Goodbye colonoscopy? A new shit test detects 90% of colorectal cancers

UNIGE scientists have created a artificial intelligence Method that can diagnose colorectal cancer by analyzing the intestinal microbiota.

Colorectal cancer classifies the second most common cause of death -related deaths in the world. When captured early, it can generally be treated effectively, but the expenses and discomfort of colonoscopies, the standard diagnostic procedure, often delay detection. Researchers from the University of Geneva (UNIGE) have now taken a step in the major by using automatic learning To map all human intestinal bacteria at an unprecedented level of detail, allowing them to better understand the physiological role of different microbial subgroups.

This complete catalog was then applied to identify colorectal cancer through bacterial signatures found in simple stool samples, offering a non -invasive and affordable alternative to traditional methods. The approach also has large potential applications, from the detection of other cancers to knowledge of knowledge of how the intestinal microbiota influences health. The results were recently published in Cell host and microbe.

Colorectal cancer is frequently diagnosed only after progressing towards advanced stages, when the processing options are limited. This highlights the urgent need for easier and less invasive screening tools, in particular given the confusing increase in cases in young adults. Although scientists have long since known that intestinal microbes have influence the development of colorectal cancer, the application of this knowledge in clinical practice has been difficult. One reason is that different strains in the same bacterial species Can act very differently – some can contribute to the disease, while others seem harmless.

“Instead of relying on the analysis of the different species making up the microbiota, which does not capture all the significant differences, nor of bacterial strains, which vary considerably from one individual to another, we have focused on an intermediate level of the microbiota, the under-species” of medicine, which led this research. “The resolution of subspecies is specific and can seize the differences in the functioning of bacteria and contributes to diseases, including cancer, while remaining enough in general to detect these changes between different groups of individuals, populations or countries.”

With the help of automatic learning

The first step was to analyze huge amounts of data. “As a bioinformatrician, the challenge was to offer an innovative approach for mass data analysis,” recalls Matija Trickovic, a doctoral student in the laboratory of Mirko Tracovski and the first author of this study. “We have succeeded in developing the first complete catalog of the subspecies of human intestinal microbiota, as well as a precise and effective method to use it both for research and in the clinic.”

By combining this catalog with existing clinical data, scientists have developed a model that can predict the presence of colorectal cancer only based on bacteria present in stool samples. “Although we had confidence in our strategy, the results were striking,” Enthusiasm Matija Trickovic. “Our method has detected 90% of cancer cases, a result very close to the detection rate of 94% reached by colonoscopies and better than all current non -invasive detection methods.”

By incorporating additional clinical data, the model has the potential to become even more precise, addressing the reliability of the colonoscopy. It could serve as a routine screening method, making it possible to identify the first cases of colorectal cancer, with colonoscopies and then used only to confirm the diagnosis in a smaller group of targeted patients.

A new world of applications

A first clinical trial is implemented in collaboration with hospitals of the University of Geneva (HUG) to determine more precisely the stages of cancer and the lesions that can be detected. However, applications go beyond colorectal cancer. By studying the differences between the subspecies of the same bacterial species, researchers can now identify the mechanisms of action by which the intestinal microbiota influences human health.

“The same method could soon be used to develop non -invasive diagnostic tools for a wide range of diseases, all based on a single microbiota analysis,” concludes Mirko Trajkovski.

Reference: “The subspecies of the human intestinal microbiota carry implicit information for the in-depth research of microbiomas” by Matija Tričković, Silas Kieser, Evgeny M. Zdobnov and Mirko Trajkovski, August 13, 2025, Cell host and microbe.
DOI: 10.1016 / J.CHOM.2025.07.015

Never miss a breakthrough: join the Scitechdaily newsletter.