The gastrointestinal tract hosts a rich ecosystem of bacteria, fungi, viruses, archaea, and parasites that colonizes the host early in life and co-evolves through adulthood. While these communities perform essential functions in nutrient utilization and homeostasis, they become dysregulated in disease. Because gut microbes are in intimate, continuous contact with the epithelial lining of the digestive tract, they are heavily implicated in the pathogenesis of cancers arising there. Mechanistic studies have revealed microbes as an integral component of the tumor microenvironment, interacting with tumor cells, immune cells, and stromal cells through microbial protein-host receptor binding, metabolite exchange, and modulation of antitumor immunity. The field has now moved beyond simple microbial profiling toward strain-specific functional characterization, spatial mapping of intratumoral bacteria, and translational applications.

Colorectal cancer, the third most common cancer worldwide and the second leading cause of cancer death, was among the earliest malignancies linked to the gut microbiome, and recent work has delivered striking surprises. Whereas most studies previously reported associations at the species level, researchers are now uncovering dramatic strain-to-strain variation in cancer-promoting effects. A comprehensive analysis of Fusobacterium nucleatum strains isolated from colorectal tumors and healthy oral cavities, using PacBio long-read sequencing, revealed that the subspecies F. nucleatum sub-species animalis is predominantly enriched in colorectal cancer. This subspecies splits into two clades, and only one, Fna C2, is enriched in tumors. Fna C2 expresses the adhesin fap2, which aids its adherence to and invasion of the colon, and its superior protumorigenic effect was validated in mouse models.

Similar precision has been applied to pks-positive Escherichia coli, a bacterium carrying the genetic island needed to synthesize the genotoxin colibactin. Comparing the non-pathogenic Nissle 1917 strain with the pathogenic 11G5 strain, researchers found that although both harbor the pks island, only 11G5 drove tumorigenesis in mice. The difference lay in a variant of the FimH adhesin, which allows 11G5 to bind colon epithelial cells and inflict DNA damage. Critically, targeting this bacterium-host cell interaction prevented 11G5-induced tumorigenesis, pointing to a therapeutic strategy. Fungi are also entering the picture: Aspergillus rambellii promotes colorectal cancer growth in vitro and in vivo, and the commensal yeast Candida albicans has been shown to synergize with F. nucleatum, ferrying the bacterium to the colonic mucosa through a Flo9-RadD interaction, evidence of bacteriome-mycobiome interplay in cancer.

Spatial profiling technologies have added a remarkable new dimension. Integrating spatial host multiomics with microbiota profiles has shown that intratumoral bacteria are heterogeneously distributed and concentrated in microniches that are infiltrated by immunosuppressive cells, with elevated neutrophils and exclusion of CD3-positive T cells. Bacteria-colonized niches show lower tumor cell proliferation but increased cell migration, suggesting that promoting growth may not be the dominant mechanism of intratumoral bacteria. A follow-up spatial imaging study found that intratumoral bacteria such as F. nucleatum exist in the extracellular milieu, where they disrupt cell-cell contacts among tumor cells and trigger a quiescent, immune-evasive phenotype that promotes chemoresistance. On the diagnostic front, stool-based single- and multi-bacterial gene biomarkers, alone or combined with the fecal immunochemical test, discriminate colorectal cancer from healthy controls with high accuracy and improve detection of adenomas, while several probiotic strains have suppressed tumorigenesis in animal models.

Perhaps the most clinically consequential findings concern therapy. Wong and Yu’s group systematically analyzed baseline gut microbiome profiles as predictors of chemotherapy response and identified Bacteroides fragilis as the top enriched bacterium in non-responders. B. fragilis drives chemoresistance through its surface protein SusD/RagB, which activates NOTCH1 signaling in colorectal cancer cells. To counter this, the team isolated a novel bacteriophage, VA7, that selectively eliminates B. fragilis in the mouse colon and reverses the chemoresistant phenotype. On the immunotherapy side, immune checkpoint blockade works only in microsatellite instability-high colorectal cancer, leaving the more than 85 percent of patients with microsatellite stable tumors largely unresponsive. The researchers discovered that the tumor-resident probiotic Clostridium butyricum potentiates anti-PD1 efficacy in both settings. Its surface protein secD binds the tumor cell receptor GRP78, suppressing the PI3K-AKT-NF-kappa-B pathway and reducing secretion of the immunosuppressive cytokine IL-6, which reactivates cytotoxic CD8-positive T cells and suppresses tumor-associated macrophages, an effect validated in humanized mice and organoid co-cultures.

In a paradoxical twist, F. nucleatum, normally a cancer-promoting pathogen, was found to enhance anti-PD1 response in microsatellite stable colorectal cancer. Intratumoral F. nucleatum secretes abundant butyrate, which inhibits histone deacetylases in CD8-positive T cells and epigenetically represses PD-1 expression, alleviating T cell exhaustion. Intratumoral F. nucleatum may therefore serve as a biomarker predicting anti-PD1 response in this patient subgroup. These findings position gut microbes as promising adjuvants for improving colorectal cancer immunotherapy.

Gastric cancer research is likewise moving beyond its long-standing obsession with Helicobacter pylori, a class I carcinogen long assumed to be the sole trigger of the Correa cascade from chronic gastritis to adenocarcinoma. Yet most infected people never develop cancer, H. pylori eradication only modestly reduces gastric cancer risk and fails in those with advanced precancerous lesions, and H. pylori abundance actually declines as lesions progress. Metagenomic sequencing across the gastric tumorigenesis sequence revealed enrichment of oral pathogens, including Streptococcus anginosus, Parvimonas micra, Peptostreptococcus stomatis, Prevotella intermedia, and F. nucleatum, whose persistence after H. pylori eradication correlated with inflammation, atrophy, and intestinal metaplasia. Causation was established when fecal microbiota transplantation from gastric cancer patients promoted precancerous lesions in germ-free mice, and when S. anginosus gavage reproduced the full gastritis-atrophy-metaplasia-dysplasia cascade in mice. The bacterium’s surface protein TMPC binds Annexin A2 on gastric cells, activating MAPK and PI3K oncogenic signaling, and additional mechanisms involving ornithine and methionine production and NLRP3 inflammasome induction are emerging. Translational prospects include S. anginosus and oral microbiota signatures as non-invasive diagnostic biomarkers, protective bacteria such as Lactobacillus paracasei, and phage-based strategies against the oral pathogen network.

Liver cancer presents a different geography of microbial mischief. The liver is not directly exposed to gut microbes in health, but the gut-liver axis channels microbial metabolites and products through the portal vein, and a leaky gut barrier in dysbiosis allows lipopolysaccharides, lipoteichoic acid, and secondary bile acids to inflame the liver. Fecal microbiota transplantation from hepatocellular carcinoma patients spontaneously triggered liver inflammation, fibrosis, and dysplasia in mice, and accelerated hepatocarcinogenesis in disease models. Strikingly, part of the translocated community, especially Klebsiella pneumoniae, survived as live bacteria colonizing the mouse liver; its surface protein PBP1B binds TLR4 on liver cancer cells and drives proliferation. A second pathogen, Catenibacterium mitsuokai, attaches to hepatocytes via Gtr1/RagA binding to gamma-catenin and secretes quinolinic acid to activate the TIE2-PI3K/Akt cascade. Metabolites matter too: in a high-fat, high-cholesterol diet model of metabolic dysfunction-associated liver cancer, the microbial metabolite 3-indolepropionic acid was depleted and shown to suppress tumor cell proliferation, while obesity-driven deoxycholic acid and Clostridium-mediated bile acid biotransformation impair antitumor immunosurveillance by natural killer T cells. Probiotics such as Bifidobacterium pseudolongum and Lactobacillus acidophilus, which produce acetate and valeric acid respectively, offer prophylactic promise.

Significant challenges remain before microbiome medicine becomes routine. Most human studies are still correlational, and the chicken-and-egg question, whether dysbiosis causes cancer or cancer reshapes the microbiome, remains unresolved, though recent evidence implicating colibactin-induced driver mutations as an early event in colorectal cancer supports a causal role. Fecal transplantation faces pathogen transmission risks and regulatory uncertainty, next-generation probiotics lack a unified marketing framework, and phage therapy suffers from a shortage of evidence-based clinical research. Microbiome diagnostics must also prove robust across geographically, ethnically, and dietary diverse populations, demanding large international validation cohorts. Still, with strain-level resolution, spatial mapping, and mechanistic dissection converging, the authors argue that well-designed clinical trials could soon translate gut microbes into diagnostics, chemopreventives, and immunotherapy adjuvants capable of meaningfully reducing the global burden of gastrointestinal cancers.

Subject of Research: The role of the gut microbiome in the development, diagnosis, and treatment of gastrointestinal cancers

Article Title: Advances in microbiome in gastrointestinal cancer

Article References: Wong, C. C., & Yu, J. (2026). Advances in microbiome in gastrointestinal cancer. Holistic Integrative Oncology, 5(1), Article 22. https://doi.org/10.1007/s44178-026-00241-4

Image Credits: AI Generated

DOI: 10.1007/s44178-026-00241-4

Keywords: gut microbiome, gastrointestinal cancer, colorectal cancer, gastric cancer, liver cancer, gut dysbiosis, Fusobacterium nucleatum, Streptococcus anginosus, immunotherapy, probiotics, bacteriophage therapy, microbiome diagnostics

The research team, led by Jasmine Kaur of Gian Sagar Medical College and Hospital in India, together with colleagues from institutions in Georgia, Nigeria, and India, followed rigorous PRISMA 2020 and MOOSE reporting guidelines, with a protocol registered in advance on PROSPERO. They searched PubMed, Embase, Web of Science, and Scopus for human studies published between January 2020 and February 2025 that examined three notorious oral bacteria: Fusobacterium nucleatum, Enterococcus faecalis, and Porphyromonas gingivalis. These species are common culprits in endodontic infections, the deep infections of the tooth pulp and root canal system, and each carries biological credentials that make the cancer hypothesis plausible.

From an initial pool of 150 unique records, 20 studies survived screening and full-text review, encompassing roughly 20,000 breast cancer cases. The studies spanned North America, Europe, Asia, and Africa, with sample sizes ranging from just 30 participants to more than 600,000. Exposure assessment varied widely: some studies used 16S rRNA gene sequencing or quantitative PCR to detect bacterial DNA, others relied on clinical periodontal examinations, antibody titers, or diagnosis codes extracted from insurance claims databases. Study quality was graded using the Newcastle-Ottawa Scale, with studies scoring seven or higher deemed high quality.

Only three studies were statistically compatible enough to pool in a meta-analysis, and the results were strikingly discordant. A large United States cohort of approximately 50,000 women found essentially no association between periodontal disease and breast cancer incidence, reporting a hazard ratio of 1.02 with a confidence interval spanning 0.94 to 1.11. A Greek case-control study, by contrast, reported that women with advanced periodontal disease had 79 percent higher odds of breast cancer, with an odds ratio of 1.79. Most paradoxically, an analysis of a US claims database found that prior Enterococcus infection was associated with a 40 percent reduction in breast cancer incidence, an odds ratio of 0.60. Pooled together, these studies yielded a summary odds ratio of exactly 1.00, with a 95 percent confidence interval of 0.62 to 1.58, meaning no overall association whatsoever.

The statistical heterogeneity between those three studies was extreme, with an I-squared value of 98 percent, indicating that nearly all of the observed variation came from differences between studies rather than chance. Subgroup analyses hinted at a familiar epidemiological pattern: associations appeared stronger in retrospective case-control designs and vanished in prospective cohorts, a signature of selection and recall bias. In case-control studies, patients who already have cancer may remember or report oral health problems differently, or the disease and its treatment may itself alter oral conditions. Sensitivity analyses confirmed the robustness of the null finding, as removing the lowest-quality study barely moved the pooled estimate, shifting it to an odds ratio of 1.02. With only three baseline studies available, the authors acknowledged that more sophisticated subgroup modeling was simply not feasible.

Where the epidemiology faltered, the laboratory evidence grew more compelling. Five studies examined whether F. nucleatum and P. gingivalis physically colonize breast tumors, and several confirmed by quantitative PCR and fluorescence in situ hybridization that F. nucleatum DNA was significantly more abundant in tumor tissue than in adjacent normal tissue. These bacteria were associated with more advanced disease stage and lymph node metastasis. Immunohistochemistry and spatial transcriptomics, techniques that map gene activity within intact tissue architecture, revealed upregulation of oncogenic signaling and immune-suppressive pathways in bacteria-positive tumors, with particularly notable findings in triple-negative breast cancer, the most aggressive subtype.

Animal models added mechanistic weight to these observations. Oral exposure to F. nucleatum and P. gingivalis promoted breast cancer metastasis in mice, apparently through inflammatory cytokine signaling involving interleukin-1 beta, interleukin-6, and CCL2. Tumors colonized by these bacteria showed reduced infiltration by cytotoxic T cells, the immune cells responsible for killing cancer cells, alongside an expansion of myeloid-derived suppressor cells, which dampen anti-tumor immunity. At the molecular level, the implicated pathways include activation of Toll-like receptor 4, engagement of the NF-kappa-B inflammatory cascade, and induction of epithelial-mesenchymal transition, the cellular program that allows stationary epithelial cells to become migratory and invasive. Only three studies tracked patient outcomes, but one linked the presence of intratumoral F. nucleatum to poor response to immunotherapy and lower recurrence-free survival.

The authors are careful to enumerate the caveats. Observational studies remain vulnerable to residual confounding by smoking, antibiotic use, diet, and socioeconomic factors, all of which influence both the oral microbiome and cancer risk. Bacterial detection methods varied so widely across studies that direct comparison is hazardous, and low-biomass tissue samples like breast tissue are notoriously susceptible to contamination, meaning some detected microbes may be laboratory artifacts rather than genuine tumor residents. Most included studies lacked longitudinal follow-up, making it impossible to establish whether bacterial colonization preceded cancer progression or followed it. The predominance of data from high-income countries further limits generalizability to populations with different oral health profiles and cancer risks.

Nevertheless, the convergence of microbiological, immunological, and genomic data points in one direction: F. nucleatum in particular emerges as a potential contributor to breast cancer progression, even if causality for cancer incidence remains unproven. The authors recommend large prospective cohorts with serial oral microbiome sequencing and rigorous confounder control, Mendelian randomization studies to probe causality, spatial multi-omics mapping of bacteria-immune interactions in human tumors, and standardized laboratory protocols for low-biomass tissue analysis with strict contamination controls. They also propose exploring oral pathogens or their signatures in saliva and blood as biomarkers for risk stratification, prognosis, or immunotherapy response prediction.

Clinical translation remains speculative but not far-fetched. Although no clinical trials of antimicrobial therapy in breast cancer yet exist, preliminary evidence from colorectal cancer, where F. nucleatum is a well-established modulator of tumor behavior and chemoresistance, suggests that modulating this bacterium could improve treatment responses. Future randomized trials might evaluate bacteriophage therapy, probiotics, or targeted antibiotics to test whether reshaping the oral and intratumoral microbiota alters the course of breast cancer. For now, the message for patients is measured: there is no evidence that a root canal infection causes breast cancer, and no reason for alarm. But the mouth, it turns out, may still have something to say about how the disease behaves, and researchers are listening closely.

Subject of Research: Association between root canal-associated oral bacteria and breast cancer incidence and progression

Article Title: Association of root canal bacteria with breast cancer: a systematic review and meta-analysis

Article References: Association of root canal bacteria with breast cancer: a systematic review and meta-analysis. (n.d.). https://doi.org/10.1007/s44178-026-00254-z

Image Credits: AI Generated

DOI: 10.1007/s44178-026-00254-z

Keywords: breast cancer, root canal bacteria, Fusobacterium nucleatum, Porphyromonas gingivalis, Enterococcus faecalis, oral microbiome, meta-analysis, systematic review, tumor microbiome, immune evasion, periodontal disease, cancer progression

The scale of the oral ecosystem alone is striking. Based on 16S rDNA profiling, the dominant bacterial phyla in a healthy mouth, Firmicutes, Actinobacteria, Proteobacteria, Fusobacteria, Bacteroidetes, and Spirochaetes, together account for roughly 96 percent of oral bacteria. These microbes occupy distinct ecological niches, from supragingival and subgingival plaque to the tongue, buccal mucosa, saliva, and tonsils, each hosting its own characteristic community assembly. Notably, the oral microbiome shows significant variation between individuals but remarkable stability within a single person over time, exhibiting less temporal fluctuation than communities at sites like the gut or skin. That stability, the authors suggest, makes the mouth an ideal biological system for microbiome research, and potentially a reliable window into systemic health. Salivary gland function, oral clearance rates, pH, and secretory immunoglobulin A levels all act as regulators of this balance, and when any of them falter, the stage is set for dysbiosis.

Under healthy conditions, the oral community functions as a defensive barrier, resisting colonization by exogenous pathogens through competition and the inhibitory effects of microbial metabolites. But the review emphasizes that the boundary between commensal and pathogen is porous. Opportunistic species such as Porphyromonas gingivalis and Fusobacterium nucleatum, ordinarily tolerated members of the oral flora, can turn inflammatory and carcinogenic when the delicate equilibrium of the community is disrupted. In a dysbiotic state, these organisms and their metabolic products activate extracellular matrix degradation pathways and interfere with immune-related signaling, contributing to a chronic pro-inflammatory state that favors tumorigenesis. The review also connects this imbalance to acute systemic conditions, noting that oral dysbiosis has been associated with acute respiratory distress syndrome and sepsis through oral-intestinal axis-mediated changes in gut flora, underscoring that the mouth’s influence extends well beyond oncology.

Central to the review’s argument is the concept of oral-systemic axes, the routes by which oral microbes establish communication with microbial communities in distant organ systems. More than half of microbial genera can be detected in both the oral cavity and the gut, and oral organisms can migrate directly to the intestinal mucosa via the digestive tract or disseminate through the bloodstream. One vivid example involves periodontitis: pathologically activated Th17 cells acquire gut tropism, migrate to inflamed intestines, are specifically reactivated by colonized oral pathogens, and subsequently induce colitis. In the tumor context, dysbiosis, colonization, and translocation of oral bacteria upregulate cytokines and inflammatory mediators that reshape the tumor microenvironment. Metabolites play a critical role as well. Lactate can recruit immunosuppressive cell types including regulatory T cells, tumor-associated macrophages, and myeloid-derived suppressor cells, while dietary tryptophan metabolites activate the aryl hydrocarbon receptor in myeloid cells, fostering an immunosuppressive milieu that supports pancreatic ductal adenocarcinoma growth.

The mechanistic detail is deepest for colorectal cancer, where Fusobacterium nucleatum has become something of a molecular celebrity. The bacterium’s outer membrane protein Fap2 binds to the Gal-GalNAc carbohydrate abundant on colorectal tumor cells, mediating bacterial enrichment within tumors, while its adhesin A binds E-cadherin on intestinal epithelial cells and activates beta-catenin signaling, driving uncontrolled cell proliferation. F. nucleatum also induces secretion of IL-8 and CXCL1 to promote cancer cell migration, recruits tumor-infiltrating immune cells via CEACAM1 to evade immune surveillance, and targets the long non-coding RNA ENO1-IT1 to enhance glycolysis and tumor growth. Strikingly, Fusobacterium strains isolated from colorectal tumors closely resemble strains from the same patients’ saliva, supporting the hypothesis that colonic dysbiosis is at least partly seeded from the mouth. Germ-free mice transplanted with human saliva have allowed researchers to track how oral bacteria colonize the gastrointestinal tract through continuous swallowing, hematogenous spread during periodontal bacteremia, and environmental changes in the gut that permit stable colonization.

Similar patterns appear across other malignancies. In gastric cancer, oral taxa such as Rothia, Eikenella corrodens, Bergeyella, and Capnocytophaga are enriched in the gastric mucosa of patients, and Helicobacter pylori is detected in the oral cavity as well as the stomach. Actinomyces, an emerging opportunistic pathogen, promotes autophagy and upregulates TLR4 expression in gastric carcinogenesis, while acid-producing bacteria drive excessive short-chain fatty acid and lactate production that fuels tumor angiogenesis. In pancreatic cancer, epidemiological studies consistently link poor oral hygiene and periodontal disease to elevated risk, and intracellular P. gingivalis has been shown to promote oncogenic behavior in pancreatic cancer cells by activating Akt signaling. Perhaps most provocatively, pancreatic cyst fluid from tumor patients harbors a microbial community overlapping substantially with the oral microbiota, and the hypoxic, immunosuppressive pancreatic tumor environment appears to offer favorable conditions for oral anaerobes. In lung cancer, reduced oral alpha-diversity is negatively associated with risk, and P. gingivalis colonization, encouraged by long-term smoking and alcohol consumption, promotes malignant progression.

The relationship runs in both directions, and cancer treatment itself reshapes the oral ecosystem. Chemotherapy commonly induces oral mucositis and xerostomia, disrupting the homeostasis between host defenses and commensal bacteria and opening portals for opportunistic infection; studies report increases in Streptococcus viridans group species and, after two weeks of treatment, in anaerobes such as F. nucleatum and Prevotella intermedia. Radiotherapy for head and neck cancers raises the abundance of Gram-negative bacteria and Candida, increases Lactobacillus across multiple oral sites, and reduces bacterial alpha-diversity, with the most frequent complication being mucositis whose severity tracks with microbial shifts. Encouragingly, early clinical evidence suggests probiotics may mitigate this damage, with Bacteroides and Bifidobacterium genera shown to boost immune cell number and activity, potentially enhancing anti-CTLA-4 and anti-PD-L1 immunotherapy while reducing treatment toxicity.

Prognostic and diagnostic implications are already taking shape. The Fusobacterium genus has been independently linked to poorer outcomes in pancreatic, oral squamous cell, and colorectal cancers, while salivary Candida carriage correlates with unfavorable prognosis in oral cancer and Malassezia enrichment with favorable prognosis. Higher oral microbial diversity within pancreatic tumors and richer, more even oral communities in non-small cell lung cancer patients have both been associated with longer survival, and diversity indices may serve as independent predictors of overall and relapse-free survival. On the diagnostic front, the numbers are eye-catching: salivary microbiome analysis achieved diagnostic accuracy exceeding 90 percent in a cohort of 47 oral squamous cell carcinoma patients, salivary screening for gastric cancer yielded area-under-curve values of 91 and 97 percent in separate studies, and combining oral with fecal microbiome profiles improved colorectal cancer detection to 95 percent specificity and 88 percent sensitivity, outperforming the standard fecal immunochemical test.

The authors are careful to temper enthusiasm with candor. Current research remains largely observational, limited by heterogeneous study designs, insufficient causal evidence in human cohorts, and persistent challenges in standardizing oral microbiome sampling and analysis, and the prognostic literature remains inconclusive given variations in population, geography, and sample size. Yet the translational trajectory is clear. Future work, they argue, should move beyond descriptive associations toward identifying specific microbial signatures and functional pathways, prioritizing standardized sampling, longitudinal designs, and multi-omics integration with clinical outcomes. Targeted antimicrobials, probiotics, and microbiota-regulating interventions could eventually reshape the oral ecosystem in ways unfavorable to tumor development or supportive of anti-tumor immunity. As a non-invasive, easily collected, and low-cost specimen, saliva may soon join blood and tissue as a routine medium for cancer screening, prognostic evaluation, and personalized medicine, transforming the humble dental checkup into a window on one of medicine’s most formidable diseases.

Subject of Research: The role of oral microbiota dysbiosis in tumorigenesis, cancer prognosis, and early detection

Article Title: Research advances in the correlation between oral microbiota and tumors

Article References: Liu, X., Liu, S., & Guo, Z. (2026). Research advances in the correlation between oral microbiota and tumors. Holistic Integrative Oncology, 5(1), Article 48. https://doi.org/10.1007/s44178-026-00270-z

Image Credits: AI Generated

DOI: 10.1007/s44178-026-00270-z

Keywords: oral microbiota, tumorigenesis, Fusobacterium nucleatum, Porphyromonas gingivalis, oral-gut axis, colorectal cancer, pancreatic cancer, dysbiosis, salivary biomarkers, chronic inflammation, microbiome, cancer screening

ETBF is a subtype of Bacteroides fragilis that carries the gene for Bacteroides fragilis toxin, known as BFT. The bacterium has been linked to diarrheal disease, inflammatory bowel disease and colorectal cancer, and colonization is known to be associated with long-term disturbances of intestinal motility. What remained unclear was precisely how a toxin famous for cleaving epithelial junctions could leave the gut’s movement machinery permanently impaired. The Shanghai team approached the question using mouse colonization models in which animals were inoculated with ETBF and followed over weeks, allowing the researchers to track both the microbial load and the structural integrity of the colonic myenteric plexus, the network of neurons embedded in the gut wall that orchestrates peristalsis.

The answer was stark. ETBF colonization triggered the loss of colonic myenteric neurons, and the effect depended entirely on BFT: mice colonized with a genetically engineered ETBF strain lacking the bft gene were largely protected. The toxin did not merely stun the neurons; it drove them into apoptosis, a programmed cell death routine. Critically, the neuronal loss and the accompanying motility dysfunction persisted even after the bacteria themselves had been cleared from the gut, a finding that echoes the clinical phenomenon of post-infectious gut disorders, in which symptoms linger long after the offending pathogen has disappeared. The researchers measured this dysfunction using gastrointestinal transit time and ex vivo recordings of colonic strip contractions, both of which deteriorated in colonized animals.

Having established that BFT kills enteric neurons, the team dissected the molecular pathway inside the cells. The toxin’s lethal signal, they found, runs through NOD1, an intracellular pattern-recognition receptor, which recruits its adaptor protein RIPK2 and then activates a caspase cascade: caspase-9, the initiator of the intrinsic apoptotic pathway, followed by caspase-3, the executioner. When the researchers knocked out Nod1 in mice, the BFT-induced neuronal apoptosis and dysmotility were markedly attenuated, confirming that this receptor, better known for sensing bacterial peptidoglycan and driving inflammatory responses, also functions as a death switch in enteric neurons when provoked by BFT. The pathway is notable because NOD1 has previously been identified as a functional receptor for BFT in cancer contexts, suggesting the toxin may have co-opted a host surveillance system for destructive purposes across multiple cell types.

The second half of the study addresses a question that has intrigued microbiome researchers for years: why do ETBF and Fusobacterium nucleatum so often appear together in diseased guts? Analyses of patient cohorts with inflammatory bowel disease and colorectal cancer revealed a positive correlation between the abundance of the two organisms, hinting at a cooperative rather than coincidental relationship. To test this directly, the researchers co-colonized mice with both species. The result was dramatic: the presence of F. nucleatum significantly aggravated the ETBF-induced loss of myenteric neurons and worsened gut dysmotility, and the effect again required bft, indicating that the oral pathobiont was not adding its own toxin but rather turning up the volume on ETBF’s.

The mechanism behind this microbial collusion proved to be a small molecule. F. nucleatum produces 2-hydroxybutyric acid, or 2HB, a metabolite generated through lactate dehydrogenase activity. When ETBF is exposed to 2HB, the toxin gene bft is transcribed at far higher levels. The team showed that 2HB physically binds to RprY, a regulatory protein in ETBF that normally represses bft transcription. By occupying RprY, the metabolite lifts this repression, derepressing the toxin gene without altering other virulence genes controlled by the regulator. In other words, one bacterium’s metabolic waste product acts as a molecular key that unlocks another bacterium’s most dangerous weapon. Administering 2HB to mice colonized with ETBF alone reproduced the aggravating effect, while F. nucleatum strains engineered to overproduce the metabolite intensified neurotoxicity further, and the researchers confirmed that multiple F. nucleatum subspecies produce 2HB in culture.

The clinical implications extend beyond motility. The study found that the enteric neuron loss and dysmotility induced by ETBF increased susceptibility to dextran sulfate sodium-induced colitis in mice, linking the neural damage to a weakened mucosal defense. This connects with a growing body of evidence that the enteric nervous system is not merely a passive conduit for brain signals but an active participant in intestinal immunity and barrier maintenance. Previous work has shown that other pathogens, from Shigella to Clostridioides difficile, can target enteric neurons, and that neuronal loss in conditions such as Chagas disease and Hirschsprung disease produces profound motility defects. The new study adds a bacterial cooperation model to this landscape, in which the composition of the microbiome determines how much toxin a colonizing pathogen actually delivers.

Human relevance was reinforced by tissue analyses. In colonic specimens from ulcerative colitis patients, the researchers detected the bft gene in mucosal tissue and observed myenteric neurons positive for NOD1, RIPK2, cleaved caspase-9 and cleaved caspase-3, the same molecular signature seen in the mouse model, with the staining pattern more prominent in patients suffering from constipation. While such observational data cannot prove causation in humans, the concordance between the animal pathway and the human tissue findings strengthens the case that the mechanism operates in clinical disease.

Perhaps the most immediately actionable finding concerns therapy. The researchers tested Bifico, a commercially available probiotic formulation containing Bifidobacterium species, in their mouse model. Treatment attenuated the neuronal toxicity, improved intestinal motility and reduced the severity of subsequent colitis. Bifidobacterium longum alone recapitulated much of the protective effect, and the probiotic did not simply eradicate ETBF; rather, it appeared to interfere with the processes that drive toxin production and neuronal death. Given that probiotics are inexpensive, widely available and generally safe, the finding suggests a plausible strategy for preventing the long-term neurological consequences of ETBF colonization, although the authors and independent observers caution that mouse models do not always translate to human therapy and that controlled clinical trials would be needed.

The study, published in Nature Microbiology with the DOI 10.1038/s41564-026-02497-y, is likely to resonate across several fields at once. For microbiome researchers, it provides a textbook example of metabolic cross-feeding with pathological consequences, in which a diffusible small molecule from one species reprograms virulence gene expression in another. For neurogastroenterologists, it identifies a specific toxin-receptor-caspase axis that destroys the neurons governing bowel movement, offering potential drug targets in NOD1, RIPK2 or the caspases themselves. And for clinicians treating patients with post-infectious bowel dysfunction, irritable bowel symptoms or inflammatory bowel disease, it raises the possibility that screening for ETBF and F. nucleatum co-colonization, and perhaps manipulating the microbiome to reduce 2HB production or boost protective Bifidobacterium, could one day prevent the slow, silent loss of the gut’s own nervous system. The data and bacterial strains underlying the work have been deposited in public repositories, including a Zenodo archive of the neuronal RNA-sequencing data, allowing other laboratories to build on the findings immediately.

Subject of Research: Bacterial cooperation between Fusobacterium nucleatum and enterotoxigenic Bacteroides fragilis in enteric neuron loss and intestinal dysmotility

Article Title: Fusobacterium nucleatum enhances enterotoxigenic Bacteroides fragilis-mediated neuron loss and intestinal motility dysfunction

Article References: Zhang, Y., Zhao, Y., Zhang, L., Xuan, B., Wang, Z., Yu, B., Li, W., Huang, X., Zhou, Y., Ning, L., Ding, J., Jiang, Y., Hu, M., Shao, Y., Li, L., Gao, X., Chen, S., Chen, H., Wang, F., … Hong, J. (2026). Fusobacterium nucleatum enhances enterotoxigenic Bacteroides fragilis-mediated neuron loss and intestinal motility dysfunction. Nature Microbiology. https://doi.org/10.1038/s41564-026-02497-y

Image Credits: AI Generated

DOI: 10.1038/s41564-026-02497-y

Keywords: Fusobacterium nucleatum, enterotoxigenic Bacteroides fragilis, Bacteroides fragilis toxin, enteric nervous system, gut dysmotility, 2-hydroxybutyric acid, NOD1, apoptosis, gut microbiota, inflammatory bowel disease, probiotics, Bifidobacterium

The scale of the problem provides the urgency. Gastric cancer alone affects more than one million people each year and accounts for roughly 5.7 percent of all cancers worldwide, yet a large proportion of patients are diagnosed only at advanced stages, when survival prospects are grim. Helicobacter pylori remains the best-established infectious risk factor, but it is no longer viewed as the sole microbial culprit. Chronic mucosal inflammation triggered by a dysbiotic community can reshape the gastric environment, and repeated activation of the host immune system by gut bacteria creates a persistent inflammatory state in which normally benign residents, so-called pathobionts, can cooperate with dietary and environmental carcinogens to push cells toward malignancy. Worryingly, studies show that cancerous tissues in the digestive tract harbor more diverse and abundant microbial populations than adjacent stromal tissue, and that microbial diversity is higher in advanced gastric cancer than in early disease.

The oral cavity has emerged as a surprising reservoir of cancer-associated organisms. Housing roughly 500 to 700 bacterial species distributed across distinct niches such as saliva, dental plaque, and the tongue, the mouth hosts microbes that appear linked to cancers far downstream. Porphyromonas gingivalis, a notorious periodontal pathogen, makes oral cancer cells more aggressive and resistant to chemotherapy by stimulating Toll-like receptor 4 and myeloid differentiation factor 88, activating the nuclear factor-κB pathway and releasing pro-inflammatory cytokines. Its lipopolysaccharide also appears to promote pancreatic cancer through MyD88-dependent signaling that biases dendritic cells toward T helper 2 differentiation. Meanwhile, molecular surveys using 16S rRNA gene sequencing have detected oral species such as Parvimonas micra, Streptococcus anginosus, Slackia exigua, and Peptostreptococcus stomatis within gastric tumors, likely because a reduction in gastric acidity, sometimes exacerbated by proton pump inhibitors, opens a colonization window for microbes that would otherwise be destroyed.

In the colon, the densest microbial reservoir in the human body, specific organisms have been interrogated at the mechanistic level. Fusobacterium nucleatum, a commensal of the mouth, uses its FadA adhesin to trigger transcriptional changes in colorectal cancer cells and drives proliferation and migration through the TLR4/MyD88/NF-κB axis, including modulation of the microRNA miR-21. Enterotoxigenic Bacteroides fragilis promotes intestinal tumorigenesis in mouse models via inflammatory STAT3 and T helper 17 mechanisms. Peptostreptococcus anaerobes, found at higher abundance in the stool of colorectal cancer patients, bind α2/β1 integrins through a surface protein, activating focal adhesion kinase phosphorylation and the PI3K/AKT pathway, which fuels cell proliferation and inflammation. Other bacteria are genotoxic in a more direct sense: certain Escherichia coli strains produce colibactin, while H. pylori strains carrying the cytotoxin-associated gene A deliver a virulence factor that manipulates host signaling, and both belong to a class of toxins called cyclomodulins that disrupt the cell cycle and render their producers highly carcinogenic.

What unifies these observations is the recognition that dysbiosis-driven carcinogenesis is not the product of a single pathway but a networked collapse of multiple regulatory systems. Bacterial metabolites, altered bile acid metabolism, weakening of the mucosal barrier, and activation of signaling cascades such as NF-κB, STAT3, and Wnt/β-catenin act in concert to create a tumor-permissive microenvironment. Microbial products also include reactive oxygen species, reactive nitrogen species, hydrogen sulfide, nitrosamines, and acetaldehyde, all of which inflict genotoxic stress on epithelial cells. Crucially, the gut microbiota also metabolizes environmental chemicals and food compounds, either detoxifying them or converting them into more dangerous intermediates, which means that the functional activity of the community, not merely its taxonomic composition, determines cancer risk. This functional perspective is reshaping how researchers design microbiome studies.

Nowhere is the mechanistic detail richer than in H. pylori-driven gastric cancer. The bacterium sustains STAT3 activation by upregulating interleukin 6, by shuttling CagA into epithelial cells where it acts through the SHP-2 pathway, and by engaging TLR2 to amplify inflammatory responses. Recent work shows H. pylori drives overexpression of DAB2 through the SRC-YAP1 axis in a STAT3-dependent manner, and activates fibroblast growth factor receptor 4 via STAT3, both processes linked to carcinogenesis. On the NF-κB front, CagA directly stimulates the transcription factor and increases its nuclear shuttling, while the bacterium activates the IκB kinase complex to boost production of tumor necrosis factor-α and interleukin 1β. One study traced an NF-κB-driven PIEZO1/YAP1/CTGF pathway through which H. pylori remodels the gastric tumor microenvironment. The Wnt/β-catenin pathway is similarly hijacked: CagA promotes the accumulation and nuclear translocation of β-catenin, H. pylori-induced LRP8 expression facilitates that nuclear entry, and the resulting expression of epithelial-to-mesenchymal transition transcription factors confers stem-like properties and invasive behavior on tumor cells. The bacterium also modulates the three major MAPK branches, JNK, ERK, and p38, and selectively downregulates other JAK-STAT components as an immune-evasion strategy.

Barrier failure adds another layer of vulnerability. The intestinal mucosal barrier, composed of epithelial cells, tight junctions, the mucus layer, and the mucosal immune system, normally keeps microbial products out of the circulation. Dysbiosis, particularly in aging, reduces microbial diversity, weakens intercellular junctions, and increases epithelial permeability, producing a leaky gut through which lipopolysaccharides and peptidoglycans flood the lamina propria and bloodstream. There they engage pattern recognition receptors and trigger chronic systemic inflammation, oxidative stress, and DNA damage, conditions that favor tumor initiation, progression, and metastasis. Barrier dysfunction also impairs antitumor immunity and can blunt the efficacy of anticancer therapies, especially immunotherapy, making the restoration of barrier integrity and microbial balance an emerging therapeutic strategy.

Even blood, long considered sterile, may carry diagnostic echoes of the gut. Culture-independent sequencing has detected microbial DNA in blood, although the existence of a stable blood-resident microbiome remains controversial because low biomass makes contamination a serious methodological concern. Nonetheless, several studies report that the blood microbiome of gastric cancer patients differs significantly from that of healthy individuals, with elevated levels of Bacteroides, Haemophilus parainfluenzae, and Acinetobacter, and that serum microbial profiles correlate with tumor size, invasion depth, metastasis, and even the distinction between lymphatic and non-lymphatic spread. In advanced colorectal cancer, patients who responded to immunochemotherapy combined with adoptive T-cell therapy carried higher concentrations of Bifidobacterium, Lactobacillus, and Enterococcus in their blood than non-responders, hinting that microbial signatures might predict treatment outcomes.

The biomarker potential extends across the digestive tract. In the esophagus, a shift from a predominantly streptococcal community toward one enriched with Actinomyces, Veillonella, Neisseria, Fusobacterium, and Porphyromonas marks an unhealthy state, and several-fold increases in Clostridiales species, F. nucleatum, Erysipelotrichales, or P. gingivalis have been proposed as prognostic indicators for esophageal squamous cell carcinoma. Gastric cancer patients show high levels of Lactobacillus, Lactococcus, and the Lachnospiraceae family alongside oral taxa such as Haemophilus and Campylobacter, and distinct subtype-specific signatures have been reported, with Proteobacteria and Acidobacteria dominating adenocarcinomas while Fusobacteria and Bacteroidetes prevail in signet ring cell carcinomas. In saliva, pancreatic cancer patients show reduced Streptococcus mitis and Neisseria elongata and an elevated Leptotrichia-to-Porphyromonas ratio, while F. nucleatum in stool is being explored as a noninvasive early-detection marker for colorectal cancer.

The authors are careful to note that the evidence is not yet ready for the clinic. Individual differences in diet, medication use, geography, sampling methods, and sequencing protocols confound current findings, and no microbial marker has been validated for routine clinical application. What the field needs, they argue, are longitudinal and multicenter studies with standardized sampling and multiomic approaches that independently validate microbial markers, alone and in combination. If that work succeeds, a detailed understanding of the interplay between the microbiome, epithelial cells, the immune system, and the tumor microenvironment could yield new targets for prevention, earlier detection, and treatment of gastrointestinal cancers, transforming the trillions of microbes we carry from an invisible background into a usable map of risk. For now, the message is clear: the balance of our inner ecosystem may be one of the most consequential frontiers in cancer medicine.

Subject of Research: The role of gastrointestinal microbiome dysbiosis in digestive cancer development and biomarker potential

Article Title: Gastrointestinal Microbiome Dysbiosis in Cancer Development: Mechanisms and Biomarker Potential

Article References: Gastrointestinal Microbiome Dysbiosis in Cancer Development: Mechanisms and Biomarker Potential. (n.d.). https://doi.org/10.1002/mbo3.70404

Image Credits: AI Generated

DOI: 10.1002/mbo3.70404

Keywords: microbiome, dysbiosis, gastric cancer, colorectal cancer, Helicobacter pylori, Fusobacterium nucleatum, NF-κB, biomarkers, microbial translocation, gut barrier, oral microbiome, cancer immunotherapy

At the center of this story are short-chain fatty acids, or SCFAs, metabolites produced when gut bacteria ferment dietary fiber. Butyrate, acetate, and propionate do far more than nourish the intestinal lining. Butyrate acts as a histone deacetylase inhibitor, meaning it chemically modifies chromatin to switch on genes involved in anti-inflammatory pathways and cytotoxic T-cell function. Studies cited in the review show that butyrate enhances the differentiation of regulatory T cells, suppresses pro-inflammatory cytokine production through inhibition of NF-κB signaling, and boosts the tumor-killing activity of CD8-positive T cells and natural killer cells by upregulating interleukin-12 and interferon-gamma. Bacteria such as Faecalibacterium prausnitzii and Clostridium butyricum, which produce butyrate, have been repeatedly associated with improved outcomes in patients receiving immune checkpoint inhibitors.

The review also dissects how specific bacterial species act as biological adjuvants for immunotherapy. Akkermansia muciniphila, a mucin-degrading bacterium, has been linked to superior responses to anti-PD-1 therapy, apparently because it recruits dendritic cells to the gut barrier and promotes the infiltration of CCR9-positive, CXCR3-positive CD4-positive T lymphocytes into tumors. Bifidobacterium longum, meanwhile, enhances dendritic cell maturation and antigen presentation, and preclinical studies show that mice supplemented with this bacterium mount stronger anti-PD-L1 responses. In landmark clinical work, melanoma patients with higher abundance of Akkermansia muciniphila responded significantly better to anti-PD-1 treatment than those lacking the species.

Not all microbes are allies, however. The review devotes considerable attention to dysbiosis, the imbalance of gut microbial communities that undermines therapy. Fusobacterium nucleatum, overrepresented in many colorectal cancers, promotes immune evasion by expanding myeloid-derived suppressor cells and regulatory T cells, suppressing the cytotoxic CD8-positive T cells that checkpoint inhibitors rely upon. In pancreatic cancer, Gammaproteobacteria express the enzyme cytidine deaminase, which degrades the chemotherapy drug gemcitabine into an inactive form, rendering treatment ineffective. Certain bacteria also reactivate drug metabolites in ways that amplify toxicity, as with microbial beta-glucuronidase reactivating irinotecan, illustrating that the microbiome can cut both ways.

The mechanistic picture extends beyond metabolites. Microbe-associated molecular patterns such as lipopolysaccharides and flagellins engage Toll-like receptors on immune cells, triggering type I interferons and interleukin-12 that sharpen antigen presentation and T-cell priming. The aryl hydrocarbon receptor, activated by microbial indole derivatives from Lactobacillus and Bacteroides species, modulates cytokine production and T-cell differentiation, with context-dependent effects on tumor immunity. Perhaps most striking is molecular mimicry: some microbial peptides resemble tumor-associated antigens closely enough to prime cross-reactive T cells, effectively training the immune system to recognize cancer.

These insights have spawned a therapeutic pipeline. Fecal microbiota transplantation, in which stool from immunotherapy responders is transferred to non-responders, has produced remarkable results in early trials. In landmark studies, melanoma patients who had previously failed anti-PD-1 therapy experienced significant tumor shrinkage after receiving fecal transplants from responsive donors, with restored microbial diversity and reinvigorated T-cell activity. Probiotics, prebiotics, and synbiotics are being tested as gentler alternatives, while engineered bacteria capable of secreting interleukin-12 directly into tumors represent a synthetic biology frontier. Small molecules that mimic butyrate or modulate tryptophan metabolism are also under investigation.

The review emphasizes that antibiotics may be an underappreciated threat to immunotherapy success. Broad-spectrum antibiotic use before or during checkpoint inhibitor treatment has been associated with lower response rates and shorter progression-free survival, presumably because it eliminates beneficial species such as Faecalibacterium, Bifidobacterium, and Akkermansia and disrupts SCFA production. Preclinical models confirm that antibiotic-treated mice show impaired tumor regression following anti-PD-1 therapy, prompting calls for microbiome-prescribing caution in oncology.

Personalized microbiome profiling is positioned as the next step. Techniques including 16S rRNA gene sequencing, shotgun metagenomics, and metabolomics can already identify whether a patient’s gut community is likely to support or sabotage immunotherapy. Machine learning models that integrate microbiome data with genomic and clinical markers are being developed to predict which patients will benefit from specific interventions, potentially allowing clinicians to reshape a patient’s microbiome before starting treatment. Microbiome-based diagnostic panels measuring species such as Akkermansia muciniphila and Faecalibacterium prausnitzii, alongside SCFA levels, could guide these decisions.

Significant obstacles remain before microbiome medicine becomes routine. Inter-individual variability in gut communities, driven by diet, geography, genetics, and antibiotic exposure, makes standardized interventions difficult. Regulatory frameworks for probiotics, prebiotics, and fecal transplantation in oncology are still immature, and safety concerns, including opportunistic infections and immune complications such as colitis in checkpoint inhibitor patients, demand rigorous screening protocols. Discrepancies between sequencing methods further complicate the comparison of studies and the validation of biomarkers.

Nevertheless, the trajectory is clear. The review argues that the microbiome-immune interface represents a paradigm shift in oncology, moving the field toward precision immunotherapy in which the gut ecosystem is treated as a modifiable organ of the immune system. With large-scale randomized trials underway and engineered microbial therapeutics advancing through preclinical development, the prospect of converting immunotherapy non-responders into responders by reprogramming their gut bacteria is no longer science fiction but an active clinical frontier.

Subject of Research: The role of the gut microbiome and its metabolites in modulating immune responses and therapeutic outcomes in cancer

Article Title: The mechanisms and therapeutic potential of the microbiome-immune interface in cancer

Article References: Christina, B., Poongkuzhali, S., Muninathan, N., Bhaskaran, K., & Suresh, A. (2025). The mechanisms and therapeutic potential of the microbiome-immune interface in cancer. Discover Biotechnology, 2(1), Article 25. https://doi.org/10.1007/s44340-025-00031-0

Image Credits: AI Generated

DOI: 10.1007/s44340-025-00031-0

Keywords: microbiome, cancer immunotherapy, immune checkpoint inhibitors, gut microbiota, short-chain fatty acids, Akkermansia muciniphila, Fusobacterium nucleatum, fecal microbiota transplantation, probiotics, dysbiosis, tumor microenvironment, personalized oncology

The stakes are enormous. Cancer affects roughly 20 million people each year, and projections suggest the annual burden could climb to 30.5 million by 2050. While tumor-intrinsic factors such as mutations and dysregulated signaling pathways remain central to oncology, researchers increasingly recognize that tumor-extrinsic factors, including everything non-cancerous within the tumor microenvironment and the broader tumor macroenvironment, powerfully influence disease trajectories. The cancer microbiome, spanning both the gut microbiome and the tumor-associated microbiome, has emerged as one of the most intriguing of these factors. Early studies focused on colorectal cancer simply because of anatomical proximity, but evidence now shows the gut microbiome acts systemically, modulating host immunity and metabolism across the body, and has been implicated in tumorigenesis, progression, treatment response, and immune-related adverse events.

The tumor-associated microbiome tells a different story. Unlike the gut’s rich microbial communities, intratumoral microbes exist in low abundance, sparse populations that vary dramatically by cancer type. Tumors exposed to the external environment, such as colorectal, gastric, and oral cancers, harbor relatively more microbial biomass, while pancreatic, liver, lung, and breast tumors are considered low-biomass settings. Evidence from experimental models and human datasets has linked these intratumoral communities to cancer progression, prognosis, and treatment response, and the microbes can localize both outside and inside cancer and immune cells, sometimes with distinct spatial organization. They influence their surroundings through infection, inflammation, and the production of metabolites, yet separating genuine microbial signals from laboratory contaminants remains one of the field’s hardest problems.

Computationally, the field has traveled a long road. Early studies relied on classical statistical tools such as differential abundance methods, including LEfSe and metagenomeSeq, to identify taxa associated with cancer risk, survival, and treatment response. But microbiome data are notoriously difficult: high-dimensional, sparse, zero-inflated, and compositional, properties that can reduce reproducibility. Network-based approaches like SparCC, CoNet, and SPIEC-EASI extended the toolkit by modeling microbial interactions and identifying community structures linked to cancer processes, yet they struggle with complex nonlinear relationships. Deep learning has now entered the picture, enabling integration of microbiome and multi-omics data to model higher-order interactions and improve biomarker discovery and clinical outcome prediction. The catch is that most AI models must work with high-dimensional but small-sample datasets, raising overfitting risks and threatening biomarker stability, especially without strong external or prospective validation.

Generating reliable data is the first battleground. Cancer microbiome studies produce diverse data types, each capturing different aspects of microbial composition, function, and host interaction. Partial 16S rRNA sequencing is affordable but usually resolves taxonomy only to the genus level; full-length 16S improves resolution to species; shotgun metagenomics captures all genes in a sample, enabling both taxonomy and functional prediction. Metatranscriptomics captures real-time gene expression but is technically demanding, limited by RNA instability and stringent handling requirements. Metaproteomics, which profiles expressed proteins, offers a more direct functional view but has barely touched cancer: a PubMed search as of June 2026 identified only nine cancer microbiome metaproteomics studies. Metabolomics rounds out the picture, measuring the small molecules microbes produce, with databases such as MiMeDB, the Natural Products Atlas, and MASST helping to attribute metabolites to microbial origins, while spatial metabolomics now maps region-specific metabolic changes within tumors.

Detecting intratumoral microbes demands special tools. Researchers have reanalyzed bulk RNA-seq and whole-genome sequencing data to infer microbial signals from non-human reads, but this approach is vulnerable to contamination, and a recent large-scale tumor whole-genome analysis found that after host subtraction and decontamination, detectable microbiome signals were largely restricted to orodigestive cancers. Specialized pipelines have emerged to help: CSI_Microbe extracts microbial reads from The Cancer Genome Atlas sequencing data, SAHMI denoises microbial signals from single-cell RNA sequencing, and INVADEseq adds a primer targeting the conserved 16S region to map microbes within individual human cells. Spatial technologies such as imaging mass cytometry with mass-tagged antibodies and desorption electrospray ionization mass spectrometry imaging can visualize microbial presence alongside host immune and tumor cells. The review also lays out practical standards for credible signals: negative controls, conservative host-read subtraction, evaluation of batch structure, and orthogonal validation through qPCR, culture, in situ hybridization, or spatial imaging.

Once data are trustworthy, AI modeling begins in earnest, and the review offers a sobering lesson: bigger is not always better. Benchmarking studies show that classical, regularized models remain strong baselines. In a 16S rRNA benchmark with 490 subjects and 6,920 features, L2-regularized logistic regression matched random forest performance while training faster and remaining more interpretable. A larger benchmark across 83 gut microbiome cohorts and 20 diseases found ridge regression and random forest among the best performers, with neural networks and gradient boosting not consistently outperforming them. Deep learning architectures, including multilayer perceptrons, transformers, graph neural networks, and autoencoders, expand modeling capacity for nonlinear and structure-aware analysis, and foundation models pretrained on large-scale microbiome data, such as MGM, GenomeOcean, and Evo2, promise transferable representations, but fine-tuning on small cancer cohorts still risks overfitting, and interpretability remains limited.

The applications are already impressive. Multi-view deep learning frameworks distinguish metastatic from non-metastatic colorectal cancer using gut microbial features, while methods like GDmicro combine graph convolutional networks with domain adaptation to improve cross-cohort robustness against differences in region, diet, and sequencing protocols. Integration frameworks such as VTrans use large-scale pretraining and selective co-attention to combine microbiome features with host transcriptomics and copy-number profiles, enhancing survival risk stratification in small cohorts. Interpretability methods, from SHAP values and integrated gradients to graph-based community explanations like Micah, are evolving from simple feature ranking toward direction-aware, network-level insights. Looking ahead, causal AI frameworks such as DAG-deepVASE, which combines deep networks with knockoff features to identify nonlinear causal relationships, could move the field beyond pure association, though the review stresses that such findings remain hypothesis-generating until validated.

Mechanistic evidence is accumulating for real biological effects. Intratumoral Fusobacterium nucleatum has been implicated in promoting tumor progression, metastasis, and chemoresistance through immune modulation, autophagy activation, and oncogenic signaling, while enterotoxigenic Bacteroides fragilis promoted tumorigenesis and metastasis in breast cancer models. Gut microbes directly metabolize therapeutic drugs: bacterial β-glucuronidase reactivates the inactive metabolite of irinotecan in the gut, causing diarrhea, and inhibiting this enzyme can preserve drug efficacy while limiting toxicity. In immunotherapy, antibiotic use before immune checkpoint inhibitor treatment is associated with poorer response and survival, fecal microbiota transplantation from responders enhances anti-PD-1 responses in preclinical models, and findings from the CIAO clinical trial showed that total intratumoral bacterial abundance was the only microbiome-related feature predicting immune checkpoint blockade response in head and neck cancer, with higher abundance linked to an immunosuppressive microenvironment.

Translating these discoveries into the clinic is the final and steepest climb. The review emphasizes a three-stage evidentiary hierarchy, analytical validation, clinical validation, and demonstrated clinical utility, that most cancer microbiome biomarkers have not yet climbed. Fecal metagenomic classifiers for colorectal cancer are the most mature, retaining accuracy around an AUC of 0.8 in independent cohorts, while tumor-intrinsic signatures have faced serious challenges over host-read misclassification and normalization artifacts. Interventional strategies show encouraging but preliminary signals, with responder-derived fecal transplants reinstating anti-PD-1 responses in some ICI-refractory melanoma patients, yet a fatal transmission of a drug-resistant bacterium during fecal transplantation underscores the safety stakes. The path forward, the authors argue, requires standardized reporting under frameworks like STORMS, contamination-aware pipelines, prospective multicenter validation, and AI systems treated as prioritization tools rather than oracles. Emerging paradigms, including agentic AI systems like Eubiota, human-in-the-loop frameworks, and digital twin approaches, may eventually knit microbiome data into iterative clinical translation, but only if every model is paired with uncertainty estimation and external validation. The era of the cancer microbiome is no longer a question of whether microbes matter in oncology, but of whether the field can prove it rigorously enough for patients to benefit.

Subject of Research: Integration of multi-omics technologies and artificial intelligence for decoding the cancer microbiome and translating discoveries into clinical oncology applications

Article Title: Decoding the cancer microbiome: multi-omics, AI, and translational opportunities

Article References: Decoding the cancer microbiome: multi-omics, AI, and translational opportunities. (n.d.). https://doi.org/10.1186/s13059-026-04284-8

Image Credits: AI Generated

DOI: 10.1186/s13059-026-04284-8

Keywords: cancer microbiome, tumor-associated microbiome, gut microbiome, multi-omics, artificial intelligence, deep learning, biomarker discovery, immunotherapy response, Fusobacterium nucleatum, metagenomics, precision oncology, clinical translation