Scientists have identified a set of biological markers that could significantly improve how gastrointestinal diseases (GIDs) are detected and treated. These conditions include gastric cancer (GC), colorectal cancer (CRC), and inflammatory bowel disease (IBD).
Their findings show that specific gut bacteria and compounds known as metabolites are closely linked to each disease. This increases the possibility of diagnosing these conditions earlier and in a less invasive manner. Some of these markers may indicate risk across multiple diseases.
AI reveals common gut biomarkers across diseases
To uncover these patterns, researchers used advanced machine learning and AI-based tools to analyze microbiome and metabolomic data from patients with GC, CRC, and IBD. They found that by comparing data between diseases, models trained in one condition were often able to predict markers in another condition. For example, models based on GC data were able to identify IBD biomarkers, and CRC models were able to accurately predict GC-related markers.
The research was carried out by a team from the University of Birmingham-Dubai (part of the Health Data Science MSc programme), the University of Birmingham, Birmingham, UK, and Birmingham University Hospitals NHS Foundation Trust. Their results are Translational Medicine Journal.
Lead co-author Dr Animesh Acharjee, from the University of Birmingham, said: ‘Current diagnostic methods such as endoscopy and biopsies are effective, but they are invasive, expensive and can miss the disease in its early stages.
“Our analysis provides a deeper understanding of the underlying mechanisms driving disease progression and identifies key biomarkers for targeted therapy. These biomarkers can help identify the disease earlier and more accurately, potentially leading to better, more personalized treatments.”
Disease-specific and overlapping intestinal features
The study also revealed distinct microbial and metabolic patterns for each disease, as well as significant overlap.
Bacteria from the Firmicutes, Bacteroidetes, and Actinobacteria groups were commonly found in GC. The researchers also observed changes in metabolites such as dihydrouracil and taurine. Some of these markers are also associated with IBD, indicating common biological features. However, although they were useful in identifying IBD, they were less effective in detecting CRC.
In the case of CRC, important indicators included bacteria such as Fusobacterium and Enterococcus, as well as metabolites such as isoleucine and nicotinamide. Some of these also appear in GC, suggesting that these diseases may share underlying biological pathways.
In IBD, bacteria from the family Lachnospiraceae played an important role, along with metabolites such as urobilin and glycerate. Of note, some of these markers are also involved in cancer-related processes, reinforcing the idea that these conditions are interconnected.
Simulations show clear differences between healthy and diseased states
The researchers also simulated how gut bacteria grow and how metabolites flow through biological systems. These simulations revealed clear metabolic differences between healthy and diseased individuals, further supporting the role of these biomarkers in diagnosis.
“The cross-disease analysis of our study highlighted the potential for microbial and metabolic biomarkers identified in one GID to be used to predict another GID,” added Dr. Acharjee. “This innovative approach could lead to the development of universal diagnostic tools that could revolutionize the diagnosis and treatment of multiple gastrointestinal diseases.”
Towards non-invasive testing and personalized treatment
The researchers will now consider how these findings can be applied in clinical settings. This includes the development of non-invasive diagnostic tests and more targeted treatments based on identified biomarkers.
They also plan to validate the model using a larger and more diverse group of patients and investigate whether these biomarkers can help predict additional related diseases in the future.
