Gut microbiome-mediated primary and acquired resistance to immune checkpoint inhibitors in MSI-H/dMMR colorectal cancer: narrative review
A narrative review maps mechanisms, biomarkers, and therapeutic strategies through which the gut microbiome may mediate resistance to ICIs in MSI-H/dMMR colorectal cancer, without generating its own quantitative effect estimates.
| Population | Adults with dMMR/MSI-H colorectal cancer treated with immune checkpoint inhibitors |
|---|---|
| Intervention | Gut microbiome composition and function as a biological determinant of response/resistance |
| Comparator | Not applicable (narrative review without formal comparator) |
| Outcome | Microbiome-mediated primary resistance to ICIs; Microbiome-mediated acquired resistance to ICIs; Microbial metabolite modulation of tumor microenvironment; Microbial predictive biomarkers of ICI response in dMMR/MSI-H CRC; FMT efficacy to overcome ICI resistance in CRC; Fusobacterium nucleatum association with worse ICI response in CRC; Short-chain fatty acid impact on antitumor immune response |
Summary of findings
| Outcome | Effect | 95% CI | Certainty | Clinical relevance | Notes |
|---|---|---|---|---|---|
| Microbiome-mediated primary resistance to ICIs | no pooled estimate available; narrative synthesis only | — | Very low | — | |
| Microbiome-mediated acquired resistance to ICIs | no pooled estimate available; narrative synthesis only | — | Very low | — | |
| Microbial metabolite modulation of tumor microenvironment | no pooled estimate available; preclinical and extrapolated data only | — | Very low | — | |
| Microbial predictive biomarkers of ICI response in dMMR/MSI-H CRC | no validated clinical biomarker identified; in the quantitative estimate | — | Very low | — | |
| FMT efficacy to overcome ICI resistance in CRC | no RCT data in dMMR/MSI-H CRC; ongoing trials only | — | Very low | — | |
| Fusobacterium nucleatum association with worse ICI response in CRC | directional association reported in preclinical/translational studies; in the HR/OR with 95% CI from this review | — | Very low | — | |
| Short-chain fatty acid impact on antitumor immune response | mechanistic evidence from preclinical models; in the clinical effect estimate in dMMR/MSI-H CRC | — | Very low | — |
Context
Up to one-third of dMMR/MSI-H CRC patients show primary resistance to PD-1 inhibitors, and a substantial proportion of initial responders develop acquired resistance. Established molecular biomarkers (dMMR, TMB, POLE/POLD1) do not fully explain response heterogeneity. The gut microbiome has emerged as a modulator of the immune axis with potential translational impact.
What the study showed
The review synthesizes evidence that microbial communities (e.g., Fusobacterium nucleatum, enterotoxigenic Bacteroides fragilis, pks+ E. coli) and metabolites (short-chain fatty acids, bile acids, inosine, tryptophan, polyamines) modulate T-cell activation, antigen presentation, and the tumor microenvironment. Most quantitative evidence was derived from preclinical models or non-CRC solid tumors (melanoma, NSCLC, RCC); dMMR/MSI-H CRC-specific data are scarce. No effect estimators (RR, OR, HR, 95% CI) were generated by this study.
How it was done
Structured narrative review with searches in PubMed/MEDLINE, Scopus, and Web of Science, supplemented by ClinicalTrials.gov. English-language literature through May 2026 was considered. Inclusion criteria prioritized RCTs, prospective observational studies, and translational investigations; no formal data extraction or standardized risk-of-bias assessment was performed.
Effect magnitude
No pooled effect estimate was produced; the study does not permit calculation of effect size, 95% CI, or any quantitative metric of microbiome-response association in dMMR/MSI-H CRC.
Risk of bias
Narrative review without pre-registered protocol or formal risk-of-bias tool (AMSTAR-2 not applied), susceptible to selection and confirmation bias in study inclusion. The majority of mechanistic evidence derives from preclinical models or other solid tumors, with extrapolation to MSI-H CRC not validated in robust clinical trials. Methodological heterogeneity precludes quantitative synthesis.
What this study does NOT prove
This study does not prove causality between any microbial taxon or metabolite and ICI resistance in dMMR/MSI-H CRC. It also does not validate microbial biomarkers for clinical use or demonstrate benefit from any microbiome intervention in this population.
In clinical practice
No change in clinical practice is supported by this review. Microbiome-ICI associations should be interpreted as hypothesis-generating, not as indications for therapeutic microbiome modulation outside experimental settings. Strategies such as fecal microbiota transplantation and probiotics to enhance ICIs in MSI-H CRC remain under clinical investigation.
Limitations
Narrative review without pre-registered protocol or formal risk-of-bias tool (AMSTAR-2 not applied), susceptible to selection and confirmation bias in study inclusion. The majority of mechanistic evidence derives from preclinical models or other solid tumors, with extrapolation to MSI-H CRC not validated in robust clinical trials. Methodological heterogeneity precludes quantitative synthesis.
What is still missing
Prospective RCTs with standardized microbiome characterization (shotgun metagenomics) in exclusively dMMR/MSI-H CRC cohorts are needed to validate biomarkers and test microbiome interventions in combination with ICIs.
Technical appendix
Version history
- 1.0 · 2026-10-13 — Auto-generated under Evidence Standard v1.0