Mouth rinse study finds stomach and colorectal cancer signals

A small rinse of the mouth produced a measurable cancer signal in samples from 507 volunteers studied by researchers at Yonsei University in Seoul. The cohort included healthy people, people with metabolic disorders, and patients with gastric or colorectal cancer; for the cancer patients, samples were taken before treatment.

The team used gene sequencing to compare microbes in each person’s mouth and stool. Its mouth-to-feces index measured how often identical microbial sequence variants appeared in both places. That index was significantly elevated in patients with gastric or colorectal cancer, but not in people with metabolic syndrome, hypertension, hyperlipidemia or type 2 diabetes. The association remained after accounting for alcohol consumption, regular exercise and body mass index.

The most practical result came from the mouth itself. Models built only from oral microbial features distinguished cancer patients from healthy people across several independent cohorts, even though only a small fraction of oral bacterial variants were also detected in the gut. In a comparison with fecal occult blood tests, a standard stool-based colorectal screening tool, the oral samples showed higher sensitivity.

And what does that change in practice? Not diagnosis yet. If larger clinical studies confirm the signal, a mouth rinse could be easier to collect and generally less burdensome than a stool sample, helping identify people who should be prioritized for established procedures such as upper endoscopy or colonoscopy.

The Yonsei researchers are planning a prospective study in people who do not already have a cancer diagnosis, along with higher-resolution metagenome sequencing. They also want to determine whether oral microbes contribute to cancer development, whether cancer allows those microbes to persist in the gastrointestinal tract, or whether both processes occur. Until those studies are completed, the result remains a promising analysis of patient samples rather than a clinical screening test.