A Comparative Analysis of Tissue and Digestive Microbiota Composition in Patients With Oral Squamous Cell Carcinoma

Abstract

Introduction

As the most prevalent malignant tumor in oral cancer, oral squamous cell carcinoma (OSCC) remains incompletely understood, especially its complex interactions with microbiota. This study investigated microbial signatures across tumor tissue and digestive tract niches in patients with OSCC.

Methods

This retrospective cohort study included 28 patients with OSCC who underwent surgical treatment at a tertiary cancer center between 2019 and 2022, along with 14 healthy controls. 16S rRNA gene sequencing was performed to characterize microbial profiles in tumor tissues, as well as oral and gut samples.

Results

Tumor tissues exhibited largely similar microbial compositions to adjacent tissues and lymph nodes. However, tumor tissues showed a relatively higher abundance of Fusobacterium and a depletion of Streptococcus. Gut microbiota in patients exhibited dysbiosis, characterized by enrichment of Prevotella, Klebsiella, Lactobacillus, and Streptococcus, while oral microbiota showed enrichment of Streptococcus and Haemophilus. In addition, increased abundance of Lactobacillus in the gut and Bacteroidales in the oral cavity was associated with adverse clinical outcomes (ACO) in OSCC patients.

Conclusion

The tissues and digestive microbiota of OSCC patients exhibit heterogeneity. These findings suggest that specific microbial signatures may serve as potential biomarkers for disease detection and progression; however, these associations should be interpreted cautiously and require validation in larger cohorts and mechanistic studies.

1. Introduction

Oral squamous cell carcinoma (OSCC) is the most common malignant tumor of the oral cavity, with approximately 370,000 new cases and 170,000 deaths in 2020.^1,2 The primary risk factors for OSCC include tobacco use, alcohol consumption, betel quid chewing, areca nut exposure, and human papillomavirus (HPV) infection.^1,3-5 However, recent systematic reviews suggest that the increasing incidence of OSCC among younger patients cannot be fully explained by traditional risk factors such as tobacco and alcohol use.^6,7 Approximately 15% of oral cancers remain unexplained by these major risk factors, ⁸ suggesting that additional factors may contribute to OSCC initiation, progression, and metastasis. Recent studies indicate that established risk factors may alter the oral microbiota, leading to dysbiosis,^9,10 which suggests a potential association between microbiota and OSCC.

Since the establishment of the link between carcinogenesis and bacteria—most notably the role of Helicobacter pylori in gastric cancer ¹¹ —increasing attention has been paid to the role of microbiota in various cancers. This growing interest has been facilitated by advances in genomic technologies. Gut microbial dysbiosis has been shown to be associated with the development and progression of colorectal cancer, with enrichment of oral pathogens such as Fusobacterium nucleatum, Parvimonas micra, and Peptostreptococcus stomatis. ¹² Alterations in microbial composition have also been reported in other cancers, such as bladder cancer, where Enterobacteriaceae are reduced, while Actinotignum and Streptococcus exhibit sex-specific differences. ¹³ Tumor-associated microbiota signatures have been linked to clinical outcomes, with specific bacterial taxa associated with long-term survival. ¹⁴ These findings collectively suggest that microbiota may contribute to tumor development and progression, and may also be involved in the pathogenesis of OSCC.

However, it is worth noting that findings from current studies remain inconsistent.^15,16 Consistent conclusions have not been reached regarding alterations in microbial diversity and relative abundance, partly due to heterogeneity in sample collection methods (e.g., tumor tissue, saliva, or oral rinse swabs) and participant characteristics (e.g., age and dietary habits). ¹⁷ Previous studies have reported conflicting results, with some demonstrating higher microbial richness and diversity in tumor tissues, while others report greater diversity in control samples. ¹⁸ Furthermore, previous research on OSCC has predominantly focused on microbial composition in either the oral cavity or specific tissues, with limited investigation into interactions with the gut microbiome. These limitations highlight the need for comprehensive studies investigating microbiota across multiple anatomical sites to better understand host–microbiome interactions in OSCC.

In our study, conducted in a Southern Chinese population, we not only analyzed microbial composition differences between tumor tissues (TT), adjacent tissues (AT), and lymph nodes (LN) but also compared microbiota profiles between OSCC patients and healthy controls (HC) across both oral and gut environments. Differential microbial composition analysis and functional prediction were performed to systematically characterize multi-site microbiota alterations, providing insights into host–microbiome interactions and suggesting potential biomarkers for disease progression.

2. Methods

2.1. Patient Recruitment

This retrospective cohort study conducted at a tertiary cancer center included two cohorts: 28 patients with squamous cell carcinoma (OSCC) who underwent surgical treatment between January 1, 2019 and December 31, 2022, and were consecutively included from eligible cases, and 14 healthy participants. TT, AT, and LN samples were collected from 28, 26, 7 OSCC patients, respectively. Clinical characteristics, including age, sex, and tumor stage, were obtained from the institutional clinical database. Clinical outcomes were defined as Adverse Clinical Outcomes (ACO), including tumor recurrence or death during follow-up. Pathological types and lymph node metastasis (LNM) status were determined by experienced pathologists based on surgical specimens. Before sample collection, the health status of all participants, including patients and HC, was assessed. Exclusion criteria included: (1) oral diseases (e.g., Sjogren’s syndrome, Meige’s syndrome, atrophic glossitis, oral leukoplakia, and oral infections); (2) intestinal diseases (e.g., inflammatory bowel disease, including Crohn’s disease and ulcerative colitis); (3) a history of other malignancies; and (4) use of antibiotics or probiotics within one month prior to sample collection. HC were confirmed to have no current or prior history of OSCC. Detailed information regarding smoking status, alcohol consumption, dietary habits, and oral hygiene was not systematically collected in this study; therefore, these factors could not be fully adjusted for in the analysis. However, all participants underwent a standardized pre-sampling health assessment, and strict inclusion and exclusion criteria were applied to minimize potential confounding effects from major oral and systemic conditions.

Written informed consent was obtained from all participants. All patient data were de-identified prior to analysis to ensure confidentiality. This study was conducted in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines. ¹⁹

2.2. Collection of Tissue Specimens

TT, AT, and LN were collected during surgery under sterile conditions. All specimens were immediately stored at −80°C until further analysis. Additionally, intestinal swabs (n = 26) and throat swabs (n = 23) were collected from patients with OSCC and HC in 2022 by trained nursing staff following standardized procedures. Samples were placed in sterile tubes and temporarily stored at −20°C before transfer to −80°C for DNA extraction and taxonomic analysis.

2.3. Sample Sequencing

16S rRNA gene sequencing was performed to characterize microbial communities in tissue, oral, and intestinal samples. For tissue samples, five hypervariable regions (V2, V3, V5, V6, and V8) were amplified, whereas the V3–V4 region was targeted for oral and intestinal samples. Sequencing was conducted on Illumina NovaSeq platforms following standard protocols. Raw sequencing data were processed using quality filtering, read merging, and chimera removal, and amplicon sequence variants (ASVs) were generated using the DADA2 algorithm. Taxonomic annotation was performed using reference databases including Greengenes and the NCBI 16S rRNA database. For tissue samples, the Short MUltiple Regions Framework (SMURF) was applied to integrate multi-region sequencing data. Functional prediction was performed using PICRUSt2. Detailed methods are provided in the Supplementary Material.

2.4. Statistical Analysis

Statistical analyses were performed using R (v4.2.1) and MicrobiomeAnalyst. Continuous variables were compared by Wilcoxon rank-sum test, and categorical variables by Fisher’s exact or chi-square test. Alpha diversity (Shannon, Chao1, observed features) and beta diversity (Bray-Curtis, PERMANOVA with 999 permutations) were assessed using the Wilcoxon rank-sum test. Benjamini-Hochberg FDR correction was applied for multiple comparisons in alpha diversity and genus-level analyses (adjusted *q* < 0.05). LEfSe identified differentially abundant taxa (LDA > 2.0, nominal P < 0.05, Wilcoxon rank-sum test); as an exploratory tool, no additional multiple-testing correction was performed. PICRUSt2-predicted pathways were compared using the Wilcoxon rank-sum test with FDR correction. All tests were two-sided, and a nominal P < 0.05 was considered significant.

3. Results

3.1. Clinical Characteristics

The clinical characteristics of the study subjects are summarized in Table 1. The median age of the patients was 53.5 years, with a predominance of males (25; 89.3%). Clinical stages were classified as stage I–II for 11 (39.3%) patients and stage III–IV for 17 (64.7%) patients. The most common tumor location was the tongue (17; 60.7%), followed by the buccal area (4; 14.3%). Additionally, 3 patients had tumors in the floor of the mouth, and 2 had tumors in the gums. Ten patients (35.7%) exhibited LNM, as indicated by positive lymph node biopsies, and 14 patients (50%) experienced ACO, including recurrence or death.

Table 1.

Baseline Characteristic.

Variables	(N=28)
Age, years, median [IQR]	53.5 [52]
Sex (%)
Male	25 (89.3)
Female	3 (10.7)
Sex ratio(M:F)	8.3
Clinical Stage, n(%)
Ⅰ-Ⅱ	11(39.3)
Ⅲ-Ⅳ	17(64.7)
Tumor site, n(%)
Tongue	17(60.7)
Cheek	4(14.3)
The floor of mouth	3(10.7)
Gingiva	2(7.1)
Others	2(7.1)
Lymph node metastasis, n(%)
Yes	10(35.7)
No	18(64.3)
Adverse clinical outcomes, n(%)
Yes	14(50.0)
No	14(50.0)

Adverse clinical outcomes are defined as recurrence or death. The criteria of clinical stage are based on the American Joint Committee on Cancer(AJCC) Cancer Staging Manual.

3.2. Differences in Microbiota Across TT, at and LN

Figure 1A and B illustrate the primary microbiota composition at the phylum and genus levels across three types of tissues. At the phylum level, the predominant bacteria in TT and AT were Firmicutes (37.81% and 41.18%, respectively), Proteobacteria (26.76% and 28.55%), Bacteroidetes (14.97% and 12.58%), and Fusobacteria (12.06% and 8.97%). In LN, the dominant phyla were Proteobacteria (49.14%), Firmicutes (26.25%), Actinobacteria (11.12%), and Bacteroidetes (7.18%) (Figure 1A, Table S1). At the genus level, the dominant genera included Neisseria, Streptococcus, Fusobacterium (Figure 1B, Table S2). However, the microbial composition across the three tissue types was similar, and there was no statistically significant difference in microbial diversity (Figure 1C and D, Supplementary Figure 1).

Figure 1.

Microbiota dysbiosis in different tissues associated with OSCC. LEfSe analysis identifies discriminative bacterial genera between tumor and adjacent tissues (E), and between tumor and lymph node tissues (F). (G–J) Comparison of microbiota in tumor and adjacent tissues between early-stage and late-stage patients (G, H), and between patients with adverse and benign clinical outcomes (I, J). (K) Differences in lymph node microbiota in patients with lymph node metastasis. Bar lengths represent log10 Linear Discriminant Analysis (LDA) scores

We compared microbiota in TT and AT between early-stage (ESTT, ESAT) and late-stage (LSTT, LSAT) patients, as well as between patients with adverse clinical outcomes (ACO) and benign clinical outcomes (BCO). Here, TT-ACO and AT-ACO refer to tumor and adjacent tissues from patients with ACO, respectively, while TT-BCO and AT-BCO refer to tumor and adjacent tissues from patients with BCO, respectively.

LEfSe testing identified discriminative bacterial taxa between TT and AT (Figure 1E, Table S3), as well as between TT and LN (p < 0.05, LDA score > 10⁴) (Figure 1F, Table S3). Notably, the genus Fusobacterium was enriched in TT compared to both AT and LN (LDA score: 4.35 and 4.59, respectively). While Streptococcus (LDA score: 4.24) and Enterococcus (LDA score: 4.52) were significantly depleted in TT. Late-stage patients, as well as those with ACO, exhibited higher relative abundances of the phylum Proteobacteria in both TT (LDA scores: 4.31 and 4.51, respectively) and AT (LDA scores: 3.47 and 4.56, respectively) (Figure 1G–J, Table S4). The relative abundance of Actinobacteria (LDA score = 4.66) was significantly higher in LN samples compared to primary TT in OSCC patients with lymph node metastasis (p < 0.05, LDA score > 10⁴) (Figure 1K, Table S4).

Overall, the microbial composition of TT, AT and LN was similar. However, we observed a higher abundance of the genus Fusobacterium and a relative depletion of Streptococcus and Proteobacteria in TT. Additionally, the increased proportion of phylum Proteobacteria in TT and AT may be associated with late-stage tumors and ACO.

3.3. Intestinal Microbiota Dysbiosis in OSCC Patients

The three most dominant bacterial phyla in OSCC patients and HC were Firmicutes (46.23% vs. 55.13%), Bacteroidota (24.71% vs. 23.90%), and Proteobacteria (18.48% vs. 8.77%) (Figure 2A, Table S5). At the genus level, 30 genera were identified, including Prevotella, Bacteroides, Escherichia-Shigella, Fenollaria, Dialister, and Ralstonia (Figure 2B, Table S6). Alpha diversity analysis based on different indices all showed that the gut microbiota diversity in OSCC patients was significantly lower than in HC (all p < 0.001, including Chao1 index, Good’s coverage, observed operational taxonomic units (OTUs), and Shannon index) (Figure 2C). NMDS analysis, based on Jaccard (stress = 0.15), Unweighted UniFrac (stress = 0.12), and Weighted UniFrac (stress = 0.13), revealed distinct clustering differences between OSCC patients and healthy controls (Figure 2D). Adonis analysis (Table 2) further confirmed significant differences in the gut microbiome composition between the two groups.

Figure 2.

Gut microbiota dysbiosis in OSCC patients compared with healthy controls. The taxonomic composition of gut microbiota in OSCC, comparing TP and HC, is shown, with average relative abundance at the phylum level (A) and genus level (B) Panel (C) provides a comparison of gut microbiota alpha diversity between TP and HC, utilizing Chao1, Good’s coverage, observed OTUs, and the Shannon index. Non-metric multidimensional scaling (NMDS) analysis is employed to evaluate beta diversity based on Jaccard, Unweighted UniFrac, and Weighted UniFrac distance metrics (D). LEfSe identifies discriminative gut bacterial genera between TP and HC in OSCC, presented in a cladogram (E) and a bar chart (F). The bar length represents log10 LDA score. LEfSe reveals the distribution of core species displaying the abundance significantly higher (LDA > log10³) among cohorts. MDS: Multidimensional scaling, HC: Healthy control, TP: Tumor patients

Table 2.

Adonis Analysis Results For Oral And Anal Microbiota In Tumor Patients Versus Healthy Controls.

Anal microbiota	Df	Sum of Sqs	R2	Pseudo-F
Bray curtis	1	0.70	0.05	1.83
Jaccard	1	1.23	0.09	3.37
Unweighted unifrac	1	0.91	0.14	5.69
Weighted unifrac	1	0.19	0.03	1.10

Compared to HC, 77 discriminative intestinal bacterial genera were identified in OSCC patients, with 18 taxa enriched and 59 taxa depleted (p < 0.05, LDA score > 10³) (Figure 2E and F, Table S7). Notably, significant differences were primarily clustered in the genus Prevotella, Klebsiella, Lactobacillus, Streptococcus (LDA score: 3.91, 3.93, 3.87, and 3.58, respectively). Brevundimonas and Sutterella (both belonging to the phylum Proteobacteria) (LDA score: 3.61 and 3.51, respectively) were markedly reduced in OSCC patients. Therefore, these findings suggest intestinal microbiota dysbiosis in OSCC patients, characterized by the genera Prevotella, Klebsiella, Lactobacillus, Streptococcus.

3.4. Alterations in Oral Microbiota in OSCC Patients

The predominant bacterial phyla in OSCC patients and HC were Proteobacteria (36.66% vs. 82.63%), Firmicutes (37.17% vs. 6.67%), and Bacteroidota (14.32% vs. 3.17%) (Figure 3A, Table S9). At the genus level, 31 genera were detected, with notable examples being Ralstonia (3.15% vs. 55.26%), Streptococcus (24.52% vs. 2.12%), Haemophilus (14.93% vs. 0.44%), and Methyloversatilis (0.80% vs. 9.67%) (Figure 3B, Table S10). There were significant differences in the alpha diversity of the oral microbiota between OSCC patients and HC (Figure 3C). NMDS analysis (Figure 3D) and Adonis analysis (Table 2) highlighted substantial β-diversity differences between the two groups.

Figure 3.

Oral microbiota dysbiosis in OSCC patients compared with healthy controls. The taxonomic composition of oral microbiota in OSCC, comparing TP and HC, is shown, with average relative abundance at the phylum level (A) and genus level (B) Panel (C) shows a comparison of oral microbiota alpha diversity between TP and HC, based on metrics such as Pielou’s evenness, the Shannon index, and the Simpson index. Beta diversity is assessed via NMDS analysis based on Bray-Curtis, Jaccard, Unweighted UniFrac, and Weighted UniFrac distances (D). LEfSe identifies discriminative oral bacterial genera between TP and HC, presented in a cladogram (E) and a bar chart (F). The bar length represents the log10 LDA score. LEfSe reveals the distribution of core species displaying the abundance significantly higher (LDA > log10⁵) among cohorts. MDS: Multidimensional scaling, HC: Healthy control, TP: Tumor patients

LEfSe analysis identified differentially abundant bacteria in oral swab samples between OSCC patients and HC. Notably, 41 taxa were significantly enriched in oral swab samples of OSCC patients, while 21 taxa were significantly depleted (p < 0.05, LDA score > 10⁵) (Figure 3E and F, Table S11). For instance, Streptococcus (LDA score: 5.04), Haemophilus (LDA score: 4.85), Neisseria, Actinobacillus, and Fusobacterium were more prevalent in OSCC patients. While Ralstonia was markedly reduced (LDA score: 5.43), which was also observed in intestinal microbiota alterations.

In conclusion, there is a significant difference in the composition of the oral microbiome between OSCC patients and HC. The composition of the oral bacterial community in OSCC patients is characterized by an abundance of Streptococcus, and Haemophilus.

3.5. Intestinal and Oral Microbiota Dysbiosis in OSCC Is Associated With Clinical Features

The relationship between alterations in intestinal and oral microbiota and clinical features was also investigated, focusing on disease stage (Figure 4A and E), LNM (Figure 4D and F) and ACO (Figure 4B and C) (p < 0.05, LDA score > 10⁴). In patients with late-stage or ACO, there was a notable increase in Lactobacillus in the intestinal microbiome (LDA scores: 4.05 and 4.08, respectively) (Figure 4A–C and Table S8). Additionally, Lactobacillus was identified as an important biomarker for OSCC patients in the previous analysis (Figure 2E and F). Conversely, in the oral cavity, patients with late-stage or LNM showed an increase in Bacteroidales (LDA scores: 4.60 and 4.75, respectively). Additionally, increased abundance of Fusobacterium, Leptotrichia, and Capnocytophaga was associated with LNM, with an LDA score of 4.5, 4.14 and 4.11, respectively (Figure 4E–G and Table S12).

Figure 4.

Dysbiosis of anal and oral microbiota in OSCC patients is associated with clinical features. LEfSe identifies discriminative anal bacterial genera between early-stage and late-stage OSCC patients, presented in a cladogram (A) and a bar chart (B). LEfSe also identifies discriminative anal bacterial genera associated with clinical outcomes (C) and lymph node metastasis (D). Additionally, LEfSe identifies representative oral bacterial genera in OSCC patients diagnosed with late-stage disease (E) and lymph node metastasis (F, G). The bar length represents the log10 LDA score. LEfSe reveals the distribution of core species that display significantly higher abundance (LDA > log10³) among cohorts. ES: Early stage, LS: Late stage, LNM: Lymph node metastasis, NLNM: Non-lymph node metastasis, ACO: Adverse clinical outcomes, BCO: Benign clinical outcomes

In conclusion, gut microbiota dysbiosis, particularly Lactobacillus enrichment, may be associated with ACO in OSCC patients. In the oral cavity, Bacteroidales, Fusobacterium, Leptotrichia, and Capnocytophaga were associated with disease progression and LNM in OSCC.

4. Discussion

The primary known risk factors for OSCC include tobacco and alcohol consumption, betel quid, areca nut, and HPV infection.^1,3-5 However, for OSCC patients lacking these traditional risk factors, chronic infection and inflammation, particularly involving the resident microbiota, are increasingly recognized as contributing factors in carcinogenic feedback loops. ²⁰ In this study, we aimed to explore the microbial profiles of tissue in OSCC patients. Our findings revealed specific microbial signatures in TT, notably the significant enrichment of Fusobacterium and depletion of Streptococcus compared to AT. We also utilized oral rinse and anal rinse to characterize the digestive microbiota. In OSCC patients, intestinal microbiota dysbiosis was characterized by genera such as Prevotella, Klebsiella, Lactobacillus, and Streptococcus. The oral bacterial communities of these patients were dominated by Streptococcus and Haemophilus, with a notable absence of Ralstonia. Additionally, Lactobacillus in the gut was associated with ACO in OSCC patients. The phylum Bacteroidales and genera Fusobacterium, Leptotrichia, and Capnocytophaga in the oral cavity were also associated with rapid progression and LNM of OSCC.

In TT, Fusobacterium levels were increased, whereas Streptococcus levels were decreased compared to AT. This conclusion is consistent with the findings of previous studies.^21-23 Fusobacterium, a Gram-negative anaerobic bacterium, is closely related to the pathogenesis of OSCC in various ways. Firstly, it can invade human epithelial cells through Fap2, FadA, and lipopolysaccharide, which promote mucosal permeability.^16,24,25 Functional prediction in our study revealed that genes involved in bacterial motility proteins, flagellar assembly, and bacterial secretion systems, which are associated with various pathological processes, were significantly increased in TT (Supplementary Figure 2). Furthermore, its adhesion protein FadA binds to E-cadherin, thereby activating the Wnt/β-catenin pathway and inducing proliferation of oral epithelial cells. ²⁶ Secondly, it can shield tumor cells from circulating immune cells, thereby facilitating tumor invasion and metastasis.^27-29 Thirdly, Fusobacterium produces volatile sulfur compounds, which have been implicated in chronic inflammation and cell proliferation, migration, invasion, and tumor angiogenesis.^30,20

Unlike TT, our analysis found that Streptococcus dominated over Fusobacterium in oral swab samples, which was also reported in previous studies.^31,32 Streptococcus species are associated with the production of acetaldehyde, a metabolic byproduct that can induce DNA damage and excessive epithelial proliferation, ²⁰ which may explain our finding. However, some previous studies reported a significant reduction in Streptococcus abundance and an increase in Fusobacterium in the oral microbiome of OSCC patients.^27,33 According to the study by Yang, C.Y. et al, the abundance of Fusobacterium increased significantly with the progression of oral cancer, while the abundance of Streptococcus mitis was inversely associated with OSCC progression. ³⁴ In our study, the higher proportion of late-stage patients may contribute to this disparity (Table 1). Furthermore, our study indicates a strong correlation between high abundance of Fusobacterium in the oral cavity and poorer clinical outcomes, aligning with findings from previous research.^35,36

In addition to disease stage distribution, other factors may also contribute to the observed discrepancies across studies. Differences in sampling strategies (e.g., oral swabs versus saliva or tissue specimens) may influence the detected microbial composition due to distinct microbial habitats and contamination profiles. Furthermore, regional population characteristics, including dietary habits and genetic background, particularly in Southern Chinese cohorts, may also shape the oral and gut microbiome profiles. Variations in sequencing platforms and bioinformatics pipelines across studies may further contribute to inconsistencies in reported microbial signatures. Collectively, these factors highlight the inherent heterogeneity in OSCC microbiome studies and underscore the importance of standardized methodologies and multi-center validation in future research.

Interestingly, Lactobacillus serves not only as a biomarker distinguishing the gut microbiome of OSCC patients from HC, but also is associated with ACO in OSCC patients.^37,38,39 Pignatelli et al reported that Lactobacillus was uniquely enriched in oral rinse swab of OSCC. ³⁴ Lactobacillus plays controversial roles in the development of OSCC. On one hand, it can produce acids, such as acetic, butyric, isobutyric, isovaleric, and isocaproic acids, which lower the environmental pH and may facilitate cancer growth and spread. ⁴⁰ On the other hand, it can suppress the immune system’s anti-tumor responses by overproducing reactive oxygen species in host cells. Additionally, it may enhance tumor suppressor gene expression and DNA repair by inhibiting histone deacetylases.^20,41 In the pre-tumor stage, lactic acid bacteria can bind and degrade carcinogens, produce anti-carcinogenic compounds, and promote and regulate both innate and adaptive immune responses, thereby contributing to tumorigenesis prevention.^20,42,43 Overall, our study suggests that Lactobacillus predominantly plays a pro-carcinogenic role in OSCC patients and may serve as a potential target for prevention and therapy.

The strengths of this study lie in the characteristic selection of our samples, primarily focusing on OSCC patients in South China. The findings reflect region-specific characteristics. Additionally, we analyzed changes in microbial communities across multiple anatomical sites, providing a more comprehensive perspective. We also placed particular emphasis on our sampling methods. Our tissue samples were fresh and untreated, ensuring the accuracy of subsequent microbiome sequencing analyses, as RNA degradation may occur in processed samples. Prior studies often relied on saliva and oral wash samples as a proxy for the oral cancer microbiome, which present challenges due to their passage through different environments.^44-46 For example, mouth rinsing and eating can alter the composition of the salivary microbiome. Therefore, our study employed oral and intestinal swabs (mucosal sampling) to sample the surface microbiome. ¹⁷ In summary, this study is, to our knowledge, the first to systematically describe the microbiota differences in OSCC patients across tissues, the oral cavity, and the gut, providing valuable insights for future research.

There were some limitations in our study. First, potential confounding factors such as smoking, alcohol consumption, dietary habits, and oral hygiene were not fully controlled, which may influence microbial communities in cancer development and affect the results. Moreover, the relatively small sample size (28 OSCC patients and 14 healthy controls) may limit the statistical power and generalizability of the findings, particularly for subgroup analyses stratified by clinical stage, LNM, and ACO, where the effective sample size was further reduceds. Importantly, 16S rRNA gene sequencing has inherent limitations, including limited taxonomic resolution at the species or strain level and reliance on predictive functional inference (e.g., PICRUSt2) rather than direct measurement of microbial activity. In addition, due to the lack of sufficient follow-up duration, we were unable to perform survival analyses on the patient cohort, which would further substantiate the correlation between biomarkers and ACO. Additionally, this study revealed correlations but not causality between oral microbiota dysbiosis and OSCC pathogenesis. Further research is needed to explore the carcinogenic mechanisms of the microbiome biomarkers identified in our study.

3. Conclusion

This study revealed intricate microbiota–host interactions across tissue and digestive microbiota in OSCC patients. The results showed a higher abundance of Fusobacterium and a depletion of Streptococcus in TT. Conversely, the oral bacterial community exhibited a predominance of Streptococcus over Fusobacterium. Additionally, Lactobacillus in the gut was associated with ACO, suggesting that it may serve as a potential biomarker for clinical outcomes and a possible target for prevention and therapy. These findings should be interpreted with caution. Further research involving longitudinal sampling and mechanistic studies is required to clarify the roles of the microbiota.

Appendix.

Abbreviations

AT

Adjacent tissue

LN

Lymph node

TT

Tumor tissue

ESTT

Early-stage tumor tissue

LSTT

Late-stage tumor tissue

ESAT

Early-stage adjacent tissue

LSAT

Late-stage adjacent tissue

T

Tumor tissue with lymph node metastasis

N

Lymph node with metastasis

TT-ACO

Tumor tissue from patients with adverse clinical outcomes

TT-BCO

Tumor tissue from patients with benign clinical outcomes

AT-ACO

Adjacent tissue from patients with adverse clinical outcomes

AT-BCO

Adjacent tissue from patients with benign clinical outcomes.

Ethical Considerations

This study protocol was reviewed and approved by the Medical Ethics Committee of Sun Yat-sen University Cancer Center (B2023-559-01).

Consent to Participate

All procedures performed during the study adhere to the principles of respect for persons, beneficence, and justice as outlined in the Declaration of Helsinki. Informed consent was obtained from all participants.

Data Availability Statement

The data that support the findings of this study are available from the author, Zhongyuan Yang (yangzhy@sysucc.org.cn), upon reasonable request.