Researchers found that certain gut bacteria and metabolites are linked to each disease — suggesting that these biomarkers could help in early, less invasive diagnosis of GIDs, with some markers indicating risk across multiple diseases.

They used advanced machine learning and AI-based algorithms to analyse microbiome and metabolome datasets from patients with GC, CRC, and IBD. Their cross-disease analysis revealed that models trained on GC data could successfully predict IBD biomarkers, and CRC models could predict GC biomarkers with high accuracy.

The research team, based at the University of Birmingham Dubai (Part of Health Data Science MSc Programme), University of Birmingham, Birmingham, UK, and University Hospitals Birmingham NHS Foundation Trust published its findings in Journal of Translational Medicine.

Lead co-author Dr Animesh Acharjee, from the University of Birmingham, commented: “Current diagnostic methods like endoscopy and biopsies are effective but can be invasive, expensive, and sometimes miss diseases at early stages.

“Our analysis offers a better understanding of the underlying mechanisms driving disease progression and identifies key biomarkers for targeted therapies. These biomarkers could help identify diseases earlier and more accurately, leading to better, more personalised treatment.”

The study reveals that, for example, in GC bacteria from the Firmicutes, Bacteroidetes, and Actinobacteria groups were common, and changes in certain metabolites like dihydrouracil and taurine were noted. Some of these biomarkers were also relevant for IBD, suggesting overlap between the diseases but, while these markers worked well for detecting IBD, they were less effective for CRC.

For CRC, bacteria such as Fusobacterium and Enterococcus, and metabolites like isoleucine and nicotinamide, were significant — sometimes overlapping with those for GC, indicating possible shared pathways in disease development.

In IBD, bacteria from the Lachnospiraceae family and metabolites like urobilin and glycerate were important, with some of these markers also involved in cancer pathways, showing how these diseases are interconnected.

The research team simulated gut microbial growth and metabolite fluxes, revealing significant metabolic differences between healthy and diseased states.

“Our study’s cross-disease analysis emphasised the potential of using microbial and metabolic biomarkers identified in one GID to predict another,” added Dr Acharjee. “This innovative approach could lead to the development of universal diagnostic tools to revolutionise the diagnosis and treatment of for multiple gastrointestinal conditions.”

The research team now plans to further explore the clinical applications of their findings, including the development of non-invasive diagnostic tests and targeted therapies based on the identified biomarkers. They also aim to validate their models in larger, diverse patient cohorts and investigate these biomarkers’ potential in predicting other related diseases.