Head and neck squamous cell carcinoma, abbreviated HNSCC, is one of the most common cancers worldwide, arising from the flat cells that line the mucosal surfaces of the larynx, pharynx, oral cavity and related structures. Because these tumors develop in a region that is continuously bathed in saliva and colonized by hundreds of microbial species, researchers have long suspected that the oral microbiome might play an active role in the disease. The concept has gained traction under the name of the oncobiome: the idea that microorganisms within the tumor microenvironment are not passive bystanders but may influence how tumors form, how they respond to therapy, and ultimately how patients fare. The new study provides some of the most concrete evidence yet that this influence extends to survival itself.

The research was designed as a prospective study, meaning that patients were enrolled and sampled according to a predefined protocol rather than retrospectively selected from archived material. The cohort included 86 patients with laryngeal carcinoma and 38 with pharyngeal carcinoma, alongside the 66 healthy controls. Saliva was collected with informed consent under ethical approvals from the Hospital Universitario Central de Asturias and the regional ethics committee of the Principality of Asturias. The team then applied 16S rRNA gene sequencing, a technique that reads a specific genetic marker found in all bacteria, allowing researchers to identify which genera and species are present in each sample and in what relative proportions, without needing to culture the organisms in the laboratory.

The first major finding was that cancer patients and healthy controls carried clearly distinct microbial community profiles. Just as striking was the discovery of pronounced site-specific dysbiosis, meaning that the pattern of microbial imbalance differed depending on where in the upper aerodigestive tract the tumor had arisen. Pharyngeal tumors and laryngeal tumors were associated with different salivary bacterial landscapes, a detail that matters because most previous studies have treated head and neck cancer as a single entity. The anatomical context of the tumor, in other words, appears to shape the ecology of the mouth around it, and any biomarker derived from saliva would need to respect that context to be clinically useful.

Treatment itself also left a microbial fingerprint. The researchers compared microbiome composition before and after therapy and found that chemoradiotherapy, the combination of chemotherapy and radiation used in many HNSCC patients, induced a marked remodeling of the oncobiome. The most notable post-treatment change was an enrichment of Lactobacillus, a genus of lactic acid bacteria that is often abundant in altered oral environments. Whether this shift is a consequence of tissue damage, changes in saliva production and pH, or a direct selective effect of the treatment on bacterial growth remains an open question, but it demonstrates that the salivary microbiome is a dynamic system that responds to clinical intervention rather than a fixed attribute of the patient.

The centerpiece of the study, however, is a ten-bacteria salivary signature associated with poorer survival outcomes. Using statistical models, the team identified a specific combination of bacterial taxa whose collective abundance in saliva was linked to worse disease-specific survival, worse overall survival, and worse disease-free survival, the three standard endpoints used in oncology. The strength of the association was remarkable. In multivariable Cox regression analysis, which adjusts for established clinicopathological variables such as tumor stage and site, the ten-bacterial signature remained independently associated with poor prognosis, with a hazard ratio of 6.215 for disease-specific survival, 5.417 for overall survival, and 6.842 for disease-free survival, all with p values below 0.001. Hazard ratios of this magnitude indicate that patients carrying the adverse signature faced a roughly five- to seven-fold higher risk of death or recurrence compared with patients whose salivary microbiome lacked it.

Beyond the unified signature, the study uncovered tumor site-specific prognostic taxa, bacteria whose association with survival depended on where the tumor was located. In pharyngeal carcinomas, the presence of putative species belonging to the genera Prevotella, Porphyromonas and Fusobacterium correlated significantly with improved disease-specific, overall and disease-free survival. This is a counterintuitive result, because Fusobacterium in particular has been implicated in promoting colorectal tumors, yet here its presence in the saliva of pharyngeal cancer patients was associated with a favorable outcome. In laryngeal carcinomas, a different set of organisms told a different story: putative Stomatobaculum longum and Dialister were associated with favorable survival outcomes. The authors interpret these findings as support for the concept of an anatomically contextualized oncobiome, in which the prognostic meaning of a given bacterium cannot be separated from the anatomical site of the tumor it accompanies.

Technically, the study relied on ANCOM-BC, a modern differential abundance method that corrects for the compositional nature of sequencing data, to compare microbial communities between groups. Prognostic associations were then tested with univariable and multivariable Cox proportional hazards models, the standard framework for survival analysis in medicine, supplemented by Kaplan-Meier style survival analyses. The word putative in the species-level assignments reflects a limitation of 16S rRNA sequencing: because the method reads only a marker gene rather than whole genomes, it can often identify a bacterium only to the genus level, with species-level calls made by inference. Higher-resolution approaches such as shotgun metagenomic sequencing would be needed to confirm the exact species and strains involved, and the authors are explicit that validation in independent cohorts is required before any clinical implementation.

Even with those caveats, the implications are considerable. Current prognostic assessment in HNSCC rests largely on tumor staging, grading, anatomical site and a handful of molecular markers, all of which are typically obtained through invasive biopsies and imaging. A salivary microbiome signature offers something different: a non-invasive, dynamic biomarker that can in principle be sampled repeatedly over time, before treatment, during follow-up, and at the first suspicion of recurrence. Because saliva collection is cheap, painless and acceptable to patients, a validated microbial risk score could be integrated into routine surveillance, helping clinicians decide which patients need intensified follow-up imaging or adjuvant therapy and which can be monitored less aggressively. It could also, in the longer term, open the door to microbiome-targeted interventions aimed at shifting the oral ecology toward a more favorable state.

The study was conducted by researchers from the Instituto de Investigación Sanitaria del Principado de Asturias in Oviedo, the University of Oviedo, the Spanish Biomedical Research Network in Cancer, the Dairy Research Institute of Asturias, the FISABIO Foundation in Valencia, and Dartmouth’s Geisel School of Medicine, with funding from the Instituto de Salud Carlos III and co-financing from the European Union and regional institutions. The work was published open access on 9 October 2026 in Microbiome under DOI 10.1186/s40168-026-02548-w. For now, the ten-bacteria signature remains a research finding rather than a bedside test, and the authors caution that larger, independent and higher-resolution studies must confirm the result. But the message is clear: the bacteria in a patient’s saliva are not just spectators to head and neck cancer. They form a readable, quantifiable signature of the disease’s trajectory, and learning to read it may reshape how survival risk is assessed in one of the world’s most challenging cancers.

Subject of Research: Salivary microbiome signatures as prognostic biomarkers for survival and recurrence in head and neck squamous cell carcinoma

Article Title: Salivary prognostic bacterial signatures associated with survival outcomes and recurrence risk in head and neck squamous cell carcinoma

Article References: Granda-Díaz, R., Molinero, N., Sánchez-Canteli, M., Álvarez-Teijeiro, S., Mira, A., Allonca, E., Quesada, V., Margolles, A., Martínez-Camblor, P., Martín-Guillermo, E., Delgado, S., Rodrigo, J. P., & García-Pedrero, J. M. (2026). Salivary prognostic bacterial signatures associated with survival outcomes and recurrence risk in head and neck squamous cell carcinoma. Microbiome. https://doi.org/10.1186/s40168-026-02548-w

Image Credits: AI Generated

DOI: 10.1186/s40168-026-02548-w

Keywords: salivary microbiome, head and neck squamous cell carcinoma, oncobiome, 16S rRNA sequencing, prognostic biomarkers, dysbiosis, Lactobacillus, Fusobacterium, chemoradiotherapy, survival analysis, recurrence risk, Microbiome journal