Researchers at the University of Geneva have catalogued gut bacteria at the subspecies level and, with machine learning, trained a stool-based test that detects 90% of colorectal cancers—almost on par with colonoscopy (94%) but cheaper, faster, and far less invasive. While still in clinical trials, this approach could reshape CRC screening by reserving colonoscopy for confirmation, not first-line detection. Beyond cancer, the method opens the door to diagnosing a range of diseases from a single stool sample.
Keep Reading
Add A Comment