Abstract
Background
Oral squamous cell carcinoma accounts for over 90% of oral neoplasms. Despite therapeutic advances, the lack of reliable, non-invasive biomarkers and delayed diagnosis continues to impede effective clinical management. By combining paired lesion and non-lesion sampling with predictive metagenomics analysis, our study addresses this gap and advances the current understanding of microbiome‒tumor interactions.Methods
We analyzed 92 buccal swab samples from 39 OSCC patients and 14 healthy controls using 16S rRNA gene (V3-V4) sequencing. Taxonomic profiling was conducted using QIIME2 and SILVA/eHOMD databases, functional pathways were predicted using PICRUSt2, and hub taxa were identified through co-abundance network analysis.Results
Microbial community structure differed significantly across lesion, non-lesion, and healthy sites (PERMANOVA, p = 0.001). Lesions were enriched with Selenomonas infelix and Treponema vincentii, while healthy controls harbored Streptococcus oralis and Gemella haemolysans. Co-abundance network analysis revealed lesion-specific hub species, notably T. vincentii, strongly correlated with predicted activation of pyrimidine biosynthesis pathways (r = 0.69, q Prevotella melaninogenica and Segatella oulorum.Conclusion
Our findings define a lesion-specific microbial signature of OSCC characterized by the depletion of health-associated taxa, enrichment of pro-inflammatory pathobionts, and predicted associations with metabolic pathways implicated in carcinogenesis. These alterations reflect a predicted functionally altered tumor microenvironment.