The Cancer in Your Stool: One Bacterial Signature Betrays Colorectal Cancer Worldwide — and Fiber Can Turn It Down

The largest gut-microbiome meta-analysis of colorectal cancer to date (6,779 stool samples plus 906 tumor-tissue samples) shows that CRC carries a single, robust bacterial signature that is essentially identical in young and old patients and detectable across continents and sequencing methods. The same signature tracks inversely with dietary fiber and is nudged downward by fiber-rich diets.

For two decades, researchers have known that the bacteria living in our gut look different in people with colorectal cancer. The problem was that every study seemed to find a slightly different set of culprits, because each used different patients, different DNA-extraction kits, and different sequencing machines. The noise drowned out the signal.

A team led by Georg Zeller, working across the European Molecular Biology Laboratory in Heidelberg and Leiden University Medical Center, set out to fix that. Rather than running yet another small cohort, they re-processed the raw data from 27 published stool studies and seven tumor-tissue studies through one uniform computational pipeline — 6,779 fecal samples from 17 countries, plus 906 tissue samples. It is, to their knowledge, the largest single-disease gut-microbiome dataset ever assembled.

The headline finding is consistency. The same handful of bacterial genera — Fusobacterium, Peptostreptococcus, Parvimonas, and Porphyromonas — are enriched in cancer patients no matter where they live or which sequencing technology was used. A machine-learning classifier trained on this unified signature distinguished cancer patients from healthy controls with an accuracy (AUC) of 0.84, and a model trained on one sequencing method still worked when applied to the other.

The second surprise concerns age. Colorectal cancer is rising alarmingly in people under 50, and several earlier (mostly Chinese) studies had claimed that “early-onset” disease has its own distinct microbial signature. With far more data spanning the globe, this study finds almost no such difference: the microbiome of a 40-year-old patient looks like that of a 70-year-old patient. Cancer status, not age, drives the change — explaining 8 to 16 percent of the variation in key bacteria, versus less than 1 percent for age.

Crucially, the team showed the signature is not fixed. Across multiple dietary studies, people who ate more fiber — vegetarians, vegans, Mediterranean-diet eaters, and participants given fiber-rich snacks — carried a measurably lower “cancer-like” microbiome score. The bacteria most strongly associated with cancer also carry virulence genes (adhesins and toxins) concentrated in the Fusobacterium nucleatum clade, with intriguing geographic differences in which strains dominate.

The work does not prove that bacteria cause cancer, nor that fiber prevents it through the microbiome. But it nails down, at unprecedented scale, that the CRC microbiome signature is real, universal, and at least partly modifiable.

Actionable Insights (with effect sizes)

The one practical, repeatable takeaway from this paper is dietary fiber. Eating more of it shifts your gut bacteria in the right direction, but the effect measured here is real yet modest in size.

The researchers gave each person a “CRC microbiome score” — higher scores mean the gut community looks more like a cancer patient’s. People who ate more fiber tended to have lower scores. The strength of that link was a Spearman correlation of about -0.13 to -0.20 across the different study groups (for example, -0.18 with a p-value of 0.00026 in the 414-person Yachida group, and -0.20 with a p-value of 0.013 in overweight adults). A correlation that size means fiber intake accounts for only about 2 to 4 percent of the differences in people’s scores — a consistent, repeatable signal, but a small one for any single person.

The actual diet experiments showed a stronger effect within each study, but those studies were small. Switching healthy people from a low-fiber Western diet to a high-fiber Mediterranean diet lowered their score (p = 0.002, 15 people). Fiber supplements did the same (p = 0.049, 34 people), as did fiber-rich snacks (p as low as 0.000059). People eating plant-based diets scored lower than meat-eaters (p = 0.043).

Bottom line: a high-fiber, plant-heavy diet measurably moves your gut bacteria away from the cancer pattern. Just treat it as one helpful nudge you can stack with other healthy habits, not a single fix with a large effect on its own.

Source:

  • Open Access Paper: Meta-analysis reveals microbiome signatures for colorectal cancer that are universal across age groups and sequencing methods
  • Institution: European Molecular Biology Laboratory (EMBL), Heidelberg, in collaboration with Leiden University Medical Center (LUMC) and partners in Dresden, Greifswald, and Bern.
  • Country: Germany (lead), with the Netherlands and Switzerland.
  • Journal: Cell Host & Microbe (Cell Press / Elsevier), Volume 34, July 8, 2026.
  • Impact Evaluation: The impact score of this journal is 18.7 (Journal Impact Factor, JCR 2024), evaluated against a typical high-end range of 0–60+ for top general and life-science journals, therefore this is a High impact journal.

Related Reading:

1 Like