"Detecting Cancer Signals with a Single Mouth Rinse"...Remarkable Technology Developed by Korean Researchers [Reading Science]

Researchers from Yonsei University and Others Quantify Migration of Oral Microbes to the Gut, Markedly Elevated in Cancer Patients

Analysis of 507 Individuals and Validation Using 2,031 International Samples Show Potential for Non-Invasive Cancer S

A new diagnostic method has been developed that can detect signs of gastric and colorectal cancer simply by rinsing the mouth once. This approach is based on the discovery that oral microbes, which have difficulty settling in the intestines in healthy individuals, are much more likely to migrate to the gut in cancer patients. The research team successfully quantified this "microbial aberrant transfer" as an index to identify cancer.

Photo by Getty Images Bank

On August 21, the National Research Foundation of Korea announced that a joint research team including Professor Jihyun Kim of the Department of Systems Biology at Yonsei University; Postdoctoral Researcher Raegun Jang; Professor Seonha Ji of the Graduate School of Public Health at Yonsei University; Professors Yongchan Lee and Taeil Kim of Yonsei University College of Medicine; and Professor Jiwon Heo of Chungnam National University, developed an "Oral-Fecal Microbiota Transfer Index" and presented a microbiome-based diagnostic strategy for distinguishing gastric and colorectal cancer utilizing this index.

Strategies for Gastric and Colorectal Cancer Diagnosis Utilizing Microbial Transfer Information from the Mouth to the Intestines. The research team analyzed oral and fecal microbiota of healthy individuals and patients with gastric and colorectal cancer to develop the "Oral-Fecal Microbiota Transfer Index (MF Index)." Cancer patients had higher MF Index values than healthy individuals, and transfer patterns varied according to cancer type and stage. A machine learning model based on this was able to distinguish gastric and colorectal cancer patients using only oral or fecal samples, and its performance was reproduced in an external clinical cohort. Provided by research team

The research results were published online the same day in the international academic journal "Cell Host & Microbe," which covers interactions between hosts and microbes. A related technology patent has also been filed.

Humans swallow about 1 liter of saliva per day. During this process, approximately 1 trillion oral microorganisms also flow into the digestive tract. However, the gastrointestinal tract of a healthy person has strong defenses such as gastric acid and the immune system, making it difficult for oral microbes to reach and settle in the intestines. For this reason, the microbial ecosystems of the mouth and gut are typically clearly differentiated.

The research team focused on the possibility that diseases such as cancer could disrupt this "microbial barrier." They hypothesized that measuring the extent of microbial transfer from the mouth to the gut could capture disease signals.

The 'Oral-Gut Microbial Barrier' Collapses with Cancer

The team collected and analyzed 1,010 paired oral and fecal samples from 507 individuals—including healthy people, patients with metabolic diseases, and those with gastric, colon, or rectal cancer—who visited Severance Hospital and the Severance Health Checkup Center. By tracking sequence variants of microbes found in both oral and fecal samples, they calculated the proportion of microbes in the feces likely originating from the mouth. This was formulated as the "Oral-Fecal Microbiota Transfer Index."

Analysis showed that cancer patients exhibited statistically significantly higher transfer indices than healthy individuals. This difference remained evident even after accounting for factors such as alcohol consumption, exercise, and body mass index (BMI) that could affect the microbiome.

The pattern of microbial transfer also varied according to the cancer type and stage. The research team explained that the characteristics of microbial migration from the mouth to the gut not only indicate the presence of cancer, but could also serve as an indicator of its progression.

The team built a model distinguishing gastric and colorectal cancer patients using the "random forest" artificial intelligence analysis method. Its performance was not only validated with domestic patient data, but also through 2,031 samples from seven external clinical cohorts around the world. This was to confirm whether cancer-related microbial signals could be reproduced in diverse countries and populations.

Notably, cancer-related signals could be detected even with only an oral sample, without the need for fecal samples. This demonstrates the possibility that a simple mouth-rinsing procedure could become a future screening tool for gastric and colorectal cancers.

The Yonsei Signature research team that developed a digestive cancer diagnosis model using the oral-gastrointestinal axis microbiome. Photo assisted by AI. (From left) Jiwon Heo, Professor at Chungnam National University; Raegun Jang, Postdoctoral Researcher at Yonsei University; Jihyun Kim, Professor at Yonsei University; Yongchan Lee, Professor at Yonsei University College of Medicine; Taeil Kim, Professor at Yonsei University College of Medicine; Seonha Ji, Professor at Yonsei University. Courtesy of Yonsei University

Higher Sensitivity than the Fecal Occult Blood Test... Further Validation Needed for Commercialization

The research team also compared their model's performance to the fecal occult blood test, which is the standard first-line screening tool for colorectal cancer. The microbiome-based model showed higher performance in terms of "sensitivity," meaning it more accurately identified true cancer cases.

However, these results do not mean that rinsing one's mouth can immediately provide a definitive diagnosis of gastric or colorectal cancer. To be used in actual health screenings, the method requires larger-scale, prospective clinical trials to verify its accuracy, as well as the possibility of false positives and false negatives. It is also necessary to determine how best to combine this method with existing techniques such as endoscopy and fecal occult blood tests. At this stage, it is closer to being a non-invasive "screening test" candidate technology for conveniently identifying high-risk groups who need further testing.

Professor Kim commented, "The fact that cancer-related signals can be detected simply by rinsing the mouth is significant in lowering the bar for cancer screening," and added, "Our future goal is to advance this approach into a personalized disease risk prediction model by integrating genetic and clinical information."

This content was produced with the assistance of AI translation services.