Colorectal cancer is one of the most common cancers worldwide and remains a leading cause of cancer-related deaths. Despite being largely preventable through timely detection and treatment, screening rates continue to be lower than recommended. A major reason is the invasive nature of colonoscopy, the current gold standard for diagnosis. But scientists at the University of Geneva have created the first detailed catalogue of gut bacteria at the sub-species level, unlocking powerful new ways to detect colorectal cancer.
By applying machine learning to stool samples, they created an inventory, which was then used to detect the presence of colorectal cancer according to the bacteria present in simple stool samples. They achieved a 90 per cent detection rate, which is closest to results from colonoscopies. “This can revolutionise early cancer screening and also make the tests affordable,” says Dr Ankit Jain, senior consultant, Medical Oncology, Indraprastha Apollo Hospitals, New Delhi.
The human gut is home to trillions of micro-organisms, collectively known as the microbiome. These microbes play an essential role in digestion, immunity and overall health. Increasing evidence shows that disruptions in the microbiome are linked to several diseases, including colorectal cancer.
Researchers are now able to study the microbiome in far greater detail, down to the subspecies level. This allows the identification of very specific bacterial patterns that may serve as early warning signs of cancer. By applying artificial intelligence to stool samples, it becomes possible to recognise these patterns with a high degree of accuracy.
A stool test that could reliably detect cancer would offer multiple benefits over colonoscopy:
Non-invasive: No sedation, instruments, or hospital visits would be required.
Convenient: Samples could be collected at home with minimal disruption.
Affordable: Costs would be lower compared to colonoscopy, making screening more accessible.
Scalable: Large-scale public health screening programmes could be expanded more easily.
Such a test could encourage higher participation in screening programmes, enabling earlier detection and improving survival rates.
While early findings are promising, stool-based tests powered by artificial intelligence are not yet ready for widespread clinical use. These technologies remain in the research stage and require validation in larger, more diverse populations before they can be integrated into standard medical practice.
It is also important to note that colonoscopy offers advantages beyond detection. The procedure allows physicians to directly visualize the colon and remove pre-cancerous polyps during the same session. Stool-based tests, on the other hand, can only signal the presence of abnormalities and would still need to be followed by colonoscopy for confirmation and treatment.
Other stool-based tests, such as the faecal immunochemical test (FIT), are already available and widely used. These tests are less invasive but must be repeated more frequently and are less sensitive than colonoscopy.
Medical guidelines currently recommend that adults aged 45 and above undergo routine colorectal cancer screening. Those with a family history or additional risk factors may need to begin earlier.