data collection was carried out under an ethics committee-approved waiver of informed consent, and CEIm-G authorized the retrospective collection and use of anonymized clinical data.
Data collection
Clinical data were retrieved from electronic medical records by local investigators at each participating center, including demographics, ECOG PS, tumor characteristics [primary site, differentiation grade, MMR loss pattern, and molecular features when available—rat sarcoma viral oncogene homolog (RAS)/B-Raf proto-oncogene, serine/threonine kinase (BRAF) for CRC and human epidermal growth factor receptor 2 (HER2) for gastroesophageal adenocarcinoma (GEA)], prior treatments, metastatic burden (number and sites), ICI regimen and dosing, best overall response per RECIST v1.1, survival outcomes, and adverse events per Common Terminology Criteria for Adverse Events (CTCAE) v5.0. Radiological assessments were carried out according to routine clinical practice at each participating center. A detailed list of collected variables is provided in Supplementary Table S1, available at https://doi.org/10.1016/j.esmorw.2025.100676.
Endpoints
The primary endpoint was OS, defined as the time from ICI initiation to death from any cause. Secondary endpoints included PFS, defined as the time from ICI initiation to documented progression or death, whichever occurred first; objective response rate (ORR), defined as the proportion of patients who achieved complete or partial response (PR); and disease control rate (DCR), defined as complete response (CR), PR, or stable disease (SD) as best response. All tumor responses were assessed according to RECIST v1.1.23 Patients without documented progression or death were censored at the date of last follow-up. Safety was evaluated as the incidence and severity of adverse events (AEs), including immune-related AEs (irAEs), graded according to CTCAE v5.0.24
Statistical analysis
Baseline characteristics were summarized using descriptive statistics. Categorical variables were expressed as frequencies and percentages, and continuous variables as medians with interquartile ranges. Survival outcomes were estimated by the Kaplan–Meier method and compared across subgroups using the log-rank test. Hazard ratios (HRs) and 95% confidence intervals (CIs) were calculated using the Cox proportional hazards regression. Variables with P < 0.10 in univariate analyses were entered into multivariable models. No data imputation was carried out; all analyses were based on available cases. Statistical significance was set at P < 0.05 (two-sided). Analyses were conducted with SPSS version 27 (IBM Corp., Armonk, NY).
Results
Baseline characteristics
A total of 122 patients were included, of whom 67 (54.9%) had CRC and 55 (45.1%) had non-CRC GI tumors, including GEA in 42 patients (34.4%) and other GI malignancies in 13 patients (10.7%). In the CRC subgroup, primary laterality was right-sided in 46/67 (68.7%), left-sided in 17/67 (25.4%), and rectal in 4/67 (5.9%). Primary tumor resection had been carried out in 81 (66.4%) patients.
The median age at ICI initiation was 70.4 years (range 29-89 years), and 64 (52.5%) were male. ECOG PS was 0-1 in 97 (79.5%) patients and ≥2 in 24 (19.7%). Most tumors were poorly differentiated (61/122; 50.0%). Limited metastatic burden was observed in the majority of cases, with 0-2 metastatic sites in 102 (83.6%). Liver metastases were present in 37 (30.3%) and peritoneal metastases in 54 (44.3%).
Molecular profiling in CRC showed BRAF mutations in 25/67 (37.3%), RAS mutations in 15/67 (22.4%), and RAS/BRAF wild-type in 23/67 (34.3%), while the status was unknown in 4/67 (6.0%). All tested GEAs were HER2-negative. Regarding the mismatch repair defect, loss of MutL homolog 1 and/or postmeiotic segregation increased 2 was observed in 101 (82.8%) and loss of MutS homolog 2 and/or MutS homolog 6 in 19 (15.6%) patients, and the MMR loss pattern was unknown in 2 (1.6%) patients.
With respect to prior treatments, 34 patients (27.9%) were treatment-naive in the metastatic setting, 23 (18.9%) had received adjuvant therapy only, and 65 (53.2%) had received one or more systemic lines for advanced disease before ICI initiation. Almost all patients received anti-PD-1 monotherapy (121; 99.2%), while one (0.8%) received the combination of anti-PD-1 plus anti-cytotoxic T-lymphocyte-associated protein 4. Fixed dosing was used in 71 (58.2%), while 51 (41.8%) received weight-based (mg/kg) dosing. Baseline patient, tumor, and molecular characteristics, as well as ICI regimens, are summarized in Table 1.
Table 1.
Baseline patient and tumor characteristics
| Characteristic | |
|---|---|
| Age, median (range), years | 70.4 (29-89) |
| Sex | |
| Male | 64 (52.5) |
| Female | 58 (47.5) |
| ECOG PS | |
| 0-1 | 97 (79.5) |
| 2 | 24 (19.7) |
| Unknown | 1 (0.8) |
| Primary tumor resection | |
| Yes | 81 (66.4) |
| No | 41 (33.6) |
| Tumor type | |
| CRC | 67 (54.9) |
| Right-sided | 46 (68.7) |
| Left-sided | 17 (25.4) |
| Rectum | 4 (5.9) |
| GEA | 42 (34.4) |
| Other GI tumors | 13 (10.7) |
| Pancreatic adenocarcinoma | 4 (7.3) |
| Intrahepatic cholangiocarcinoma | 2 (3.6) |
| Esophageal squamous-cell carcinoma | 2 (3.6) |
| Duodenal adenocarcinoma | 2 (3.6) |
| Jejunal adenocarcinoma | 2 (3.6) |
| Unknown primary adenocarcinoma | 1 (1.8) |
| Tumor grade | |
| Low grade (G1-2) | 48 (39.3) |
| High grade (G3) | 61 (50.0) |
| Unknown | 13 (10.7) |
| CRC molecular profile (n = 67) | |
| RAS/BRAF wild-type | 23 (34.3) |
| KRAS/NRAS mutated | 15 (22.4) |
| BRAF mutated | 25 (37.3) |
| Unknown | 4 (6.0) |
| MMR loss pattern | |
| MLH1 and/or PMS2 | 101 (82.8) |
| MSH2 and/or MSH6 | 19 (15.6) |
| Unknown | 2 (1.6) |
| Previous treatment | |
| None | 34 (27.9) |
| Adjuvant ChT | 23 (18.9) |
| ≥1 line metastatic ChT | 65 (53.2) |
| Number of metastatic sites | |
| 0-2 | 102 (83.6) |
| ≥3 | 20 (16.4) |
| Liver metastases | |
| Yes | 37 (30.3) |
| No | 85 (69.7) |
| Peritoneal metastases | |
| Yes | 54 (44.3) |
| No | 68 (55.7) |
| ICI regimen | |
| Anti-PD-1 | 121 (99.2) |
| Anti-PD-1 + Anti-CTLA-4 | 1 (0.8) |
| ICI dosing | |
| Fixed dose | 71 (58.2) |
| Weight-based (mg/kg) | 51 (41.8) |
Data are presented as n (%) unless otherwise specified. Percentages are calculated based on the total cohort unless indicated. Missing data were not imputed, and analyses were carried out on available cases only.
B-RAF, B-Raf proto-oncogene, serine/threonine kinase; ChT, chemotherapy; CRC, colorectal cancer; CTLA-4, cytotoxic T-lymphocyte-associated protein 4; ECOG, Eastern Cooperative Oncology Group; GEA, gastroesophageal adenocarcinoma; GI, gastrointestinal; ICI, immune checkpoint inhibitor; MLH1, MutL homolog 1; MMR, mismatch repair; MSH2, MutS homolog 2; MSH6, MutS homolog 6; PD-1, programmed cell death protein 1; PMS2, postmeiotic segregation increased 2; PS, performance status; RAS, rat sarcoma viral oncogene homolog.
Efficacy
Best overall response was evaluable in 112 patients, including 61 with CRC, 38 with GEA, and 13 with other GI tumors. CR was achieved in 35 patients (31.3%), PR in 52 (46.4%), SD in 15 (13.4%), and progressive disease (PD) in 10 (8.9%), resulting in an ORR of 87/112 (77.7%) and a DCR of 102/112 (91.1%) (Table 2). Response outcomes were broadly comparable across tumor subgroups, with an ORR of 75.4% in CRC, 78.9% in GEA, and 84.6% in other GI tumors, and a DCR of 90.2%, 92.1%, and 92.3%, respectively.
Table 2.
Efficacy outcomes of immune checkpoint inhibitors in advanced dMMR/MSI-H gastrointestinal tumors, overall and by tumor subtype
| Outcome | Total (N = 122) | CRC (n = 67) | GEA (n = 42) | Other GI tumors (n = 13) |
|---|---|---|---|---|
| Best overall response (n evaluable) | 112 | 61 | 38 | 13 |
| Complete response (CR), n (%) | 35 (31.3) | 17 (27.9) | 13 (34.2) | 6 (46.2) |
| Partial response (PR), n (%) | 52 (46.4) | 29 (47.5) | 17 (44.7) | 5 (38.4) |
| Stable disease (SD), n (%) | 15 (13.4) | 9 (14.8) | 5 (13.2) | 1 (7.7) |
| Progressive disease (PD), n (%) | 10 (8.9) | 6 (9.8) | 3 (7.9) | 1 (7.7) |
| Objective response rate (ORR), n (%) | 87 (77.7) | 46 (75.4) | 30 (78.9) | 11 (84.6) |
| Disease control rate (DCR), n (%) | 102 (91.1) | 55 (90.2) | 35 (92.1) | 12 (92.3) |
| Progression-free survival (PFS) | ||||
| Median, months (95% CI) | 46.0 (30.6-61.4) | 50.7 (42.7-58.6) | 33.1 (14.3-51.8) | 30.7 (0.1-113.2) |
| 24-month PFS rate (%) | 60.6 | 62.1 | 57.4 | 64.0 |
| 36-month PFS rate (%) | 57.1 | 59.4 | 54.9 | 64.0 |
| Overall survival (OS) | ||||
| Median, months (95% CI) | 53.2 (42.6-63.7) | 60.9 (NR) | 34.0 (21.6-46.4) | 56.8 (NR) |
| 24-month OS rate (%) | 70.3 | 76.2 | 63.3 | 59.9 |
| 36-month OS rate (%) | 62.8 | 69.2 | 48.8 | 59.9 |
Survival rates at 24 and 36 months were estimated using the Kaplan–Meier method. Subgroup analyses are exploratory and were not powered for formal statistical comparisons.
CI, confidence interval; CRC, colorectal cancer; dMMR, mismatch repair deficiency; GEA, gastroesophageal adenocarcinoma; GI, gastrointestinal; MSI-H, high microsatellite instability; NR, not reached.
At a median follow-up of 33.0 months, the median PFS was 46.0 months (95% CI 30.6-61.4 months). The estimated 24- and 36-month PFS rates were 60.6% and 57.1%, respectively. The median OS was 53.2 months (95% CI 42.6-63.7 months), with 24- and 36-month OS rates of 70.3% and 62.8%, respectively (Figure 1).
Figure 1.
Kaplan–Meier survival curves for the overall cohort. (A) Progression-free survival (PFS). (B) Overall survival (OS). Survival probabilities were estimated using the Kaplan–Meier method; patients without events were censored at the date of last follow-up. Median follow-up was 33 months. ICI, immune checkpoint inhibitor.
In exploratory subgroup analyses, median PFS was 50.7 months in CRC, 33.1 months in GEA, and 30.7 months in other GI tumors, while median OS was 60.9 months, 34.0 months, and 56.8 months, respectively (Table 2; Figure 2). No statistically significant differences in survival outcomes were observed across tumor subgroups, consistent with the limited sample size of non-CRC cohorts (Table 2; Figure 2C and D).
Figure 2.
Kaplan–Meier survival curves according to prognostic factors. (A, B) OS and PFS by ECOG PS (0-1 versus 2). (C, D) OS and PFS in CRC patients by primary tumor laterality (right versus left/rectum). (E, F) OS and PFS by best overall response (CR/PR versus SD/PD). Subgroup analyses are exploratory and were not powered for formal statistical comparisons. Survival probabilities were estimated using the Kaplan–Meier method, with censoring at last follow-up. CR, complete response; CRC, colorectal cancer; ECOG, Eastern Cooperative Oncology Group; ICI, immune checkpoint inhibitor; OS, overall survival; PD, progressive disease; PFS, progression-free survival; PR, partial response; PS, performance status; SD, stable disease.
Safety
The safety profile of ICIs was consistent with previous reports, and treatment was generally well tolerated. AEs of any grade occurred in 58/122 (47.5%) patients, while grade ≥3 events were reported in 13/122 (10.7%) (Table 3). Among patients experiencing any AE, the most frequent toxicities of any grade were asthenia (reported in 58 patients; 47.5%), anemia (29/122; 23.8%), and diarrhea (17/122; 13.9%).
Table 3.
Treatment-related adverse events (AEs)
| Adverse event | All grades, n (%) | Grade ≥3, n (%) |
|---|---|---|
| Asthenia | 58 (47.5) | 3 (2.5) |
| Anemia | 29 (23.8) | — |
| Diarrhea | 17 (13.9) | 1 (0.8) |
| Hepatic toxicity (AST/ALT increase) | 15 (12.3) | 4 (3.3) |
| Cutaneous toxicity | 10 (8.2) | — |
| Nephritis | 7 (5.7) | 4 (3.3) |
| Hypothyroidism | 6 (4.9) | 1 (0.8) |
| Neutropenia | 6 (4.9) | 1 (0.8) |
| Arthritis | 6 (4.9) | 2 (1.6) |
| Thrombocytopenia | 3 (2.5) | — |
| Pneumonitis | 2 (1.6) | 2 (1.6) |
Adverse events were collected retrospectively from electronic medical records and graded according to the Common Terminology Criteria for Adverse Events (CTCAE), version 5.0. Due to the retrospective nature of the study, underreporting of low-grade or late-onset events cannot be excluded. Percentages are calculated based on the total cohort (n = 122).
ALT, alanine aminotransferase; AST, aspartate aminotransferase.
Among irAEs, the most commonly reported were hypothyroidism (6/122; 4.9%), arthritis (6/122; 4.9%), and nephritis (7/122; 5.7%). Severe events (grade ≥3) were infrequent, with hepatic toxicity (4/122; 3.3%) and nephritis (4/122; 3.3%) being the most notable. Treatment discontinuation due to toxicity occurred in 6/122 (4.9%) patients, and no treatment-related deaths were observed.
Prognostic factors
Univariable analyses for OS and PFS are shown in Table 4. ECOG PS emerged as the most consistent prognostic factor. Patients with an ECOG PS of 0 achieved markedly longer OS and PFS compared with those with an ECOG PS of 1 or 2 (P = 0.001 and P = 0.019, respectively; Figure 2A and B). Primary tumor resection was associated with longer OS (HR 2.13, 95% CI 1.16-3.92, P = 0.012) and showed a trend toward better PFS (P = 0.087). Patients with three or more metastatic sites had significantly shorter OS than those with 0-2 sites (corresponding to an HR of 0.48 for the reference category, 95% CI 0.25-0.93, P = 0.026), although the difference in PFS was not significant. The presence of liver metastases showed only a trend toward inferior OS and PFS (P = 0.061 and P = 0.093, respectively), whereas peritoneal involvement had no impact.
Table 4.
Univariable and multivariable Cox regression analysis of baseline clinical factors associated with overall survival (OS) and progression-free survival (PFS)
| Variable | Comparison (ref) | HR OS univariable (95% CI) | P OS univariable | HR PFS univariable (95% CI) | P PFS univariable | HR OS multivariable (95% CI) | P OS multivariable | HR PFS multivariable (95% CI) | P PFS multivariable |
|---|---|---|---|---|---|---|---|---|---|
| ECOG PS | 1 versus 0 | 0.16 (0.05-0.54) | 0.001 | 0.24 (0.09-0.69) | 0.019 | 0.16 (0.05-0.54) | 0.003 | 0.24 (0.09-0.69) | 0.008 |
| ECOG PS | 2 versus 0 | 0.43 (0.20-0.93) | — | 0.43 (0.20-0.93) | — | 0.43 (0.20-0.93) | 0.032 | 0.52 (0.25-1.09) | 0.085∗ |
| Primary tumor resection | Yes versus no | 0.47 (0.26-0.86) | 0.012 | 0.61 (0.35-1.09) | 0.087∗ | 0.46 (0.19-1.10) | 0.081∗ | 0.50 (0.23-1.07) | 0.078∗ |
| No. of metastatic sites | >3 versus 0-2 | 2.08 (1.07-3.97) | 0.026 | 1.64 (0.86-3.23) | 0.126 | 0.76 (0.34-1.69) | 0.476 | 0.90 (0.39-2.07) | 0.772 |
| Liver metastases | Yes versus no | 1.79 (0.97-3.27) | 0.061∗ | 1.61 (0.91-2.83) | 0.093∗ | NA | NA | NA | NA |
| Peritoneal metastases | Yes versus no | 1.16 (0.64-2.12) | 0.608 | 1.11 (0.63-1.90) | 0.733 | NA | NA | NA | NA |
| Tumor type | CRC versus non-CRC | 0.57 (0.31-1.04) | 0.063∗ | 0.84 (0.48-1.47) | 0.541 | 0.85 (0.36-1.97) | 0.701 | 0.64 (0.30-1.35) | 0.243 |
| CRC laterality | Right versus left/rectum | 0.27 (0.12-0.64) | 0.002 | 0.38 (0.18-0.83) | 0.012 | NA | NA | NA | NA |
| Differentiation grade | High versus low | 0.84 (0.44-1.61) | 0.597 | 0.87 (0.47-1.60) | 0.664 | NA | NA | NA | NA |
| BRAF status (CRC) | Mutated versus WT | 0.78 (0.25-2.43) | 0.662 | 0.67 (0.24-1.85) | 0.437 | NA | NA | NA | NA |
| MMR loss pattern | MLH1/PMS2 versus MSH2/MSH6 | 1.26 (0.56-3.00) | 0.595 | 1.21 (0.54-2.71) | 0.637 | NA | NA | NA | NA |
| Best response to ICI | SD/PD versus CR/PR | 6.15 (3.09-12.26) | <0.001 | 6.62 (3.56-12.45) | <0.001 | NA | NA | NA | NA |
HRs were estimated using Cox proportional hazards models. Univariable analyses were carried out for all listed variables; variables with P < 0.10 in univariable analysis were entered into the multivariable models. The reference category for ECOG PS was ECOG 0. Best response to ICI therapy was analyzed as a landmark variable and therefore was not included in multivariable models. No imputation was carried out for missing data; analyses were conducted on available cases only.
CI, confidence interval; CR, complete response; CRC, colorectal cancer; ECOG PS, Eastern Cooperative Oncology Group performance status; HR, hazard ratio; ICI, immune checkpoint inhibitor; OS, overall survival; MLH1, MutL homolog 1; MMR, mismatch repair; MSH2, MutS homolog 2; MSH6, MutS homolog 6; NA, not applicable; PD, progressive disease; PFS, progression-free survival; PMS2, postmeiotic segregation increased 2; PR, partial response; PS, performance status; SD, stable disease; WT, wild-type.
P values between 0.05 and 0.10 are reported as statistical trends.
Tumor-related features were less predictive. The type of primary tumor (CRC versus non-CRC) and differentiation grade were not associated with survival. Within the CRC subgroup, however, primary tumor laterality was relevant, with right-sided tumors showing longer OS and PFS compared with left-sided or rectal primaries (P = 0.002 for OS; P = 0.012 for PFS; Figure 2C and D). Neither BRAF status nor the specific pattern of MMR loss influenced OS or PFS. Response to ICI was strongly correlated with prognosis, as patients achieving CR or PR had significantly longer OS and PFS compared with those with SD or PD (both P < 0.001; Figure 2E and F).
In the multivariable analysis (Table 4; Figure 2), ECOG PS remained the only independent prognostic factor for both OS and PFS (P = 0.003 and P = 0.008, respectively). Patients with an ECOG PS of 0 consistently achieved the most favorable outcomes, while those with an ECOG PS of 1-2 showed significantly inferior survival. Other variables, including primary tumor resection, metastatic burden, and tumor type (CRC versus non-CRC), did not retain independent prognostic value after adjustment.
Discussion
ICIs have transformed the therapeutic landscape of dMMR/MSI-H GI tumors, providing durable benefit across tumor types.3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17 In this multicenter real-world cohort, ICI therapy achieved an ORR of 77.7% and a DCR of 91.1%, with a median PFS of 46.0 months and median OS of 53.2 months, outcomes that mirror those of pivotal clinical trials5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17 and confirm their reproducibility in routine clinical practice across different GI malignancies.
Most available evidence in this scenario comes from clinical trials in CRC, whereas data in non-CRC GI tumors remain scarce.10, 11, 12 In the present series, GEA accounted for more than one-third of cases and was analyzed separately from other non-CRC tumors, allowing a more granular assessment of efficacy beyond CRC. Outcomes in GEA and other non-CRC GI malignancies were comparable to those observed in CRC, supporting the use of ICIs across the spectrum of advanced dMMR/MSI-H GI tumors. These findings should be interpreted as confirmatory and exploratory, given the modest sample size of non-CRC subgroups, but are consistent with the tumor-agnostic activity reported in KEYNOTE-158 and subsequent studies.10, 11, 12
ECOG PS emerged as the most robust prognostic determinant, with patients with an ECOG PS of 2 showing markedly worse outcomes compared with those with an ECOG PS of 0-1. This finding highlights the clinical importance of baseline functional status when selecting patients for ICIs in the real-world setting. In addition, primary tumor resection and the number of metastatic sites showed associations with OS in univariable analysis, although these did not retain independent significance in multivariable models. Within the CRC subgroup, primary tumor laterality was associated with survival, with right-sided tumors achieving superior outcomes, a finding that aligns with prior observations suggesting biological and immunological differences according to tumor location.25, 26, 27 Best overall response was also strongly correlated with long-term survival, reinforcing the concept that achieving CR or PR represents a surrogate marker of durable benefit with ICIs, as previously reported in other tumor types.28
Treatment was generally well tolerated, with grade ≥3 AEs occurring in 10.7% of patients, similar to rates reported in pivotal trials.5, 6, 7,13 The predominance of ICI monotherapy in this cohort reflects the regulatory approvals and reimbursement landscape in Spain during much of the study period, particularly for MSI-H/dMMR gastroesophageal cancer, and limits the extrapolation of these results to current chemo-immunotherapy combinations. The most frequently reported immune-related toxicities involved hepatic, renal, and endocrine systems and were manageable with standard approaches. The overall incidence of AEs was somewhat lower than that reported in prospective trials, likely reflecting underreporting inherent to retrospective real-world data, a limitation previously described in other observational series.29
Several additional limitations warrant consideration. The retrospective design introduces potential selection and reporting bias, and MSI/MMR testing was carried out locally without centralized confirmation. Although loss of MSH2 and/or MSH6 suggests a possible association with Lynch syndrome, systematic germline testing was not available, precluding formal classification of hereditary cases. Data on co-mutational landscapes and subsequent lines of therapy were limited, restricting mechanistic interpretation and precluding robust analyses of treatment sequencing. Moreover, although treatment was administered across different lines, the relatively small sample size limited the ability to evaluate treatment line as an independent prognostic factor. Finally, some subgroups—particularly non-CRC histologies and patients with an ECOG PS of 2—remain underpowered, and results should therefore be interpreted with caution.
In conclusion, this multicenter real-world study confirms the effectiveness and safety of ICIs in advanced MSI-H/dMMR GI tumors and extends trial evidence to routine clinical practice. ECOG PS emerged as the main independent prognostic factor, while tumor subtype, primary tumor resection, metastatic burden, CRC laterality, and treatment response provided additional exploratory insights. These findings support the broad use of ICIs across GI malignancies while highlighting the need for larger, prospectively collected real-world datasets to refine patient selection and contextualize emerging combination strategies.
Conclusions
In this multicenter real-world cohort, ICIs demonstrated clinically meaningful efficacy and durable benefit in advanced MSI-H/dMMR GI tumors, with outcomes consistent with those reported in pivotal clinical trials. ECOG PS emerged as the only independent prognostic factor for survival, while primary tumor resection, metastatic burden, CRC laterality, and best overall response provided additional exploratory insights. Together, these findings confirm the reproducibility of ICI benefit in routine clinical practice across different GI malignancies and underscore the importance of baseline clinical factors in patient selection. Larger and prospectively collected real-world studies are warranted to further refine prognostic stratification and to contextualize emerging combination strategies.
Acknowledgements
The authors thank all members of the Galician Research Group on Gastrointestinal Tumors (GITuD) and the Hospital Universitario Marqués de Valdecilla for their collaboration in patient management and data collection.
Funding
None declared.
Disclosure
NML reports honoraria, advisory roles, travel grants, or research funding from Amgen, Astellas, AstraZeneca, Bayer, Be One, BMS, Deciphera, Ipsen, Kiowa Kirin, Leopharma, MSD, Novartis, Nutricia, Servier, Takeda. AC reports honoraria, advisory roles, travel grants, or research funding from AstraZeneca, BMS, Merck, Servier. RM received payment or honoraria from Merck Sharp & Dohme, Merck, Servier; received support for attending meetings and/or travel from Merck Sharp & Dohme, Merck, Servier. AG reports honoraria, advisory roles, travel grants, or research funding from Merck, Novocure, Servier. AFM has received fees or honoraria for lectures, presentations, speaker bureaus, manuscript writing, or educational events from Servier, Lilly, Sanofi, AstraZeneca, MSD, Roche, Celgene, Amgen, and Astellas; she has also participated in advisory boards or data safety monitoring boards for MSD, Bristol Myers Squibb (BMS), Astellas, and Bayer; additionally, she has received compensation as a consultant for Amgen S.A. CL received consulting fees from Merck Sharp & Dohme, Roche, Merck, Servier, and Takeda; received payment or honoraria from Merck Sharp & Dohme, Roche, Merck, Servier, Takeda, Ipsen, Sanofi, Amgen, and Bristol Myers Squibb; received support for attending meetings and/or travel from Merck Sharp & Dohme, Roche, Merck, Servier, Ipsen, Sanofi, Amgen, and Bristol Myers Squibb; and participated on a data safety monitoring board or advisory board for Merck Sharp & Dohme, Roche, Merck, Servier, and Takeda. All other authors have declared no conflicts of interest.
Data sharing
The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request.
Supplementary Data
Supplementary Figure.
References
- 1.Boland C.R., Goel A. Microsatellite instability in colorectal cancer. Gastroenterology. 2010;138(6):2073–2087.e3. doi: 10.1053/j.gastro.2009.12.064. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Bonneville R., Krook M.A., Kautto E.A., et al. Landscape of microsatellite instability across 39 cancer types. JCO Precis Oncol. 2017;2017 doi: 10.1200/PO.17.00073. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Le D.T., Uram J.N., Wang H., et al. PD-1 blockade in tumors with mismatch-repair deficiency. N Engl J Med. 2015;372(26):2509–2520. doi: 10.1056/NEJMoa1500596. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Le D.T., Durham J.N., Smith K.N., et al. Mismatch repair deficiency predicts response of solid tumors to PD-1 blockade. Science. 2017;357(6349):409–413. doi: 10.1126/science.aan6733. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Le D.T., Kim T.W., Van Cutsem E., et al. Phase II open-label study of pembrolizumab in treatment-refractory, microsatellite instability-high/mismatch repair-deficient metastatic colorectal cancer: KEYNOTE-164. J Clin Oncol. 2020;38(1):11–19. doi: 10.1200/JCO.19.02107. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Overman M.J., McDermott R., Leach J.L., et al. Nivolumab in patients with metastatic DNA mismatch repair-deficient or microsatellite instability-high colorectal cancer (CheckMate 142): an open-label, multicentre, phase 2 study. Lancet Oncol. 2017;18(9):1182–1191. doi: 10.1016/S1470-2045(17)30422-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Overman M.J., Lonardi S., Wong K.Y.M., et al. Durable clinical benefit with nivolumab plus ipilimumab in DNA mismatch repair-deficient/microsatellite instability-high metastatic colorectal cancer. J Clin Oncol. 2018;36(8):773–779. doi: 10.1200/JCO.2017.76.9901. [DOI] [PubMed] [Google Scholar]
- 8.Overman M.J., Bergamo F., McDermott R.S., et al. Nivolumab in patients with dMMR/MSI-H mCRC: long-term survival according to prior line (CheckMate-142) J Clin Oncol. 2021;39(15_suppl):3573. [Google Scholar]
- 9.Lenz H.J., Lonardi S., Zagonel V., et al. Nivolumab plus low-dose ipilimumab as first-line treatment for MSI-H/dMMR mCRC: clinical update. J Clin Oncol. 2022;40(14):1616–1624. [Google Scholar]
- 10.Marabelle A., Le D.T., Ascierto P.A., et al. Efficacy of pembrolizumab in non-CRC MSI-H/dMMR cancers: KEYNOTE-158. J Clin Oncol. 2020;38(1):1–10. doi: 10.1200/JCO.19.02105. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Maio M., Ascierto P.A., Manzyuk L., et al. Pembrolizumab in MSI-H/dMMR cancers: updated KEYNOTE-158. Ann Oncol. 2022;33(9):929–938. doi: 10.1016/j.annonc.2022.05.519. [DOI] [PubMed] [Google Scholar]
- 12.Marabelle A., O’Malley D.M., Hendifar A.E., et al. Pembrolizumab in MSI-H/dMMR advanced solid tumors: updated KEYNOTE-158. Nat Cancer. 2025;6(2):253–258. doi: 10.1038/s43018-024-00894-y. [DOI] [PubMed] [Google Scholar]
- 13.André T., Shiu K.K., Kim T.W., et al. Pembrolizumab in MSI-H advanced CRC. N Engl J Med. 2020;383(23):2207–2218. doi: 10.1056/NEJMoa2017699. [DOI] [PubMed] [Google Scholar]
- 14.Diaz L.A., Jr., Shiu K.K., Kim T.W., et al. Pembrolizumab versus chemotherapy for microsatellite instability-high or mismatch repair-deficient metastatic colorectal cancer (KEYNOTE-177): final analysis of a randomised, open-label, phase 3 study. Lancet Oncol. 2022;23(5):659–670. doi: 10.1016/S1470-2045(22)00197-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.André T., Shiu K.K., Kim T.W., et al. Pembrolizumab versus chemotherapy in MSI-H/dMMR mCRC: 5-year analysis of KEYNOTE-177. Ann Oncol. 2025;36(3):277–284. doi: 10.1016/j.annonc.2024.11.012. [DOI] [PubMed] [Google Scholar]
- 16.André T., Lonardi S., Wong K.Y.M., et al. Nivolumab plus low-dose ipilimumab versus standard of care in first-line MSI-H/dMMR mCRC: checkMate-8HW. Lancet. 2024;404(10319):1893–1905. [Google Scholar]
- 17.André T., Elez E., Van Cutsem E., et al. CheckMate-8HW Investigators Nivolumab plus ipilimumab in MSI-H mCRC. N Engl J Med. 2024;391:2014–2026. doi: 10.1056/NEJMoa2402141. [DOI] [PubMed] [Google Scholar]
- 18.Benson A.B., Venook A.P., Al-Hawary M.M., et al. NCCN guidelines insights: colon cancer, Version 2.2018. J Natl Compr Canc Netw. 2018;16(4):359–369. doi: 10.6004/jnccn.2018.0021. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Cervantes A., Adam R., Roselló S., et al. Metastatic colorectal cancer: ESMO Clinical Practice Guideline for diagnosis, treatment and follow-up. Ann Oncol. 2023;34(1):10–32. doi: 10.1016/j.annonc.2022.10.003. [DOI] [PubMed] [Google Scholar]
- 20.Lilenbaum R., Horn L., Gaspar L.E., et al. Modernizing clinical trial eligibility criteria: recommendations of the ASCO–friends of cancer research performance status, prior therapies, and Organ Function Working Group. J Clin Oncol. 2017;35(33):3745–3751. [Google Scholar]
- 21.Chow L.Q.M., Haddad R., Gupta S., et al. Antitumor activity of pembrolizumab in biomarker-unselected patients with recurrent and/or metastatic head and neck squamous cell carcinoma: results from the phase Ib KEYNOTE-012 expansion cohort. J Clin Oncol. 2016;34(32):3838–3845. doi: 10.1200/JCO.2016.68.1478. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Facchinetti F., Mazzaschi G., Barbieri F., et al. First-line immunotherapy in non-small cell lung cancer patients with poor performance status: results of a real-world study. Lung Cancer. 2020;145:95–101. [Google Scholar]
- 23.Eisenhauer E.A., Therasse P., Bogaerts J., et al. New response evaluation criteria in solid tumours: revised RECIST guideline (version 1.1) Eur J Cancer. 2009;45(2):228–247. doi: 10.1016/j.ejca.2008.10.026. [DOI] [PubMed] [Google Scholar]
- 24.National Cancer Institute Common terminology criteria for adverse events (CTCAE), version 5.0. 2017. https://ctep.cancer.gov/protocoldevelopment/electronic_applications/docs/ctcae_v5_quick_reference_5x7.pdf Available at.
- 25.Franko J., Shi Q., Goldman C.D., et al. Treatment of colorectal peritoneal carcinomatosis with systemic chemotherapy: a pooled analysis of North Central Cancer Treatment Group phase III trials N9741 and N9841. J Clin Oncol. 2012;30(3):263–267. doi: 10.1200/JCO.2011.37.1039. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Missiaglia E., Jacobs B., D’Ario G., et al. Distal and proximal colon cancers differ in terms of molecular, pathological, and clinical features. Ann Oncol. 2014;25(10):1995–2001. doi: 10.1093/annonc/mdu275. [DOI] [PubMed] [Google Scholar]
- 27.Tumeh P.C., Hellmann M.D., Hamid O., et al. Liver metastasis and treatment outcome with anti-PD-1 monoclonal antibody in patients with melanoma and non-small cell lung cancer. Cancer Immunol Res. 2017;5(5):417–424. doi: 10.1158/2326-6066.CIR-16-0325. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Hodi F.S., Wolchok J.D., Schadendorf D., et al. Durable cancer regression and long-term benefit with nivolumab in previously treated advanced melanoma. J Clin Oncol. 2018;36(8):773–779. [Google Scholar]
- 29.Khozin S., Abernethy A.P., Nussbaum N.C., et al. Characteristics of real-world oncology practice and outcomes: the FDA Project CancerLINQ pilot. JCO Clin Cancer Inform. 2017;1:1–13. [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.