FMT combined with ICIs
While the application of ICIs has revolutionized the pharmacological treatment paradigm for GI cancers, their efficacy is universally challenged by primary, adaptive, and acquired resistance. This resistance manifests as a complex, multifactorial, and multilayered network. At the intrinsic tumor-cell level, tumors directly evade T-cell recognition through mechanisms such as impaired antigen presentation and upregulation of multiple immune checkpoints.96 Concurrently, mutations and persistent activation of diverse oncogenic signaling pathways reinforce the immune evasion capacity of tumor cells. Within the TME, at the cellular level, extensive infiltration of immunosuppressive cells, including Tregs, M2 macrophages, and suppressive neutrophils, forms a robust inhibitory network.97 At the functional level, the TME undergoes profound metabolic reprogramming, characterized by sustained hypoxia, lactate accumulation, and depletion of critical nutrients such as glucose, directly impairing T-cell function and inducing exhaustion.98 Simultaneously, networks of inhibitory cytokines such as TGF-β further undermine antitumor immunity. Critically, these mechanisms do not operate in isolation but are embedded within and modulated by broader host systemic factors. Among these, gut microbiota dysbiosis plays a particularly important role, driving systemic immune dysfunction. Beyond the microbiome, diet, environmental exposures, and lifestyle collectively constitute the macro-level context that ultimately shapes the immunotherapy response.99
FMT may reverse ICI resistance by reshaping the gut microbiota, thereby modulating immune responses and restoring therapeutic efficacy. Preclinical studies indicate that specific gut microbial communities or FMT from ICI responders can induce tumor regression, potentiate T-cell responses, and enhance the antitumor efficacy of ICIs. In preclinical CRC models, FMT from microsatellite-stable (MSS) CRC patients with high Fusobacterium nucleatum (Fn) abundance or from anti-PD-1-responsive mice conferred sensitivity to anti-PD-1 therapy, whereas FMT from low-Fn or non-responsive donors proved ineffective.100,101 Furthermore, oral administration of Bifidobacterium catenulatum potentiated anti-PD-1 efficacy; this effect was transferable by FMT and accompanied by enhanced intratumoral CD8+ T-cell infiltration.102 Currently, numerous clinical studies have evaluated FMT combined with ICI therapy across different cancer types, though efficacy data from large-scale randomized trials remain lacking.103 Melanoma has been one of the earliest cancer types investigated in FMT–ICI combination research, with early studies showing preliminary feasibility and safety of FMT in overcoming ICI resistance. Two single-arm trials provided preliminary clinical signals suggesting that responder-derived FMT combined with anti-PD-1 therapy resensitized a subset of refractory melanoma patients to immunotherapy.27,104 Routy et al.64 further reported in a phase I trial that FMT from healthy donors combined with anti-PD-1 therapy achieved an objective response rate (ORR) of 65% (13 of 20) in previously untreated patients with advanced melanoma, with an unexpected median overall survival (mOS) of 52.8 months compared with RCTs and real-world data (30.0–39.6 months mOS).105 Although these findings require validation in larger controlled trials and safety considerations warrant ongoing attention, they offer valuable exploratory insights that may inform future investigation of FMT in GI tumors, where clinical evidence remains emerging.
FMT combined with ICI therapy is increasingly being investigated in GI tumors, with preliminary clinical signals suggesting potential value, particularly in ICI-resistant patients. While the current evidence is largely derived from early-phase trials and small-cohort studies, clinical signals across diverse GI cancer subtypes, such as colorectal, gastric, esophageal, and hepatocellular carcinoma, provide preliminary support for the therapeutic potential of FMT in combination with anti-PD-1 therapy with or without anti-angiogenic agents. These studies report disease control rates (DCRs) generally ranging from 40% to 95%, with some patients achieving marked tumor regression or durable long-term disease stability. In a study by Kim et al.,53 13 patients with metastatic GI cancers received responder-derived FMT combined with continued nivolumab treatment, with an ORR of 7.7% and a DCR of 46.2%. Four cases of ESCC and one of HCC achieved stable disease, while one HCC patient achieved a partial response (PR) with a maximum tumor reduction of 47.7% and a progression-free survival (PFS) of 8.7 months.53 In a phase I trial reported by Zhang et al.,106 healthy-donor FMT combined with nivolumab was evaluated in patients with anti-PD-1-refractory MSS GI cancer (n = 10). The study demonstrated an ORR of 20% overall and a DCR of 40%, while the ORR was 25% and the DCR was 50% when restricted to the eight gastric cancer patients, with two patients achieving PR. Responses were accompanied by donor microbial engraftment and systemic immune activation.106 In a phase II trial (RENMIN-215) involving 20 patients with MSS metastatic CRC, the combination of FMT, tislelizumab, and fruquintinib as third-line or later therapy achieved an ORR of 20%, a DCR of 95%, and an mOS of 13.7 months. A post hoc subgroup analysis suggested that patients without liver metastases might derive more pronounced benefit, with median PFS not reached.107 Furthermore, Cheng et al.108 reported a case of a patient with proficient mismatch repair and MSS stage IVB colon cancer who, after failing multiple lines of chemotherapy and targeted therapy, received a chemotherapy-free triple regimen comprising tislelizumab, bevacizumab, and fecal microbiota capsules. The tumor regressed markedly, enabling surgical resection, and postoperative pathology confirmed a pathological complete response.108 These early-phase findings provide hypothesis-generating evidence for FMT–ICI combinations in GI cancers, with DCRs generally ranging from 40% to 95% across heterogeneous study designs and patient populations. Nevertheless, the available studies are limited by small sample sizes, lack of randomization, insufficient follow-up data, and inadequate safety monitoring for FMT-specific risks. Furthermore, the current literature is heavily weighted toward positive findings, with limited reporting of null results or cases of disease progression following FMT. Several ongoing clinical trials are further investigating the efficacy of FMT combined with ICIs in GI cancers (NCT05750030; NCT05690048; NCT05273255; NCT04729322).
After FMT–ICI combination treatment, the gut microbiota of GI cancer patients may undergo remodeling, especially in CRC. In preclinical studies, multi-omics analysis of CRC mouse models suggested that FMT combined with anti-PD-1 therapy enriched beneficial bacteria such as Bacteroides thetaiotaomicron and B. fragilis, while suppressing potentially detrimental species like Bacteroides ovatus and Lactobacillus murinus. These alterations were accompanied by shifts in microbial functional profiles and changes in the host plasma metabolome, though their direct translational relevance to human disease requires further validation.109 At the phylum level, Zhao et al.107 found that the proportion of Proteobacteria was significantly higher in responders, while Actinobacteriota and the genus Bifidobacterium were more abundant in non-responders. At the family and genus levels, Lachnospiraceae emerged as a critical beneficial group, and its genera Roseburia and Lachnospira were significantly increased in responders. Peng et al.110 also observed that after treatment, patients’ gut microbiota composition shifted toward that of their corresponding healthy donors, and those who achieved clinical benefit showed significantly higher similarity to the donor. Through exploratory analysis, a group of species potentially associated with clinical benefit were identified in responders, such as Bacteroides coprocola, Bacteroides stercoris, and Parabacteroides goldsteinii, which were hypothesized to have originated from successful donor colonization.106
FMT has been proposed to potentially reduce the incidence of irAEs, though this association requires further validation in larger prospective studies. In available early-phase trials and retrospective analyses, most irAEs or treatment-related adverse events (AEs) associated with FMT–ICI combination were grade 1 or 2, primarily involving the GI tract and presenting as nausea and constipation.53,106 However, it should be noted that FMT itself carries inherent safety risks, including pathogen transmission, bacteremia, and potential immune dysregulation, particularly in immunocompromised oncology patients. ICI-associated colitis is a common and potentially severe adverse reaction, typically presenting as diarrhea but capable of progressing to fever, hematochezia, bowel obstruction, megacolon, peritonitis, intestinal perforation, and death.111 Emerging evidence suggests that FMT may represent a promising, though not yet established, treatment approach for ICI-associated colitis. Three case reports suggest that FMT is associated with symptom alleviation, restoration of mucosal integrity, and remodeling of gut microbial composition and local immune homeostasis.112,113 Additionally, two larger clinical analyses reported similarly notable clinical remission rates.114,115
FMT combined with traditional therapies
FMT combined with chemotherapy
Preclinical and emerging clinical evidence suggests that chemotherapeutic agents can rapidly reduce tumor volume and remodel the TME by inducing immunogenic cell death, promoting CTL infiltration, and depleting or reprogramming immunosuppressive cells.116 The immunological effects of chemotherapy depend on the drug type, dosage, and administration regimen. High-dose chemotherapy may suppress immune effector cells such as CD8+ T cells and NK cells, whereas low-dose chemotherapy has shown the potential to enhance immune responses by reducing the number of MDSCs and Tregs, although it is often insufficient to achieve complete responses alone.117 While inhibiting the growth, division, or DNA replication of cancer cells, chemotherapy also affects normal cells that are actively dividing, such as those in the bone marrow, GI tract, and hair follicles, leading to adverse effects such as nausea, vomiting, alopecia, myelosuppression, and immunosuppression. Prior studies have reported that chemotherapy can also damage the intestinal epithelium, disrupt microbiota balance, and impair immune homeostasis.118
The interaction between the gut microbiota and chemotherapy is bidirectional. Gut microbiota influence drugs through multiple processes, including bacterial translocation, reduced diversity, drug deactivation and reactivation, immune modulation, and biotransformation.119 These mechanisms may act in concert to ultimately shape therapeutic efficacy and toxicity by altering drug activity, the TME, and host metabolic status. For instance, the antitumor efficacy of cyclophosphamide (CTX) observed in preclinical mouse models is potentially mediated by the activities of specific gut commensal bacteria. CTX promotes translocation of Gram-positive bacteria such as Enterococcus hirae, which is associated with induction of Th1/Th17 immune responses.120 CTX also appears to facilitate the accumulation of Gram-negative bacteria such as Barnesiella intestinihominis in the colon.121 Notably, these beneficial effects are markedly attenuated when the gut microbiota is disrupted by antibiotics. Similarly, the efficacy of platinum-based drugs such as oxaliplatin may also depend on the gut microbiota. In mice with disrupted microbiota, treatment failure was associated with a blunted ROS response by tumor-infiltrating myeloid cells and suggested a correlation between specific bacterial genera such as Alistipes and improved therapeutic outcomes.122 More recently, Xu et al.123 further showed that Akkermansia muciniphila may enhance oxaliplatin sensitivity in gastric cancer via its metabolite pentadecanoic acid, which putatively targets FUBP1 to inhibit tumor glycolysis. In pancreatic ductal adenocarcinoma mouse models, the microbiota-derived metabolite indole-3-acetic acid (3-IAA), produced by Bacteroides species, potentially enhanced FIRINOX or gemcitabine/nab-paclitaxel chemotherapy through ROS accumulation and autophagy downregulation in cancer cells. The same study further observed that in two patient cohorts, higher serum 3-IAA correlated with better response and longer survival, though the mechanism in humans remains to be confirmed.124 However, the gut microbiota may also contribute to potentiating the toxicity of chemotherapeutic agents. Wallace et al.125 provided preclinical evidence that irinotecan-induced diarrhea is mechanistically driven by bacterial β-glucuronidase, which reactivates the glucuronide conjugate SN-38G back to the active metabolite SN-38, thereby directly damaging the intestinal mucosa. Sun et al.126 further found that the ratio of SN-38 to SN-38G in intestinal tissue correlates most closely with diarrhea severity. Furthermore, certain bacterial species directly undermine therapeutic efficacy by activating resistance pathways in cancer cells. Bacteroides fragilis promotes resistance to 5-fluorouracil and oxaliplatin in CRC mice by directly binding to the Notch1 receptor on cancer cells via its surface proteins, which in turn activates Notch1 signaling and induces epithelial-mesenchymal transition and stemness.127 Collectively, these findings suggest that the gut microbiota may function as a contributor to chemotherapy outcomes, influencing both efficacy and toxicity through distinct microbial mechanisms. While causal relationships in humans remain largely unestablished, the microbiota-dependent nature of these processes raises the possibility that modulating microbial composition could represent one avenue to improve therapeutic responses, a hypothesis that has prompted investigation into microbiome-directed strategies, including FMT.
Preclinical studies have provided initial evidence that FMT may both enhance chemotherapy efficacy and alleviate chemotherapy-induced toxicity through modulation of gut microbiota composition, restoration of intestinal barrier integrity, and attenuation of pro-inflammatory signaling. In rodent models of chemotherapy-induced intestinal injury, FMT has demonstrated beneficial effects. Le Bastard et al.128 showed that FMT after 5-fluorouracil and antibiotics restored microbial diversity and butyrate-producing commensals, recovered metabolic functions, and suppressed pathogen outgrowth. Wardill et al.,129 using methotrexate, further delineated that autologous pre-treatment FMT, but not post-chemotherapy FMT, reduced diarrhea, with benefit linked to Muribaculaceae engraftment and mucosal recovery. Li et al.130 also found that healthy-donor FMT alleviated weight loss and colon shortening in 5-FU-treated mice, while FMT from 5-FU-treated donors transferred these injuries to vancomycin-pretreated recipients. These findings were further corroborated in tumor-bearing CRC models, where FMT not only alleviated chemotherapy-induced mucosal injury but also enhanced antitumor efficacy. Arshad et al.131 showed that FMT synergized with capecitabine to reverse microbial dysbiosis and potentiate antitumor immunity. Similarly, Chang et al.77 demonstrated that FMT safely mitigated FOLFOX-induced intestinal mucositis and diarrhea without compromising the antitumor efficacy of chemotherapy, while Unrug-Bielawska et al.132 reported that FMT might potentiate FOLFOX’s antitumor activity in specific CRC patient-derived xenograft models. Moreover, several animal studies have shown that FMT can transmit intestinal mucosal protection or chemotherapy sensitization conferred by gut microbiota-modulating agents to recipient animals.133-136 Whether these findings translate into clinical benefit remains uncertain. A clinical study of 62 CRC patients reported that FMT alleviated refractory FOLFIRI-induced diarrhea, along with reduced markers of intestinal permeability.137 In a randomized, double-blind, phase II trial involving 24 cachectic patients with advanced gastroesophageal cancer, administration of allogenic FMT from healthy obese donors prior to first-line chemotherapy significantly improved the DCR (83% vs. 42%) and showed a trend toward improved mOS and PFS compared with autologous FMT, although the primary endpoint of improving cachexia was not met.138 Given the small sample sizes and limited cancer types investigated so far, larger-scale studies are needed to replicate these findings. Several ongoing clinical trials include ChiCTR2400094513, NCT06405113, and NCT06346093.
FMT combined with radiotherapy
Radiotherapy, as one of the cornerstone treatments for malignant tumors, is administered to approximately 50% of cancer patients during their disease course, with about 40% achieving cure through this modality.139 It directly kills tumor cells and inhibits their proliferation by inducing DNA damage while exerting dual immunomodulatory effects. On one hand, radiotherapy activates antitumor immune responses through immunogenic cell death, tumor antigen release, and enhanced antigen presentation, remodeling the TME and potentially mediating the abscopal effect on distant metastases, although the reverse abscopal effect may also occur via independent pathways.140 On the other hand, it may drive immunosuppression by impairing immune cells, expanding MDSCs and Tregs, upregulating TGF-β expression, and reducing CD8+ T-cell infiltration, thereby undermining treatment durability.141 The cytotoxic effects of radiotherapy inevitably extend to surrounding healthy tissues, particularly rapidly proliferating epithelia such as those of the GI tract, oral mucosa, and skin. Approximately 90% of patients receiving abdominal, pelvic, or rectal radiotherapy develop RE.142 RE is characterized by acute intestinal mucositis and chronic fibrosis affecting the small intestine and rectum, with clinical manifestations including abdominal pain, diarrhea, and hematochezia, significantly impairing quality of life and potentially necessitating treatment discontinuation.143
A bidirectional interaction exists between the gut microbiota and radiotherapy. Ionizing radiation rapidly and dose-dependently alters the composition and function of the microbiota and may even directly damage microbial DNA or induce horizontal gene transfer among gut microbes.144,145 A preclinical study also reported that radiation may reshape the gut virome. Dysregulation of the viral community can lead to excessive activation of RIG-I and Notch signaling pathways, further impairing intestinal stem cell regeneration and differentiation, and potentially contributing to exacerbated intestinal injury in mouse models.146 While ionizing radiation alters the composition and function of the gut microbiota, the microbiota in turn modulates the efficacy and toxicity of treatment. A healthy gut microbiota can play a key protective role in mitigating radiotherapy-induced toxicity by modulating host immune and oxidative stress responses. Dysbiosis may weaken radiotherapy responses, amplify radiotherapy-induced toxicity, and even indirectly increase the risk of tumor progression or metastasis. A prospective study of 172 patients with esophageal carcinoma showed that pathological complete responders maintained stable gut microbial diversity during neoadjuvant chemoradiotherapy, whereas diversity significantly declined in non-responders. Additionally, lower pre-surgery diversity independently predicted worse PFS.147 Neoadjuvant concurrent chemoradiotherapy (nCRT) is the standard treatment for locally advanced rectal cancer (LARC), yet both therapeutic response and toxicity vary considerably among individuals. Two small exploratory studies have provided initial evidence of compositional and dynamic changes in the gut microbiome during nCRT. At baseline, Clostridium sensu stricto 1 and Shuttleworthia were enriched in responders, whereas genera such as Murimonas and Faecalibacterium were more abundant in non-responders.148,149 During nCRT, microbial diversity showed a declining trend in patients with poor response, and a longitudinal decrease in Intestinimonas was predictive of a good pathological response.149 In addition, Shi et al.148 observed that genera including Bifidobacterium and Clostridia were more abundant in patients with less severe diarrhea. A larger study of 126 LARC patients found that non-responders showed a significant post-treatment enrichment of Bacteroides vulgatus and enhanced microbial nucleotide biosynthesis activity in feces and tumor tissues. The study further demonstrated in preclinical models that B. vulgatus-derived nucleosides can be taken up by tumor cells via nucleoside transporters and enhance DNA damage repair, thereby directly contributing to chemoradiotherapy resistance.150 Huang et al.151 further found in mouse models that Bacteroides enrichment may exacerbate radiation proctitis by depleting colonic NAD+ and impairing mucosal proliferative capacity, thereby amplifying normal tissue injury.
The combination of radiotherapy and FMT may represent a strategy to potentiate antitumor immunity, mitigate radiation-induced intestinal injury, and restore gut microbial homeostasis, thereby improving both the efficacy and tolerability of radiotherapy. Studies in murine models have shown that the protective effects of low-intensity exercise against radiation-induced intestinal injury can be transferred to recipient mice via FMT. In the same work, direct administration of Akkermansia muciniphila similarly reduced radiation-induced gut toxicity.152 Li et al.153 reported that transplanting gut microbiota from radiotherapy responders into antibiotic-treated HCC mouse models via FMT restored the antitumor efficacy of radiotherapy through the cGAS–STING signaling pathway, in which microbiota-derived c-di-AMP and radiotherapy-induced dsDNA from damaged tumor cells synergistically triggered IFN-β production and CTL activation. FMT also reshapes gut microbiota structure in irradiated mice, improves GI tract function and intestinal epithelial integrity, and thereby significantly increases survival rates and alleviates radiation-induced GI toxicity.79 However, direct clinical investigation of FMT combined with radiotherapy in GI cancers remains virtually absent, and the translational relevance of these preclinical findings has yet to be established in controlled human studies.
Preliminary clinical evidence suggests a potential reparative role of FMT in radiation-induced intestinal injury, though current data remain limited and largely derived from gynecological cancer patients. Three clinical case studies involving cervical cancer patients provided initial evidence for the efficacy of FMT. Wang et al.154 and Zheng et al.155 reported that multi-course FMT significantly alleviated the clinical symptoms of chronic RE and improved long-term outcomes, while the gut microbiota profiles of these patients were shifted toward those of healthy donors. Furthermore, a case report confirmed that FMT with the patient’s son as the donor effectively alleviated symptoms in a patient with recurrent RE 18 years after radiotherapy, underscoring the therapeutic potential of FMT for RE at different disease stages.156 In addition, Ding et al.157 administered washed microbiota transplantation to five patients with refractory chronic RE; three achieved a clinical response at 8 weeks post-transplantation, with improvement in intestinal symptoms and mucosal injury, and no severe treatment-related AEs were reported. Microbiological analysis showed that the gut microbiota profiles of responders shifted toward the donor lineage, accompanied by enriched beneficial taxa such as Phascolarctobacterium, Lachnoclostridium, and Blautia. Furthermore, Cui et al.158 conducted a prospective cohort study in 45 patients with RE complicated by intestinal obstruction. Compared with conventional treatment, perioperative FMT combined with nutritional support significantly shortened the time to postoperative GI function recovery and hospital stay, and effectively reduced the incidence of postoperative inflammatory obstruction. More importantly, this combined regimen significantly improved postoperative nutritional parameters and GI quality-of-life scores, though the study design and sample size again limit the strength of conclusions that can be drawn.158
HDMX-based donor screening strategy for XenoFMT
In the clinical application of FMT in GI oncology, donor selection has largely been restricted to safety-based screening criteria with taxonomic characterization, encompassing pathogen exclusion and infectious disease testing, with limited systematic attention to microbiota characteristics potentially relevant to therapeutic efficacy. In the ICI setting, two primary donor sources have been utilized in studies, namely healthy donors and treatment-responding donors who have achieved a complete or PR to ICI, with the former being more commonly reported. In GI cancers, post hoc analyses of donors associated with clinical benefit in FMT–ICI combination therapy suggest that therapeutic efficacy may depend on the enrichment of immunostimulatory bacterial strains and the exclusion of immunosuppressive strains,53,107 and may also be linked to the establishment of functionally cooperative microbial consortia.106 However, donor analyses to date have been largely retrospective and lack prospective application. Given interstudy heterogeneity and the absence of RCTs, determination of the optimal donor source remains unresolved. In clinical trials of FMT combined with chemotherapy or radiotherapy, even retrospective profiling is absent, with available studies focused exclusively on toxicity mitigation and symptomatic management,137,138,156-158 with no characterization of donor microbiota in relation to response. A preclinical study in CRC patient-derived xenograft models further suggested that identical healthy-donor FMT enhanced FOLFOX efficacy in only two of four tumor models, with differential responses potentially attributable to specific host–microbiota–tumor interactions rather than the overall extent of microbiota restructuring, supporting the idea that compositional profiling alone may be insufficient to predict chemotherapy sensitization across distinct tumor contexts.132 Current FMT-oncology clinical trials have yet to incorporate functional donor characterization as a design criterion. Few strategies incorporate functional evaluation of how the donor microbiota, as an integrated community, modulates host antitumor responses. Moreover, the efficacy of microbiome-based interventions varies substantially among individuals. Even strains within the same bacterial species may produce divergent outcomes due to genomic variation.34,159 In addition, microbiota-mediated effects are profoundly context-dependent, shaped by tumor type, treatment modality, and host immune status, rendering static taxonomic profiling insufficient as a basis for donor selection.
To address this limitation, we propose an integrated preclinical-to-clinical framework comprising two sequential components (Fig. 2). The first is the HDMX model, a preclinical platform in which recipient mice undergo antibiotic-mediated microbiota depletion, followed by transplantation of microbiota from individual donors within a prospectively assembled donor pool. After tumor inoculation and immunotherapy administration, this system evaluates the ability of each donor microbiota to modulate therapeutic efficacy and host immune responses in a disease-relevant context.
Although illustrated here in the setting of tumor immunotherapy, the HDMX framework is broadly applicable to a range of microbiota-associated diseases, including inflammatory bowel disease, irritable bowel syndrome, metabolic disorders such as type 2 diabetes and obesity, and neuropsychiatric conditions such as depression and Alzheimer’s disease. In these contexts, the sequence of microbiota transplantation and disease induction can be adjusted according to disease-specific experimental designs.
The second component is XenoFMT, in which fecal microbiota preparations from donors functionally validated through the HDMX platform are administered to eligible patients. Unlike conventional approaches that rely primarily on safety screening and taxonomic profiling, XenoFMT incorporates functional preclinical validation as a prerequisite for donor selection. Together, the HDMX–XenoFMT pipeline proposes an integrated framework linking functional donor evaluation in a standardized preclinical system with personalized microbiota delivery in the clinical setting, with the potential to inform precision microbiome-based therapeutic strategies.