Endoscopic management of familial adenomatous polyposis using artificial intelligence-assisted polypectomy

The incorporation of artificial intelligence (AI) has significantly increased the detection of colorectal neoplasia, as determined by a meta-analysis of randomized controlled trials. However, since polyposis patients were excluded from these studies, the performance of AI in colonoscopy for the detection of hereditary polyposis syndrome has never been reported.
We herein report cases of patients followed in hour hospital for familial adenomatous polyposis (FAP). They were diagnosed with FAP because of their family history and colonoscopic findings and recommended to receive standard treatment with total colectomy. Patients refused total colectomy and chose treatment option involving endoscopic treatment. They underwent repeated colonic polypectomy every 6 months, with several colorectal polyps of >5 mm in size resected in each session. A computer-aided detection system (EX-1; FUJIFILM Holdings Corp.) was used in recent polypectomy procedure. Many polyps were indicated by cyan-colored squares in a real-time single screen. Blue-light imaging mode showed a polyp with a regular vessel pattern and indicated that the detected polyps were neoplastic lesions with yellow marking on the position map. The lesions detected by the system were removed by snare polypectomy. More than 40 polyps were resected in single polypectomy session. AI helped the endoscopist detect polyps and eventually supported resection of numerous polyps. This real-time AI-assisted colonoscopy will contribute to the management of FAP.