Microbiota and head and neck cancers: Mechanistic insights and translational potential

ABSTRACT

Head and neck cancers (HNCs) impose a substantial global health burden, particularly in Asia and China. Emerging evidence highlights the pivotal role of oral, pharyngeal, and intratumoral microbiota in HNC pathogenesis. Microbial dysbiosis promotes tumor initiation and progression through interconnected mechanisms, including chronic inflammation, immune modulation, and metabolic reprogramming. Specifically, dysbiosis induces a pro-tumorigenic microenvironment by triggering persistent inflammation, disrupting immune surveillance, and generating oncogenic metabolites or depleting protective compounds. Comparative analyses reveal both shared and site-specific microbial signatures, reflecting complex host – microbe interactions. Notably, the microbiota holds promise as a source of noninvasive biomarkers for early detection and prognosis, and may influence responses to immunotherapy and chemoradiotherapy. However, challenges remain in methodological standardization and in establishing causal relationships. Integrating multi-omics approaches with functional studies will be essential to advance precision diagnostics and microbiota-targeted strategies for personalized HNC management.

GRAPHICAL ABSTRACT

Head and Neck Microbiota—Sites, Mechanisms, and Translational Opportunities. Site-specific microbiota dysbiosis drives carcinogenesis through inflammation, metabolic byproducts, and TME remodeling. These mechanistic insights provide new opportunities for biomarker discovery and microbiota-targeted clinical interventions.

Introduction

Head and neck cancers (HNCs) represent a heterogeneous group of malignancies arising from the oral cavity, pharynx, larynx, and other adjacent anatomical sites [1]. These cancers collectively constitute a major global health challenge, with a particularly high incidence and mortality rate in certain regions, notably Asia, where nasopharyngeal carcinoma (NPC) is endemic [2]. Despite significant advancements in the diagnostic techniques and treatment modalities for HNC, including surgery, radiotherapy, chemotherapy, and immunotherapy, the prognosis for many patients remains unsatisfactory. This is mainly attributed to challenges such as late diagnosis, high recurrence rates, and treatment resistance [3,4]. Traditionally, the occurrence of HNC was believed to be mainly associated with risk factors such as tobacco use, alcohol consumption, and virus infection. However, current research has gradually revealed that the pathogenesis of HNC is far more complex than previously thought which has extended to the intricate interactions between the tumor microenvironment (TME) and the host resident microbial communities. Emerging research has also found that imbalances in the oral and intestinal microbiome can be involved in carcinogenesis by promoting chronic inflammation and abnormal immune regulation [5–7].

The human body harbors trillions of microorganisms, collectively known as the microbiota, which reside in various anatomical niches, including the oral cavity, nasopharynx, and larynx [8–10]. These microbial communities, once considered mere commensals, are now understood to play profound roles in maintaining host health, influencing immune system development, and regulating metabolic processes [11,12]. Dysbiosis, an imbalance in the composition or function of these microbial communities, has been implicated in the etiology and progression of numerous systemic diseases, including various forms of cancer [13–15]. In the context of HNCs, the unique anatomical location of these tumors places them in direct contact with a rich and diverse microbiota, particularly the oral and pharyngeal microbiota, which serves as a primary interface between the host and the external environment [16,17]. Emerging evidence strongly suggests that both the commensal microbiota of the oral and pharyngeal regions and the intratumoral microbiota – microbes residing directly within cancer tissues – are critical, yet often overlooked, contributors to HNC initiation, progression, and therapeutic response [18,19].

This review aims to systematically synthesize the latest advancements in understanding the multifaceted roles of microbiota in HNCs, with a specific focus on oral squamous cell carcinoma (OSCC), NPC, and Laryngeal Cancer (LSCC). We will delve into the distinct characteristics of microbial communities associated with each of these cancer types, exploring how their dysregulation contributes to carcinogenesis through mechanisms involving chronic inflammation, immune modulation, the production of metabolic byproducts, and the intricate shaping of the TME. A comparative analysis will highlight both the shared microbial features and the unique site-specific patterns observed across these diverse HNCs, providing a comprehensive understanding of host-microbe interactions in different anatomical niches. Furthermore, we will discuss the significant translational potential of microbiota research in head and neck oncology, including their utility as noninvasive biomarkers for early detection and prognostic assessment, and their promise as novel therapeutic targets to enhance immune responses and improve the efficacy of conventional treatments such as chemotherapy and radiation therapy. We will also explore innovative strategies centered on harnessing microbes or their metabolites for precision medicine approaches tailored to individual patients. Finally, this review will critically address the current challenges that impede progress in this rapidly evolving field, such as the imperative for standardizing sample collection protocols, establishing robust causal relationships between specific microbes and cancer, and effectively accounting for the substantial interindividual variability in microbial compositions. By addressing these critical gaps, this review seeks to lay a solid theoretical foundation for future mechanistic investigations and accelerate clinical translation efforts, ultimately leading to improved patient outcomes in head and neck oncology.

Unique characteristics of microbiota in HNCs

The oral cavity, nasopharynx, and larynx each harbor distinct microbial ecosystems, which undergo significant alterations during the development and progression of cancer. Understanding these site-specific dysbiotic patterns is crucial for deciphering their roles in HNC pathogenesis.

Oral cancer (OSCC) microbiota

OSCC is the most prevalent type of HNC, and its development is strongly associated with profound alterations in the oral microbiome [20,21]. Numerous studies employing high-throughput sequencing technologies, such as 16S rRNA gene sequencing and shotgun metagenomics, have consistently revealed a shift from a healthy, diverse microbial community to a dysbiotic state characterized by reduced diversity and an enrichment of specific pathogenic bacteria in OSCC patients and tumor tissues [22,23]. For instance, Zhang et al. conducted a study on 50 OSCC patients and found that microbial richness and diversity were significantly higher in tumor sites compared to adjacent normal tissues [24], with an enrichment of six bacterial families and 13 genera, including Fusobacterium, Alloprevotella, and Porphyromonas. This finding contrasts with some earlier studies that reported decreased diversity in cancer, highlighting the complexity and potential heterogeneity in findings depending on the specific cohort, sampling site (e.g. saliva vs. tissue), and sequencing depth.

The oral mycobiome refers to the microbial community colonizing the mucosal surfaces of the oral cavity and pharynx such as saliva, gingival crevices, the tongue surface, and tonsils. The common sampling methods included saliva collection, oral swab wiping, mouthwash collection, and gingival sulcus fluid extraction. Key bacterial species frequently implicated in OSCC include Porphyromonas gingivalis (P. gingivalis), Fusobacterium nucleatum (F. nucleatum), and certain Streptococcus species [20,25,26]. P. gingivalis, a well-known periodontal pathogen, has been consistently linked to OSCC. Studies have shown its ability to promote tumor progression, increase cellular invasion, and enhance cancer stem cell (CSC) proliferation [21,27]. Similarly, F. nucleatum is frequently enriched in OSCC tumor samples and has been shown to promote epithelial-mesenchymal transition (EMT), a critical process in cancer metastasis, through pathways involving INHBA-dependent SMAD/Laminin332/PI3K/AKT signaling [28,29]. The presence of F. nucleatum has also been associated with increased inflammation and poor prognosis in HNSCC more broadly [30]. A Mendelian randomization (MR) study provided compelling evidence for a causal association [31], identifying Veillonella as deleterious and Prevotella as protective against oral cancer, a significant advantage over purely observational studies in establishing causality. Since genotypes are usually determined before the occurrence of diseases and are less susceptible to reverse causality or environmental confounding, MR Can effectively overcome the confounding bias and reverse causality problems commonly found in traditional observational studies, thereby providing more reliable causal inference evidence. However, it is important to note that MR studies rely on genetic variants as instrumental variables, and their findings, while robust, still require validation through functional experiments. Other notable bacterial shifts in OSCC include an increase in Capnocytophaga, Flavobacteriaceae, and Vibrionaceae, through 16S rDNA sequencing of oral microbiota in OSCC patients [32]. Conversely, some studies have reported a decrease in beneficial or commensal bacteria. For instance, Frank et al. observed elevated Lactobacillus spp [33]. and diminished Neisseria spp. in HNSCC, suggesting that this dysbiosis actively promotes tumorigenesis. Previous study highlighted that oral mucosal dysplasia and oral cancer are associated with dysbiotic changes in both bacteriome and mycobiome [23], suggesting a broader microbial involvement beyond bacteria. Monteiro et al. further emphasized the less understood but potentially significant role of fungi in oral cancer [34], noting the ambiguity of its role in carcinogenesis and the potential for reactivating the mycobiome during cancer treatment. These studies underscore the need for comprehensive analyses that include both bacterial and fungal communities to fully understand the oral microbiome’s impact on OSCC.

Unlike the oral microbiota, the microbiota within tumors is regarded as an inherent component of the tumor microenvironment and can directly influence the biological behavior and treatment response of tumors [19,35]. Intratumoral microbiome refers to the microbial communities present within the tumor tissue, including those that are infiltrated into the tumor cells, the tumor stroma, and the tumor microenvironment. In OSCC, they need to be obtained through tumor tissue biopsy or surgical resection. Schmidt et al. similarly found decreased Firmicutes (specifically Streptococcus) and Actinobacteria (Rothia) in cancer samples compared to normal tissue [36], with these decreases observed even in pre-cancers. Luo et al. emphasized that intralesional microbiota are important components of the TME and may contribute to carcinogenesis [18]. Specifically, Fusobacterium and Treponema species in OSCC were found to infect macrophages and aneuploid cancer epithelial cells, upregulating cancer progression pathways [37]. This cutting-edge research provides unprecedented spatial and cellular resolution, demonstrating how intratumoral bacteria directly interact with host cells and influence their transcriptional programs, thereby contributing to tumor heterogeneity.

NPC microbiota

NPC exhibits one of the most striking geographic distributions among all cancers, with over 80% of cases occurring in East Asia, Southeast Asia, and North Africa [5,38]. There are unique geographic and ethnic risks associated with NPC. The disease occurs frequently in Asia, with China having a high rate. Chinese heritage is a risk factor on its own, no matter the location of residence. People with Native Alaskan heritage have also shown a high incidence of NPC [39]. The nasopharyngeal and oral microbiota are increasingly recognized as crucial factors influencing NPC development and progression. A comprehensive review highlights the significant role of host microbiota [40], including both nasopharyngeal and oral communities, in NPC development. They specifically linked bacterial strains such as Neisseria, Staphylococcus, and Fusobacterium to NPC. Studies have consistently shown altered microbial composition and diversity in NPC patients. Hao et al. observed lower nasal cavity and nasopharynx microbial diversities and distinct oral microbiome beta diversity in NPC patients compared to healthy individuals [41]. Their study, which aimed to develop noninvasive screening tools, identified higher Granulicatella abundance in the oral cavity and lower Pseudomonas and Acinetobacter in the nasopharynx of NPC patients, achieving high accuracy in screening. This highlights the potential of oral and nasopharyngeal microbiome as diagnostic biomarkers.

The oral microbiota’s influence on NPC extends to its interaction with EBV. EBV is the main pathogenic factor of NPC. Patients with different EBV loads have varying degrees of intratumoral microenvironment and immune infiltration, which may lead to fundamental changes in the microbial community structure [42,43]. Liao et al. demonstrated a link between oral microbiota alteration and EBV reactivation in NPC [44]. Their study of 303 NPC patients and controls found lower microbial diversity and an imbalanced composition in NPC patients. Notably, Streptococcus sanguinis, enriched in patients, was associated with NPC risk and EBV reactivation. Mechanistically, S. sanguinis induced EBV lytic activation via hydrogen peroxide (H₂O₂), which in turn triggered host cell signaling pathways, including TNF-α via NF-κB, and EBV genome demethylation, potentially promoting tumorigenesis. This provides a crucial mechanistic link between specific oral bacteria and the viral etiology of NPC.

Familial aggregation of NPC is a known phenomenon, and recent research suggests a microbial component to this heritability. Liao et al. investigated familial aggregation in 127 families and 337 individuals [45], finding significant microbial similarity among family members. They identified three NPC-enriched taxa with high heritability—Gemella sp., Lautropia mirabilis, and Streptococcus sp.—suggesting their role in familial aggregation and potential for identifying high-risk individuals. This study uniquely combines host genetics and microbiome analysis, providing a novel perspective on NPC risk.

The intratumoral microbiota in NPC also presents unique characteristics. The large-scale studies offer robust evidence for the prognostic value of intratumoral microbiota in NPC. For example, a retrospective study of 491 NPC patients reveals significant differences in microbial composition [46], alpha diversity, and beta diversity between NPC and normal tissues. They identified a bacterial signature of four risk genera (Bacteroides, Alloprevotella, Parvimonas, and Dialister) that predicted shorter disease-free survival (DFS), distant metastasis-free survival (DMFS), and overall survival (OS). The strength of this study lies in its large cohort and integrated multi-omics analysis, which indicated deficient immune infiltration in NPC tissues with abundant risk bacteria, suggesting a direct role of tissue-resident microbiota in NPC prognosis. Similarly, Qiao et al. [47], in a large retrospective cohort study of 802 NPC patients, found intratumoral microbiota, particularly Corynebacterium and Staphylococcus, associated with poor prognosis. They further established that nasopharyngeal microbiota was the main source of intratumoral bacteria and that a higher bacterial load was negatively associated with T-lymphocyte infiltration, providing a crucial link to immune evasion. After adjusting for TNM staging, bacterial load remained an independent prognostic factor. Zhang et al. further characterized intratumoral microbiota in NPC [48], identifying Proteobacteria, Actinobacteria, Firmicutes, and Bacteroidetes as key phyla. Their findings indicated that these phyla are sensitive to systemic inflammation and immune status, with Proteobacteria being crucial in NPC inflammation and Bacteroidetes sensitive to tumor immune status. They also noted that Corynebacterium and Brevundimonas impacted tumor T stage, with higher Corynebacterium associated with better prognosis, which contrasts with Qiao et al.’s finding of Corynebacterium being associated with poor prognosis [47]. Such discrepancies highlight the need for further research to reconcile findings and understand the nuances of specific species within genera. Dietary habits, environmental exposure and host genetic background in different regions may lead to significant differences in the baseline nasopharyngeal microbiome. Future research should go beyond the descriptive level at the genus level and delve into the functional and molecular mechanisms of the different strain types under the same background, in order to provide a reliable basis for precise diagnosis and treatment of NPC based on the microbiome. For instance, using organoids derived from NPC patients, different Corynebacterium strains isolated were inoculated, and their direct effects on tumor cell proliferation, invasion, and epithelial-mesenchymal transition (EMT) were observed. At the same time, the intratumoral metabolites were detected to determine whether Corynebacterium produced specific metabolites, such as short-chain fatty acids and tryptophan metabolites to regulate immunity or tumor growth.

LSCC microbiota

LSCC is another significant HNC, and its association with microbiota dysbiosis is increasingly recognized. Similar to oral and nasopharyngeal cancers, the microbial communities in the larynx and adjacent oral cavity undergo considerable changes in LSCC patients. An early study compares throat and tissue microbiota in laryngeal carcinoma patients and controls [49]. They identified Firmicutes, Fusobacteria, Bacteroidetes, and Proteobacteria as main phyla, with Streptococcus, Fusobacterium, and Prevotella as predominant genera. Crucially, they found significant differences in bacterial profiles between LSCC patients and controls, suggesting a potential link. A follow-up study further detailed these alterations [50], observing higher alpha-diversity in tumor tissues and adjacent normal tissues of cancer patients compared to controls with vocal cord polyps, and identifying seven bacterial genera with differential abundance.

More recent studies have focused on specific bacterial species and their mechanistic roles. F. nucleatum has emerged as a key player in LSCC, similar to its role in OSCC. Hsueh et al. reported high enrichment of F. nucleatum in LSCC [51], associating it with poor prognosis. Mechanistically, F. nucleatum activated innate immune signaling, leading to increased miR-155-5p and miR-205-5p, which suppressed ADH1B and TGFBR2, thereby reprogramming ethanol metabolism. This reprogramming promoted the PI3K/AKT signaling, EMT, and ultimately LSCC progression and metastasis. This study provides a detailed molecular pathway linking a specific bacterium to LSCC progression, highlighting F. nucleatum levels as a potential prognostic biomarker. Ren et al. further supported the oncogenic role of F. nucleatum in LSCC [52], demonstrating that it promotes carcinogenesis by upregulating YWHAZ, which in turn activates the Akt signaling to enhance proliferation, migration, and colony formation while inhibiting apoptosis in LSCC cells. The dependence of F. nucleatum’s oncogenic effects on YWHAZ was confirmed by YWHAZ knockdown experiments, providing strong evidence for this specific molecular pathway. Yuan et al. added another layer of mechanistic understanding by showing that F. nucleatum promotes JAK3 gene expression through decreased methylation in LSCC [53], further correlating its abundance with tumor progression. These identified epigenetic regulatory mechanisms open new possibilities for mechanistic studies to pursue novel therapeutic strategies.

The oral microbiome, as a readily accessible proxy for the upper aerodigestive tract, also shows promise for LSCC diagnosis. Yu et al. demonstrated that oral rinse microbiome analysis using 16S rRNA gene sequencing could non-invasively diagnose LSCC with 85.7% accuracy [54]. Their study of 77 LSCC patients and 76 controls revealed significant differences in bacterial genera, identifying the oral rinse microbiome as a potential noninvasive biomarker. This offers a practical advantage for early detection strategies. A systematic review by Lechien synthesized findings from 10 studies on glottic and LSCCs, consistently reporting lower bacterial diversity in LSCC tissues [55]. Bacteroidetes (Prevotella) and Fusobacteriota (Fusobacterium) were frequently overrepresented in LSCC, while Firmicutes (Stomatobaculum longum, Abiotrophia, Gemella, Streptococcus) and Actinobacteria (Actinomyces, Corynebacterium, Rothia mucilaginosa) were predominant in controls. This review underscores the consistent dysbiotic patterns in LSCC, despite methodological heterogeneity across studies. Fu further strengthened the causal link through a two-sample Mendelian randomization analysis in East Asian populations [56], identifying 14 causal associations between oral microbial taxa and pharyngeal-LSCC risk, with seven taxa increasing risk and seven showing protective effects. This robust methodology provides strong evidence for the direct involvement of oral microbiota in LSCC pathogenesis. The presence and activity of intratumoral microbes can significantly impact the TME and host immune responses. Hamada et al. examined microbiome profiling in HNSCC using TCMA and The Cancer Genome Atlas [57], finding that Actinomyces, Fusobacterium, and Rothia profiles differed between normal and tumor tissue. They also observed that higher OS rates correlated with Leptotrichia over the median, suggesting a potential protective role for certain intratumoral bacteria. This highlights the dual nature of the tumor microbiome, where some species may be pro-tumorigenic while others might be associated with better outcomes. The challenges in studying intratumoral microbiota, particularly its low biomass, mean that advanced techniques like those discussed by Lu et al. are essential for robust and reproducible findings [19].

Mechanistic contributions to cancer initiation and progression

The intricate interplay between the microbiota and host cells in HNCs is mediated through several key mechanisms, including chronic inflammation, immune modulation, metabolic byproducts, and the direct shaping of the TME. These mechanisms are not mutually exclusive but rather form a complex network that collectively drives cancer initiation and progression.

Chronic inflammation and immune modulation

Chronic inflammation is a well-established hallmark of cancer, providing a fertile ground for tumor initiation and progression⁵. The dysbiotic microbiota in HNCs, particularly in the oral cavity, can trigger and sustain inflammatory responses that contribute to carcinogenesis. Pathogenic bacteria, often enriched in the HNC microenvironment, possess virulence factors that activate host immune cells, leading to the release of pro-inflammatory cytokines and chemokines (Figure 1).

Figure 1.

Microbiota-driven chronic inflammation and immune modulation in head and neck cancers. Dysbiotic oral and nasopharyngeal bacteria such as P. gingivalis, F. nucleatum, and Streptococcus sanguinis activate inflammatory pathways and perturb local immune responses. Chronic cytokine signaling and immune suppression collectively shape a pro-tumorigenic microenvironment.

P. gingivalis and F. nucleatum are two prominent oral pathogens frequently implicated in driving inflammation in OSCC. Gallimidi et al. provided early mechanistic evidence [27], demonstrating that chronic infection with P. gingivalis and F. nucleatum promotes OSCC in a murine model of periodontitis-associated oral tumorigenesis, specifically through augmented IL-6-STAT3 signaling. They showed that these pathogens stimulate tumorigenesis via Toll-like receptors (TLRs), promoting human OSCC proliferation and the expression of key tumorigenesis molecules. This highlights a direct link between bacterial infection, inflammation, and cancer development. Kamarajan et al. further elucidated this mechanism [58], showing that these periodontal pathogens promote OSCC aggressivity via TLR/MyD88-triggered activation of Integrin/FAK signaling, which enhances OSCC cell migration, invasion, and tumorigenesis. The ability of these bacteria to activate innate immune signaling pathways like TLRs is a critical step in initiating and perpetuating inflammation.

In NPC, the interaction between microbiota, inflammation, and the immune system is particularly complex due to the strong association with EBV. Liao et al. showed that Streptococcus sanguinis, enriched in NPC patients, induces EBV lytic activation via H₂O₂ [44], which then triggers host cell signaling pathways, including TNF-α via NF-κB. NF-κB is a central regulator of inflammatory and immune responses, and its constitutive activation is a hallmark of many cancers, including NPC [59]. This suggests a microbial-viral synergy in driving inflammation and oncogenesis. Zhang et al. further linked intratumoral microbiota in NPC to systemic immune inflammation [48], finding that phyla like Proteobacteria and Firmicutes differed significantly in inflammation groups, while Actinobacteria and Bacteroidetes varied with platelet-related indices and NAR groups, respectively. This indicates that the intratumoral microbial composition can reflect and potentially influence the host’s inflammatory and immune status.

The microbiota also plays a crucial role in immune modulation, influencing both innate and adaptive immune responses within the TME. Dysbiosis can lead to an immunosuppressive environment that favors tumor growth and immune evasion. Qiao et al. found that a higher intratumoral bacterial load in NPC was negatively associated with T-lymphocyte infiltration [47], suggesting that certain bacteria might directly or indirectly suppress anti-tumor immune responses. Tan et al. similarly reported deficient immune infiltration in NPC tissues with abundant risk bacteria [46], further supporting the notion that intratumoral microbiota can contribute to an immunosuppressed TME. This is a critical finding, as T-lymphocyte infiltration is a key determinant of response to immunotherapy [60–62].

Beyond direct immune cell modulation, the microbiota can influence the expression of immune checkpoints and the overall immune landscape. Bullman revealed that bacterial colonization in HNCs is associated with myeloid cell infiltration [37], MAPK and immune checkpoint upregulation, and reduced T-cell levels. This suggests that intratumoral microbes can create a TME that is conducive to immune evasion, potentially by upregulating inhibitory pathways. Hu et al. investigated tumor-immune-microbe interactions in HPV-negative HNSCC [63], finding that microbial signatures distinguished Hypoxia/Immune phenotypes and that certain microbes correlated with immune processes, impacting survival differences. This highlights the potential of microbiota to influence the immune contexture of tumors, which is particularly relevant for HPV-negative HNSCC, known for worse outcomes.

The impact of microbiota on immune responses extends to therapeutic outcomes. Matson et al. broadly reviewed how commensal microbiota influences cancer [64], immune responses, and immunotherapy, noting that both pro- and anti-tumor effects exist and that microbial composition can affect immunotherapy efficacy. Lim et al. specifically highlighted that microbiota composition and tumor-infiltrating microbiota numbers influence NPC prognosis and treatment response [65], including to chemotherapy and immunotherapy. This suggests that modulating the microbiota could be a strategy to enhance anti-tumor immunity and improve therapeutic efficacy.

Metabolic byproducts and carcinogenesis

Microbial metabolism generates a diverse array of byproducts, many of which can directly influence host cell behavior, contributing to carcinogenesis or modulating tumor progression. These metabolites can act as signaling molecules, genotoxins, or modulators of host metabolic pathways. Duan et al. emphasized that metabolites from both gut and intratumoral microbiota significantly impact cancer initiation and progression [66], although their dual roles (pro- or anti-tumorigenic) often hinder clinical translation.

One prominent example is the production of acetaldehyde. Smędra et al. proposed that oral auto-brewery syndrome due to microbiota dysbiosis leads to acetaldehyde accumulation [67], which is a known carcinogen and a risk factor for OSCC. This hypothesis suggests that even in the absence of alcohol consumption, dysbiotic oral microbiota can produce acetaldehyde from nonalcoholic foods or drinks, contributing to cancer development. This mechanism highlights a direct chemical pathway by which microbes can promote carcinogenesis.

Short-chain fatty acids (SCFAs), such as acetate, propionate, and butyrate, are well-known microbial metabolites, primarily produced by gut bacteria, but also by oral microbes. Their roles in cancer are complex and context-dependent. Liu et al. investigated microbiota and metabolomic profiling in NPC patients stratified by radiotherapy response [68]. They found elevated acetate levels in radiotherapy-non-responsive (NR) groups, suggesting that specific microbial metabolites might influence treatment outcomes. This implies that targeting the microbiota-metabolite axis could enhance radiotherapy efficacy. Praveen et al. found that oral cancer patients had higher Streptococcus and Parvimonas abundance, lower Corynebacterium and Prevotella [69], and higher fatty acid oxidation enzyme levels, including CPT1A, compared to controls. Conversely, healthy controls had increased SCFAs and CD4+ T-helper cell counts. This suggests a shift in metabolic profiles associated with oral cancer, where reduced SCFAs might contribute to an environment less conducive to anti-tumor immunity. Nouri et al. similarly observed reduced SCFAs and FFAR2 expression, along with increased TNFAIP8, IL-6, and STAT3 levels, in cancer patients, further linking oral microbiota alterations to cancer development through inflammatory path ways influenced by metabolic changes [70].

Beyond SCFAs, other microbial metabolites can play critical roles. Guo et al. demonstrated that F. nucleatum promotes radioresistance and tumor growth in NPC by slowing mitochondrial dysfunction and inhibiting PANoptosis via the SLC7A5/leucine-mTORC1 axis [71]. Crucially, leucine restriction mitigated Fn-induced radioresistance and tumor progression. This study identified a specific amino acid metabolite (leucine) and a microbial-host metabolic pathway that can be targeted therapeutically. This is a significant advancement as it moves beyond identifying mere associations to elucidating specific metabolic dependencies. Guo et al. also showed that exosomes derived from F. nucleatum-infected colorectal cancer cells facilitate tumor metastasis by selectively carrying specific miRNAs (miR-1246/92b-3p/27a-3p) and CXCL16 [72]. While this study focused on colorectal cancer, it illustrates a general mechanism by which F. nucleatum can influence host cell behavior through exosomal cargo, which could be relevant to HNCs given the bacterium’s prevalence.

The interaction between microbial metabolites and host genetics and epigenetics is also a burgeoning area. Cai et al. conducted an integrative analysis of oral microbiota dysbiosis and host genetic and epigenetic aberrations in OSCC [28]. They found that seven enriched bacterial species in the TME correlated with 206 upregulated host genes. Their analysis identified 20 dysregulated host genes with inverted CpG methylation, and F. nucleatum was confirmed to play a role in cellular invasion via E-cadherin/β-catenin signaling, TNFα/NF-κB pathway, and extracellular matrix remodeling. This multi-omics approach highlights how microbial presence can epigenetically modify host gene expression, leading to oncogenic changes.

Shaping the TME

The TME is a complex ecosystem comprising cancer cells, immune cells, stromal cells, extracellular matrix, and signaling molecules, all of which interact to influence tumor progression and response to therapy [73–75]. The microbiota, both commensal and intratumoral, is increasingly recognized as a crucial component that actively shapes the TME, influencing its cellular composition, metabolic landscape, and immune suppressive properties (Figure 2).

Figure 2.

Microbiota–tumor microenvironment interactions. Commensal and intratumoral bacteria modulate immune suppression, metabolic reprogramming, stromal remodeling, and cancer stemness within the TME. These microbe-driven alterations enhance tumor progression, invasion, and therapeutic resistance.

One of the primary ways microbiota shapes the TME is by inducing chronic inflammation, as discussed previously. This inflammation recruits various immune cells, including myeloid-derived suppressor cells (MDSCs) and tumor-associated macrophages (TAMs), which often adopt pro-tumorigenic and immunosuppressive phenotypes [76–78]. Bullman demonstrated that bacterial colonization in HNCs is associated with myeloid cell infiltration [37], which can contribute to an immunosuppressive TME. Upadhyay et al. found that HPV-negative HNSCC tumors had higher M0 and M2 macrophages [79–81], which are typically pro-tumorigenic and immunosuppressive, while HPV-positive tumors had more T regulatory cells and CD8+ T-cells. While this study focused on HPV status, the presence of specific microbial species was also significantly associated with survival, suggesting that the microbiota could influence the balance of these immune cell populations.

The microbiota can also directly impact the physical and metabolic characteristics of the TME. Deluca et al. noted that oral cavity cancers are often more hypoxic than tumors in other subsites [82], and distinct microbial populations are over-represented in these hypoxic tumors. Hypoxia is a common feature of the TME that promotes angiogenesis, metastasis, and resistance to radiotherapy and chemotherapy [83]. The presence of specific anaerobic bacteria, which thrive in low-oxygen environments, could exacerbate or be a consequence of tumor hypoxia, creating a vicious cycle that favors tumor progression. Ashique et al. also listed tumor hypoxia as one of the mechanisms by which host microbiota contributes to NPC development [40].

Beyond immune cells and hypoxia, the microbiota can influence the stromal components of the TME. Cai et al. showed that F. nucleatum contributes to extracellular matrix remodeling in OSCC [28], a process critical for tumor invasion and metastasis. This suggests that microbial activity can directly alter the structural components of the TME, facilitating cancer progression. The study by Gong et al. using comprehensive single-cell sequencing in NPC revealed distinct features in the TME compared to nonmalignant environments [84], including phenotypic abundance, genetic alterations, and immune dynamics. While this study focused on host stromal cells, it provides a high-resolution map of the TME that future studies can integrate with microbial data to understand how microbes influence these stromal dynamics.

The microbiota’s influence on the TME extends to its impact on Cancer stem cells (CSCs) and tumor plasticity. Salem et al. highlighted the pivotal role of CSCs in HNSCC tumorigenicity, metastasis [85], therapy resistance, and relapse, and specifically considered the oral microbiota’s contribution to CSC plasticity. Saikia et al. demonstrated that F. nucleatum and its LPS activate a niche defense (TSD) phenotype in oral cancer CSCs in vitro [21], contributing to oral cancer progression and relapse. This suggests that specific microbial components can directly interact with CSCs, enhancing their stemness and resistance to therapy, thereby profoundly shaping the TME’s aggressive potential. Mehta et al. developed 3D stem-like spheroids-on-a-chip for personalized combinatorial drug testing in oral cancer [86], which could serve as valuable preclinical models for studying microbiota-CSC interactions in a more realistic TME representation.

The concept of the “cancer ecology theory” for NPC, proposed by Luo [87], views cancer as a multidimensional spatiotemporal “unity of ecology and evolution” pathological ecosystem. In this framework, NPC cells are invasive species, and metastasis is multidirectional ecological dispersal. The tumor-host interface is considered an ecological transition zone with significant edge effects. This ecological perspective provides a holistic view of how the microbiota, as a key component of this ecosystem, interacts with cancer cells and other host elements to drive disease progression. Understanding these ecological dynamics is crucial for developing effective interventions that target the TME.

Comparative analysis of microbiota across HNC types

While the oral cavity, nasopharynx, and larynx are anatomically contiguous, the specific microbial communities and their roles in carcinogenesis exhibit both shared features and distinct patterns across OSCC, NPC, and LSCC. A comparative analysis helps to identify universal mechanisms and site-specific vulnerabilities (Figure 3).

Figure 3.

Comparative microbial signatures across OSCC, NPC, and LSCC. Site-specific microbial patterns, major risk exposures, and distinctive pathogenic pathways characterize OSCC, NPC, and LSCC. Shared oncogenic taxa such as F. nucleatum coexist with unique microbe–host interactions shaped by each anatomical niche.

Shared microbial signatures and commonalities

Despite the anatomical and etiological differences among OSCC, NPC, and LSCC, certain microbial signatures and mechanisms appear to be common across these HNC types. The most consistently implicated bacterium across all three HNCs is F. nucleatum. In OSCC, F. nucleatum is enriched in tumor tissues and promotes EMT and invasion [28,29,88]. In NPC, it is linked to development and poor prognosis [40]. In LSCC, F. nucleatum is highly enriched, associated with poor prognosis, and promotes progression and metastasis through ethanol metabolism reprogramming and YWHAZ upregulation [51,53]. This consistent presence and oncogenic role of F. nucleatum across diverse HNCs suggest a fundamental, shared mechanism by which this bacterium contributes to head and neck carcinogenesis, likely involving chronic inflammation, immune modulation, and metabolic alterations. Qiao et al. conducted a multi-omics integration study using The Cancer Microbiome Atlas (TCMA) and found that Fusobacterium was enriched in HNSC tissues overall [30], linked to increased inflammation and poor prognosis, further supporting its pan-HNC oncogenic role. It emphasizes that the carcinogenic effect of Fusobacterium nucleatum is universal across subtypes, which provides a theoretical basis for the development of microbiome-targeted therapeutic strategies for head and neck cancer as a whole.

Another common theme is the general concept of dysbiosis, characterized by a shift from a diverse, balanced microbial community to one dominated by a few pathogenic species, often accompanied by a reduction in overall diversity. This pattern has been observed in OSCC [22,24], NPC (lower diversity in nasal cavity and nasopharynx) [41,44], and LSCC (lower bacterial diversity in tissues) [55]. This reduction in diversity and enrichment of pathobionts is a broad indicator of an unhealthy microbial ecosystem that may contribute to chronic inflammation and immune dysregulation, creating a permissive environment for cancer development. Frank et al. broadly stated that a dysbiotic oral microbiota [33], including elevated Lactobacillus spp. and diminished Neisseria spp., promotes HNSCC, suggesting a generalizable pattern of dysbiosis across the head and neck region.

The involvement of P. gingivalis is also a recurring theme, particularly in OSCC, where it promotes tumor progression and enhances CSC proliferation [21,27,89]. While its role is most strongly established in OSCC, its general association with periodontal disease, a risk factor for HNCs, suggests its potential indirect involvement across the region. Irfan et al. broadly linked P. gingivalis to oro-digestive cancer [20], indicating its broader oncogenic potential.

Mechanistically, chronic inflammation and immune modulation appear to be shared pathways. The activation of the TLRs, NF-κB signaling, and the subsequent release of pro-inflammatory cytokines are common responses to dysbiotic microbiota across HNCs, contributing to a pro-tumorigenic microenvironment [27,28,44]. The impact on immune cell infiltration and function, leading to immunosuppression, is also a shared characteristic, as seen in NPC where high bacterial load correlates with reduced T-lymphocyte infiltration [46,47]. This suggests that targeting inflammatory pathways or modulating immune responses could be a universally applicable therapeutic strategy across HNCs.

The concept of intratumoral microbiota influencing prognosis and treatment response is also a shared feature. While specific genera may differ, the general principle that bacteria residing within the tumor can impact disease outcome and immune infiltration has been demonstrated in both NPC and OSCC [18,37,46,47]. This highlights the importance of considering the tumor’s intrinsic microbial ecosystem, not just the surrounding commensal communities, in understanding HNC biology.

Site-specific microbial patterns and distinctions

Despite shared commonalities, each HNC type also exhibits distinct microbial patterns and unique interactions that reflect the specific anatomical niche and etiological factors.

In OSCC, the oral cavity’s direct exposure to environmental carcinogens and its unique microbial diversity leads to specific dysbiotic signatures. While F. nucleatum and P. gingivalis are prominent, other species like Veillonella (deleterious) and Prevotella (protective) have been causally linked to oral cancer through Mendelian randomization, highlighting specific roles for these genera [31]. The enrichment of Capnocytophaga, Flavobacteriaceae, and Vibrionaceae in OSCC patients, and the dysregulation of metabolic pathways associated with them, as identified by Wei et al. [32], point to unique metabolic interactions within the oral TME. Furthermore, the interplay between oral microbiota and host genetic and epigenetic aberrations, such as the correlation between enriched bacteria and upregulated host genes or altered CpG methylation, as shown by Cai et al. [28], is particularly well-studied in OSCC, offering a deeper understanding of host-microbe co-evolution in this specific cancer. The role of oral fungal infections, such as Cladosporium cladosporioides, in enhancing SCC development and altering bacterial compositions, as explored by Li et al. [90], also points to a unique mycobiome involvement that might be more pronounced in the oral cavity due to its direct exposure.

NPC, with its strong association with EBV, presents unique microbial interactions. The oral microbiota’s role in mediating EBV reactivation, specifically Streptococcus sanguinis inducing lytic activation via H2O2, is a distinct mechanism observed in NPC [44]. This highlights a microbial-viral synergy that is less prominent in other HNCs. The familial aggregation of NPC, influenced by heritable oral bacteria like Gemella sp., Lautropia mirabilis, and Streptococcus sp., is another unique aspect of NPC etiology [45]. The intratumoral microbiota in NPC also shows specific prognostic signatures, such as the four-genera risk signature (Bacteroides, Alloprevotella, Parvimonas, and Dialister) identified by Tan et al. [46], and the association of Corynebacterium and Staphylococcus with poor prognosis by Qiao et al. [47]. While Corynebacterium was also found in other HNCs, its specific prognostic implications and the source from nasopharyngeal microbiota are distinct for NPC. The broader context of the nasopharyngeal microbiota as an “important window of opportunity” for understanding health and disease, as discussed by Berrington et al. [91], underscores the unique anatomical and immunological environment of the nasopharynx that shapes its microbial communities and their interactions with NPC.

In LSCC, the specific anatomical location and its exposure to inhaled irritants (tobacco smoke, alcohol) contribute to distinct microbial patterns. While F. nucleatum is a common oncopathogen, its specific mechanism in LSCC involving ethanol metabolism reprogramming via miR-155-5p/miR-205-5p and the SLC7A5/leucine-mTORC1 axis, leading to radioresistance, is a unique mechanistic insight [51,71]. The upregulation of YWHAZ by F. nucleatum in LSCC cells, promoting proliferation and metastasis, also points to a specific molecular pathway in this cancer type [52]. The diagnostic potential of the oral rinse microbiome for LSCC, achieving high accuracy in identifying specific bacterial genera, offers a noninvasive diagnostic tool that is particularly relevant for this site [54]. The systematic review by Lechien consistently highlighted the overrepresentation of Bacteroidetes (Prevotella) and Fusobacteriota (Fusobacterium) in LSCC [55], while Firmicutes and Actinobacteria were predominant in controls, providing a clear, albeit broad, site-specific microbial signature. The causal associations between specific oral microbial taxa and pharyngeal-LSCC risk identified by Fu further emphasize the distinct microbial contributions to this cancer type in East Asian populations [56]. Publication bias of populations represents an important consideration in this field. Studies reporting positive associations between specific taxa and cancer outcomes are more likely to be published than null findings. Our literature search identified only 3 studies reporting no significant microbiome-cancer associations, which may reflect underreporting of negative results. This bias could lead to overestimation of the strength and consistency of microbiome-cancer relationships. Future systematic reviews should consider incorporating unpublished data and preprints to mitigate this limitation.

Overall, while there are shared oncogenic bacteria and general principles of dysbiosis and inflammation across HNCs, the specific species involved, their precise mechanistic interactions with host cells and viruses EBV, and their prognostic or diagnostic utility often exhibit site-specific nuances. These distinctions underscore the importance of studying each HNC type individually while also seeking overarching principles. The comparative analysis reveals that while F. nucleatum is a pan-HNC oncopathogen, the specific molecular pathways it activates, or the co-factors (e.g. EBV in NPC, ethanol metabolism in LSCC), can vary significantly, leading to distinct disease phenotypes and therapeutic vulnerabilities.

Translational potential in head and neck oncology

The burgeoning understanding of the microbiota’s role in HNCs has opened exciting avenues for clinical translation, offering promising applications in early detection, prognostic assessment, therapeutic modulation, and personalized medicine.

Microbiota as biomarkers for early detection and prognosis

One of the most immediate and impactful translational applications of microbiota research is the development of noninvasive biomarkers for early cancer detection and prognostic assessment. The accessibility of the oral cavity and nasopharynx makes their microbial communities ideal candidates for such applications (Figure 4).

Figure 4.

Clinical utility of the microbiota for early detection and prognosis. Oral and nasopharyngeal microbiota can be sampled, sequenced, and profiled to derive diagnostic and prognostic microbial signatures. These biomarkers support risk stratification and guide clinical decision-making in head and neck cancer management.

For early detection, several studies have demonstrated the potential of oral and nasopharyngeal microbiota as diagnostic tools. Hao et al. developed noninvasive NPC screening tools based on nasopharyngeal and oral microbiomes [41], achieving 88% and 77.2% accuracies, respectively. They identified higher Granulicatella abundance in the oral cavity and lower Pseudomonas and Acinetobacter in the nasopharynx as key indicators. This study’s strength lies in its focus on screening tools, which could significantly improve early detection and reduce public health costs, especially in endemic regions. Similarly, Yu et al. showed that oral rinse microbiome analysis could diagnose LSCC with 85.7% accuracy [54], identifying specific bacterial genera as key features. The noninvasive nature of oral rinse samples makes this approach highly practical for widespread screening. Lim et al. developed an oral microbiome biomarker panel for predicting oral cavity and oropharyngeal cancers from oral rinse samples [92], achieving 100% sensitivity and 90% specificity. This high performance, even in a pilot study, suggests significant potential for clinical diagnostic workflows. The use of saliva metabolic profiling, as demonstrated by Song et al. [93], which achieved 86.7% accuracy in OSCC diagnosis, further supports the utility of noninvasive biological fluids for early detection, although this focuses on metabolites rather than the microbes themselves. The advantage of these approaches is their non-invasiveness and potential for high throughput, making them suitable for population-level screening.

For prognostic assessment, specific microbial signatures have been linked to patient outcomes, including DFS, DMFS, and OS. In NPC, Tan et al. identified a four-genera bacterial signature (Bacteroides, Alloprevotella, Parvimonas, and Dialister) within NPC tissues that predicted shorter DFS, DMFS, and OS [46]. The large retrospective cohort (491 patients) and integrated multi-omics analysis provide robust evidence for this prognostic utility. Qiao et al. [47], in an even larger cohort (802 NPC patients), found that high intratumoral bacterial load, particularly of Corynebacterium and Staphylococcus, was associated with poor prognosis (lower DFS, DMFS, and OS). These findings highlight the potential of intratumoral microbiota as powerful prognostic biomarkers, offering clinicians valuable information for risk stratification and treatment planning.

In OSCC, Praveen et al. found that higher abundance of Streptococcus and Parvimonas correlated with poorer DFS and OS rates [69]. This suggests that specific oral microbial shifts can serve as predictive markers for disease progression and survival. Similarly, in LSCC, F. nucleatum levels have been identified as a potential prognostic biomarker, with high enrichment associated with poor prognosis [51]. Hamada et al. also observed that higher OS rates correlated with Leptotrichia over the median in HNSCC [57], suggesting that certain intratumoral bacteria might be associated with better outcomes. These studies collectively demonstrate that both commensal and intratumoral microbiota can provide valuable prognostic information, complementing traditional clinical and pathological staging. The advantage of these microbial biomarkers is their potential to offer noninvasive or minimally invasive assessment of disease aggressiveness and patient prognosis, which can guide personalized treatment strategies. It is impressive to carry out the prospective, multicenter validation and standardized before clinical implementation.

Microbiota-targeted therapies to enhance treatment efficacy

Beyond biomarkers, the microbiota holds immense promise as a therapeutic target to enhance the efficacy of conventional cancer treatments, including chemotherapy, radiation therapy, and immunotherapy. Modulating the microbial environment can potentially sensitize tumors to therapy, reduce side effects, and improve overall patient outcomes.

Enhancing immune responses

The microbiota’s profound influence on the host immune system makes it an attractive target for improving immunotherapy responses. Matson et al. reviewed how commensal microbiota can influence anti-tumor immune responses and affect immunotherapy efficacy [64], suggesting that microbiome-based interventions could enhance checkpoint blockade. While specific interventions for HNCs are still emerging, the general principle applies. For instance, the observation that intratumoral microbiota can lead to deficient immune infiltration in NPC [46,47] suggests that strategies to reduce the load of pro-tumorigenic bacteria or introduce beneficial microbes could potentially “reinvigorate” the anti-tumor immune response, making tumors more susceptible to immunotherapies like PD-1 blockade. Gong et al. showed that anti-CD70 + anti-PD-1 therapy improved tumor-killing efficacy in preclinical NPC models [94], and tumor-restricted CD70 correlated with Treg abundance and improved anti-PD1 response.

Improving conventional treatments (chemotherapy and radiation therapy)

Microbiota dysbiosis can contribute to treatment resistance and exacerbate side effects, offering opportunities for targeted interventions. (1) Radiotherapy: Guo et al. demonstrated that F. nucleatum promotes radioresistance in NPC [71]. Crucially, leucine restriction or mitochondria-targeted antibiotics reduced Fn load and re-sensitized tumors to radiotherapy. This provides a clear, actionable therapeutic strategy: targeting F. nucleatum or its metabolic pathways can overcome radioresistance, suggesting that interventions based on the microbiome represent a promising frontier [95]. Liu et al. identified specific bacterial taxa (Acidosoma, Propionibacterium acnes, and Clostridium magna) and elevated acetate levels in NR NPC patients [68]. This suggests that targeting these microbes or their metabolic byproducts could enhance radiotherapy efficacy. Lim et al. also noted that specific microbial profiles predict varied responses to chemotherapy and immunotherapy in NPC [65], and that modulating microbiota may enhance treatment efficacy. (2) Chemotherapy: Chemotherapy-induced oral mucositis (OM) is a common and debilitating side effect in HNC patients, often exacerbated by microbiota shifts. Hong et al. found that mucositis severity correlated with bacterial dysbiosis [96], including depletion of health-associated commensals (Streptococcus salivarius) and enrichment of Gram-negative bacteria (F. nucleatum). They showed that F. nucleatum was pro-inflammatory and pro-apoptotic, suggesting that controlling oral bacterial dysbiosis could prevent OM. Bruno et al. reviewed bacterial shifts in cancer patients with OM [97], emphasizing the potential for microbial target therapy to maintain resident bacteria as a promising OM treatment approach. This highlights the dual benefit of microbiota modulation: improving cancer treatment efficacy while mitigating adverse side effects. (3) General Treatment Response: Lim et al. comprehensively reviewed how altered microbiota composition and diversity in NPC patients [65], particularly in oral, nasal, and gut communities, are associated with disease progression and treatment response. They suggested that modulating microbiota could alleviate side effects, improve quality of life, and enhance treatment efficacy, positioning microbiota as a therapeutic target and biomarker for NPC.

Harnessing microbes for precision medicine approaches

The concept of precision medicine in oncology aims to tailor treatments to individual patients based on their unique biological characteristics. The microbiota, with its high interindividual variability and dynamic nature, offers a new dimension for personalized interventions in HNCs (Figure 5).

Figure 5.

Harnessing the microbiome for precision medicine in head and neck cancer. engineered bacteria, microbiota modulation, and microbiota-guided therapy selection offer emerging avenues for individualized treatment. Integrating host genomics, tumor profiling, and microbiome features enables more precise therapeutic planning.

Engineered bacteria

A cutting-edge approach involves bioengineering natural bacteria to enhance cancer therapy. Wang et al. reviewed the promising dawn of engineering oral bacteria for TME therapy [98]. These engineered bacteria can be designed to deliver therapeutic agents, stimulate immune responses, or selectively target cancer cells, offering new insights and strategies for prevention and treatment [99]. This approach leverages the inherent ability of bacteria to colonize tumors and interact with the TME, providing a highly localized and potentially less toxic therapeutic modality.

Probiotics and microbiota transplantation

Modulating the microbiota through probiotics or fecal microbiota transplantation (FMT) is another promising strategy. Bostanghadiri et al. discussed emerging strategies like probiotics and microbiota transplantation for oral disorders [16], including cancer. Nami et al. identified potential probiotic lactobacilli strains from traditional dairy products that exhibited anti-oral cancer properties in vitro [100], inducing apoptosis in cancer cells without harming normal cells. This suggests that specific probiotic strains could be developed as adjunct therapies or preventive agents. Kamarajan et al. showed that Nisin, a bacteriocin, inhibited the pro-tumorigenic effects of periodontal pathogens in OSCC [58], suggesting its potential as an antimicrobial and anti-tumorigenic agent. Ji et al. explored the impact of oral healthcare interventions [101,102], such as Yikou gargle, on oral microbiome dynamics in HNC patients, finding that it reduced pathobionts and was associated with fewer dental complications. This highlights the potential of simple microbial-modulating interventions to improve patient outcomes.

Microbiota-guided drug selection

The microbial profile of a tumor or host could guide the selection of optimal therapies. For example, if a specific bacterium is known to confer radioresistance (e.g. F. nucleatum in NPC), then a patient harboring high levels of this bacterium might benefit from a combination therapy that includes an anti-microbial agent or a metabolic inhibitor alongside radiotherapy [71]. Similarly, if certain microbial profiles predict a better response to immunotherapy, these could be used to identify patients most likely to benefit, thereby personalizing treatment decisions. The findings by Liu et al. on microbiota and metabolomic predictors of radiotherapy response in NPC exemplify this potential [68].

Preventive strategies

Understanding the role of microbiota in familial aggregation of NPC, as shown by Liao et al. [45], could lead to personalized prevention strategies for high-risk individuals. Identifying heritable bacteria associated with increased NPC clustering could enable targeted interventions or surveillance programs. Similarly, early diagnosis and treatment of periodontitis, which is linked to oral cancer, could be a crucial preventive measure [103].

The overarching goal is to integrate multi-omics data (microbiome, host genomics, transcriptomics, metabolomics) with functional and clinical studies to develop a comprehensive understanding of each patient’s unique host-microbe-tumor ecosystem [104,105]. This holistic view will enable the development of truly personalized diagnostic and therapeutic strategies, moving beyond a one-size-fits-all approach in HNC oncology [106].

Challenges and future directions

Despite the exciting advancements in understanding the microbiota’s role in HNCs, several significant challenges currently hinder progress and clinical translation. Addressing these gaps is crucial for realizing the full potential of microbiome research in oncology.

Methodological standardization and causal inference

One of the foremost challenges in microbiome research is the lack of methodological standardization across studies. This encompasses various aspects, from sample collection and processing to sequencing techniques and bioinformatics analysis. (1) Sample Collection Protocols: Variations in sample collection methods (e.g. oral rinse, saliva, oral swab, tissue biopsy) can significantly impact the microbial profiles obtained. For instance, Caselli et al. showed that oral rinse microbiome represented whole site-specific microbiomes better than saliva [107], highlighting the importance of sample type. The specific anatomical site of collection (e.g. tumor tissue vs. adjacent normal tissue vs. distant mucosal site) also influences findings, as demonstrated by Zhang et al. who found significant differences in microbial diversity and function between saliva and dental plaque in oral cancer patients [108]. Standardized protocols are essential to ensure comparability and reproducibility of results across different research groups and clinical settings. (2) Sequencing Techniques: While 16S rRNA gene sequencing is widely used, its resolution is limited, often only allowing identification to the genus level. Metagenomic sequencing offers higher resolution, enabling species-level identification and functional profiling, but is more costly and computationally intensive. Lu et al. emphasized the need for a multiple-modality integrated framework [19], including single-cell sequencing, to overcome the challenges of low biomass in intratumoral microbiota. Fuks et al. proposed combining 16S rRNA gene variable regions to enhance resolution [109], offering a potential compromise. The choice of sequencing platform, primer sets, and sequencing depth can all introduce biases. (3) Bioinformatics Analysis: The downstream analysis of sequencing data, including quality control, taxonomic assignment, and statistical methods, also lacks universal standardization. Different bioinformatics pipelines and statistical tools (e.g. ANCOM-BC for bias correction) can yield varying results [110,111], making cross-study comparisons difficult. The use of robust statistical methods and transparent reporting of bioinformatics pipelines are critical.

Establishing robust causal relationships

A substantial proportion of current research in this field remains observational, primarily identifying associations between microbial dysbiosis and cancer development or progression. Although these studies are critical for hypothesis generation, they fall short of establishing causality, and the longstanding “chicken-or-egg” question – whether microbial alterations drive carcinogenesis or emerge as a consequence of tumor development – continues to pose a major challenge. Observational studies reporting bacterial enrichment within tumor tissues or associations between specific microbial profiles and patient prognosis offer compelling correlational insights but necessitate further experimental substantiation [22,24,46,47]. To advance causal inference, more rigorous methodological frameworks are required. Mendelian randomization (MR) studies, such as those by Wu et al. in oral cancer and Fu in pharyngeal – LSCC [31,56], exploit genetic variants as instrumental variables to infer causality, thereby reducing confounding and reverse-causation bias and offering a distinct advantage over traditional observational approaches. Preclinical models further strengthen causal claims; for instance, Frank et al. showed that antibiotic depletion of the microbiota delayed oral tumorigenesis in HNSCC mouse models [33], whereas microbiota transplantation from tumor-bearing mice accelerated carcinogenesis, providing robust experimental evidence of causative microbial influence. Functional validation through mechanistic in vitro and in vivo studies is essential to elucidate how specific microbes and their metabolites modulate cancer cell behavior and the TME. Recent investigations detailing the ability of F. nucleatum to promote epithelial – mesenchymal transition and radiotherapy resistance through defined signaling pathways exemplify this approach. Moreover [28,29,51,71], patient-derived organoids or mini-colon systems, as demonstrated by Lorenzo-Martín et al. [112], represent promising platforms for mechanistic interrogation under physiologically relevant conditions, particularly when recapitulating native TMEs. Finally, prospective longitudinal studies that repeatedly sample the microbiome in high-risk individuals will be indispensable for tracking microbial shifts preceding disease onset, thereby strengthening evidence for causality and clarifying temporal relationships in microbiome – cancer dynamics.

Accounting for interindividual variability

The human microbiome exhibits remarkable diversity, with substantial interindividual variability shaped by host genetics, lifestyle, diet, geography, and environmental exposures, presenting a significant challenge in establishing universal microbial biomarkers or therapeutic targets. Host genetic factors contribute to oral microbiome composition; Liu et al. identified five genetic loci associated with oral microbial variation and demonstrated that human genetics accounted for at least 10% of oral microbiome diversity [113], suggesting that genetic background influences susceptibility to dysbiosis and microbiota-associated carcinogenesis. Nevertheless, Shaw et al. reported that shared environmental factors [114], rather than inherited genetics, were more strongly associated with salivary microbiome composition in a large family-based cohort, underscoring the complex interplay between genetic and environmental determinants. Environmental exposures – including diet, tobacco and alcohol use, and oral hygiene practices – are major risk factors for HNCs and critically shape the oral microbiota; Brnigen et al. demonstrated that poor oral hygiene [115], tobacco smoking, and tooth loss correlate with significant shifts in oral bacterial communities, while Liao et al. reported that family environment, cigarette smoking, age, and gender contributed to microbiome variation in NPC, and Huang et al. linked dietary patterns with NPC susceptibility [45,116]. Further complicating microbial heterogeneity, geographical and ethnic differences play an important role, as the incidence and biological characteristics of HNCs – particularly NPC – vary markedly across populations [2,5]. Microbial signatures identified in population-specific studies, such as those focusing on East Asians for NPC [45–47], provide valuable mechanistic insights but may not be fully generalizable across diverse ethnic and geographic groups. Accordingly, future investigations should prioritize large-scale, multi-ethnic cohort designs to delineate globally conserved microbial features while also characterizing population-specific patterns. Addressing interindividual variability will require expanded cohort sizes to capture microbial diversity comprehensively, integration of detailed clinical metadata – including lifestyle, dietary, genetic, and treatment-related information – to control for confounding factors, and the development of precision medicine strategies that incorporate individual microbial profiles alongside host genetic and tumor characteristics to guide personalized intervention approaches [106].

Integrating multi-omics and functional validation

To advance beyond descriptive observations and comprehensively elucidate host – microbe interactions in HNCs, there is an urgent need to integrate multi-omics strategies with rigorous functional validation. Integrative multi-omics approaches that combine microbial profiling (16S rRNA sequencing, metagenomics) with host genomics, transcriptomics, proteomics, and metabolomics offer a systems-level understanding of the tumor ecosystem. For example, Cai et al. employed integrative genomic and epigenomic analyses to link oral microbiota dysbiosis in OSCC to host genetic and epigenetic alterations [28], demonstrating that microbial enrichment is associated with increased expression of cancer-related genes and changes in CpG methylation, thereby delineating how microbial communities can modulate host transcriptional programs. Similarly, metabolomic integration is emerging as a powerful tool: Liu et al. combined microbiota and metabolomic profiling in NPC and identified bacterial taxa associated with elevated acetate levels and radiotherapy response [68], while Duan et al. emphasized that microbial metabolomics integrated with multi-omics surpasses conventional research paradigms [66], particularly in unraveling the dual functions of microbial metabolites in tumor biology. Advances in spatial omics and single-cell sequencing, such as spatial transcriptomics, spatial proteomics, and cutting-edge tools like INVADEseq, are enabling high-resolution characterization of intratumoral microbial niches and their interactions with distinct cellular compartments. Bullman leveraged these technologies to map intratumoral microbes within discrete microniches and elucidate their contributions to spatial and cellular heterogeneity [37], identifying host cell-associated bacteria that shape transcriptional networks [117,118]. In addition, Shi et al. developed SpatialTME [119], an online platform integrating histopathology with single-cell and spatial transcriptomic data, providing an invaluable resource for interrogating the TME and its microbial constituents. However, omics-derived correlations require experimental validation to establish causality. Functional studies utilizing 2D cultures, 3D spheroids, and patient-derived organoids (e.g. mini-colons), as underscored by Becker et al., are essential to dissect direct interactions among microbes, cancer cells, and immune components in patient-specific contexts [112,120]. Complementary in vivo approaches – including genetically engineered, germ-free, and humanized mouse models – are needed to determine the causal influence of specific microbes on tumor initiation, progression, metastasis, and treatment response. Ultimately, translation of preclinical discoveries into clinical settings is critical; Dai et al. highlighted emerging microbiome-based biomarkers and therapeutic strategies in oncology and underscored the imperative for clinical trials to evaluate microbiota-targeted interventions in patients [121].

By rigorously addressing these challenges through standardized methodologies, robust causal inference, comprehensive accounting for variability, and integrated multi-omics with functional validation, the field of microbiota and HNCs can accelerate its trajectory toward innovative diagnostics and personalized therapeutic strategies. The absence of meta-analysis represents an intrinsic limitation of this review. Quantitative synthesis was not feasible due to inconsistent reporting of microbial abundance, heterogeneous outcome measures (OS, DFS, recurrence), and insufficient raw data availability for re-analysis. We acknowledge that this limits the ability to generate pooled effect estimates and assess statistical heterogeneity formally. In addition, this review focused primarily on the bacteriome, with incomplete integration of the mycobiome and virome. Emerging evidence suggests that fungal and viral communities may interact with bacteria and host immunity to influence cancer outcomes, warranting more comprehensive multi-kingdom analyses. Finally, mechanistic understanding of how systemic microbiota interact with local tumor microbiota and immune cells to modulate therapy responses remains incomplete. Most studies report associations without elucidating underlying molecular pathways. Future research should prioritize: (1) standardized sampling and sequencing protocols; (2) prospective longitudinal designs with pre- and post-treatment sampling; (3) multi-omics integration (metagenomics, metatranscriptomics, metabolomics); (4) gnotobiotic animal models and organoid systems for mechanistic validation; (5) diverse, multi-ethnic cohorts to improve generalizability; and (6) interventional trials targeting the microbiome to enhance cancer therapy efficacy.

Conclusion

The intricate relationship between the microbiota and HNCs has emerged as a critical area of research, profoundly reshaping our understanding of cancer pathogenesis and opening new avenues for clinical intervention. This review has systematically summarized recent advances, highlighting the unique characteristics of both commensal and intratumoral microbiota in oral, nasopharyngeal, and LSCCs. We have seen how dysbiosis within these microbial communities contributes to cancer initiation and progression through complex mechanisms involving chronic inflammation, immune modulation, the production of metabolic byproducts, and the active shaping of the TME.

Specific microbial signatures have been consistently identified across different HNC types. F. nucleatum, for instance, stands out as a pan-HNC oncopathogen, implicated in promoting proliferation, invasion, EMT, and radioresistance across OSCC, NPC, and LSCC [28,29,51,52,71]. Other key players include P. gingivalis in OSCC, which enhances tumor progression and stemness [21,27], and specific Streptococcus species in NPC, which can mediate EBV reactivation and familial aggregation [44,45]. Comparative analyses reveal both shared microbial features, such as a general pattern of reduced diversity and enrichment of pathobionts, and distinct site-specific patterns that reflect the unique anatomical niches and etiological factors of each cancer type. For example, the oral microbiota’s role in mediating EBV reactivation is particularly relevant for NPC [44], while specific metabolic reprogramming pathways driven by F. nucleatum are distinct in LSCC [51].

Beyond advancing our mechanistic understanding, the microbiota holds substantial translational potential in head and neck oncology. As noninvasive biomarkers, oral and nasopharyngeal microbial profiles show promise for early detection and prognostic assessment, offering high accuracy in identifying HNC patients and predicting survival outcomes [41,46,47,54]. Furthermore, the microbiota represents a novel therapeutic target. Strategies aimed at modulating microbial communities, such as using antibiotics to reduce pro-tumorigenic bacteria like F. nucleatum to overcome radioresistance [71], or employing probiotics to mitigate treatment side effects like OM [96,97], are gaining traction. The emerging field of engineered bacteria for targeted cancer therapy also offers exciting prospects for precision medicine approaches, allowing for tailored interventions based on individual patient microbial profiles [98].

However, the field faces critical challenges that must be addressed to accelerate progress and clinical translation. Standardizing sample collection protocols, sequencing techniques, and bioinformatics analyses is paramount to ensure the comparability and reproducibility of research findings. Establishing robust causal relationships between specific microbes and cancer requires moving beyond correlational studies to employ rigorous methodologies such as Mendelian randomization, preclinical animal models, and comprehensive functional validation in vitro and in vivo [31,33,51]. Moreover, accounting for the significant interindividual variability in microbial composition, influenced by host genetics, lifestyle, diet, and geographical factors, is crucial for developing universally applicable biomarkers and therapies [45,113].

Future directions in this rapidly evolving area of research must prioritize the integration of multi-omics approaches, combining microbiome data with host genomics, transcriptomics, proteomics, and metabolomics, to provide a holistic understanding of the host-microbe-tumor ecosystem [28,66]. The application of advanced technologies like single-cell and spatial omics will be instrumental in elucidating the precise spatial organization and cellular interactions of intratumoral microbiota within the TME [37,119]. Furthermore, robust functional validation in patient-derived models and well-designed clinical trials are essential to translate promising preclinical findings into effective clinical interventions. By systematically addressing these gaps and fostering interdisciplinary collaboration, the field of microbiota and HNCs is poised to unlock innovative strategies for personalized diagnostics, prognostics, and microbiota-targeted therapies, ultimately leading to improved outcomes for patients suffering from these challenging malignancies.

Funding Statement

The author(s) reported there is no funding associated with the work featured in this article.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Ethics approval and consent to participate

Not applicable.

Data availability statement

All data are presented in the manuscript.