Summary
A medRxiv preprint reports that a risk score based on 28 gut microbial species was associated with cancer diagnoses during up to 10 years of follow-up. The finding comes from an observational analysis and represents a candidate biomarker for future validation, not a clinical screening test.
A gut microbiome score derived from stool samples was associated with the risk of receiving a cancer diagnosis years later, according to a medRxiv preprint posted on September 13, 2026. The researchers built the score from 28 microbial species and examined its relationship with cancer diagnoses during up to 10 years of follow-up.
The analysis used shotgun metagenomic sequencing of stool samples from 5,997 participants in the population-based Lifelines cohort and the Dutch Microbiome Project. The microbiome data were linked to longitudinal cancer registry records, allowing the researchers to study incident cancer—cancer diagnosed after the microbiome sample was collected—rather than only the microbial features of people who already had cancer.
How the microbiome score was developed
Shotgun metagenomic sequencing reads microbial DNA in a sample, allowing researchers to estimate which organisms are present and in what relative abundance. The researchers used these measurements to derive a composite score based on 28 microbial species. A higher score represented a microbiome profile associated with greater cancer risk in the study analysis.
Among the 5,997 participants, each one-standard-deviation increase in the score was associated with a higher rate of cancer diagnosis during follow-up. After adjustment for established cancer risk factors, the hazard ratio was 1.62, with a 95% confidence interval of 1.44 to 1.82 and P=1.54e-15.
A hazard ratio compares the rate at which an outcome occurs over time between groups or across different levels of an exposure. Here, the estimate describes a relative association per standard-deviation increase in the score; it is not an individual’s absolute probability of developing cancer.
Findings across additional datasets
The researchers also examined the score in independent external datasets containing 1,879 patients who had cancer when their microbiome was sampled and 5,341 cancer-free controls. In these data, the score distinguished cancer cases from controls, providing a separate test of whether the microbial signature was associated with cancer status.
In the population-based FINRISK cohort, the score showed a consistent trend toward reduced cancer-free survival. That analysis included 1,105 people who developed cancer during follow-up and 5,670 people who did not.
Together, the results indicate that measurable gut microbial signatures were associated with cancer risk before diagnosis in the main longitudinal analysis. The authors describe the score as a potential biomarker for early cancer-risk stratification and suggest that the findings may help investigate the gut microbiome’s possible role in cancer development.
The work is an observational study reported as a preprint, so the score is best understood as a candidate research biomarker rather than an established clinical test. The external case-control analysis included patients who already had cancer when their microbiome was sampled, while the Lifelines and Dutch Microbiome Project analysis examined later diagnoses. Whether using such a score in screening improves early detection or patient outcomes, and whether microbiome features contribute causally to cancer development, remain questions for further research.