With the global rise of obesity and diabetes, MASLD has become one of the fastest-growing etiological contributors to HCC in the United States and Europe. In this context, gut microbiota dysbiosis and the resulting immunosuppressive microenvironment may synergistically increase the risk of hepatocarcinogenesis^[9]. Following intestinal barrier dysfunction, pro-inflammatory microbial products such as LPS and potentially deleterious metabolites such as TMAO can enter the liver through the portal circulation; LPS activates TLR4/NF-κB signaling, whereas TMAO contributes to oxidative and inflammatory stress^[10]. The microbiota can also influence hepatic metabolic homeostasis by regulating bile acid metabolism and intestinal hormones such as glucagon-like peptide 1 (GLP-1) and fibroblast growth factor 15/19 (FGF15/19)^[11].

SCFAs, especially acetate, propionate, and butyrate, which are the primary metabolites of dietary fiber fermentation by the intestinal microbiota, have been implicated in regulating hepatic lipid metabolism and inflammation. Experimental and translational studies suggest that SCFAs may reduce hepatic steatosis by activating G protein-coupled receptors in the liver and intestine, inhibiting lipid synthesis, enhancing fatty acid oxidation, and reducing oxidative stress and the release of inflammatory factors; depletion of SCFA-producing bacteria is commonly associated with MASLD progression^[12]. In preclinical studies, the polysaccharide extracted from Ganoderma lucidum promoted the growth of beneficial intestinal bacteria, repaired the intestinal barrier, and increased the production of SCFAs, changes that were associated with improvements in metabolic and inflammatory phenotypes, and FMT experiments provided additional support for a microbiota-dependent component^[13].

The intestinal microbiota may also affect hepatic lipid metabolism in the host, inflammation, and fibrosis by regulating bile acid metabolism. Experimental evidence indicates that reducing intestinal bacterial bile salt hydrolase (BSH) activity can alter the conjugated bile-acid pool and may mitigate selected metabolic and inflammatory abnormalities^[14]. Bile acids regulate hepatic lipid metabolism by activating the nuclear receptor farnesoid X receptor (FXR); dysbiosis in MASLD models has been reported to increase secondary bile acids such as deoxycholic acid (DCA), which may activate the TLR4-inflammasome pathway in hepatocytes through the gut-liver axis, stimulate hepatic stellate cells (HSCs) to produce fibrogenic factors, and interact with DNA damage and immune suppression to favor hepatocarcinogenesis even in non-cirrhotic settings^[15]. FXR agonists have reduced triglyceride levels in experimental settings by inhibiting key genes of hepatic lipid synthesis, such as Scd1, Lpin1 and Dgat2, and by reducing intestinal lipid absorption^[16]. These findings provide a rationale for investigating the gut microbiota-liver axis and its metabolites as potential targets for prevention research in MASLD-related HCC.

In MASLD-HCC, the broadest consensus is that metabolic dysfunction, barrier impairment, reduced SCFA-producing taxa, and altered bile acid signaling coexist and may reinforce hepatic inflammation. The central controversy is whether dysbiosis is an upstream driver, a marker of metabolic liver injury, or both, because most human data remain observational and many mechanistic claims are derived from diet- or toxin-based animal models [Table 1].

Alcohol-associated liver disease and HCC (ALD-HCC) is a major contributor to HCC and has been estimated to account for approximately 19-30% of HCC deaths globally^[17]. Thus, ALD has become a major public health concern, increasingly contributing to the global burden of disease and premature mortality, and requiring a combination of metabolic risk-factor intervention and screening of high-risk populations.

The pathogenesis of ALD-related HCC is a multifactorial, multistage process involving oxidative stress-induced DNA damage, lipid metabolism disorders, gut microbiota dysbiosis, immune-inflammatory responses, and programmed cell death pathways such as ferroptosis, pyroptosis, and necroptosis^[18]. Ethanol is oxidized to acetaldehyde in the liver, which not only forms complexes with DNA and proteins but also triggers inflammatory and fibrotic signaling; CYP2E1-mediated reactive oxygen species (ROS) release further induces lipid peroxidation and cell death^[19]. Patients with ALD, particularly severe alcohol-associated hepatitis, have been reported to exhibit reduced microbial diversity and enrichment of potentially pathogenic taxa, including cytolysin-positive Enterococcus faecalis; whether these features constitute an ALD-related HCC-specific signature remains unclear^[20]. In addition, fungal dysbiosis (the gut mycobiome), especially expansion of Candida species and exposure to fungal products such as β-glucan and candidalysin, can worsen alcohol-associated liver injury by activating pattern-recognition receptors and inflammatory pathways, providing a plausible link between ALD, barrier failure, and hepatocarcinogenesis^[20-22]. Bacterial toxins, fungal products, DCA, and other microbial molecules may enter the portal circulation, activate NOD-like receptor protein 3 (NLRP3) inflammasome signaling and hepatocyte death via TLR/NF-κB/mitogen-activated protein kinase (MAPK) signaling, and reinforce a pro-fibrotic niche^[23]. Apoptotic hepatocytes release exosomes rich in miR-155, miR-122, and miR-27a, which remodel the microbiota and strengthen M2 polarization and TGF-β/Smad-driven fibrosis, suggesting an ethanol-microbiota-exosome feedback loop that may promote fibrosis and malignant transformation^[24]. In animal models, bacteriophage therapy targeting cytolysin-positive Enterococcus faecalis reduced ethanol-induced liver injury^[25].

For ALD-HCC, there is broad agreement that alcohol, barrier disruption, bacterial and fungal products, and inflammatory-fibrotic signaling interact along the gut-liver axis to promote disease progression. However, pathogen-specific interventions such as phage therapy are supported mainly by ALD models rather than HCC trials, so their clinical relevance for preventing ALD-HCC should be interpreted with caution [Table 1].

Viral hepatitis, particularly hepatitis B virus (HBV) and hepatitis C virus (HCV) infection, is a major global cause of HCC^[26]. With broader HBV vaccination and increasingly effective antiviral therapy for HBV and HCV, the aging of chronically infected individuals and metabolic-viral dual hits are emerging as new epidemiological features of virus-associated HCC^[27].

Chronic hepatitis virus infection not only triggers local immune disorders in the liver but also indirectly affects viral replication and immune response by disrupting the intestinal microbial homeostasis. In the progression of virus-associated HCC, human studies frequently report a “pro-inflammatory, low-diversity” microbial signature, characterized by a significant reduction in protective butyrate-producing bacteria, such as Lachnospiraceae, Butyricimonas, and Faecalibacterium, and an increase in pro-inflammatory bacteria, such as Enterobacteriaceae, Veillonella, Klebsiella, and Streptococcus, accompanied by a decrease in SCFA levels, an increase in plasma LPS, and an imbalance of secondary bile acids^[28]. These changes are associated with altered peripheral T-cell composition and increased expression of inhibitory receptors, including PD-1, T-cell immunoreceptor with Ig and ITIM domains (TIGIT), and T-cell immunoglobulin and mucin domain-containing protein 3 (TIM-3); these findings are compatible with dysfunctional or exhaustion-associated T-cell states^[29,30]. In HBV-associated chronic liver disease, microbiota-associated features and bacterial products have been linked to alterations in peripheral immune responses, including T helper 1 (Th1)/Th17-related pathways^[31]. The intestinal microbiota may participate in immune and metabolic processes associated with viral persistence and liver injury, although its independent causal contribution to viral hepatocarcinogenesis remains uncertain. Targeted probiotics, FMT, or metabolite regulators are therefore candidate interventions for restoring barrier function and reshaping the immune microenvironment, providing possible avenues for precision prevention and treatment research in virus-associated HCC. However, the unresolved issue is directionality: antiviral treatment, cirrhosis severity, diet, and metabolic comorbidities may shape the microbiota, making it difficult to prove that microbial changes independently accelerate viral hepatocarcinogenesis [Table 1].

Gut microbiota-derived metabolites are increasingly implicated in cancer pathogenesis. SCFAs are core metabolites generated by bacterial fermentation of dietary fiber and mainly include acetate, propionate, and butyrate. These metabolites sustain intestinal epithelial energy supply, strengthen barrier integrity, induce regulatory T-cell (Treg) differentiation, and suppress excessive inflammation by activating G protein-coupled receptors such as GPR41 (FFAR3), GPR43 (FFAR2) and GPR109A (HCAR2), inhibiting histone deacetylases (HDACs), or acting as ligands of transcription factors^[32]. Accumulating evidence suggests that SCFAs may indirectly regulate metabolism-related HCC through the gut-liver axis; in experimental settings, butyrate has been shown to reshape tumor-cell chromatin, upregulate C-X-C motif chemokine ligand 11 (CXCL11), recruit natural killer (NK) cells and inhibit HCC progression^[33]. Song et al. reported in MASLD-HCC models that acetate produced by Bifidobacterium pseudolongum reached the liver via the portal vein, activated GPR43 in hepatocytes, blocked IL-6/Janus kinase 1 (JAK1)/signal transducer and activator of transcription 3 (STAT3) carcinogenic signaling, and inhibited both metabolic dysfunction-associated steatohepatitis (MASH) and tumor formation; GPR43 knockout or acetate depletion abolished this protective effect in that model^[34]. In MASLD, intestinal barrier damage and bile acid metabolic disorders reduce SCFA-producing bacteria, including butyrate producers such as Faecalibacterium prausnitzii and Eubacterium rectale, and propionate-associated taxa, including members of Bacteroidetes and genera such as Prevotella and Phascolarctobacterium; acetate-producing Bifidobacterium may initially increase and later decrease, but the direction and mechanisms of these changes remain uncertain^[35]. Mcbrearty et al. further found in HBV X protein (HBx) transgenic mice that oral supplementation with a physiological mixture of sodium acetate, propionate and butyrate upregulated the tumor suppressor DAB2, inhibited Ras-GTP activity, and downregulated HBx-associated pathways including phosphoinositide 3-kinase (PI3K), NF-κB, and vascular endothelial growth factor (VEGF), thereby delaying dysplastic changes and liver-tumor development in HBx-transgenic mice^[36]. Taken together, these findings suggest that SCFAs are key metabolic messengers linking the gut microbiota to HCC-prevention biology, although therapeutic translation still requires clinical validation.

Bile acid diversity results from host-microbial co-metabolism, and disruption of bile acid-microbiota homeostasis may contribute to HCC development^[37]. Shen et al. found that gut BSH-producing bacteria, such as Bifidobacteriales, were markedly depleted in HCC patients and DEN-induced hepatocarcinoma mice, resulting in a sharp decline in secondary bile acids, especially the conjugated deoxycholic acids glycodeoxycholic acid (GDCA) and taurodeoxycholic acid (TDCA); antibiotic-mediated removal of these bacteria further reduced GDCA levels and promoted tumor growth^[38]. Moreover, abnormal bile acid metabolism has been shown in experimental models to promote HCC development through Yes-associated protein (YAP) activation and FXR inhibition, forming a positive feedback loop associated with cholestasis, inflammation, and hepatocarcinogenesis; targeting the FXR-HDAC1-YAP axis can reverse this process in experimental models^[39]. TMAO, a host-microbial co-metabolite generated predominantly by hepatic flavin-containing monooxygenase 3 (FMO3) oxidation of microbiota-derived trimethylamine (TMA), has been reported to induce oxidative stress, inflammation, epithelial-mesenchymal transition (EMT), and insulin resistance by activating pathways such as MAPK, PERK-FoxO1, and NLRP3, and may promote HCC cell proliferation while inhibiting apoptosis^[40]. In HCC models, TMAO has been linked to tumor-promoting inflammatory signaling, including POSTN-associated pathways^[41].

In summary, SCFAs often appear protective, whereas dysregulated bile acids, LPS-related signaling, TMAO, and selected microbial oncometabolites may promote inflammation and tumor growth. Therefore, targeting the intestinal microbiota and its metabolites represents a promising research direction for HCC prevention and treatment. However, questions regarding dosage, timing, host conditions, and the ability of metabolite manipulation to reproduce animal-model effects in heterogeneous HCC patients still need to be resolved [Figure 1].

The role of the intestinal microbiota and its metabolites in HCC pathogenesis through the gut-liver axis has been extensively investigated in recent years. A decrease in butyrate-producing and bile acid-transforming bacteria, together with ethanol-acetaldehyde exposure in ALD and selected MASLD models, may contribute to barrier breakdown, PAMP translocation, and viable bacterial translocation^[42]. Subsequently, microbial products and metabolites reach the liver via the portal vein and act on Kupffer cells, HSCs, and recruited MDSCs. Endotoxin- or ethanol-related signals activate the TLR4-NF-κB/MAPK pathway, inducing Kupffer cells to secrete TNF-α and IL-6 and promoting fibrotic activation of HSCs^[43]. In experimental models, DCA together with selected microbial products such as lipoteichoic acid (LTA) has been linked to HSC senescence and SASP-associated immunosuppression. Activated HSCs can recruit MDSCs through the CXCL12 (SDF-1)/C-X-C chemokine receptor type 4 (CXCR4) axis in an orthotopic HCC model^[44]. Consequently, gut-derived signals may reshape the hepatic immune network, contributing to an inflammatory-fibrotic-immunosuppressive microenvironment and potentially promoting progression from cirrhosis to HCC. Across different experimental systems, SCFAs, bile-acid receptor agonists and other gut-liver-axis interventions have each been reported to affect selected barrier or immune endpoints, including Treg/Th17 balance, macrophage polarization and lymphocyte recruitment^[45,46] [Figure 2].

Immune-regulation studies converge on barrier failure, PAMP translocation, Kupffer-cell/HSC activation, myeloid-cell recruitment, and T-cell suppression as plausible links between dysbiosis and HCC. Yet most pathway-level evidence remains experimental, and clinical studies have not yet established which immune nodes represent actionable biomarkers or therapeutic targets in routine HCC care.

In recent years, the mechanisms by which intestinal microbiota and its metabolites regulate the self-renewal of liver cancer stem cells (LCSCs) through the “gut-liver axis” have attracted increasing attention. Experimental studies indicate that bile-acid and microbial-product signaling can activate HSCs and alter secretion of cytokines and mitogens, including IL-6, prostaglandin E2 (PGE2) and epiregulin, and enhance sphere formation and serial-transplantation tumorigenicity of LCSCs^[47]. Microbiota-derived genotoxins have been shown to induce DNA damage and accelerate the accumulation of mutations in tumor-initiating cells^[48]. In contrast, SCFAs such as butyrate may enhance CD8⁺ T-cell IFN-γ secretion and weaken the maintenance of stemness^[49]. In addition, urolithin A has been reported to activate silent mating type information regulation 2 homolog 1 (SIRT1)/3, block Wnt/β-catenin signaling and decrease the proportion of CD133⁺ EpCAM⁺ stem cells^[50] [Figure 2].

The microbiota-LCSC concept offers a testable hypothesis for how microbial metabolites and immune signaling might contribute to recurrence and treatment resistance, especially through bile-acid, genotoxin, and immune-mediated pathways. The current controversy is that LCSC assays and tumor-microenvironment models remain mostly preclinical, so evidence for direct microbiota control of stem-like cells in human HCC remains preliminary.

A growing body of evidence has revealed an association between the gut microbiota, HCC development, and immunotherapy response. Preclinical studies and early clinical datasets from HCC or other solid tumors suggest that probiotics, defined microbial consortia, or fecal microbiota products may influence PD-1/programmed death-ligand 1 (PD-L1) efficacy, but these interventions remain investigational^[51-56]. In HCC models and in studies of other solid tumors, taxa including Akkermansia, Bifidobacterium and Faecalibacterium have been associated with immune features such as enhanced antigen presentation and CD8⁺ T-cell activity. Conversely, enrichment of Proteobacteria and Veillonellaceae and accumulation of harmful secondary bile acids may sustain TLR4 signaling, IL-6 and TNF-α production, M2 tumor-associated macrophage (TAM) and Treg expansion, and suppression of the CXCL16-CXCR6 axis, thereby promoting immune escape. Probiotics, prebiotics, and FMT should currently be regarded as investigational adjuncts rather than validated HCC treatments [Figure 3].

Recent HCC therapy has shifted from tyrosine kinase inhibitor (TKI) monotherapy to immune-based combinations. Atezolizumab plus bevacizumab and the STRIDE regimen of tremelimumab plus durvalumab are widely used first-line options for eligible patients with advanced HCC^[57,58]. This shift makes microbiome-response research clinically more relevant: a prospective multi-kingdom study of 80 immune checkpoint inhibitor (ICI)-treated patients with HCC identified response-associated differences in bacterial, fungal and metabolite profiles, developed an 18-bacterial-species model for durable clinical benefit, and identified bacterial-metabolite signatures associated with survival^[59]. These data support fecal metagenomics, mycobiome profiling, and metabolomics as candidate companion biomarkers, while underscoring that antibiotics, diet, and liver function reserve must be controlled in future trials.

The strongest clinical signal for microbiota-treatment interactions currently comes from immunotherapy cohorts, where baseline microbial diversity and specific taxa correlate with response. Consensus is limited to association and biomarker potential; controversy remains over causality, confounding by antibiotics, diet, and liver function, and whether microbiota modulation improves ICI outcomes in HCC.

Although immune-based combinations have become the preferred first-line systemic therapies for advanced HCC, sorafenib and lenvatinib remain alternative options for selected patients who are ineligible for these regimens. Notably, most available evidence concerning interactions between the gut microbiota and targeted therapy in HCC is sorafenib-specific and predominantly preclinical; these findings should therefore not be extrapolated directly to lenvatinib or other TKIs. Prohep, a probiotic mixture, has been reported to enhance the antitumor activity of low-dose sorafenib in experimental models by increasing SCFA production, activating AMPK signaling, and suppressing PI3K/mechanistic target of rapamycin (mTOR) signaling^[60]. In vitro, sodium butyrate reduced the viability of sorafenib-resistant HCC cells, increased apoptosis, and modulated sorafenib-resistance-associated miRNAs, including miR-7641 and miR-199^[61]. However, this study did not directly demonstrate that microbiota-derived butyrate induces ferroptosis. A separate nanomedicine study used butyrate as a surface-modifying ligand to improve the intestinal absorption and tumor targeting of nanoparticles co-loaded with sorafenib and salinomycin^[62]. In this system, ferroptosis was primarily induced by the released sorafenib and salinomycin rather than by butyrate itself. More direct microbiota-related evidence came from a study showing that Enterococcus faecium-derived exopolysaccharides promoted IFN-γ-producing CD8⁺ T cells, activated JAK-STAT1 signaling, suppressed solute carrier family 7 member 11 (SLC7A11) expression, and thereby enhanced sorafenib-associated ferroptosis in experimental HCC models^[63] [Figure 3].

Targeted-therapy research should also extend beyond sorafenib. Lenvatinib, regorafenib, cabozantinib, and ramucirumab act through partly distinct VEGF, fibroblast growth factor receptor (FGFR), MET, AXL receptor tyrosine kinase (AXL) or VEGFR2-related axes, so microbiome findings from sorafenib should therefore not be generalized across targeted agents. For future studies, baseline fecal microbiota, bile acid profiles, SCFA levels, and antibiotic exposure should be prospectively collected in patients receiving TKIs alone or in TKI-ICI combinations, because diarrhea, anorexia, bile acid disruption, and systemic inflammation may both reflect and reshape the intestinal ecosystem.

TACE is a standard treatment for intermediate-stage HCC and is also used in selected advanced or combination settings. Its efficacy depends on local ischemia and chemotherapy-induced tumor killing, and may also relate to microecological remodeling along the gut-liver axis. Clinical studies suggest that TACE treatment is associated with dynamic microbiota changes in HCC patients: in the early post-procedural period (1-5 days), dysbiosis may be temporarily aggravated by ischemia-reperfusion injury and chemotherapy toxicity; however, by approximately 10 days after TACE, some studies have reported decreases in intestinal-permeability markers and circulating LPS together with increases in butyrate-producing bacteria^[64]. These observations have led to the hypothesis that post-TACE microbiota changes may be associated with attenuation of TLR4/NF-κB-related inflammatory signaling^[64]. Clinical studies have shown that baseline Limosilactobacillus reuteri and indole-3-lactic acid (ILA) levels were associated with post-TACE liver injury and clinical outcomes^[65]. TACE-TKI-ICI has shown clinical activity in unresectable HCC^[66]. However, no prospective longitudinal microbiome study has demonstrated that this combination reproducibly remodels the gut microbiota or that microbiota changes mediate treatment efficacy; these questions remain to be established. Future prospective studies could evaluate microbiota-targeted interventions alongside TACE, with prespecified safety, biological and efficacy endpoints [Figure 3].

Chemotherapy-related microbiome effects in HCC are most relevant in locoregional regimens such as TACE, hepatic arterial infusion chemotherapy, because conventional systemic cytotoxic chemotherapy has limited use in cirrhosis-associated HCC. Locoregional regimens may themselves alter intestinal and systemic physiology, making it difficult to distinguish treatment-induced microbiome changes from changes secondary to tumor response or liver injury. A 2026 prospective TACE microbiome study reported that lower Prevotella and higher Bacteroides/Faecalibacterium were associated with better progression-free survival, while higher Phascolarctobacterium or Veillonella correlated with worse overall survival^[67].

For TACE and other locoregional therapies, available studies suggest that microbiota features may track treatment response, liver injury, and prognosis. The controversy is whether observed remodeling is a cause of response, a consequence of embolization-induced tumor and liver changes, or a composite marker of host reserve.

FMT is being explored as a microbiota-modulating strategy for solid tumors and chronic liver diseases, but its role in HCC remains investigational^[74]. Compared with single-strain probiotic supplementation, FMT may alter gut-derived inflammatory signaling, including exposure to microbial products such as LPS, and may offer therapeutic potential in selected chronic liver diseases^[78]. In HCC-related models or early translational settings, FMT may improve dysbiosis and the immune microenvironment by restoring butyrate-producing bacteria, inhibiting pro-inflammatory metabolites, enhancing CD8⁺ T-cell function, regulating Treg balance, and potentially synergizing with PD-1/PD-L1 blockade^[79,80]. In animal models and preliminary clinical trials of MASLD or liver fibrosis, FMT has been reported to improve steatosis, inflammation, and fibrosis by restoring intestinal barrier function, modulating the bile acid-FXR axis, and reducing systemic inflammation, with possible heterogeneity between lean and obese MASLD populations^[81,82]. Therefore, FMT is conceptually attractive because it can restore ecological function rather than targeting one organism at a time. Nonetheless, donor variability, engraftment unpredictability, infection risk, and the lack of HCC-specific randomized data mean that FMT should remain a trial-based intervention rather than routine prevention or treatment [Figure 4].

Probiotics may remodel the intestinal microbiota and attenuate inflammatory-carcinogenic signaling, offering multi-targeted and relatively low-toxicity prospects for HCC prevention research and adjuvant-therapy development^[83,84]. Probiotics may enhance anticancer immune responses by strengthening the intestinal barrier, promoting SCFA production, and activating CD8⁺ T-cells, whereas prebiotics may indirectly improve host immune status by selectively promoting beneficial bacteria; this statement is supported mainly by review-level evidence and should not be attributed to a single HCC-specific clinical trial^[75]. Lau et al. reported that Lactobacillus acidophilus supplementation suppressed tumorigenesis in a MASLD-HCC model; produced pentanoic acid, which engaged GPR41/43 on hepatocytes, inhibited carcinogenic Rho-GTPase signaling, and induced cell-cycle arrest and apoptosis^[85]. Song et al. found that Bifidobacterium was significantly depleted in MASLD-HCC mouse models and patients, and Bifidobacterium pseudolongum supplementation suppressed tumorigenesis in this MASLD-HCC model through acetate-dependent signaling^[34]. Moreover, prebiotics, as non-digestible dietary components that selectively promote beneficial bacteria, have been shown in animal models to restore intestinal microecological balance, strengthen the intestinal barrier, reduce systemic inflammation, and inhibit the development of HCC^[86]. Probiotics and prebiotics have the practical advantage of scalability and established safety, and animal models support immune and barrier effects. The key controversies concern strain specificity, dose, durability, and whether benefits observed in MASLD-HCC models translate to patients with established cirrhosis or advanced HCC [Figure 4].

Diet is closely linked to HCC risk. Several systematic reviews and prospective studies have shown that Western dietary patterns are associated with higher HCC risk, chronic inflammation, insulin resistance and microbiota dysbiosis are plausible mediators of this association, whereas mediterranean, plant-based, high-fiber, and low-glycemic-load diets may reduce risk by improving metabolic status, enhancing intestinal barrier function, and regulating immune responses^[87]. Mechanistically, monounsaturated fatty acids (MUFAs), vitamin E, folic acid, and β-carotene may interfere with the chronic hepatitis-fibrosis-carcinogenesis axis through antioxidant activity, suppression of HSC activation, and effects on DNA methylation; conversely, dietary patterns rich in saturated fat, sugar and processed meats may increase exposure to N-nitroso compounds (NOCs) and may influence host insulin-like growth factor 1 (IGF-1) signaling, thus may increase carcinogenesis risk^[88]. Experimentally, mice fed a high-fat, high-cholesterol diet show increased abundance of mucin-producing genera, Desulfovibrio, and Anaerostipes, reduced Deltaproteobacteria or Bacteroides, and increased hepatic lipid accumulation, inflammation, and cell proliferation^[89]. In experimental settings, phytic acid, a component of high-fiber foods such as soybeans and nuts, ameliorated these changes by reducing hepatic lipid synthesis via SREBP-1c/FAS inhibition, suppressing NF-κB inflammatory signaling, and limiting harmful bacteria^[90]. Similarly, dietary carbohydrate restriction (DCR) improved the hepatic metabolic-inflammatory environment and restored the microbiota-bile acid-immune axis, and was associated with inhibition of HCC progression in experimental settings^[91] [Figure 4].

Clinical intervention evidence is also beginning to extend beyond animal models. A 2026 single-center randomized trial evaluated a multicomponent nutritional and microbiota-directed intervention comprising high-fiber dietary advice, probiotics, prebiotics and increased unsaturated-fat intake; the bundled design precludes attribution of clinical effects to any single component^[92].

Dietary intervention has the strongest population-level plausibility because diet simultaneously influences obesity, diabetes, inflammation, and microbial metabolism. However, clinical evidence in HCC remains difficult to interpret because bundled lifestyle interventions, treatment adherence, cancer stage, and concurrent therapy can confound microbiota-mediated effects.

Changes in metabolites during liver disease progression suggest potential targets for small-molecule drug development. For example, interventions that inhibit microbial TMA formation or pharmacologically activate FXR have shown metabolic improvement or antitumorigenic effects in animal models of MASLD and liver fibrosis, although HCC-specific clinical evidence remains limited^[93,94]. Butyrate has been reported to reduce the expression of lipid-synthesis enzymes such as fatty acid synthase (FASN), acetyl-CoA carboxylase (ACC), ATP-citrate lyase (ACLY), and stearoyl-CoA desaturase (SCD) by inhibiting the PI3K/Akt/mTOR-SREBP-1 axis, thereby reducing hepatic lipid deposition and inflammatory responses; small-molecule drugs targeting the lipid-metabolism axis, such as the FASN inhibitor TVB-2640 and ACC inhibitor ND-654, have shown anti-HCC potential in animal models or in early clinical development^[95]. In addition, Zhang et al. found that the tumor and fecal samples from HCC patients were enriched with Catenibacterium mitsuokai, which was reported to translocate to the liver and adhere to HCC cells through an interaction between its surface protein Gtr1/RagA and host-cell γ-catenin, while secreted QA bound TEK receptor tyrosine kinase (TIE2) and activated oncogenic PI3K/AKT signaling. TIE2 inhibitors blocked the QA-mediated tumor-promoting effect in vivo, suggesting that the microbiota-QA-TIE2 axis is a candidate actionable metabolic-signaling pathway^[76]. Metabolite-targeted therapy offers a more druggable route to microbiota translation by focusing on defined enzymes or receptors. However, many metabolites exert dual, context-dependent effects, and most candidate pathways still require pharmacodynamic biomarkers and HCC-specific clinical testing [Figure 4].

Bacteriophages, an important component of the intestinal virome, have shown experimental value as precise microbiome-editing tools in liver disease models. The proposed strategy is to eliminate cancer-promoting or liver-injury-promoting pathogens and thereby potentially interrupt gut-liver axis-driven malignant transformation^[96,97]. Duan et al. found that Enterococcus faecalis is significantly enriched in the intestinal tract of patients with alcoholic hepatitis and that its secreted cytolysin worsens ethanol-induced liver injury; oral phage cocktail therapy targeting this bacterium reduced its abundance and attenuated liver disease progression in mice^[25]. Gan et al. showed that phages targeting high-alcohol-producing Klebsiella pneumoniae inhibited MASLD progression in mice by blocking endogenous ethanol production and reversing hepatic steatosis, inflammation, and fibrotic cascades^[98]. Additional translational rationale comes from studies of other inflammatory diseases showing that designed phage consortia can selectively suppress target commensals, but comparable HCC-specific clinical evidence remains unavailable^[99]. Bacteriophage therapy may offer greater taxonomic specificity than broad-spectrum antibiotics for selected microbial targets, although resistance, delivery and host immune clearance remain major limitations^[100,101]. Phage therapy illustrates the potential for precision microbiome editing, especially when a defined harmful bacterium or toxin is present. The major uncertainty is whether HCC has sufficiently stable pathogen targets for phage design, and whether resistance, delivery, and immune clearance can be managed safely in cirrhosis or cancer therapy settings [Figure 4].