Grant: Project Grant

Colorectal cancer (CRC) is the second leading cause of cancer deaths in Canada, with more than 25,000 new cases expected each year. Most CRCs develop from polyps (small growths in the colon) that can be detected and removed during colonoscopy. A crucial step is measuring polyp size, because size determines how the polyp should be removed and how often a patient should return for follow-up. Unfortunately, doctors currently rely on visual estimates or simple tools such as forceps or snares, which are often inaccurate. This can result in incomplete removal, incorrect follow-up, or missed opportunities to prevent cancer. Our team is developing an artificial intelligence (AI) tool that can automatically and accurately measure polyp size during colonoscopy, without extra equipment. The system will be trained on a large video databank at the Centre hospitalier de l'Université de Montréal (over 7,500 patients and 6,000 polyps recorded) and expanded with a new nationwide study across eight Canadian hospitals. Our group has pioneered a new "gold standard" method for measuring polyp size using digital microscopy of fresh specimens, ensuring that the AI model learns from the most accurate reference available. This project will test the AI tool's ability to classify polyps into key size groups, provide exact millimeter measurements, and guide correct follow-up schedules. It will also evaluate whether the tool performs equally well across patients of different sexes, genders, and ethnic backgrounds, and across different endoscopy platforms. By giving endoscopists reliable, real-time size estimates, this AI tool could make colonoscopy safer, more accurate, and more effective. It has the potential to prevent thousands of cancers, reduce unnecessary procedures, and improve care for Canadians. Beyond this project, the technology could also be applied to other gastrointestinal conditions where accurate measurement is critical.