Researchers have been working toward using the gut microbiome as a clinical target for screening for colorectal cancer. This would ideally allow simple stool tests to be used for early detection of the disease in place of expensive procedures such as colonoscopies. An international study that has just been published in Nature Medicine focused precisely on this subject. The research is coordinated by the University of Trento. The paper was first-authored by Gianmarco Piccinno, with Nicola Segata as head of project and scientific coordinator, both Department of Cellular, Computational and Integrative Biology-Cibio. Segata is also principal investigator at the European Institute of Oncology (IEO) in Milan.
The study includes over 1,600 newly sequenced samples, with the largest group of new colorectal cancer cases (204/525, more than 38%) provided by the Czech COLOBIOME cohort. This cohort was initiated by Dr. Eva Budinská from the RECETOX research institute at the Faculty of Science, Masaryk University, during her parallel affiliation with the Masaryk Memorial Cancer Institute (MMCI), where a prospective observational study was conducted under an AZV national grant. The samples were later integrated into the ONCOBIOME project, in which Dr. Budinská now serves as the principal investigator for the Czech Republic, thanks to the support of the MMCI’s Bank of Biological Material, which enabled long-term storage of the study biospecimens.
An additional Czech dataset (IIGM-CZ), coordinated by the Institute of Experimental Medicine of the Czech Academy of Sciences (IEM CAS) and Charles University in Prague and Pilsen in collaboration with Italian Institute for Genomic Medicine, contributed a further 124 samples (of which 59 cancer cases). Together, these Czech cohorts represent the largest national contribution to the newly generated data and were pivotal in building and validating the study’s predictive models.
Dr. Roman Hrstka, Director of the Bank of Biological Material at Masaryk Memorial Cancer Institute, commented:
“We are pleased that biospecimens provided by our biobank played a key role in such an impactful study. This is a clear example of how long-term investments into high-quality biobanking infrastructure can significantly contribute to international research efforts with direct clinical implications.”
About the study
The team of researchers has identified a "microbial signature", that is, a group of about a dozen bacteria that tend to be much more present and abundant in the intestines of people with colorectal cancer. Among them, Fusobacterium nucleatum has already been extensively studied, but this study has highlighted others such as Parvimonas micra, Gemella morbillorum, Peptostreptococcus stomatis, to name a few.
Nicola Segata explains: "We think that probably, for reasons not fully known yet, these species are able to colonize the so-called ‘tumor microenvironment’ in the guts of patients with colorectal cancer. These bacteria,” he continues, “are normally found in the mouth of healthy individuals, but we find them in sequenced stool samples (that reflect the composition of the gut microbiome) almost only belonging to individuals with colorectal cancer."
The identification of this microbial signature could allow the development of accurate disease screening approaches based exclusively on gut metagenomics, which is an entirely non-invasive method. According to the two scholars, the accuracy of this approach is close to 90 percent. But there’s more to it. The microbial signature they have identified is also correlated with the clinical stage of colorectal cancer, i.e., its severity and anatomical location. They have seen how some bacteria of the microbial signature increase with tumor development and could therefore provide interesting hypotheses on the mechanisms of the disease and possibly on its clinical development.
Dr. Eva Budinská, from RECETOX MU, principal investigator for the Czech Republic in the ONCOBIOME consortium, commented:
“We are proud that the Czech cohorts played such a vital role in this study. Our COLOBIOME cohort provides one of the most comprehensive clinical resources to date, with detailed staging and tumor localization for over 200 patients. This depth of information significantly strengthened the study’s ability to link microbiome signatures with cancer progression.”
“This study is a great example of how international collaboration and data sharing can accelerate progress in complex diseases like colorectal cancer. By combining cutting-edge metagenomics analysis with large, well-characterized cohorts across countries, we are moving closer to real-world applications of microbiome research in early detection and personalized medicine. It has been a privilege for our team to contribute meaningfully to this important global effort.” says Dr. Vymetálková from collaborating Institute of Experimental Medicine of CAS.
The contribution of machine learning
The researchers achieved these results thanks to a methodology based on machine learning that they developed and used. "This technology, applied to the metagenomic analysis approaches that we use in our laboratory,” underlines Segata, “provides a predictive model that says if there is a high probability or not that a person has colorectal cancer."
The question that remains open and which others are trying to answer is whether and which of these microbial species and strains contribute to the onset of the cancer itself. "Whether or not the microbiome in the tumor is a cause of the tumor is almost indifferent if the primary objective is to develop a screening approach for individuals in the general population," clarifies Gianmarco Piccinno. In fact, the study focuses on the potential use as a non-invasive screening tool, so that diseases are identified as early as possible. Nevertheless, the definitive clinical and preventive value of this metagenomic test remains to be verified via future registered clinical trials.
However, according to Piccinno and Segata, the results achieved, once validated, could contain enough information to be used as an important additional form of screening. This would therefore allow a more targeted use of colonoscopy, which remains the necessary examination for a final diagnosis.
The link between the gut microbiome and colorectal cancer
The first connection was established years ago: some specific bacteria in the microbiota have the ability to disrupt our own cellular biology and are thought to contribute to cancer over time. In some cases, this is due to the production of toxins that can damage our DNA leading to mutations- which are one of the factors that lead to cancer – thus allowing the gut microbiota to perhaps generate an environment that could facilitate the development of the disease. But this is only one of the potential ways through which the microbiome is linked with colorectal cancer, which is the third most frequent type of cancer and the second most lethal in the world, and only 40 percent of patients are diagnosed before metastasis.
This study was primarily funded by the European Commission with the collaboration "ONCOBIOME" led by prof. Laurence Zitvogel which has the overall aim of finding links not only between the human microbiome and colorectal cancer, but also the role of the microbiome in supporting cancer treatment, especially the latest development in immunotherapy approaches for cancer such as metastatic melanoma, small cell lung cancer, and metastatic breast cancer.
What prompted the researchers to investigate this connection further was the increase in incidence of the disease in young adults, particularly in individuals under 50 years of age. Early onset cancers were one of the challenges put to the research community in 2023 by Cancer Grand Challenges, the research initiative co-founded by Cancer Research UK and the National Cancer Institute, with team PROSPECT, led by Andy Chan of Massachusetts General Hospital and Yin Cao of Washington University in St. Louis, subsequently awarded up to $25m funding in 2024 to unravel the mechanisms behind the rise in early onset colorectal cancer.
Segata and Piccinno also joined the multidisciplinary international PROSPECT team together with Professor Curtis Huttenhower of the Harvard T.H. Chan School of Public Health. Segata, Piccinno, Huttenhower, Chan, and Zitvogel are all co-authors of the study together with several members of their laboratories.
Many questions remain regarding the role of the microbiome in cancer, including the full range of mechanisms involved and which microbial species play a causal role, if any.
A worldwide network
The study is part of major international projects involving research partners from all over the world: ONCOBIOME, funded by the European Commission, as well as two Cancer Grand Challenges teams (team PROSPECT, funded by Cancer Research UK, the National Cancer Institute, The Bowelbabe Fund for Cancer Research UK and The French National Cancer Institute, and team OPTIMISTICC, funded by Cancer Research UK).
Thanks to this collaboration network, the research group of UniTrento was able to access more than 3,700 samples, collected and sequenced from 18 different studies and from almost all continents (excluding Africa, South America and Oceania).
Together, the two Czech cohorts contributed 328 newly sequenced samples, including 260 colorectal cancer cases—representing nearly 50% of all CRC samples in the new datasets. Their inclusion was critical in achieving the level of statistical power necessary for machine learning models to reach high predictive accuracy.
The ultimate objective of these studies is to verify whether there is an association between colorectal cancer and the microbiome compared to external factors such as pollution, diet, lifestyle, and to understand which the most effective treatments are given the unique composition of each patient's microbiome.
“We are especially grateful to the Czech patients who generously participated in this study, and to our clinical collaborators – in particular the Masaryk Memorial Cancer Institute, University Hospital and Faculty of Medicine in Pilsen (Charles University), Thomayer Hospital and the First Faculty of Medicine (Charles University), and the Institute for Clinical and Experimental Medicine in Prague – whose support was essential for collecting high-quality, clinically annotated samples,” added Dr. Budinská
“Cross-stage, strain-level, pooled analysis of the gut microbiome in 3,741 individuals from 18 colorectal cancer cohorts” was published in Nature Medicine and is available at: https://www.nature.com/articles/s41591-025-03693-9 (DOI: 10.1038/s41591-025-03693-9)
This text was taken and adapted from the original press release of the University of Trento.
