LAUSR

Colorectal tumors with superficial submucosal invasion, which cannot be removed by snaring, are one of the most optimal indications for colorectal endoscopic submucosal dissection (ESD). Therefore, estimation of the invasion depth is the first key to successful colorectal ESD. Although estimation of the invasion depth based on the gross morphology may be useful in selected cases, its diagnostic accuracy could not reach the clinical requirement. The Japan Narrow-band Imaging (NBI) Expert Team (JNET) classification of NBI magnifying endoscopy findings is a useful method for histologic prediction and invasion depth estimation. However, magnifying chromoendoscopy is still necessary for JNET type 2B lesions to reach a satisfactory diagnostic accuracy. Endocytoscopy with artificial intelligence is a promising technology in invasion depth estimation; however, more data are needed for its clinical application.