News

Abstract

　CT colonography (CTC), originally known as virtual colonoscopy, was introduced in the 1990s as a non-invasive alternative to optical colonoscopy for the diagnosis of colorectal cancer. Despite its clinical promise, widespread adoption has been limited by the need for cathartic bowel preparation that patients find burdensome. This talk traces the evolution of CTC over the past quarter century, and presents recent work on laxative-free CTC, where deep learning and generative AI models are used to perform virtual cleansing of residual bowel contents and computer-aided detection of polyps, thereby eliminating the traditional need for a thorough cathartic colon preparation. We show that these AI-powered approaches resolve the long-standing challenge of detecting polyps in a minimally prepared colon, bringing truly patient-friendly yet accurate colorectal cancer screening to clinical reality.

Contact

Prof. Kenji Suzuki（AI-Based Information Integration Research Core）